# About Hats Network > **AS203314 - Global BGP Backbone** > > Hats Network Inc. operates **AS203314**, an eyeball ISP and IP transit network. We provide peering and transit services across the locations listed in our operational references. > > * **Founded:** 16 Aug 2022 > * **ASN:** AS203314 > * **Peering policy:** Open, subject to the [published requirements](/docs/peering) > * **Network type:** Eyeball ISP and transit provider ## What We Operate Our work covers backbone routing, interconnection and customer handoffs. The service pages describe delivery options; the peering policy sets the requirements for exchanging routes with us. Use the network map for location context, PeeringDB for current peering coordinates, and the dated latency dataset for published measurements. These sources describe different aspects of the network and should not be treated as interchangeable. ## Network Resources - [Global PoP Map](/#network): View our interactive global network map and Points of Presence locations across Asia, Europe, and the Americas. - [GeoFeed](https://hatsnet.io/geofeed/all.csv): Subscribe to our Geographic Feed (RFC 8805) for the latest updates on IP geographical mappings. - [PeeringDB](https://www.peeringdb.com/asn/203314): View our complete peering information, including IX participation and contact details. - [BGP Tools](https://bgp.tools/as/203314): Explore our routing information, prefixes, and ASN connectivity on BGP.tools. ## Global Presence > Interactive content is available on the canonical HTML page. > The map shows the network's published locations. Availability, port capacity and handoff details must be confirmed for an order. [PeeringDB](https://www.peeringdb.com/asn/203314) is the operational reference for peering. ### Core Locations | Region | City | IATA | Capacity | IX Exchanges | Status | | ----------- | --------- | -------------- | -------- | -------------------------------- | ------ | | Hong Kong | Hong Kong | HKG1/HKG2/HKG3 | 50 Gbps | Equinix HK, HKIX | Active | | Japan | Tokyo | TYO2 | 100 Gbps | Equinix Tokyo, JPIX, BBIX, JPNAP | Active | | Singapore | Singapore | SIN1 | 200 Gbps | Equinix SG, SGIX | Active | | Netherlands | Amsterdam | AMS | 40 Gbps | AMS-IX, NL-IX | Active | | Germany | Frankfurt | FRA | 100 Gbps | DE-CIX Frankfurt | Active | | UK | London | LON | 100 Gbps | LINX | Active | ### Edge Locations | Region | City | IATA | Capacity | IX Exchanges | Status | | ------------ | ------------ | --------- | ---------- | ---------------------- | ------- | | Taiwan | Taipei | TPE1/TPE2 | 20 Gbps | STUIX, TPIX-TW, TWIX | Active | | Germany | Berlin | BER | Not listed | DE-CIX, BCIX | Active | | France | Calais | CQF1 | 1 Gbps | Not listed | Active | | France | Marseille | MRS | 80 Gbps | France-IX | Active | | USA | Los Angeles | LAX | 100 Gbps | SFMIX | Active | | USA | Ashburn | IAD | Not listed | Not listed | Active | | USA | New York | NYC | 150 Gbps | NYIIX, DE-CIX New York | Active | | USA | Seattle | SEA | 100 Gbps | SIX Seattle | Active | | Canada | Toronto | YYZ1 | Not listed | Not listed | Active | | Australia | Melbourne | MEL | Not listed | HE / Superloop | Active | | South Africa | Johannesburg | JNB | Not listed | HE Only | Planned | | Brazil | São Paulo | GRU | Not listed | NTT / Ascenty | Planned | ### Backbone Latency Matrix Key routes (representative sample, updated periodically): | Route | RTT (ms) | Region Pair | | ----------------------- | -------- | -------------- | | Hong Kong → Taipei | \~15 | Intra-Asia | | Hong Kong → Singapore | \~32 | Intra-Asia | | Amsterdam → Frankfurt | \~5 | Intra-Europe | | Amsterdam → London | \~6 | Intra-Europe | | Ashburn → New York | \~6 | Intra-NorthAm | | New York → London | \~92 | Trans-Atlantic | | Hong Kong → Los Angeles | \~145 | Trans-Pacific | | Singapore → Melbourne | \~89 | Asia-Oceania | For dated measurements and the full set of origins, see the [Backbone Latency Matrix](/docs/network/latency). The approximate values above are reference examples. ### Featured Market Notes #### Los Angeles (LAX) Los Angeles is a US West Coast handoff market. The published site information lists SFMIX and 100 Gbps committed capacity. Check the [Los Angeles latency page](/docs/network/latency/lax-los-angeles) for the current dataset rather than using a fixed RTT estimate. #### New York (NYC) New York is an East Coast handoff market with 150 Gbps committed capacity and NYIIX and DE-CIX New York peering listed here. The [New York route measurements](/docs/network/latency/nyc-new-york) include both transatlantic and domestic comparisons. #### Seattle (SEA) Seattle provides a Pacific Northwest location with SIX Seattle peering and HE and Cogent upstreams. The listed committed capacity is 100 Gbps. [Seattle's measured routes](/docs/network/latency/sea-seattle) separate regional and trans-Pacific RTTs. #### Marseille (MRS) Marseille is a Mediterranean location with France-IX peering and 80 Gbps committed capacity listed here. See the [Marseille route measurements](/docs/network/latency/mrs-marseille) for dated RTT values. #### Calais (CQF1) Calais announces `2a0c:9a40:95ea::/48` through MoeDove (AS44324). Its location code is `253`; it supplies an additional European path for the North American backbone. #### Toronto (YYZ1) Toronto is the Canadian edge location, with a North American path toward Europe. It announces `2a0c:9a40:95eb::/48` and `2a13:a5c7:2526::/48` through ParadoxNetworks (AS52025). The primary node address is `2a13:a5c7:2526:1001::1`, and the location code is `304`. #### Melbourne (MEL) Melbourne is an Australian location with HE and Superloop upstream connectivity. Use the [Melbourne latency page](/docs/network/latency/mel-melbourne) to compare the published routes to Singapore and other destinations. ## Contact Us > **Need Help?** > > Send orders to Sales, peering requests to the Peering team, and network incidents to the NOC. Support handles general enquiries and abuse reports. ### Contact Information | Team | Email Address | Purpose | | ----------- | ----------------------------------------------- | ------------------------------------ | | **Sales** | [sales@hatsnet.io](mailto:sales@hatsnet.io) | IP-Transit orders, pricing inquiries | | **Peering** | [peering@hatsnet.io](mailto:peering@hatsnet.io) | Peering requests, PNI arrangements | | **NOC** | [noc@hatsnet.io](mailto:noc@hatsnet.io) | Network issues, outages | | **Support** | [support@hatsnet.io](mailto:support@hatsnet.io) | Abuse reports, general support | ## Quick Links - [Peering Policy](/docs/peering): Learn about our open peering policy and how to establish a session with AS203314. - [IP Transit](/docs/transit): Explore our IP-Transit services in Asia and Europe. - [BGP Communities](/docs/community): View our BGP community strings for route management. --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/about). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Quick Start Guide > **Hats Network Inc. (AS203314)** > > Hats Network operates **AS203314**. This documentation covers service handoffs, peering requirements, routing controls and the published latency dataset. Start with the peering policy to exchange routes, the transit pages to order a handoff, or the community reference to control route announcements. - [About Hats Network](/docs/about): Company overview, PoP locations, geofeed, and contact information. - [Our Services](/docs/services): Enterprise private networks, dedicated lines, premium user access, and performance routing profiles. - [Peering](/docs/peering): Open peering policy, IX presence, and peering coordinates. - [IP Transit](/docs/transit): Transit availability, pricing, and service coverage. - [BGP Communities Reference](/docs/community): Community strings for route control and traffic engineering. - [Need to Report Abuse?](/docs/legal/report-abuse): Please report any violations by our end-users, including DMCA, SPAM, DDoS attacks and other such activities. ## Before Configuring a Session Confirm the location and session addresses against our [PeeringDB record](https://www.peeringdb.com/asn/203314). Configuration examples use placeholder addresses and documentation prefixes; replace them with the values agreed for your session. For path comparisons, use the [dated latency matrix](/docs/network/latency). Its PoP measurements do not include a customer's access network or establish an SLA. --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Control BGP Communities Attach control communities to the prefixes you announce to AS203314. They request export restrictions or changes to path attributes as listed below. > **Traffic Engineering** > > The actions below affect propagation beyond Hats Network. ## Blackhole & Filtering | Community | Action | | -------------- | -------------------------------------- | | `203314:0:666` | Do not announce this route (blackhole) | **Usage**: Stop announcement of a prefix (e.g., during DDoS mitigation or IP address deprecation). **Note**: Equivalent to standard blackhole community `65535:65281`. ## Propagation Control Select the peer types that may receive your routes: | Community | Action | | ------------ | ------------------------- | | `203314:1:0` | Do not send to Upstream | | `203314:1:1` | Do not send to Peering | | `203314:1:2` | Do not send to Downstream | > **Example Use Cases**: > > * Use `203314:1:0` to prevent transit from carrying your prefix > * Use `203314:1:1` to keep traffic limited to IX route servers > * Use `203314:1:2` to prevent customer leakage ## Geographic Restrictions Restrict export by location: ### Regional | Community | Action | | --------------- | --------------------------------------------------- | | `203314:2:1` | Do not send to other regions (Disable Cross Region) | | `203314:2:2` | Do not send to other PoPs (Disable Cross PoP) | | `203314:201:XX` | Do not send to Region `XX` | **Examples**: ```text # Keep traffic within Asia only 203314:2:1 # Disable cross-region 203314:201:300 # Exclude North America 203314:201:200 # Exclude Europe ``` ### Pop | Community | Action | | ---------------- | ------------------------ | | `203314:202:XXX` | Do not send to PoP `XXX` | **Examples**: ```text # Exclude specific PoPs 203314:202:301 # Do not announce via Seattle 203314:202:131 # Do not announce via Tokyo 203314:202:102 # Do not announce via Hong Kong HKG2 ``` ## Path Prepending Request AS-path prepending on external advertisements: > **Scope**: Path prepending only affects routes announced to external peers (upstream, peering, customers)-not internal routing. | Community | Action | | -------------- | ------------------------------- | | `203314:220:1` | Prepend 1x outside Hats Network | | `203314:220:2` | Prepend 2x outside Hats Network | | `203314:220:3` | Prepend 3x outside Hats Network | | `203314:220:4` | Prepend 4x outside Hats Network | | `203314:220:5` | Prepend 5x outside Hats Network | **Usage**: Deprioritize specific prefixes for inbound traffic engineering. ## Community Decision Flow ## Usage Examples ### Example 1: Local Traffic Only Announce a prefix only within a specific region: ```text # Announce to Asia West only 203314:2:1 # Disable cross-region ``` ### Example 2: Deprioritize Backup Prefix Make a backup prefix less attractive: ```text # Deprioritize via path prepending 203314:220:3 # Prepend 3x AS203314 ``` ### Example 3: Selective Peering Announce only to specific peer types: ```text # Announce to peering and customers only 203314:1:0 # Do not send to upstream ``` ### Example 4: Exclude Location Prevent announcement via specific location: ```text # Exclude PoP-specific announcement 203314:202:301 # Do not send via Seattle ``` ### Example 5: Combined Control Combine multiple communities for complex policies: ```text # Deprioritized, Asia-only, no upstream 203314:1:0 # No upstream 203314:2:1 # Asia only 203314:220:2 # Prepend 2x ``` ## Implementation Notes > **Best Practices**: > > 1. **Test before production**: Use `203314:220:1` (1x prepend) to verify communities work > 2. **Monitor impact**: Check route servers and looking glasses after applying changes > 3. **Document your policies**: Keep track of which prefixes use which communities > 4. **Use conservative prepends**: 3x prepend is usually sufficient; 5x is extreme ## Configuration Examples (JunOS / BIRD2) These are **BGP Large Communities** under RFC 8092. JunOS uses the `large:` prefix, for example `large:203314:0:666`. BIRD2 uses `bgp_large_community` tuples such as `(203314, 0, 666)`. > **Export Policy Safety Default (Strongly Recommended)** > > These are export filters. End BIRD2 filters with `reject;` and JunOS policies with `term REJECT-ALL then reject` to prevent unmatched routes from being announced. ### Junos **Example: Blackhole (`203314:0:666`)** ```bash set policy-options community HATS-BLACKHOLE members large:203314:0:666 set policy-options policy-statement EXPORT-TO-HATS term BLACKHOLE from route-filter 203.0.113.0/24 exact set policy-options policy-statement EXPORT-TO-HATS term BLACKHOLE then community add HATS-BLACKHOLE set policy-options policy-statement EXPORT-TO-HATS term BLACKHOLE then accept set policy-options policy-statement EXPORT-TO-HATS term REJECT-ALL then reject ``` **Example: No Upstream (`203314:1:0`) + Prepend 3x (`203314:220:3`)** ```bash set policy-options community HATS-NO-UPSTREAM members large:203314:1:0 set policy-options community HATS-PREPEND-3X members large:203314:220:3 set policy-options policy-statement EXPORT-TO-HATS term BACKUP from route-filter 203.0.113.0/24 exact set policy-options policy-statement EXPORT-TO-HATS term BACKUP then community add [ HATS-NO-UPSTREAM HATS-PREPEND-3X ] set policy-options policy-statement EXPORT-TO-HATS term BACKUP then accept set policy-options policy-statement EXPORT-TO-HATS term REJECT-ALL then reject ``` **Example: Keep in-region + Exclude a PoP (`203314:2:1` + `203314:202:253`)** ```bash set policy-options community HATS-NO-CROSS-REGION members large:203314:2:1 set policy-options community HATS-NO-AMS members large:203314:202:253 set policy-options policy-statement EXPORT-TO-HATS term LOCAL-ONLY from route-filter 203.0.113.0/24 exact set policy-options policy-statement EXPORT-TO-HATS term LOCAL-ONLY then community add [ HATS-NO-CROSS-REGION HATS-NO-AMS ] set policy-options policy-statement EXPORT-TO-HATS term LOCAL-ONLY then accept set policy-options policy-statement EXPORT-TO-HATS term REJECT-ALL then reject ``` ### Bird2 **Example: Blackhole (`203314:0:666`)** ```bird2 filter export_to_hats_v4 { if net = 203.0.113.0/24 then { bgp_large_community.add((203314, 0, 666)); accept; } reject; # Safety default } ``` **Example: No Upstream (`203314:1:0`) + Prepend 3x (`203314:220:3`)** ```bird2 filter export_to_hats_v4 { if net = 203.0.113.0/24 then { bgp_large_community.add((203314, 1, 0)); bgp_large_community.add((203314, 220, 3)); accept; } reject; # Safety default } ``` **Example: Keep in-region + Exclude a PoP (`203314:2:1` + `203314:202:253`)** ```bird2 filter export_to_hats_v4 { if net = 203.0.113.0/24 then { bgp_large_community.add((203314, 2, 1)); bgp_large_community.add((203314, 202, 253)); accept; } reject; # Safety default } ``` ## Related Information * **[PoP Codes](/docs/community/pop-codes)** - Reference for PoP numbers (XXX) in geographic restrictions * **[Region Codes](/docs/community/region-codes)** - Reference for region numbers (XX) in geographic restrictions * **[Internal Communities](/docs/community/internal-communities)** - Communities we attach to routes we announce > **Important**: Control communities are applied to routes **you announce to AS203314**. They do not affect routes we announce to you. --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/community/control-communities). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # BGP Communities > **AS203314 BGP Communities** > > AS203314 uses BGP Large Communities to label route origins and accept customer routing controls. Internal tags describe received routes; control communities request an export action. ## Community System Overview ## Quick Reference | Category | Community | Purpose | Details | | ------------------- | ---------------- | ------------------------- | ------------------------------------------------------------------------------------- | | **Origin** | `203314:110:XX` | Route source type | [See Origin →](/docs/community/internal-communities#route-origin) | | **PoP** | `203314:120:XXX` | Learned at PoP | [See PoP-Based →](/docs/community/internal-communities#pop-based-communities) | | **PoP Pass** | `203314:125:XXX` | Passed through PoP | [See PoP-Based →](/docs/community/internal-communities#pop-based-communities) | | **Region** | `203314:130:XX` | Learned in region | [See Region-Based →](/docs/community/internal-communities#region-based-communities) | | **Region Pass** | `203314:135:XX` | Passed through region | [See Region-Based →](/docs/community/internal-communities#region-based-communities) | | **Blackhole** | `203314:0:666` | Do not announce | [See Blackhole →](/docs/community/control-communities#blackhole--filtering) | | **No Upstream** | `203314:1:0` | Do not send to upstream | [See Propagation →](/docs/community/control-communities#propagation-control) | | **No Peering** | `203314:1:1` | Do not send to peering | [See Propagation →](/docs/community/control-communities#propagation-control) | | **No Customer** | `203314:1:2` | Do not send to downstream | [See Propagation →](/docs/community/control-communities#propagation-control) | | **No Cross-Region** | `203314:2:1` | Keep within region | [See Geo Restrictions →](/docs/community/control-communities#geographic-restrictions) | | **No Cross-PoP** | `203314:2:2` | Keep within PoP | [See Geo Restrictions →](/docs/community/control-communities#geographic-restrictions) | | **No Region** | `203314:201:XX` | Exclude region | [See Geo Restrictions →](/docs/community/control-communities#geographic-restrictions) | | **No PoP** | `203314:202:XXX` | Exclude PoP | [See Geo Restrictions →](/docs/community/control-communities#geographic-restrictions) | | **Prepend 1x** | `203314:220:1` | Prepend AS 1 time | [See Path Prepending →](/docs/community/control-communities#path-prepending) | | **Prepend 2x** | `203314:220:2` | Prepend AS 2 times | [See Path Prepending →](/docs/community/control-communities#path-prepending) | | **Prepend 3x** | `203314:220:3` | Prepend AS 3 times | [See Path Prepending →](/docs/community/control-communities#path-prepending) | | **Prepend 4x** | `203314:220:4` | Prepend AS 4 times | [See Path Prepending →](/docs/community/control-communities#path-prepending) | | **Prepend 5x** | `203314:220:5` | Prepend AS 5 times | [See Path Prepending →](/docs/community/control-communities#path-prepending) | ## Community Categories - [Internal Communities](/docs/community/internal-communities): Route origin types, PoP locations, and region codes for identifying where routes are learned. - [Control Communities](/docs/community/control-communities): Traffic engineering communities for blackhole, propagation control, and path prepending. - [PoP Codes](/docs/community/pop-codes): 3-digit codes for each Point of Presence used in location-based communities. - [Region Codes](/docs/community/region-codes): 2-digit geographic region codes for area-based routing policies. ## Usage Examples > ### Common Scenarios > > * Keep traffic local: > > ```bird2 > bgp_large_community.add((203314, 2, 1)); # Do not send to other regions > ``` > > * Deprioritize upstream: > > ```bird2 > bgp_large_community.add((203314, 220, 3)); # Prepend 3x to upstream routes > ``` > > * Exclude specific PoP: > > ```bird2 > bgp_large_community.add((203314, 202, 253)); # Do not send to AMS > ``` ## Configuration Examples (JunOS / BIRD2) These examples attach AS203314 large communities to prefixes announced to us. > **Export Policy Safety Default (Strongly Recommended)** > > These are export policies. End BIRD2 filters with `reject;` and JunOS policies with `term REJECT-ALL then reject` so unmatched routes are not exported. > **Documentation Prefixes** > > The examples use RFC 5737 and RFC 3849 documentation prefixes, including `203.0.113.0/24` and `2001:db8::/32`. Substitute only prefixes you are authorized to announce. ### Junos **Keep traffic local (`203314:2:1`)** ```bash set policy-options community HATS-NO-CROSS-REGION members large:203314:2:1 set policy-options policy-statement EXPORT-TO-HATS term LOCAL from route-filter 203.0.113.0/24 exact set policy-options policy-statement EXPORT-TO-HATS term LOCAL then community add HATS-NO-CROSS-REGION set policy-options policy-statement EXPORT-TO-HATS term LOCAL then accept set policy-options policy-statement EXPORT-TO-HATS term REJECT-ALL then reject ``` **Deprioritize a backup prefix (`203314:220:3`)** ```bash set policy-options community HATS-PREPEND-3X members large:203314:220:3 set policy-options policy-statement EXPORT-TO-HATS term BACKUP from route-filter 203.0.113.0/24 exact set policy-options policy-statement EXPORT-TO-HATS term BACKUP then community add HATS-PREPEND-3X set policy-options policy-statement EXPORT-TO-HATS term BACKUP then accept set policy-options policy-statement EXPORT-TO-HATS term REJECT-ALL then reject ``` **IPv6 variant** ```bash set policy-options community HATS-NO-CROSS-REGION members large:203314:2:1 set policy-options policy-statement EXPORT-TO-HATS term LOCAL-V6 from route-filter 2001:db8:203:113::/48 exact set policy-options policy-statement EXPORT-TO-HATS term LOCAL-V6 then community add HATS-NO-CROSS-REGION set policy-options policy-statement EXPORT-TO-HATS term LOCAL-V6 then accept set policy-options policy-statement EXPORT-TO-HATS term REJECT-ALL then reject ``` ### Bird2 **Keep traffic local (`203314:2:1`)** ```bird2 filter export_to_hats_v4 { if net = 203.0.113.0/24 then { bgp_large_community.add((203314, 2, 1)); accept; } reject; # Safety default } ``` **Deprioritize a backup prefix (`203314:220:3`)** ```bird2 filter export_to_hats_v4 { if net = 203.0.113.0/24 then { bgp_large_community.add((203314, 220, 3)); accept; } reject; # Safety default } ``` **IPv6 variant** ```bird2 filter export_to_hats_v6 { if net = 2001:db8:203:113::/48 then { bgp_large_community.add((203314, 2, 1)); accept; } reject; # Safety default } ``` ## Related Resources - [Peering Policy](/docs/peering): Learn how to establish BGP peering with AS203314. - [IP Transit](/docs/transit): Get IP-Transit with BGP community support. --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/community). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Internal BGP Communities AS203314 attaches internal communities to describe how and where routes entered or crossed the network. Match these tags when selecting or filtering routes received from us. > **Community Format:** `203314:XXX:YY` where `XXX` is the category and `YY` is the specific value. ## Route Origin The following tags identify the relationship through which a route was learned: | Community | Description | Unique | | ---------------- | ------------------------------------ | ------ | | `203314:110:10` | Route learned from Upstream | Yes | | `203314:110:20` | Route learned from Downstream | Yes | | `203314:110:30` | Route learned from Peering | Yes | | `203314:110:110` | Route originated within Hats Network | Yes | **Unique** means this route is only announced via this method-if you see `203314:110:30`, the route is exclusively available via peering. ## PoP-Based Communities PoP tags distinguish the entry point from a transit location: ### Learned Routes learned **at** a specific PoP (entry point): | Community | Description | Unique | | ---------------- | -------------------------- | ------ | | `203314:120:XXX` | Route learned at PoP `XXX` | Yes | **Example**: `203314:120:101` means the route was learned at Hong Kong (HKG1). ### Passed Routes that **passed through** a PoP (transit): | Community | Description | Unique | | ---------------- | ------------------------------ | ------ | | `203314:125:XXX` | Route passed through PoP `XXX` | No | **Example**: `203314:125:101` means the route passed through Hong Kong (HKG1). ## Region-Based Communities Region tags distinguish the learning region from a region traversed: ### Learned Routes learned **within** a specific region: | Community | Description | Unique | | --------------- | ---------------------------- | ------ | | `203314:130:XX` | Route learned in Region `XX` | Yes | **Example**: `203314:130:100` means the route was learned in Asia (West). ### Passed Routes that **passed through** a region: | Community | Description | Unique | | --------------- | -------------------------------- | ------ | | `203314:135:XX` | Route passed through Region `XX` | No | **Example**: `203314:135:100` means the route passed through Asia (West). ## Usage Examples ### Identify Route Source Determine if a route comes from peering, upstream, or our customers: ```bird2 # These large communities are attached by AS203314 on routes you receive. # You typically match them in your import policy: filter import_prefer_peering_v4 { if (203314, 110, 30) ~ bgp_large_community then bgp_local_pref = 200; # prefer peering-learned routes if (203314, 110, 10) ~ bgp_large_community then bgp_local_pref = 80; # de-prioritize upstream-learned routes accept; } ``` ### Identify Entry Location Determine where a route entered our network: ```bird2 filter import_prefer_asia_pops_v4 { if (203314, 120, 101) ~ bgp_large_community then bgp_local_pref = 200; # entered at HKG1 (Hong Kong) if (203314, 120, 131) ~ bgp_large_community then bgp_local_pref = 180; # entered at TYO2 (Tokyo) accept; } ``` ### Filter by Region Accept routes only from specific regions: ```bird2 # Only accept routes learned in Asia West filter import_only_asia_west_v4 { if (203314, 130, 100) ~ bgp_large_community then accept; reject; } # Reject routes that passed through Europe (both West=200 and East=250) filter import_reject_europe_transit_v4 { if (203314, 135, 200) ~ bgp_large_community then reject; if (203314, 135, 250) ~ bgp_large_community then reject; accept; } ``` ## Policy Examples (JunOS / BIRD2) Internal tags are BGP Large Communities. Import policies can match them to set local preference or filter received routes. ### Junos **Prefer routes learned in Asia West (`203314:130:100`)** ```bash set policy-options community HATS-REGION-ASIA-W members large:203314:130:100 set policy-options policy-statement IMPORT-FROM-HATS term ASIA-W from community HATS-REGION-ASIA-W set policy-options policy-statement IMPORT-FROM-HATS term ASIA-W then local-preference 200 set policy-options policy-statement IMPORT-FROM-HATS term ASIA-W then accept ``` **Reject routes learned from Upstream (`203314:110:10`)** ```bash set policy-options community HATS-ORIGIN-UPSTREAM members large:203314:110:10 set policy-options policy-statement IMPORT-FROM-HATS term NO-UPSTREAM from community HATS-ORIGIN-UPSTREAM set policy-options policy-statement IMPORT-FROM-HATS term NO-UPSTREAM then reject ``` ### Bird2 **Prefer routes learned in Asia West (`203314:130:100`)** ```bird2 filter import_from_hats_v4 { if (203314, 130, 100) ~ bgp_large_community then { bgp_local_pref = 200; accept; } accept; } ``` **Reject routes learned from Upstream (`203314:110:10`)** ```bird2 filter import_from_hats_v4 { if (203314, 110, 10) ~ bgp_large_community then reject; accept; } ``` ## Related Information * **[PoP Codes](/docs/community/pop-codes)** - Reference for PoP numbers (XXX) * **[Region Codes](/docs/community/region-codes)** - Reference for region numbers (XX) * **[Control Communities](/docs/community/control-communities)** - Traffic engineering communities > **Tip**: Internal communities are automatically attached to routes we announce to you. Use them to make informed routing decisions based on route origin and location. --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/community/internal-communities). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # PoP Codes PoP identifiers occupy the final field of location communities. For example, `203314:120:XXX` records a learning location and `203314:202:XXX` excludes a location from export. > **Usage**: These codes identify specific PoP locations in internal communities (`120:XXX`, `125:XXX`) and control communities (`202:XXX`). ## PoP Code Reference | Location | IATA | Generation | Type | PoP Code | Notes | | --------- | ---- | ---------- | ---- | -------- | ------ | | Hong Kong | HKG1 | Gen2 | Core | 101 | Unique | | Hong Kong | HKG2 | Gen2 | Core | 102 | Unique | | Hong Kong | HKG3 | Gen2 | Core | 103 | | | Taiwan | TPE1 | Gen2 | Core | 121 | | | Taiwan | TPE2 | Gen2 | Core | 122 | | | Tokyo | TYO1 | Gen2 | Edge | 111 | | | Tokyo | TYO2 | Gen2 | Core | 131 | Unique | | Singapore | SIN1 | Gen2 | Core | 151 | Unique | | Seattle | SEA | Gen2 | Core | 301 | | | Ashburn | IAD | Gen2 | Core | 302 | | | San Jose | SJC | Gen2 | Edge | 303 | | | Toronto | YYZ1 | Gen2 | Edge | 304 | Unique | | Moscow | MOW | Gen2 | Core | 201 | | | Frankfurt | FRA | Gen2 | Edge | 252 | | | Zurich | ZRH | Gen2 | Edge | 251 | | | Calais | CQF1 | Gen2 | Edge | 253 | Unique | | Berlin | BER | Gen2 | Edge | 254 | Unique | | London | LON | Gen2 | Edge | 255 | Unique | | Melbourne | MEL | Gen2 | Edge | 451 | | ## Type Definitions * **Core PoP**: Major interconnection points with multiple upstream providers * **Edge PoP**: Smaller presence focused on local connectivity ## Generation Definitions * **Gen2**: Second-generation infrastructure (current deployment) ## Using PoP Codes ### In Internal Communities Identify where a route was learned: ```bird2 # Learned-at tags (set by AS203314 on routes you receive): # (203314, 120, 101) -> HKG1 (Hong Kong) # (203314, 120, 131) -> TYO2 (Tokyo) # (203314, 120, 302) -> IAD (Ashburn) filter prefer_hkg1_routes_v4 { if (203314, 120, 101) ~ bgp_large_community then bgp_local_pref = 200; accept; } ``` ### In Control Communities Exclude specific PoPs from receiving announcements: ```bird2 bgp_large_community.add((203314, 202, 301)); # Do not announce via Seattle bgp_large_community.add((203314, 202, 131)); # Do not announce via Tokyo bgp_large_community.add((203314, 202, 102)); # Do not announce via HKG2 ``` ### Configuration Snippets (JunOS / BIRD2) To exclude a PoP, set `203314:202:XXX` with its location code. Calais uses `253`, giving `203314:202:253`. ### Junos ```bash set policy-options community HATS-NO-CQF members large:203314:202:253 set policy-options policy-statement EXPORT-TO-HATS term NO-CQF from route-filter 203.0.113.0/24 exact set policy-options policy-statement EXPORT-TO-HATS term NO-CQF then community add HATS-NO-CQF set policy-options policy-statement EXPORT-TO-HATS term NO-CQF then accept set policy-options policy-statement EXPORT-TO-HATS term REJECT-ALL then reject ``` ### Bird2 ```bird2 filter export_to_hats_v4 { if net = 203.0.113.0/24 then { bgp_large_community.add((203314, 202, 253)); # no CQF accept; } reject; # Safety default } ``` ## Related Information * **[Region Codes](/docs/community/region-codes)** - Geographic region groupings * **[Internal Communities](/docs/community/internal-communities)** - Using PoP codes in location-based communities * **[Control Communities](/docs/community/control-communities)** - Using PoP codes in traffic engineering > **Tip**: PoP codes are primarily used for granular traffic engineering when you need to control route propagation at the location level. --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/community/pop-codes). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Region Codes Region communities use a two-digit geographic code. `203314:130:XX` identifies a learning region; `203314:201:XX` requests that a region be excluded from export. > **Usage**: These codes identify geographic regions in internal communities (`130:XX`, `135:XX`) and control communities (`201:XX`). ## Region Code Reference | Region Name | Area | Code | Notes | | ----------- | ----- | ---- | ---------------------------- | | Asia | West | 100 | Hong Kong, Taiwan | | Asia | East | 150 | Japan, Singapore | | Europe | West | 200 | UK, Ireland, France | | Europe | East | 250 | Germany, Netherlands, Russia | | Americas | North | 300 | USA, Canada | | Americas | South | 350 | Brazil, Argentina | | Australia | West | 400 | Perth | | Australia | East | 450 | Melbourne, Sydney | | Antarctica | - | 500 | Research networks | ## Regional Groupings ### Asia (100, 150) **Asia West (100)**: Hong Kong, Taiwan, mainland China **Asia East (150)**: Japan, Korea, Singapore, Southeast Asia ### Europe (200, 250) **Europe West (200)**: UK, Ireland, France, Benelux, Iberia **Europe East (250)**: Germany, DACH, Netherlands, Nordics, Eastern Europe ### Americas (300, 350) **North America (300)**: United States, Canada, Mexico **South America (350)**: Brazil, Argentina, Chile, Colombia ### Oceania (400, 450) **Australia West (400)**: Western Australia **Australia East (450)**: Eastern states (NSW, VIC, QLD) ## Using Region Codes ### In Internal Communities Identify where a route was learned: ```bird2 # Learned-in tags (set by AS203314 on routes you receive): # (203314, 130, 100) -> Asia West # (203314, 130, 150) -> Asia East # (203314, 130, 300) -> North America filter prefer_asia_west_routes_v4 { if (203314, 130, 100) ~ bgp_large_community then bgp_local_pref = 200; accept; } ``` ### In Control Communities Control geographic propagation: ```bird2 bgp_large_community.add((203314, 2, 1)); # Disable cross-region bgp_large_community.add((203314, 201, 300)); # Exclude North America bgp_large_community.add((203314, 201, 200)); # Exclude Europe ``` ### Configuration Snippets (JunOS / BIRD2) The examples below match a learning-region tag or attach an export exclusion. Replace the example prefixes with your authorized announcements. ### Junos **Exclude North America (`203314:201:300`)** ```bash set policy-options community HATS-NO-NA members large:203314:201:300 set policy-options policy-statement EXPORT-TO-HATS term NO-NA from route-filter 203.0.113.0/24 exact set policy-options policy-statement EXPORT-TO-HATS term NO-NA then community add HATS-NO-NA set policy-options policy-statement EXPORT-TO-HATS term NO-NA then accept set policy-options policy-statement EXPORT-TO-HATS term REJECT-ALL then reject ``` **Prefer routes learned in Asia West (`203314:130:100`)** ```bash set policy-options community HATS-ASIA-W members large:203314:130:100 set policy-options policy-statement IMPORT-FROM-HATS term ASIA-W from community HATS-ASIA-W set policy-options policy-statement IMPORT-FROM-HATS term ASIA-W then local-preference 200 set policy-options policy-statement IMPORT-FROM-HATS term ASIA-W then accept ``` ### Bird2 **Exclude North America (`203314:201:300`)** ```bird2 filter export_to_hats_v4 { if net = 203.0.113.0/24 then { bgp_large_community.add((203314, 201, 300)); accept; } reject; # Safety default } ``` **Prefer routes learned in Asia West (`203314:130:100`)** ```bird2 filter import_from_hats_v4 { if (203314, 130, 100) ~ bgp_large_community then bgp_local_pref = 200; accept; } ``` ### Regional Filtering Examples **Accept only Asian routes**: ```text # Allow only Asia West and Asia East if (community matches 203314:130:1*) then accept else reject ``` **Prefer local region**: ```text # Higher local preference for Asia West if (community matches 203314:130:100) then localpref 200 if (community matches 203314:130:150) then localpref 150 if (community matches 203314:130:300) then localpref 100 ``` ## Related Information * **[PoP Codes](/docs/community/pop-codes)** - Specific location codes within regions * **[Internal Communities](/docs/community/internal-communities)** - Using region codes in location-based communities * **[Control Communities](/docs/community/control-communities)** - Using region codes in traffic engineering > **Tip**: Region codes are ideal for high-level geographic traffic engineering when you don't need PoP-level granularity. --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/community/region-codes). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Content and Data License **Effective date:** July 12, 2026 Selected Hats Network public data and original documentation are available under the Creative Commons licenses below. The grants cover only material for which Hats Network Inc. holds the necessary rights. ## Public Latency Data (CC BY 4.0) The public **Hats Network Backbone Latency Dataset** is licensed under [Creative Commons Attribution 4.0 International](https://creativecommons.org/licenses/by/4.0/). It includes published RTT measurements, PoP codes, city and region labels, update dates and public exports derived from those measurements. Sharing and adaptation, including commercial use, are permitted with appropriate credit, links to the license and source dataset, and an indication of any changes. This dataset includes published RTT measurements, PoP codes, city and region labels, update dates, and public exports derived from those measurements. You may share and adapt this data, including commercially, provided that you give appropriate credit, link to the license and source dataset, and indicate whether changes were made. Suggested attribution: > Hats Network Backbone Latency Dataset © Hats Network Inc., > [https://hatsnet.io/docs/network/latency](https://hatsnet.io/docs/network/latency), licensed under CC BY 4.0. ## Original Documentation (CC BY-SA 4.0) Unless a page states otherwise, original explanatory articles in Hats Network Documentation are licensed under [Creative Commons Attribution-ShareAlike 4.0 International](https://creativecommons.org/licenses/by-sa/4.0/). Sharing and adaptation, including commercial use, require appropriate credit, a license link and an indication of changes. Adaptations must use the same or a compatible license. Suggested attribution: > “\[Article title],” Hats Network Documentation © Hats Network Inc., \[original article URL], > licensed under CC BY-SA 4.0. ## Exclusions The Creative Commons grants exclude: * Hats Network names, trademarks, service marks, logos, or other brand assets; * software, source code, configuration, APIs, or infrastructure designs; * third-party text, images, data, trademarks, or other material; * customer information, private telemetry, raw operational logs, credentials, or security data; * network services, service-level commitments, commercial offers, or contractual rights. Third-party material retains its original terms. Neither license permits an implication that Hats Network Inc. endorses or is affiliated with the user. ## No Warranty Content and data are provided "as is", without warranties of accuracy, completeness, availability, fitness for a particular purpose or non-infringement. To the maximum extent permitted by law, Hats Network Inc. is not liable for losses arising from their use. Licensing enquiries: [legal@hatsnet.io](mailto:legal@hatsnet.io). --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/legal/content-and-data-license). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Legal These policies govern Hats Network's website, documentation, open data and services. Use the service-specific terms alongside the general terms below. - [Privacy Policy](/docs/legal/privacy-policy): How we collect, use, and protect personal information. - [Terms of Service](/docs/legal/terms-of-service): The terms that govern use of our website and services. - [Content & Data License](/docs/legal/content-and-data-license): Licensing for documentation content (CC BY-SA 4.0) and open data (CC BY 4.0). - [Report Abuse](/docs/legal/report-abuse): How to report abuse originating from our network. - [IP Rental](/docs/legal/ip-rental): Terms specific to IP address rental services. --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/legal). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # IP Address Rental Terms > **Effective Date:** June 1, 2024\ > **Last Revised:** August 24, 2024 ## 1. Service Description Hats Network Inc. rents IP addresses subject to these terms. Lessees must comply with applicable laws and regulations. > **What is IP Rental?** > > The service leases IPv4 or IPv6 addresses for the lessee's infrastructure. Typical uses include: > > * Hosting providers requiring additional IP space > * Research and development projects > * Temporary network expansions ## 2. Responsibilities and Penalties The lessee is responsible for all activity associated with the rented addresses. ### General Complaints > **Penalty:** $5 per incident, maximum $20 per billing period Including but not limited to: * DMCA copyright complaints * Spam complaints (unsolicited email) * Unauthorized marketing activity complaints * Minor service abuse complaints ### Severe Misconduct > **Penalty:** $15 per incident, maximum $30 per billing period > > Severe misconduct may result in heavier penalties and immediate suspension. Including but not limited to: * Network scanning activities (port scanning, vulnerability scanning) * Spreading malware or viruses * Participating in network attacks (DDoS, brute force) * Unauthorized system intrusion attempts * Phishing activities * Hosting illegal content ### IP Blacklisting > **Penalty:** \$30 per incident OR 6-month service extension The following are examples of major blacklists covered by this provision: | Blacklist | Type | | --------------------- | ----------------------------------- | | Spamhaus | Email reputation | | DNSBL | DNS-based blacklist | | SORBS | Spam and Open Relay Blocking System | | SpamCop | Spam reporting | | Cisco Talos | Security intelligence | | Microsoft SmartScreen | Email filtering | | Google Safe Browsing | Web safety | ## 3. Payment and Penalty Enforcement > **Important:** All penalty fees will be added to the lessee's next invoice. Penalties take effect immediately upon notification. Failure to pay may result in service suspension or termination. ### Payment Terms * Invoices are issued monthly * Payment is due within 14 days of invoice date * Late payments subject to 1.5% monthly service charge * All fees are non-refundable unless otherwise stated ## 4. Service Termination > **Termination Policy** > > Hats Network Inc. may terminate service **without refund** for severe or repeated violations. ### Severe Violations Include: * Multiple occurrences of severe misconduct * Repeated IP blacklisting * Engagement in illegal activities * Non-payment of fees for 30+ days ### Voluntary Termination: A lessee may terminate service with 30 days' written notice. Prepaid unused periods are not refundable. ## 5. Disclaimer Hats Network Inc. is not responsible for losses or damages arising from the lessee's use of rented addresses. The lessee agrees to indemnify and hold the company harmless from claims arising from that use. ## 6. Monitoring and Reporting > Hats Network Inc. may monitor address usage for compliance with these terms. Lessees must promptly report abuse or unauthorized use. Monitoring covers: * Unusual traffic patterns * Blacklist status changes * Abuse reports from third parties * Compliance with AUP (Acceptable Use Policy) ## 7. Amendment of Terms Hats Network Inc. may modify these terms at any time. Modified terms will be: 1. Posted on the company website 2. Emailed to the lessee's registered contact 3. Effective 30 days after posting **Continued use implies acceptance of modified terms.** ## 8. Notice of Changes Changes may be made without prior notice in the following cases: * Emergency security issues * Legal or regulatory requirements * Force majeure events ## Questions? - [Contact Sales](mailto:sales@hatsnet.io): For questions about IP rental availability and pricing. - [Contact Support](mailto:support@hatsnet.io): For abuse reports or compliance questions. *** > *By using this service, the lessee agrees to abide by these terms.* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/legal/ip-rental). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Privacy Policy > **Effective Date:** March 24, 2026\ > **Last Revised:** March 24, 2026 ## Introduction This policy applies to information handled by Hats Network Inc. (AS203314) through its website and services. It describes the information we collect, its uses, permitted disclosures and the measures used to protect it. > Read the policy before using the services. Do not use them if you do not agree with these practices. ## Information We Collect ### Information You Provide Directly * Contact information (name, email, phone, company name) via contact forms * Account information when you register for services * Payment information processed securely through third-party providers * Communication history and support inquiries ### Information Collected Automatically * IP address and device information * Browser type and operating system * Pages visited, time spent, and referring URLs * Network performance metrics and latency measurements * Cookies and similar tracking technologies ## How We Use Your Information We use collected information for these purposes: * Provide, maintain, and improve our services * Process transactions and send related information * Send technical notices and support messages * Respond to inquiries and customer support requests * Monitor and analyze usage patterns and service performance * Detect, prevent, and address technical issues and fraud * Comply with legal obligations and enforce agreements * Send marketing communications (with your consent) ## Information Sharing > We do not sell personal information to third parties. Information may be shared with: * **Service providers** - Hosting, payment processing, analytics * **Business partners** - With your consent * **Law enforcement** - When required by law * **Other parties** - In connection with company transactions ## Data Security We use appropriate technical and organizational measures against unauthorized access, alteration, disclosure and destruction. Internet transmission and electronic storage cannot be made completely secure. We do not guarantee absolute security. > Keep your account credentials confidential; this remains your responsibility. ## Cookies and Tracking Cookies support site functions and usage analysis. Browser settings allow you to control cookies. Third-party analytics providers may collect usage data to help us understand site traffic. ## Third-Party Links External sites linked from our website have their own privacy practices. This policy does not govern those sites, and we are not responsible for their practices. ## Children's Privacy > The services are not directed to people under 18. We do not knowingly collect personal information from children under 18. ## Your Privacy Rights Your jurisdiction may give you rights to: * **Access** - Request a copy of your personal data * **Correction** - Update inaccurate or incomplete information * **Deletion** - Request deletion of your personal data (subject to legal obligations) * **Portability** - Receive your data in a structured, commonly used format * **Opt-out** - Decline marketing communications Send rights requests to [privacy@hatsnet.io](mailto:privacy@hatsnet.io). ## Data Retention Information is retained for as long as needed to: * Provide services to you * Comply with legal obligations * Resolve disputes and enforce agreements * Maintain business records After the retention purpose ends, we securely delete or anonymize the information. ## Changes to Privacy Policy > We may revise this policy. Changes take effect 30 days after publication on our website. > > **Continued use of our services implies acceptance of the updated policy.** ## Contact Us - [Privacy Questions](mailto:privacy@hatsnet.io): For privacy-related inquiries. - [Account & Security](mailto:support@hatsnet.io): For account security concerns. *** *By using Hats Network Inc. services, you acknowledge that you have read and understood this Privacy Policy.* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/legal/privacy-policy). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Report Abuse ## How to Report Abuse Report abusive activity involving Hats Network infrastructure to the Security Team: - [Email Report](mailto:abuse@hatsnet.io): Send abuse reports with detailed information to our Security Team. > The Security Team investigates reports. Privacy restrictions may prevent us from sharing case-specific updates. ## Required Information An abuse report **must include**: Source IP address(es) The IP address(es) originating the abusive activity. Destination IP address(es) The IP address(es) being targeted or receiving the abuse. Destination port(s) Specific port(s) involved in the abusive activity. Exact date/time stamp and timezone Precise timing of the activity (e.g., 2026-03-24 15:30:45 UTC). > **Important Notice** > > *Incomplete reports may not be processed.* If your abuse report does not contain enough information to identify the source of the abuse, we may not follow up with you regarding your report(s). ## Important Notice ### Blacklisted Report Sources Reports from the following sources are **automatically ignored** because of their history of false reports: | Source | Status | | ------------- | ----------- | | bitninja.info | Blacklisted | | myipr.com.cn | Blacklisted | | aldimna.com | Blacklisted | ### Ignored Email Domains > **Note** > > Use a **work email** address where possible so the report can be properly investigated. Reports sent from the following email domains will not be processed: | Domain | Status | | ------------------- | ----------- | | @tsinghua.edu.org | Ignored | | @163.com / @126.com | Gray-listed | | @qq.com | Gray-listed | | @sina.com | Gray-listed | | @mail.ru | Gray-listed | ## What Constitutes Abuse? Reportable activity includes: * **DDoS Attacks** - Distributed Denial of Service attacks * **Malware Distribution** - Hosting or spreading malicious software * **Phishing** - Unauthorized attempts to collect sensitive information * **Network Scanning** - Unauthorized port scans or vulnerability assessments * **Spam** - Unsolicited bulk communications * **Copyright Infringement** - Unauthorized use of intellectual property * **Illegal Content** - Hosting of illegal material * **Intrusion Attempts** - Unauthorized access to systems ## Security Best Practices When preparing a report: 1. **Include Evidence** - Packet captures, logs, or screenshots when possible 2. **Be Specific** - Provide exact timestamps and port information 3. **One Report Per Issue** - Submit separate reports for different incidents 4. **Follow Up** - Reference your original email if sending additional information ## Investigation Timeline | Report Type | Processing Time | Notes | | -------------- | --------------- | ------------------------------------ | | Standard Email | 24-48 hours | Reviewed during business hours (UTC) | | Urgent Cases | 4 hours | Critical infrastructure threats only | > Reports are treated confidentially. Do not disclose publicly that you have filed a report with Hats Network, as doing so may compromise the investigation. *** Retain the original evidence in case the Security Team needs further details. --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/legal/report-abuse). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Terms of Service > **Effective Date:** March 24, 2026\ > **Last Revised:** July 12, 2026 ## Agreement to Terms Accessing or using Hats Network Inc. (AS203314) websites and services constitutes acceptance of this agreement. > Do not use the services if you do not accept these terms. We may modify these terms at any time. Continued use after revised terms are posted constitutes acceptance of the changes. ## Service Description The services covered by this agreement include: * IP transit with redundant upstream connections * Direct Internet Exchange (IX) peering * Low-latency routing with specialized network optimization * Network monitoring and DDoS mitigation * Custom network solutions for enterprise clients ## Acceptable Use Policy ### Prohibited Activities The following activities are prohibited: ### Common * Harassment, abuse, or threats toward individuals or organizations - Transmission of malware, viruses, or harmful code - Spam or unsolicited bulk email transmission - Copyright infringement or intellectual property violation - Phishing, spoofing, or social engineering attacks ### Severe * Network scanning, port scanning, or vulnerability assessment without authorization - Distributed Denial of Service (DDoS) attacks or similar abuse - Unauthorized access or intrusion attempts - Hosting of illegal content or facilitating illegal services - Any illegal activities or violation of laws and regulations ### Violation Penalties | Violation Type | Penalty | Consequence | | ------------------- | ----------------- | ------------------------- | | First violation | Written warning | $5-$15 penalty fee | | Repeated violations | Multiple warnings | Service suspension option | | Severe violation | Immediate action | Service termination | | Criminal activity | Legal reporting | Law enforcement involved | > **Important** > > Hats Network Inc. may terminate services **without refund** for severe or repeated violations. ## Payment Terms > **Billing Cycle:** Monthly * Invoices issued on the first day of each billing period * Payment due within 14 days of invoice date * Accepted payment methods: credit card, bank transfer, ACH * Late payments subject to 1.5% monthly service charge * Service suspension after 30 days of non-payment * All fees are exclusive of applicable taxes and levies * Refund policy: No refunds for partial months or prepaid periods ## Service Levels & Uptime Hats Network Inc. commits to **99.99% network availability** across its core infrastructure. ### SLA Details * Uptime calculated monthly * Excludes scheduled maintenance windows * Excludes force majeure events * Scheduled maintenance announced ≥14 days in advance when possible * Service credits available for downtime exceeding SLA (details in service agreement) * Network performance statistics available via portal and API ## Limitation of Liability > TO THE MAXIMUM EXTENT PERMITTED BY LAW, HATS NETWORK INC. > > SHALL NOT BE LIABLE FOR ANY INDIRECT, INCIDENTAL, SPECIAL, CONSEQUENTIAL, OR PUNITIVE DAMAGES. Liability is limited to the amount the customer paid in the 12 months before the claim. ### We are not responsible for: * Third-party network issues or connectivity problems * Data loss or corruption caused by misuse or negligence * Service interruptions due to force majeure events * Loss of revenue, profit, or business opportunities * Customer's failure to maintain backups or redundancy ## Service Termination ### Either Party May Terminate * 30 days written notice required * Customer must migrate services within 30 days * Customer must remove any IPs in use within 30 days * No refunds provided for customer-initiated termination ### Immediate Termination Permitted For: * Material breach of this agreement by either party * Illegal activity or security violations * Non-payment of invoices for 30+ days * Violation of Acceptable Use Policy ## Additional Terms ### Intellectual Property and Open Licenses Unless an express license states otherwise, Hats Network Inc. retains all rights in its services, software, branding and materials. The [Content and Data License](/docs/legal/content-and-data-license) sets out the separate terms for public latency data and original documentation. - **Intellectual Property**: Openly licensed materials follow our Content and Data License; all other rights are reserved. - **Warranties Disclaimer**: Services provided 'as-is' without warranties. We don't warrant uninterrupted service. - **Indemnification**: You indemnify us from claims arising from service use or terms violations. - **Governing Law**: Governed by applicable laws. Disputes resolved through binding arbitration. - **Entire Agreement**: This agreement and service amendments constitute the entire agreement. - **Severability**: If any provision is invalid, remaining provisions continue in full effect. ## Questions? - [Legal Questions](mailto:legal@hatsnet.io): For legal inquiries and interpretations. - [Abuse Reports](mailto:abuse@hatsnet.io): Report violations and suspicious activity. - [Technical Support](mailto:support@hatsnet.io): For service and technical support. *** > *By using Hats Network Inc. services, you acknowledge that you have read, understood, and agree to be bound by these Terms of Service.* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/legal/terms-of-service). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Peering Policy - AS203314 > **Open Peering Policy** > > AS203314 has an **open peering policy**. Requests are assessed against the requirements below, available capacity, route security, operational readiness and mutual agreement. > > Our [PeeringDB record](https://www.peeringdb.com/asn/203314) is the authoritative reference for interconnection locations, session addresses and operational peering statistics. ## Policy at a Glance | Item | Policy | | --------------------- | ---------------------------------------------------------------------------------------- | | General policy | Open | | Autonomous system | AS203314 | | IRR AS-SET | `AS203314:AS-HATS` | | Address families | IPv4 and IPv6; dual-stack sessions are preferred where both are available | | Traffic ratio | No fixed ratio requirement | | Minimum IX traffic | None | | Multiple locations | Preferred for resilience | | Bilateral IX contract | Not normally required | | Operational source | [PeeringDB](https://www.peeringdb.com/asn/203314) for current IX, facility, and NOC data | ## Technical Requirements A peering request must meet these requirements: 1. Operate a publicly routable **Autonomous System Number (ASN)**. 2. Announce at least one globally routable **IPv4 /24** or **IPv6 /48** prefix. 3. Maintain a complete and current [PeeringDB](https://www.peeringdb.com/) profile, including routing, facility, and 24×7 NOC information. 4. Maintain accurate route or route-set objects in a public **Internet Routing Registry (IRR)**. 5. Connect at a shared Internet Exchange or data centre, or use a mutually approved [Layer 2/3 tunnel](/docs/peering/via-tunnel) endpoint. 6. Provide sufficient, stable interconnection capacity and avoid sustained congestion or packet loss. We prefer dual-stack sessions where both networks support IPv4 and IPv6. A single-stack session may be accepted when the other address family is unavailable at that location. ## Routing and Security Policy ### Route registration and validation * Peers must keep their IRR objects, RPKI Route Origin Authorizations (ROAs), and PeeringDB data accurate and up to date. * AS203314 builds inbound filters from public routing data and may reject unregistered routes, RPKI-invalid announcements, bogons, default routes, malformed AS paths, and prefixes more specific than IPv4 /24 or IPv6 /48. * Peers should announce only their own routes and customer routes represented by their published AS-SET. * Route-security practices should align with [MANRS](https://www.manrs.org/) principles. RPKI-valid announcements are strongly preferred; RPKI-invalid announcements are rejected. ### Route exchange * AS203314 generally advertises its eligible routes in all peering locations. Announcements may vary for maintenance, traffic engineering, security, or service-specific reasons. * We accept standard BGP communities where operationally supported. See our [BGP Communities reference](/docs/community) for available controls. * Peers must configure a reasonable maximum-prefix limit based on current PeeringDB and IRR data, with enough headroom for normal growth. Material changes should be coordinated with our NOC. * Do not rely on Multi-Exit Discriminator (MED) handling unless it has been agreed in advance. * Neither party may use a static route, a default route, or any other mechanism to send traffic for destinations not announced to that party through BGP. * Traffic may not be forwarded through a third party over the peering interconnection unless both networks have explicitly agreed to that service. > **Filtering and session protection** > > We may suspend a session for route leaks, exceeded prefix limits, instability or other security and operational risks. We will make a reasonable effort to notify the peer's registered NOC contact. ## Operational Requirements Peers must maintain a [**24×7 NOC**](/docs/about#contact-us) that can work with us on routing incidents, performance degradation, denial-of-service attacks, abuse and security events. * NOC email and telephone details must remain current on PeeringDB. * Both parties should communicate planned maintenance that may materially affect the interconnection. * Peers must investigate route leaks, session instability, abuse, or congestion promptly. * BGP MD5 and BFD may be enabled where supported and mutually agreed. * Operational issues should be reported to [noc@hatsnet.io](mailto:noc@hatsnet.io); new peering requests should be sent to [peering@hatsnet.io](mailto:peering@hatsnet.io). ## Capacity and Traffic * There is **no minimum traffic requirement** for peering over an Internet Exchange. * A [Private Network Interconnect (PNI)](/docs/peering/via-cross-connect) may be preferred when traffic is sustained above **1 Gbps**, when an IX path is congested, or when resilience and traffic engineering justify a dedicated interconnection. * PNI speed, optics, cross-connect, and redundancy are agreed per location. 10G and 100G single-mode fibre are preferred where available. * Peers should begin capacity-upgrade planning when sustained peak utilization exceeds **50%**, and complete the upgrade before congestion affects traffic. ## How to Peer ### Step 1: Verify eligibility Check the requirements and our [PeeringDB record](https://www.peeringdb.com/asn/203314) for a common location, session addresses and expected route counts. ### Step 2: Send a request Send [peering@hatsnet.io](mailto:peering@hatsnet.io) your ASN, PeeringDB URL, requested location and address families, expected traffic, prefix count and interconnection method. ### Step 3: Configure and validate Once session details are agreed, configure route filters, maximum-prefix protection and any agreed MD5 or BFD settings. We check route exchange and reachability before marking the session operational. ## Peering Methods - [Internet Exchange](https://www.peeringdb.com/asn/203314): Public peering at our current Internet Exchange locations. - [Private Interconnect (PNI)](/docs/peering/via-cross-connect): Dedicated cross-connect or private VLAN at a supported facility. - [Tunnel Peering](/docs/peering/via-tunnel): A mutually approved Layer 2 or Layer 3 tunnel interconnection. ## Policy Administration Meeting these requirements does not guarantee peering. Hats Network may accept, decline, suspend or terminate a relationship for capacity, security, commercial, legal or operational reasons. The policy may change; where practical, material session changes will be coordinated through registered NOC contacts. --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/peering). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Peering via Cross Connect (PNI) ## Supported Data Centres The following facilities are listed for Private Network Interconnect (PNI). Confirm space, port capacity and cross-connect details with the peering team before ordering. ### 🇭🇰 Hong Kong * [Equinix HK1](https://www.equinix.com/data-centers/asia-pacific-colocation/hong-kong-colocation/hong-kong-data-centers/hk1) * [Equinix HK2](https://www.equinix.com/data-centers/asia-pacific-colocation/hong-kong-colocation/hong-kong-data-centers/hk2) * [Equinix HK3](https://www.equinix.com/data-centers/asia-pacific-colocation/hong-kong-colocation/hong-kong-data-centers/hk3) * [TGT Hong Kong DC 2](https://www.towngastelecom.com/locations/) ### 🇯🇵 Japan * [Equinix TY8](https://www.equinix.com/locations/asia-colocation/japan-colocation/tokyo-data-centers/ty8/) * [Equinix TY9](https://www.equinix.com/locations/asia-colocation/japan-colocation/tokyo-data-centers/ty9/) ### 🇸🇬 Singapore * [Equinix SG1](https://www.equinix.com/locations/singapore-colocation/singapore-data-center/sg1/) ### 🇺🇸 United States * [CoreSite LA2](https://www.coresite.com/data-center/la2-los-angeles-ca) > To arrange a cross-connect, please [contact our peering team](/docs/about#contact-us). --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/peering/via-cross-connect). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Peering via Layer 2 Tunnel A Layer 2 tunnel carries Ethernet frames between endpoints. Use it for VLAN-tagged traffic or a shared broadcast domain; choose Layer 3 when only routed IP connectivity is needed. > **Layer 2 Tunnel Overview** > > Replace every placeholder with the agreed endpoint addresses, interface names and tunnel identifiers before applying a configuration. > > * `{name}` - Tunnel interface name > * `{yourside ip}` - Your public IP address > * `{ourside ip}` - Our endpoint IP address > * `{your tunnel ip cidr}` - Your tunnel IP/subnet > * `{vni}` - VxLAN Network Identifier (24-bit value) ## Protocol Choice ## GRETAP **GRETAP** (GRE Bridging) operates at Layer 2, encapsulating Ethernet frames inside GRE. Use this when you need a transparent L2 bridge over the tunnel. ### Shell ```shell ip link add {name} type gretap local {yourside ip} remote {ourside ip} ttl 255 ip addr add {your tunnel ip cidr} dev {name} ip link set dev {name} up ``` ### Netplan 1. Create a Netplan configuration file: ```yaml title="/etc/netplan/10-{name}.yaml" network: version: 2 tunnels: { name }: mode: gretap local: { yourside ip } remote: { ourside ip } ttl: 255 addresses: - { your tunnel ip cidr } ``` 2. Apply the configuration: ```sh netplan apply ``` ### Systemd 1. Create the `.netdev` file: ```ini title="/etc/systemd/network/10-{name}.netdev" [NetDev] Name = {name} Kind = gretap [Tunnel] Local = {yourside ip} Remote = {ourside ip} TTL = 255 ``` 2. Configure the tunnel IP address: ```ini title="/etc/systemd/network/10-{name}.network" [Match] Name = {name} [Network] Address = {your tunnel ip cidr} ``` 3. Apply the configuration: ```sh systemctl restart systemd-networkd ``` ## VxLAN **VxLAN** (Virtual Extensible LAN) is a Layer 2 overlay protocol that encapsulates Ethernet frames in UDP. It's designed for large-scale multi-tenant environments and software-defined networks. > The default VXLAN destination port is **4789**. Its 24-bit VNI identifies the virtual network and ranges from 0 to 16777215. ### Shell ```shell ip link add {name} type vxlan local {yourside ip} remote {ourside ip} dstport 4789 id {vni} ttl 255 ip addr add {your tunnel ip cidr} dev {name} ip link set dev {name} up ``` ### Netplan > The Netplan example requires **netplan ≥ 0.106** (Ubuntu 23.04 or later). 1. Create a Netplan configuration file: ```yaml title="/etc/netplan/10-{name}.yaml" network: version: 2 tunnels: { name }: mode: vxlan local: { yourside ip } remote: { ourside ip } port: 4789 id: { vni } ttl: 255 addresses: - { your tunnel ip cidr } ``` 2. Apply the configuration: ```sh netplan apply ``` ### Systemd 1. Create the `.netdev` file: ```ini title="/etc/systemd/network/10-{name}.netdev" [NetDev] Name = {name} Kind = vxlan [VXLAN] VNI = {vni} Local = {yourside ip} Remote = {ourside ip} DestinationPort = 4789 TTL = 255 ``` 2. Configure the tunnel IP address: ```ini title="/etc/systemd/network/10-{name}.network" [Match] Name = {name} [Network] Address = {your tunnel ip cidr} ``` 3. Apply the configuration: ```sh systemctl restart systemd-networkd ``` ## GRETAP vs VxLAN | Feature | GRETAP | VxLAN | | ------------------ | ---------------- | ------------------- | | Protocol | IP Protocol 47 | UDP port 4789 | | Encapsulation | GRE header | UDP + VxLAN header | | MTU overhead | 38 bytes | 50 bytes | | NAT traversal | Limited | Better (UDP-based) | | Multi-cast support | Yes | Yes | | Use case | Simple L2 bridge | Data center overlay | ## Considerations > **Important Notes for Layer 2 Tunneling** > > * **MTU**: Layer 2 tunnels add significant overhead. Reduce your MTU accordingly (typically 1450-1476 bytes). > * **Broadcast domain**: The tunnel extends your broadcast domain, which may cause issues with certain protocols. > * **Spanning Tree**: Be cautious with STP over tunnels-consider using RSTP or disabling STP on the tunnel interface. > * **Performance**: Layer 2 tunneling has more overhead than Layer 3. Use only when necessary. ## Next Steps After configuring both endpoints, verify the following: 1. Verify connectivity using `ping` or `traceroute` 2. Configure your BGP daemon (BIRD, FRR, etc.) to use the tunnel interface 3. [Contact us](/docs/about#contact-us) to finalize the peering session > **Prefer Layer 3?** For most peering scenarios, Layer 3 tunnels (WireGuard, GRE) are recommended due to lower overhead and better performance. --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/peering/via-layer2-tunnel). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Peering via Layer 3 Tunnel A Layer 3 tunnel carries IP packets between endpoints without extending an Ethernet broadcast domain. It is the usual choice for routed peering; encapsulation and encryption determine its overhead. > **Layer 3 Tunnel Overview** > > Replace the placeholder addresses, interface names and keys with the agreed configuration before applying these examples. > > * `{name}` - Tunnel interface name > * `{yourside ip}` - Your public IP address > * `{ourside ip}` - Our endpoint IP address > * `{yourside port}` - Your source port (WireGuard) > * `{ourside port}` - Our destination port (WireGuard) > * `{your tunnel ip cidr}` - Your tunnel IP/subnet > * `{our public key}` - Our WireGuard public key > * `{your private key}` - Your WireGuard private key ## Tunnel Profile Selection ## WireGuard **WireGuard** is a modern, lightweight VPN protocol that provides encrypted Layer 3 tunneling. It's our recommended choice for secure peering due to its simplicity and performance. ### Wg Quick For a wg-quick deployment, use the following WireGuard configuration: ```ini title="/etc/wireguard/{name}.conf" [Interface] Address = {your tunnel ip cidr} ListenPort = {yourside port} PrivateKey = {your private key} # Disable WireGuard's built-in routing table management when using # an external routing daemon (e.g. BIRD, FRR) Table = off [Peer] PublicKey = {our public key} AllowedIPs = 0.0.0.0/0, ::/0 Endpoint = {ourside ip}:{ourside port} PersistentKeepalive = 25 ``` ### Systemd 1. Create the `.netdev` file: ```ini title="/etc/systemd/network/10-{name}.netdev" [NetDev] Name = {name} Kind = wireguard [WireGuard] ListenPort = {yourside port} PrivateKey = {your private key} [WireGuardPeer] PublicKey = {our public key} AllowedIPs = 0.0.0.0/0, ::/0 Endpoint = {ourside ip}:{ourside port} PersistentKeepalive = 25 ``` 2. Configure the tunnel IP address: ```ini title="/etc/systemd/network/10-{name}.network" [Match] Name = {name} [Network] Address = {your tunnel ip cidr} ``` 3. Apply the configuration: ```sh systemctl restart systemd-networkd ``` ### Enable Enable the interface with wg-quick: ```sh systemctl enable --now wg-quick@{name} ``` ## GRE Tunnel **GRE** (Generic Routing Encapsulation) operates at Layer 3 and is suitable for routing IPv4/IPv6 traffic over an IPv4 underlay. It's simple, widely-supported, and has minimal overhead. ### Shell ```shell ip tunnel add {name} mode gre local {yourside ip} remote {ourside ip} ttl 255 ip addr add {your tunnel ip cidr} dev {name} ip link set dev {name} up ``` ### Netplan 1. Create a Netplan configuration file: ```yaml title="/etc/netplan/10-{name}.yaml" network: version: 2 tunnels: { name }: mode: gre local: { yourside ip } remote: { ourside ip } ttl: 255 addresses: - { your tunnel ip cidr } ``` 2. Apply the configuration: ```sh netplan apply ``` ### Systemd 1. Create the `.netdev` file: ```ini title="/etc/systemd/network/10-{name}.netdev" [NetDev] Name = {name} Kind = gre [Tunnel] Local = {yourside ip} Remote = {ourside ip} TTL = 255 ``` 2. Configure the tunnel IP address: ```ini title="/etc/systemd/network/10-{name}.network" [Match] Name = {name} [Network] Address = {your tunnel ip cidr} ``` 3. Apply the configuration: ```sh systemctl restart systemd-networkd ``` ## SIT / ip6gre (IPv6 Tunneling) **SIT** (Simple Internet Transition) tunnels IPv6 traffic over an IPv4 underlay and is commonly used for 6in4 connectivity. **ip6gre** provides full GRE encapsulation for IPv6 and is preferred when you need GRE key support or multi-protocol capability. ### Sit Shell ```shell # SIT: IPv6-in-IPv4 ip tunnel add {name} mode sit local {yourside ipv4} remote {ourside ipv4} ttl 255 ip addr add {your tunnel ipv6 cidr} dev {name} ip link set dev {name} up ``` ### Ip6gre Shell ```shell # ip6gre: GRE over IPv4 carrying IPv6 ip tunnel add {name} mode ip6gre local {yourside ipv4} remote {ourside ipv4} ttl 255 ip addr add {your tunnel ipv6 cidr} dev {name} ip link set dev {name} up ``` ### Netplan 1. Create a Netplan configuration file: ```yaml title="/etc/netplan/10-{name}.yaml" network: version: 2 tunnels: { name }: # Use 'sit' for 6in4, or 'ip6gre' for GRE-encapsulated IPv6 mode: sit local: { yourside ipv4 } remote: { ourside ipv4 } ttl: 255 addresses: - { your tunnel ipv6 cidr } ``` 2. Apply the configuration: ```sh netplan apply ``` ### Systemd 1. Create the `.netdev` file: ```ini title="/etc/systemd/network/10-{name}.netdev" [NetDev] Name = {name} # Kind = sit (for 6in4) # Kind = ip6gre (for GRE over IPv4 carrying IPv6) Kind = sit [Tunnel] Local = {yourside ipv4} Remote = {ourside ipv4} TTL = 255 ``` 2. Configure the tunnel IP address: ```ini title="/etc/systemd/network/10-{name}.network" [Match] Name = {name} [Network] Address = {your tunnel ipv6 cidr} ``` 3. Apply the configuration: ```sh systemctl restart systemd-networkd ``` ## Protocol Comparison | Protocol | Encryption | IPv4 | IPv6 | Overhead | NAT Traversal | | --------- | ---------- | ---- | ---- | -------- | ------------- | | WireGuard | Yes | ✓ | ✓ | 32 bytes | Good (UDP) | | GRE | No | ✓ | ✓ | 28 bytes | Limited | | SIT | No | N/A | ✓ | 20 bytes | Limited | | ip6gre | No | N/A | ✓ | 28 bytes | Limited | > **Protocol Selection Guide** > > * **WireGuard**: Best for secure peering, supports both IPv4 and IPv6 > * **GRE**: Simple, widely-supported, good for IPv4 peering > * **SIT/ip6gre**: Use when you only need IPv6 transport over IPv4 ## MTU Considerations Encapsulation consumes part of the underlay MTU. Set the tunnel MTU to leave room for the chosen protocol's headers. | Protocol | Overhead | Recommended MTU | | --------- | -------- | --------------- | | WireGuard | 32 bytes | 1468 | | GRE | 28 bytes | 1472 | | SIT | 20 bytes | 1480 | | ip6gre | 28 bytes | 1472 | WireGuard MTU example: ```sh ip link set dev {name} mtu 1468 ``` ## Next Steps After configuring both endpoints, verify the following: 1. Verify connectivity using `ping` or `traceroute` 2. Configure your BGP daemon (BIRD, FRR, etc.) to use the tunnel interface 3. [Contact us](/docs/about#contact-us) to finalize the peering session > Once tunnel reachability is confirmed, contact the peering team to complete the AS203314 BGP session. --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/peering/via-layer3-tunnel). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Peering via Tunnel Tunnel peering can connect AS203314 with a network that has no suitable physical interconnection. Layer 2 and Layer 3 options are documented below. > **When to Use Tunnel Peering** > > A tunnel may be appropriate when: > > * You're not co-located with us at any IXP or data center > * Physical cross-connect is not available or cost-prohibitive > * You need a temporary peering arrangement before establishing physical connectivity > * You're in a region where we don't have physical presence ## Tunnel Architecture ## Layer 2 vs Layer 3 Tunneling ### L2 **Layer 2 tunnels** encapsulate Ethernet frames, creating a virtual bridge between networks. **Protocols**: GRETAP, VxLAN **Use cases**: * Carrying VLAN-tagged traffic * Running bridging protocols (STP, LLDP) * Extending a broadcast domain **Considerations**: Higher overhead than Layer 3, not suitable for long-distance peering. [Learn more about Layer 2 tunnels →](/docs/peering/via-layer2-tunnel) ### L3 **Layer 3 tunnels** encapsulate IP packets, operating at the network layer. **Protocols**: WireGuard, GRE, SIT/ip6gre **Use cases**: * Standard BGP peering * IPv6 transport over IPv4 (6in4) * Encrypted peering (WireGuard) **Considerations**: Lower overhead, more efficient for most peering scenarios. [Learn more about Layer 3 tunnels →](/docs/peering/via-layer3-tunnel) ## Configuration Examples The examples use placeholders. Replace each one with the values agreed for your session: ```yaml # Replace these placeholders with your actual values your_name: "{tunnel name}" yourside_ip: "{your public IP}" ourside_ip: "{our endpoint IP}" yourside_port: "{your source port (WireGuard)}" ourside_port: "{our destination port (WireGuard)}" your_tunnel_ip: "{your tunnel IP address}" our_tunnel_ip: "{our tunnel IP address}" tunnel_cidr: "{tunnel subnet CIDR}" vni: "{VxLAN Network Identifier}" public_key: "{our WireGuard public key}" private_key: "{your WireGuard private key}" ``` ## Protocol Comparison | Protocol | Layer | Encryption | Multi-protocol | Use Case | | --------- | ----- | ---------- | -------------- | ---------------------- | | WireGuard | L3 | Yes | IPv4/IPv6 | Secure, modern peering | | GRE | L3 | No | IPv4/IPv6 | Simple IP tunneling | | GRETAP | L2 | No | Ethernet | Bridging, VLANs | | VxLAN | L2 | No | Ethernet | Data center overlay | | SIT | L3 | No | IPv6-over-IPv4 | 6in4 connectivity | | ip6gre | L3 | No | IPv6-over-IPv4 | GRE for IPv6 | - [Layer 2 Tunnels](/docs/peering/via-layer2-tunnel): GRETAP and VxLAN for Ethernet-layer encapsulation. - [Layer 3 Tunnels](/docs/peering/via-layer3-tunnel): WireGuard, GRE, and SIT for IP-layer encapsulation. ## Next Steps 1. Choose the appropriate tunnel type for your use case 2. Follow the configuration examples 3. Contact us to finalize the peering session > **Need help deciding?** > > Use a Layer 3 tunnel for routed peering unless the handoff requires Ethernet or VLAN transport. Compare GRE and WireGuard based on encapsulation, encryption and endpoint support. --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/peering/via-tunnel). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Hats BDL | Dedicated Internet Access **Hats BDL (Business Dedicated Line)** is enterprise-grade Dedicated Internet Access with contractual bandwidth guarantees and 99.99% SLA backing. The service order defines the handoff, coverage and remedies. > **Market Reality** > > Shared broadband shares access capacity between customers. Performance can fall when that capacity is busy. > > **Hats BDL** provides dedicated fiber access with a contracted Committed Information Rate (CIR). > > Source: [CTC Technology & Energy: All Mbps Are Not Created Equal](https://www.ctcnet.us/wp-content/uploads/2014/02/CTC-ConnectivityPerformanceFactorsBrief0213141.pdf) ## Why Hats BDL Over Standard Broadband | Capability | Business Broadband | Hats BDL | | ----------------------- | --------------------------- | ------------------------------ | | **Bandwidth Guarantee** | Best effort / shared | CIR with contractual guarantee | | **SLA Availability** | None or limited | 99.99% with financial backing | | **Latency Consistency** | Variable (peak-hour spikes) | Predictable, low jitter | | **Symmetry** | Often asymmetric | Symmetric upload/download | | **Static IPs** | Dynamic or limited | Block assignment available | | **Support Response** | Standard business hours | 24/7 with escalation paths | ## Use Cases ### Primary Uplink Office connectivity for cloud applications, with bandwidth and availability terms agreed in the service order. ### Data Center Interconnect Dual-homed circuits can provide a second path between facilities. Physical diversity must be confirmed for each installation. ### Cloud Access Cloud access for AWS, Azure and Google Cloud, subject to the available handoff and route at each site. ### VoIP and UCaaS Voice quality depends on delay, jitter and loss. A dedicated access circuit controls one part of that path; it does not control the remote service. > **Benchmarks for Critical Workloads** > > > TU-T guidance commonly uses \<150ms one-way delay for highly interactive voice tasks. > > > Microsoft’s Teams connectivity test passes with UDP latency \<100ms, UDP jitter \<30ms, and UDP packet loss \<1%. > > Sources: [ITU-T Rec. G.114 (05/2003)](https://www.itu.int/rec/dologin_pub.asp?lang=e\&id=T-REC-G.114-200305-I!!PDF-E), [Microsoft 365 network connectivity test tool](https://learn.microsoft.com/en-us/microsoft-365/enterprise/office-365-network-mac-perf-onboarding-tool?view=o365-worldwide) ## Delivery Options ### Bgp **For organizations with their own ASN and routing policies.** * BGP session established to your edge router * Full routing table or default route options * Support for inbound/outbound policy coordination * Enables dual-homing and traffic engineering BGP handoff supports customer routing policy and multi-homed deployments. ### Static **For streamlined deployments without BGP complexity.** * Static route configuration at the demarcation * Faster provisioning and turn-up * Ideal for single-site deployments and branches * Reduced technical overhead for smaller IT teams ### L2 **When you prefer to terminate routing on your equipment.** * VLAN handoff to your router/firewall stack * Transparent Layer 2 transport * Compatible with HA firewall pairs and SD-WAN edge devices * Simplifies multi-site routing consistency Layer 2 handoff connects SD-WAN edges and enterprise security appliances. ## Technical Specifications | Specification | Details | | ------------------------ | ---------------------------------------- | | **Committed Rate (CIR)** | 100% of contracted bandwidth, guaranteed | | **Availability SLA** | 99.99% with service credits | | **Latency SLA** | \< 10ms to major IX points from PoP | | **Jitter Target** | \< 2ms (95th percentile) | | **Packet Loss** | \< 0.1% during normal operation | | **MTTR** | 4-hour response for critical issues | | **IPv6** | Native dual-stack support | > **Published Backbone RTT** > > For the full auto-updated RTT matrix across our PoPs (Backbone measurements), see: [Backbone > Latency Matrix](/docs/network/latency) ## Deployment Process ### Site Survey Verify fiber availability and path diversity at your location. ### Handoff Selection Choose BGP, static routes, or L2 transport based on your infrastructure. ### Provisioning Physical installation and acceptance testing. ### Go Live Production handover with 24/7 NOC escalation. ## Frequently Asked Questions ### What is CIR and why does it matter? CIR is the committed bandwidth specified in your service order. The contract defines the measurement point, applicable conditions and remedies. ### How does the 99.99% SLA work? The availability SLA covers the demarcation point through our core network. It defines uptime targets and automatic service credits; monthly uptime reports are provided. ### Can I get redundant paths? Dual-homed service is available with active/active or active/standby operation. BGP can withdraw or prefer prefixes during failover; the design depends on the customer equipment and circuit paths. ### What equipment do I need? BGP handoff requires a BGP-4-capable router. Static and Layer 2 handoffs support enterprise routers and firewalls. We confirm the interface specifications during quoting. ### How long does installation take? Installation normally takes 20-30 business days where fiber is already available. New construction depends on access and permitting. We provide milestone updates during delivery. ### Do you support SD-WAN integration? The Layer 2 handoff can carry traffic between your SD-WAN edge and the network. Confirm interface and encapsulation requirements during design. *** ## Get Connected Send the site address and required bandwidth to **[sales@hatsnet.io](mailto:sales@hatsnet.io)**. Our response target is one business day. *** > **Ready to Crank It Up?** > > Contact [sales@hatsnet.io](mailto:sales@hatsnet.io) with your locations, bandwidth and target destinations. --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/services/business-dedicated-line). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Hats EPN | SD-WAN as a Service **Hats EPN (Enterprise Private Network)** provides managed routing or Ethernet transport between customer sites. > **Why Managed WAN?** > > Hats EPN connects multiple sites with separate traffic segments and managed routing. We handle the backbone and agreed edge configuration; your team manages the applications and local networks. ## Why Managed WAN? The comparison below separates the tasks managed by Hats Network from those retained in a customer-operated deployment. | Aspect | Self-Built SD-WAN | Hats EPN (Managed) | | --------------------------- | ------------------------------------------- | ------------------------------- | | **Setup Time** | 3-6 months (hardware, config, testing) | Days to weeks | | **Expertise Required** | Senior network engineers on staff | Zero - we handle everything | | **Ongoing Ops** | 24/7 NOC, patch management, troubleshooting | Included in service | | **Hardware Investment** | CAPEX for edge appliances | OPEX only - no hardware to buy | | **Multi-vendor Complexity** | Integrating firewalls, SD-WAN, monitoring | Single unified service | | **Scaling** | Manual provisioning, capacity planning | On-demand bandwidth adjustments | | **SLA Accountability** | You own all uptime risk | 99.99% SLA with Hats Network | **Responsibility boundary:** Hats Network manages the agreed service; your team retains control of applications and local networks. ## Delivery Models ### Routed **Managed routed connectivity** for enterprises who want us to handle the core routing infrastructure while they focus on their applications. **What You Get**: * Software-defined routing with traffic segmentation * BGP peering at your edge (optional) * Complete traffic isolation between environments * Automatic failover and path optimization * Predictable latency across our private backbone **Best For**: * Multi-office WAN with centralized internet breakout * Cloud on-ramp integration (AWS, Azure, GCP) * Segmented environments (production, development, voice, security zones) * Enterprises without deep networking expertise in-house ### Switched **Transparent Ethernet transport** for customers who prefer to manage their own routing stack while we handle the underlying transport. **What You Get**: * Ethernet handoff at each site (VLAN-tagged) * Single broadcast domain across all locations * You control all IP addressing and routing protocols * Transparent transport for multicast and custom protocols * MAC learning and forwarding managed by Hats **Best For**: * High-availability router pairs with VRRP/HSRP * Customer-managed SD-WAN appliances requiring L2 handoff * Legacy applications requiring flat Layer-2 adjacency * Organizations with existing networking teams ### Underlay **Stable transport foundation** for enterprises running their own SD-WAN, SASE, or custom overlay solutions. **What You Get**: * Stable IP transport with measurable, consistent latency * No encryption overhead (your overlay handles security) * Predictable path characteristics for intelligent path selection * Dual-stack IPv4/IPv6 support * Clean handoff to your edge devices **Best For**: * SD-WAN fabric underlay (any vendor) * SASE architecture foundation * Custom overlay with proprietary encryption requirements * Hybrid WAN combining private transport with internet backup ## Architecture **Simple connection model**: We manage the complex backbone. You get a clean handoff at each site. ## Use Cases ### Multi-Site Enterprise WAN Connect offices, factories and data centers over a managed backbone. Site handoffs, routing and support responsibilities are agreed during design. * **Manufacturing**: Connect global production facilities to ERP systems * **Retail**: Link stores to centralized POS and inventory systems * **Professional Services**: Reliable video conferencing between offices ### Cloud Access Cloud connections can join the private network. Latency still depends on the cloud region, physical distance and application architecture. * **AWS Direct Connect** integration * **Azure ExpressRoute** connectivity * **Multi-cloud** architectures with consistent performance ### Critical Workloads For voice and other delay-sensitive traffic, agree latency, jitter and packet-loss requirements before provisioning. * **Financial Trading**: Sub-50ms latency between trading venues * **Healthcare**: Reliable DICOM image transfers for telemedicine * **Real-time Collaboration**: Consistent video and voice quality > **Benchmarks for Critical Workloads** > > > TU-T guidance commonly uses \<150ms one-way delay for highly interactive voice tasks. > > > Microsoft’s Teams connectivity test passes with UDP latency \<100ms, UDP jitter \<30ms, and UDP packet loss \<1%. > > Sources: [ITU-T Rec. G.114 (05/2003)](https://www.itu.int/rec/dologin_pub.asp?lang=e\&id=T-REC-G.114-200305-I!!PDF-E), [Microsoft 365 network connectivity test tool](https://learn.microsoft.com/en-us/microsoft-365/enterprise/office-365-network-mac-perf-onboarding-tool?view=o365-worldwide) ## Service Specifications | Specification | Routed | Switched | Underlay | | ------------------- | -------------------- | ------------- | ------------- | | **Handoff Type** | BGP or Static | 802.1Q VLAN | IP or BGP | | **Routing Control** | Hats-managed | You manage | Minimal | | **MTU** | 1500 or 9000 (Jumbo) | 1500 or 9000 | 1500 or 9000 | | **Multicast** | Available | Transparent | No | | **IPv6** | Native | Native | Native | | **SLA** | 99.99% uptime | 99.99% uptime | 99.99% uptime | ## Traffic Segmentation & QoS Traffic classes can be configured for the following workloads: | Traffic Class | Priority | Treatment | Typical Use | | ------------- | ----------- | -------------------- | ------------------------- | | **Real-Time** | Highest | Guaranteed bandwidth | Voice, video conferencing | | **Critical** | High | Low latency queue | Trading systems, ERP | | **Business** | Standard | Standard forwarding | Email, SaaS applications | | **Default** | Best effort | Standard forwarding | Bulk transfers, backups | **No VRF configuration needed** - traffic segmentation is handled by our platform based on your requirements. ## FAQ ### What's the difference between Routed and Switched Private Network? **Routed Private Network**: We manage the routing. You connect your edge device to our network, and we handle all the complexity of path selection, traffic segmentation, and optimization. Best for organizations that want "it just works." **Switched Private Network**: We provide transparent Ethernet transport between your sites. Your routers see each other as directly connected on the same LAN segment. You manage all IP addressing and routing. Best for organizations with existing networking teams who want control. Use Routed handoff when Hats Network should manage inter-site routing. Use Switched handoff when the design requires Layer 2 adjacency or customer-managed routing protocols. ### Can I use Hats EPN as an SD-WAN underlay? The SD-WAN Underlay option supplies IP transport beneath your existing overlay. The SD-WAN controller remains responsible for overlay policy and path selection. Underlay capabilities: * **Consistent baselines**: Your SD-WAN makes better path decisions with stable metrics * **Reduced tunnel flapping**: No internet jitter causing unnecessary failover events * **Guaranteed bandwidth**: Private backbone capacity independent of internet congestion * **Simplified ops**: We maintain the underlay; you focus on overlay policies Supported deployments include VeloCloud, Fortinet, Palo Alto, Cisco and open-source SD-WAN. Confirm the selected platform and handoff during design. ### How does traffic segmentation work? EPN separates traffic into logical domains on shared physical infrastructure. The segmentation policy specifies which domains may communicate. Common traffic segments: * **Environment Isolation**: Production, staging, and development as separate virtual networks * **Traffic Type Separation**: Voice traffic on its own segment with priority treatment * **Security Zones**: Guest networks, corporate data, and PCI-compliant traffic fully isolated * **Departmental Segments**: Different business units with controlled interconnectivity Your team defines the segments and access policy; we configure the agreed network boundaries. ### What does 'fully managed' actually mean? Hats Network manages these service components: **We Handle**: * Core network design and operation * Routing protocol management and optimization * 24/7 monitoring and incident response * Capacity planning and scaling * Software updates and security patches * Peering and upstream relationships **You Handle**: * Your edge devices (routers, firewalls) * Your internal LAN * Your application configuration Each site receives an agreed handoff. The service order defines the boundary between our managed network and your local equipment. ### What are the latency commitments? We record baseline latency during onboarding. The following inter-PoP figures are reference values, not a site-specific service commitment: > **Published Backbone RTT** > > For the full auto-updated RTT matrix across our PoPs (Backbone measurements), see: [Backbone > Latency Matrix](/docs/network/latency) | Route | Latency | | --------------------- | -------- | | Hong Kong ↔ Tokyo | 35-45ms | | Hong Kong ↔ Singapore | 35-45ms | | Tokyo ↔ Los Angeles | 90-110ms | | Frankfurt ↔ Amsterdam | 5-8ms | The EPN SLA includes latency variation and packet loss as well as availability. The applicable targets and measurement boundaries are set out in the contract. ### Can we connect to cloud providers? Cloud connectivity options include: * **AWS**: Direct Connect via our partner locations * **Microsoft Azure**: ExpressRoute connectivity * **Google Cloud**: Dedicated Interconnect * **Alibaba Cloud**: Express Connect Cloud resources can join the same private segments as your sites. Routing and access rules determine which resources are reachable. ### How quickly can we get started? Deployment stages: | Phase | Timeline | Details | | ---------------- | --------- | ---------------------------------- | | **Discovery** | 1-2 days | Requirements, topology design | | **Contracting** | 3-5 days | Service order, terms | | **Provisioning** | 5-15 days | Circuit activation, cross-connects | | **Testing** | 1-2 days | Validation, acceptance | | **Go-Live** | - | Production traffic | **Typical total: 2-4 weeks** from signature to production, subject to site readiness and the agreed design. Existing customers may be able to reuse installed infrastructure. We confirm the delivery schedule after checking capacity and the required changes. ## Getting Started ### Schedule a Consultation Tell us about your sites, bandwidth needs, and any cloud connectivity requirements. We'll design a topology that fits. ### Review Your Design We provide a detailed proposal including topology, SLAs, and pricing. No surprises. ### Provisioning We handle circuit coordination, equipment installation, and configuration. You provide the cross-connect details. ### Go Live We validate end-to-end connectivity, establish baseline performance metrics, and hand over operational documentation. > **Ready to simplify your WAN?** > > Send the proposed design to **[sales@hatsnet.io](mailto:sales@hatsnet.io)**. > > Please include: > > * Number of sites and locations > * Bandwidth requirements per site > * Preferred delivery model (Routed/Switched/Underlay) > * Cloud connectivity needs > * Target deployment regions *** > **Ready to Crank It Up?** > > Contact [sales@hatsnet.io](mailto:sales@hatsnet.io) with your locations, bandwidth and target destinations. --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/services/enterprise-private-network). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Hats GO | Gaming Routing *** ## The Problem Hats GO is a routing profile for latency-sensitive game traffic. A high access speed does not guarantee low RTT to a game server. Route length, interconnection load and the player's local network all affect the result. Loss and variable delay can disrupt timing even when the average RTT looks acceptable. An overlay may change the path, but can also add processing and another endpoint. Hats GO instead applies routing policy within the network. *** ## What Is Hats GO (Gaming Optimized)? **Hats GO** applies routing policy to game destinations. Hats GO is not a VPN or an application tunnel. The service uses BGP policy and available Internet Exchange peering to reach game infrastructure. Whether that helps depends on the destination and the route it replaces. * **Direct IX peering** at HKIX, JPIX, DE-CIX, AMS-IX * **BGP-level optimization**-no encryption overhead * **Endpoint-specific route measurements** to the game servers you use * **99.99% uptime** with automatic failover Measurements to the actual game endpoints are more useful than a general speed test. *** ## The Numbers Evaluate the route using average RTT, variation and packet loss. > **Published Backbone RTT** > > For the full auto-updated RTT matrix across our PoPs (Backbone measurements), see: [Backbone > Latency Matrix](/docs/network/latency) | Route | Standard Path (Ping / Loss) | Hats GO (Ping / Loss) | Overall Improvement | | --------------------- | --------------------------- | --------------------- | ------------------- | | Hong Kong → Singapore | 60\~200ms / 4\~8% | 30\~45ms / \<1% | \~74% | | Hong Kong → Tokyo | 80\~150ms / 3\~7% | 45\~55ms / \<1% | \~60% | | Hong Kong → Taipei | 30\~80ms / 2\~5% | 15\~30ms / \<1% | \~61% | | Hong Kong → Seattle | 150\~200ms / 8\~15% | 130\~150ms / 1\~2% | \~28% | | Hong Kong → Frankfurt | 200\~270ms / 10\~20% | 150\~170ms / 1\~3% | \~42% | | Hong Kong → Sydney | 150\~200ms / 8\~12% | 120\~150ms / 1\~2% | \~30% | > **How We Calculate Improvement** > > The comparison uses an **Effective Latency** estimate that includes a retransmission penalty for packet loss. It is a model, not a measured game response time. > > **Variables:** > > * **$E$**: Effective Latency > * **$P$**: ICMP Latency (Base Ping) > * **$L$**: Packet Loss Rate > * **$std$ / $go$**: Subscripts for Standard Path and Hats GO > > **1. Effective Latency Base Formula:** > > ```math > E = \frac{P}{1 - L} > ``` > > **2. Improvement Calculation:** > > ```math > I = \left(1-\frac{P_{go} \times (1 - L_{std})}{P_{std} \times (1 - L_{go})} \right) \times 100\% > ``` \* Measured from HKG PoP. Your mileage varies based on location and last-mile conditions. *** ## How It Works ### Direct Peering Where the game network is reachable through our exchange peers, traffic can use that interconnection. | IX Location | Publishers On-Site | | ----------- | ------------------------------- | | **HKIX** | Tencent, Garena, Riot SEA | | **JPIX** | Square Enix, Sega, Bandai Namco | | **DE-CIX** | EU game servers, Ubisoft EU | | **AMS-IX** | Valve EU, EA EU | Hop count alone does not establish latency. Compare end-to-end measurements. ### Route Engineering BGP policy selects among the routes available at each PoP. * **Community tagging** - Game traffic gets priority paths * **Local preference** - IX routes over transit, always * **AS-path control** - Inbound optimization for symmetric latency The profile does not add an encryption or tunnel-termination layer. ### Qos Variation between replies matters alongside the average RTT. | Metric | Target | | ---------------- | ------ | | Packet Loss | \<0.1% | | Jitter | \<2ms | | Latency Variance | \<5% | A stable 35 ms path may be preferable to one varying between 25 and 60 ms, depending on the game's timing and loss behavior. *** ## By Game Type The relevant thresholds depend on the game and server region. | Genre | Examples | Target Latency | Why It Matters | | ----------------- | -------------------- | -------------- | -------------------------------------------------------------------------- | | **FPS** | CS2, Valorant, Apex | Sub-30ms | Peeker's advantage. Hit reg. When every frame counts in a duel. | | **Fighting** | SF6, Tekken 8, GBVSR | Sub-16ms | Rollback netcode works best with minimal delay. Frame-perfect inputs. | | **MOBA** | Dota 2, LoL | Sub-50ms | Skill-shot timing. Team fight responsiveness. No stutter on clutch plays. | | **Battle Royale** | PUBG, Fortnite | Sub-40ms | Position updates. Hit registration at range. Circle closure reaction time. | > **References for Competitive Latency Ceilings** > > Epic describes \<=20ms as “good” latency, 20-100ms as “acceptable”, and >=100ms as “poor” for Fortnite. > > Source: [Epic Games: Understanding latency or ping in Fortnite](https://www.epicgames.com/help/fortnite-battle-royale-c-202300000001636/technical-support-c-202300000001719/fortnite-latency-and-ping-troubleshooting-a202300000010042) *** ## Coverage GO availability by PoP: **Asia-Pacific** * Hong Kong (HKG) * Tokyo (NRT) * Singapore (SIN) **Europe** * Frankfurt (FRA) * Amsterdam (AMS) **North America** * Los Angeles (LAX) * New York (NYC) Confirm the target game servers and regional availability with sales. *** ## FAQ ### Is this a VPN? No. GO uses BGP routing and available IX interconnections, without adding a VPN tunnel. The actual path still depends on the destination network. ### Will this lower my ping? It may help when our available route is better than your current path. Little improvement should be expected when the existing route is already short and stable. Request tests to your target servers first. ### Does this work with my ISP? GO is delivered through BGP peering or IP transit, usually to businesses, data centers and esports venues. Residential access depends on regional ISP arrangements. ### Which games benefit most? The profile can carry standard IP game traffic. Benefits depend on the path and the game's sensitivity to RTT, jitter and loss. ### Is this legal for competitive play? GO uses standard IP routing. It does not manipulate game packets or bypass regional restrictions; game and platform rules still apply. ### Can I try it first? Business customers can request measurements and trials. Send sales your location and target server addresses. *** > **Ready to Crank It Up?** > > Contact [sales@hatsnet.io](mailto:sales@hatsnet.io) with your locations, bandwidth and target destinations. --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/services/gaming-optimized). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Hats Network Services Hats Network provides dedicated access, managed private networks and routing profiles. Choose a service by the required handoff, bandwidth commitment and support boundary. *** > ## Service Catalog - [Hats EPN](/docs/services/enterprise-private-network): SD-WAN as a Service. Multi-site private connectivity with traffic segmentation and guaranteed SLAs. - [Hats BDL](/docs/services/business-dedicated-line): Enterprise DIA. Committed bandwidth with flexible handoff options for offices and critical endpoints. - [Hats PUL](/docs/services/premium-user-line): Performance IP. Low-jitter access for latency-sensitive professionals and distributed teams. - [Hats GO](/docs/services/gaming-optimized): Competitive Gaming Routes. Sub-30ms paths to publishers via direct IX peering. - [Hats SR](/docs/services/streaming-ready): Media-Grade Connectivity. High-throughput infrastructure for 4K/8K delivery with CDN peering. *** ## Service Selection Decision Tree Start with the number of sites and the type of traffic you need to carry: ## Quick Selection Guide ### Saas **Primary Need**: Secure, segmented connectivity between offices and cloud. | Requirement | Recommended | | ----------------------- | -------------- | | 3+ sites, private mesh | Hats EPN | | Single site, SLA-backed | Hats BDL | | Hybrid cloud handoff | Hats EPN + BGP | ### Multi Site **Primary Need**: Reliable inter-office connectivity that scales. | Site Count | Recommended | | -------------- | ------------------------- | | 2–3 locations | Hats BDL (point-to-point) | | 4+ locations | Hats EPN (full mesh) | | Remote workers | Hats PUL (endpoints) | ### Creative **Primary Need**: Low-jitter access to cloud suites and collaboration tools. | Use Case | Recommended | | --------------------- | ----------- | | Video editing teams | Hats PUL | | Animation/VFX studios | Hats BDL | | Distributed teams | Hats EPN | ### Gaming **Primary Need**: Minimal latency to game servers and tournament infra. | Organization | Recommended | | -------------------------- | --------------------- | | Esports teams & facilities | Hats GO | | Tournament organizers | Hats GO + BGP | | Gaming cafes/chains | Hats BDL + GO profile | ### Media **Primary Need**: Consistent throughput for high-bitrate video. | Scale | Recommended | | ------------------- | ----------------------- | | Individual creators | Hats PUL | | OTT platforms | Hats SR | | CDN operators | IP Transit + SR profile | *** ## Service Architecture Overview ## Why Enterprise-Grade | Feature | Best-Effort Internet | Hats Enterprise-Grade | | --------- | ------------------------ | ---------------------------- | | Bandwidth | Shared, oversubscribed | Committed CIR | | Routing | Cost-optimized, variable | BGP-controlled deterministic | | Peering | Multi-transit hops | Direct IX at 15+ exchanges | | SLA | None | 99.99% uptime with credits | | Support | Business hours | 24/7 NOC | *** ## Frequently Asked Questions ### What's the difference between Hats BDL and standard business broadband? BDL includes a **Committed Information Rate (CIR)**. The service order defines the bandwidth commitment and its measurement conditions; ordinary broadband is typically sold on a best-effort basis. BDL offers symmetric speeds, static IPs, SLA terms and optional BGP handoff. > **Sources** > > Oversubscription examples and typical ratios: [CTC Technology & Energy: All Mbps Are Not Created > Equal (Connectivity Performance > Factors)](https://www.ctcnet.us/wp-content/uploads/2014/02/CTC-ConnectivityPerformanceFactorsBrief0213141.pdf) ### Can I start with one service and upgrade later? A BDL installation can be incorporated into an EPN design as more sites are added. Confirm handoff compatibility, capacity and any required changes with your account manager. Your account manager can review the existing installation and quote the migration. ### Does Hats EPN replace my existing SD-WAN? EPN manages network-layer routing and segmentation. SD-WAN appliances can run over the backbone; appliance management and vendor support are scoped separately. ### What's the typical installation timeline? | Service | Timeline | | -------- | -------------------- | | Hats PUL | 5\~10 business days | | Hats BDL | 15\~30 business days | | Hats EPN | 30\~45 business days | Expedited installation may be available in selected markets. ### Which PoPs support GO and SR profiles? GO and SR availability depends on the interconnections at each location: **Asia-Pacific**: Hong Kong, Tokyo, Singapore\ **Europe**: Frankfurt, Amsterdam, London\ **North America**: Los Angeles, New York, Ashburn Ask sales for availability and measurements to the intended destinations. *** ## Get Started > **Contact Sales** > > Send sales your sites, required bandwidth, target destinations and preferred handoff. We can then assess the available service options. > > **[sales@hatsnet.io](mailto:sales@hatsnet.io)** - include your current setup, bandwidth needs, and any performance issues. *** > **Ready to Crank It Up?** > > Contact [sales@hatsnet.io](mailto:sales@hatsnet.io) with your locations, bandwidth and target destinations. --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/services). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Hats PUL | Gaming & Cloud Performance IP ## What It Is **Hats PUL (Premium User Line)** is Performance IP connectivity - not a VPN, not a proxy, not a tunnel. PUL routes traffic over the AS203314 backbone and available IX peers after it reaches our edge. The access connection remains part of the end-to-end path. The result depends on how this path compares with the route your ISP already uses. *** ## The Problem A high advertised access speed does not establish low latency to a particular destination. Transit and peering choices affect the path. Congestion can occur in the access network, at interconnections or near the destination. * Delay variation between successive replies * Long paths to the selected server region * Packet loss during busy periods Increasing access bandwidth will not necessarily reduce RTT or resolve loss elsewhere on the path. *** ## The Solution **Route selection after the access network.** | What We Do | What It Means | | -------------------------- | ------------------------------------------------------------------------------------------- | | **Direct IX Peering** | Traffic enters our network and exits directly at major IX points. No middlemen. No detours. | | **AS203314 Backbone** | Our own infrastructure. Our own routes. Full control over path selection. | | **Gaming-Optimized Paths** | Routes tuned for UDP consistency, not just TCP throughput. | | **Multi-Hub Redundancy** | Automatic failover between our PoPs without dropping your session. | Measure the complete path to the game or cloud region you use. The profile cannot guarantee a fixed RTT across all destinations. *** ## By The Numbers > **Published Backbone RTT** > > For the full auto-updated RTT matrix across our PoPs (Backbone measurements), see: [Backbone > Latency Matrix](/docs/network/latency) Reference measurements from core PoPs: | Route | Standard ISP | Hats PUL | Improvement | | --------------------- | ------------ | -------- | ------------ | | Singapore → Hong Kong | 45-60ms | 28-32ms | \~40% faster | | Tokyo → Seoul | 55-70ms | 32-38ms | \~45% faster | | Frankfurt → Amsterdam | 25-35ms | 12-15ms | \~55% faster | | Los Angeles → Tokyo | 140-160ms | 95-105ms | \~30% faster | These PoP-to-PoP examples exclude your access network. Use an endpoint-specific test before ordering. > **References for Competitive Latency Ceilings** > > Epic describes \<=20ms as “good” latency, 20-100ms as “acceptable”, and >=100ms as “poor” for Fortnite. > > Source: [Epic Games: Understanding latency or ping in Fortnite](https://www.epicgames.com/help/fortnite-battle-royale-c-202300000001636/technical-support-c-202300000001719/fortnite-latency-and-ping-troubleshooting-a202300000010042) *** ## Who It's For ### Competitive Gamers Compare RTT, jitter and packet loss to the game servers you use. ### Streamers & Content Creators Check upstream capacity for streaming alongside the route to the game server. ### Remote Workers Evaluate both directions for calls and file transfers, including the local access connection. ### Cloud Developers Compare paths to the specific AWS, Azure or Google Cloud region hosting your workload. *** ## How It Works 1. **Connect** - Route your traffic through our nearest PoP 2. **Optimize** - Our edge routers select the best path in real-time 3. **Win** - Lower latency, reduced jitter, consistent performance The service is delivered through network routing. We confirm the handoff and any local configuration needed for your connection. *** ## Questions ### Is this a VPN? **No.** Hats PUL is Performance IP connectivity through route optimization. We don't encrypt your traffic, change your IP location, or provide "privacy" features. We're a network infrastructure service - not a VPN provider. ### Will this work with my ISP? PUL changes routing after traffic reaches our edge. Access compatibility and delivery options depend on your ISP and region. ### Does it support all games? Standard IP game traffic can use the service. An improvement is not assured: it depends on whether the new route reduces delay, variation or loss. ### Who operates the network? Hats Network operates AS203314. The available backbone and peering paths determine which destinations the profile can improve. ### How do I get started? Ask sales to check availability, delivery requirements and measurements for your location before ordering. *** > **Ready to Crank It Up?** > > Contact [sales@hatsnet.io](mailto:sales@hatsnet.io) with your locations, bandwidth and target destinations. --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/services/premium-user-line). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Hats SR | Media-Grade Connectivity Hats SR is a connectivity profile for sustained media traffic. SR targets sustained upload and download capacity for professional media workflows. Required bandwidth depends on the codecs, bitrates and number of concurrent streams. > **Why Quality Matters** > > According to [Conviva's 2025 research](https://www.conviva.ai/2025-state-of-digital-experience-report/): > > > Every 1% degradation in Quality of Experience leads to a 1% increase in customer churn. > > Rebuffering interrupts playback. Capacity planning should account for peak demand and the behavior of the delivery platform. *** ## The Problem: Burst vs. Sustained A short speed test does not establish sustained streaming performance: ### Oversubscription **The Problem**: Consumer ISPs heavily oversubscribe bandwidth, assuming not everyone uses it simultaneously. **The Impact**: * Bitrate throttling (4K drops to 720p) * Buffering during high-action scenes * Inconsistent quality throughout playback **Our Solution**: Hats SR provisions dedicated capacity with direct CDN peering to maintain quality during peak demand periods. ### Peak Hour **The Problem**: During peak hours (7-11 PM), ISP peering links to CDNs often run at saturation. **The Impact**: * Higher latency to cache * Increased startup time * Frequent quality switches **Our Solution**: We maintain low-latency paths to major CDNs via IX peering at STUIX, NL-IX, DE-CIX, HKIX, and more. ### Burst Sustained **The Problem**: Speed tests measure burst capacity, but 8K streaming requires sustained 100Mbps+ for hours. **The Impact**: * High initial speed, then quality drops * Inability to maintain 8K streams * Degradation during long viewing sessions **Our Solution**: Our network is provisioned for sustained throughput with redundant upstream providers. *** ## What Is "Hats SR"? Hats SR provisions connectivity for sustained media workloads, with capacity and handoff requirements agreed for the deployment. Typical workloads include: * **Sustained throughput** - hours of consistent 4K/8K delivery * **Direct CDN peering** - bypassing congested transit paths * **Low jitter** - stable bitrates without quality switches * **Redundant architecture** - 99.99% uptime SLA *** ## Bandwidth Requirements Use the following bandwidth figures as planning references: | Quality | Resolution | Recommended Bandwidth | Typical Bitrate\* | | ------- | ----------- | --------------------- | ----------------- | | HD | 1080p | 10-15 Mbps | 5-9 Mbps | | Full HD | 1080p60 | 15-25 Mbps | 9-15 Mbps | | 4K | 2160p | 25-50 Mbps | 15-25 Mbps | | 4K HDR | 2160p + HDR | 40-60 Mbps | 20-35 Mbps | | 8K | 4320p | 100+ Mbps | 50-80+ Mbps | Size the service for the combined bitrate of concurrent streams, protocol overhead and recovery margin. > **Sources** > > Bandwidth/bitrate reference points: [Netflix-recommended internet > speeds](https://help.netflix.com/en/node/306), [YouTube Live - Bitrate > Recommendations](https://support.google.com/youtube/answer/2853702?hl=en) *** ## CDN Peering Available CDN interconnections include the following exchange peers: | CDN | IX Access Points | Key Benefit | | -------------------- | --------------------------- | ---------------------------------------- | | Netflix Open Connect | SIX Seattle, AMS-IX, others | 4K/8K content served locally | | Google Global Cache | Multiple IX locations | YouTube and Google services optimization | | Akamai | Multiple IX locations | Broad content delivery network access | | Cloudflare | Multiple IX locations | Content delivery + performance | | Amazon CloudFront | Multiple IX locations | AWS-originated content optimization | | Fastly | Multiple IX locations | Real-time content and API optimization | > **IX Advantage** > > IX presence can provide a path to CDN networks. The path used for a request depends on CDN steering, routing policy and current reachability. *** ## Use Cases ### Video Platforms OTT delivery requirements: * OTT platforms * Media publishers * VOD services * Corporate video delivery ### Live Production Live production requirements: * 4K60 streaming to multiple platforms simultaneously * Stable bitrates for broadcast-quality productions * Low-latency options for interactive streaming * Remote production workflows ### Enterprise Streaming Video conferencing requirements: * All-hands meetings in 4K * Executive broadcasts * Town halls and events * Microsoft Teams, Zoom, Google Meet optimization *** ## Technical Specifications | Specification | Value | | ----------------- | ---------------------------------- | | Minimum Bandwidth | 100 Mbps sustained per stream (8K) | | Global Capacity | 1 Tbps+ backbone capacity | | Jitter | \<2ms (95th percentile) | | Packet Loss | \<0.1% (normal operation) | | IX Peering | HKIX, JPIX, DE-CIX, AMS-IX, STUIX | | Uptime SLA | 99.99% with automatic failover | > **Published Backbone RTT** > > For the full auto-updated RTT matrix across our PoPs (Backbone measurements), see: [Backbone > Latency Matrix](/docs/network/latency) *** ## Regional Availability SR availability by PoP: **Asia-Pacific**: Hong Kong, Tokyo, Singapore **North America**: Los Angeles, New York **Europe**: Frankfurt, Amsterdam Ask sales about locations not listed here. *** ## FAQ ### What makes Hats SR different from business internet? SR is scoped around sustained throughput and the available CDN interconnections. Confirm committed bandwidth, contention and SLA terms in the quote. ### How do you handle peak hours? Capacity planning uses peak-hour demand projections. Multiple upstreams provide alternate paths, but failover does not guarantee identical performance. ### Can Hats SR support live production workflows? SR can support live-production uploads and multiple publishing destinations. Provide the platform list, bitrates and latency requirements so the design can be checked. ### What happens if a CDN has issues? Alternate network paths can help with a failed interconnection. Switching to a different CDN requires support from the application or delivery platform. ### Do you offer dedicated infrastructure? Custom infrastructure can be assessed for high-volume deployments. Send sales the traffic profile, locations and required handoffs. ### What is the minimum commitment? Pricing depends on location, bandwidth and delivery requirements. Request a quote for the proposed workload. *** > **Ready to Crank It Up?** > > Contact [sales@hatsnet.io](mailto:sales@hatsnet.io) with your locations, bandwidth and target destinations. --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/services/streaming-ready). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Asia Transit The locations below have different address-family and filtering options. Taipei is closed to new orders as noted in its section. Confirm the requested port and prefix policy with [sales](mailto:sales@hatsnet.io) before ordering. ## 🇭🇰 Hong Kong (HKG3) > [View HKG3 Routes in BGP Tools →](https://bgp.tools/prefix/2602:2a3:300::/41#connectivity) | Feature | Value | Notes | | ---------------- | ----- | ------------------------ | | IPv4 IP-Transit | ✅ | Do not use for Anti-DDoS | | IPv6 IP-Transit | ✅ | Connected AS174 & AS6939 | | Automated Filter | ❌ | Prefix list required | ### Upstream * Hytron Network Limited ([AS202662](https://bgp.tools/as/202662)) * Cogent Communications ([AS174](https://bgp.tools/as/174)) * Hurricane Electric ([AS6939](https://bgp.tools/as/6939)) * CDN77 ([AS60068](https://bgp.tools/as/60068)) ### Peering * 🇭🇰 HKIX (Hong Kong) *(via AS202662)* * 🇭🇰 Equinix IX (Hong Kong) *(via AS202662)* *** ## 🇸🇬 Singapore (SIN1) | Feature | Value | Notes | | ---------------- | ----- | ------------------------ | | IPv4 IP-Transit | ✅ | Do not use for Anti-DDoS | | IPv6 IP-Transit | ✅ | Connected AS2914 & AS174 | | Automated Filter | ❌ | Prefix list required | ### Upstream * Nearoute Limited ([AS51847](https://bgp.tools/as/51847)) * Cogent Communications ([AS174](https://bgp.tools/as/174)) * NTT ([AS2914](https://bgp.tools/as/2914)) ### Peering * 🇸🇬 Equinix IX (Singapore) *(via Upstream)* *** ## 🇯🇵 Tokyo, Japan (TYO2) | Feature | Value | Notes | | ---------------- | ----- | --------------------------------- | | IPv4 IP-Transit | ✅ | Do not use for Anti-DDoS | | IPv6 IP-Transit | ✅ | Connected AS2914 & AS174 & AS3491 | | Automated Filter | ❌ | Prefix list required | ### Upstream * Nearoute Limited ([AS51847](https://bgp.tools/as/51847)) * Cogent Communications ([AS174](https://bgp.tools/as/174)) * NTT ([AS2914](https://bgp.tools/as/2914)) * PCCW Global ([AS3491](https://bgp.tools/as/3491)) ### Peering * 🇯🇵 Equinix IX (Tokyo) *(via Upstream)* *** ## 🇹🇼 Taipei (TPE2) > This location is currently **Closed** for new IP-Transit orders. [View Route Map in BGP Tools > →](https://bgp.tools/prefix/2a0c:9a40:95e3::/48#connectivity) | Feature | Value | Notes | | ---------------- | ----- | ----------------------- | | IPv6 IP-Transit | ✅ | Connected AS6939 | | IPv6 to Cogent | ❌ | Unconnected AS174 | | Automated Filter | ✅ | Supports RPKI ROA / IRR | ### Upstream * Hurricane Electric ([AS6939](https://bgp.tools/as/6939)) ### Peering * 🇹🇼 STUIX (Taipei) --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/transit/asia). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Europe Transit Amsterdam and Frankfurt both list HE and iFog upstreams. Their filtering requirements differ; the handoff tables below describe each location. Confirm the requested port and prefix policy with [sales](mailto:sales@hatsnet.io) before ordering. ## 🇳🇱 Amsterdam, NL (AMS) > [View AMS Route Map in BGP Tools →](https://bgp.tools/prefix/2a0c:9a40:95e3::/48#connectivity) | Feature | Value | Notes | | ---------------- | ----- | -------------------------- | | IPv6 IP-Transit | ✅ | Direct Connected to AS6939 | | IPv6 to Cogent | ✅ | Reachable via AS34927 | | Automated Filter | ✅ | Supports RPKI ROA / IRR | ### Upstream * Hurricane Electric ([AS6939](https://bgp.tools/as/6939)) * iFog GmbH ([AS34927](https://bgp.tools/as/34927)) ### Peering * 🇳🇱 NL-IX (Amsterdam) *** ## 🇩🇪 Frankfurt, DE (FRA) > [View FRA Routes in BGP Tools →](https://bgp.tools/prefix/2a0c:9a40:95e2::/48#connectivity) | Feature | Value | Notes | | ---------------- | ----- | -------------------------- | | IPv6 IP-Transit | ✅ | Direct Connected to AS6939 | | IPv6 to Cogent | ✅ | Reachable via AS34927 | | Automated Filter | ❌ | Prefix list required | ### Upstream * Hurricane Electric ([AS6939](https://bgp.tools/as/6939)) * iFog GmbH ([AS34927](https://bgp.tools/as/34927)) ### Peering * 🇩🇪 Loc-IX (Frankfurt) --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/transit/europe). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # IP Transit Services > **IP Transit from AS203314** > > Hats Network offers IP transit in Asia and Europe. The available handoff, address families and filtering method depend on the location. > > * **Multi-homed** upstream connectivity (Cogent, NTT, HE, PCCW) > * **IPv4 and IPv6** dual-stack support > * **Competitive pricing** with flexible commit options > * **24/7 NOC** support for critical issues ## Transit Ordering Flow ## Available Locations Transit locations are listed below: | Continent | Country | City | Datacenter | Status | | --------- | ----------- | ------------- | ---------- | ------ | | Europe | Netherlands | Amsterdam | TBD | Open | | Asia | Hong Kong | Tseung Kwan O | HKG3 | Open | | Asia | Singapore | Singapore | SIN1 | Open | | Asia | Japan | Tokyo | TYO2 | Open | | Asia | Taiwan | Taipei | TPE2 | Closed | ## How to Order ### Choose Your Location Check the [Asia](/docs/transit/asia) and [Europe](/docs/transit/europe) pages for the location-specific handoffs. Confirm these requirements before requesting a quote: * **Latency requirements** to your target markets * **Upstream diversity** available at each location * **IPv4 vs IPv6** availability ### Contact Sales Send [sales@hatsnet.io](mailto:sales@hatsnet.io) the following details: * Company name and ASN * Desired location and bandwidth requirements * Estimated commit level * BGP communities or special routing requirements ### Setup & Turn-up After order confirmation, we will: 1. Provision the cross-connect or tunnel 2. Establish BGP sessions 3. Announce your prefixes 4. Provide 24/7 NOC contact information ## Location Details - [Asia Transit](/docs/transit/asia): Hong Kong, Singapore, Tokyo, and Taipei locations with connections to major Asian IXs. - [Europe Transit](/docs/transit/europe): Amsterdam and Frankfurt locations with diverse European connectivity. ## Technical Specifications > **Standard Features** > > IP transit includes: > > * **BGP Communities** for traffic engineering > * **RPKI/IRR filtering** (where available) > * **Looking glass** access for route verification > * **Real-time** traffic graphs > * **Email** notifications for maintenance windows ### Supported Protocols | Protocol | Support | Notes | | -------- | -------- | ------------------------------ | | IPv4 | Full | Available at all locations | | IPv6 | Full | Native dual-stack | | BGP-4 | Full | Standard and large communities | | BFD | Optional | For faster convergence | ### Upstream Networks Upstream providers include: * **Cogent Communications** (AS174) * **Hurricane Electric** (AS6939) * **NTT Communications** (AS2914) * **PCCW Global** (AS3491) * **iFog GmbH** (AS34927) ## Pricing > Pricing depends on the location and bandwidth commitment. Request a quote from [sales](mailto:sales@hatsnet.io). ## Support IP transit support contacts: * **Sales:** [sales@hatsnet.io](mailto:sales@hatsnet.io) * **NOC:** [noc@hatsnet.io](mailto:noc@hatsnet.io) --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/transit). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Amsterdam network latency (AMS) Amsterdam (AMS) is a Hats Network PoP in the Netherlands. This round includes domestic-region and intercontinental measurements. Measurement round: `2026-08-16T04:07:28Z`. The published round covers **19 outbound routes** from Amsterdam. Incoming routes are measured separately; neither direction describes last-mile performance. ## Location * **Location:** Amsterdam, Netherlands (Europe) * **Region:** Europe * **Role:** Backbone interconnection point for Hats Network (AS203314) * **Public city reference:** [GeoNames 2759794](https://www.geonames.org/2759794); 52.37403, 4.88969 * **Coverage:** RTT measurements to 19 other PoPs across 5 continents ### Interconnection * **Facility:** AMS1 * **Upstream transit:** HE / Cogent / Telia / Liberty Global * **Peering / IX:** NL-IX @ ANS > Interactive content is available on the canonical HTML page. ## Round Summary | Metric | Value | | --------------------- | ----------------------------------------------------------------------------------------------------------- | | Fastest Route | [🇳🇱 Amsterdam (AMS) → 🇬🇧 London (LON)](/docs/network/latency/pairs/ams-lon-rtt) (**5.2 ms**); Ultra-Low | | Slowest Route | [🇳🇱 Amsterdam (AMS) → 🇦🇺 Sydney (SYD)](/docs/network/latency/pairs/ams-syd-rtt) (**258 ms**) | | Average RTT | **112.4 ms** | | Average Jitter | **0.86 ms** | | Average Packet Loss | **0.05%** | | Best Fiber Efficiency | **83.9%** | | Intra-Region Peers | 6 PoPs in Europe | | Total Routes | 19 outbound / 19 inbound (100% coverage) | > Jitter, packet-loss, percentile, and trend fields are estimated comparative indicators. Read the [theoretical fiber latency methodology](/docs/network/latency/theoretical-fiber-latency). ## Regional Routes Amsterdam's 6 other **Europe** PoPs have RTTs from **5.2 ms** (London (LON)) to **38.1 ms** (Moscow (MOW)). Amsterdam is in the Netherlands, on the North Sea side of continental Europe. ## Outbound Latency from Amsterdam (AMS) Round-trip time in milliseconds **from Amsterdam** to all other backbone PoPs, sorted fastest to slowest. | Destination | RTT (ms) | Fiber Floor | Efficiency | Jitter | Loss | Tier | | ----------------------- | ---------------------------------------------------- | ----------- | ---------- | ------- | ----- | --------- | | 🇬🇧 London (LON) | **[5.2](/docs/network/latency/pairs/ams-lon-rtt)** | 3.51 ms | 67.4% | 0.12 ms | 0.09% | Ultra-Low | | 🇩🇪 Frankfurt (FRA) | **[5.9](/docs/network/latency/pairs/ams-fra-rtt)** | 3.57 ms | 60.6% | 0.12 ms | 0.05% | Ultra-Low | | 🇫🇷 Paris (PAR) | **[7](/docs/network/latency/pairs/ams-par-rtt)** | 4.22 ms | 60.3% | 0.12 ms | 0.03% | Ultra-Low | | 🇩🇪 Berlin (BER) | **[7.2](/docs/network/latency/pairs/ams-ber-rtt)** | 5.67 ms | 78.8% | 0.12 ms | 0.19% | Ultra-Low | | 🇫🇷 Marseille (MRS) | **[19.6](/docs/network/latency/pairs/ams-mrs-rtt)** | 9.89 ms | 50.5% | 0.12 ms | 0.03% | Ultra-Low | | 🇷🇺 Moscow (MOW) | **[38.1](/docs/network/latency/pairs/ams-mow-rtt)** | 21.1 ms | 55.4% | 0.38 ms | 0% | Excellent | | 🇺🇸 New York (NYC) | **[68.6](/docs/network/latency/pairs/ams-nyc-rtt)** | 57.56 ms | 83.9% | 0.31 ms | 0% | Excellent | | 🇺🇸 Ashburn (IAD) | **[75.7](/docs/network/latency/pairs/ams-iad-rtt)** | 60.94 ms | 80.5% | 0.75 ms | 0.05% | Excellent | | 🇺🇸 Miami (MIA) | **[106](/docs/network/latency/pairs/ams-mia-rtt)** | 73.03 ms | 68.9% | 1.03 ms | 0.01% | Good | | 🇺🇸 Seattle (SEA) | **[130.2](/docs/network/latency/pairs/ams-sea-rtt)** | 76.86 ms | 59% | 1.26 ms | 0.04% | Good | | 🇺🇸 Los Angeles (LAX) | **[133.7](/docs/network/latency/pairs/ams-lax-rtt)** | 87.74 ms | 65.6% | 0.65 ms | 0% | Good | | 🇭🇰 Hong Kong (HKG) | **[155.9](/docs/network/latency/pairs/ams-hkg-rtt)** | 91.05 ms | 58.4% | 1.21 ms | 0.05% | Fair | | 🇸🇬 Singapore (SIN) | **[157.6](/docs/network/latency/pairs/ams-sin-rtt)** | 102.86 ms | 65.3% | 0.82 ms | 0% | Fair | | 🇹🇼 Taipei (TPE) | **[168.7](/docs/network/latency/pairs/ams-tpe-rtt)** | 92.72 ms | 55% | 1.29 ms | 0.11% | Fair | | 🇿🇦 Johannesburg (JNB) | **[171.9](/docs/network/latency/pairs/ams-jnb-rtt)** | 88.05 ms | 51.2% | 1.1 ms | 0.1% | Fair | | 🇧🇷 São Paulo (GRU) | **[175.5](/docs/network/latency/pairs/ams-gru-rtt)** | 95.79 ms | 54.6% | 1.53 ms | 0% | Fair | | 🇯🇵 Tokyo (TYO) | **[198.2](/docs/network/latency/pairs/ams-tyo-rtt)** | 91.18 ms | 46% | 1.27 ms | 0.14% | Fair | | 🇦🇺 Melbourne (MEL) | **[252.5](/docs/network/latency/pairs/ams-mel-rtt)** | 161.98 ms | 64.1% | 1.77 ms | 0.03% | High | | 🇦🇺 Sydney (SYD) | **[258](/docs/network/latency/pairs/ams-syd-rtt)** | 162.93 ms | 63.2% | 2.35 ms | 0.11% | High | ## Fastest Routes to Amsterdam (AMS) The 10 fastest measured routes **to Amsterdam**, ranked by average RTT. Min / max / stdev are computed from the same measurement round (50 ICMP echo probes per route, 100 ms interval); routes without a published per-sample round show the measured average RTT only. | Rank | Source | Avg RTT | Min | Max | Stdev | | ---- | -------------------- | ------------------------------------------------------- | --------- | --------- | ------- | | 1 | 🇬🇧 London (LON) | **[5.2 ms](/docs/network/latency/pairs/lon-ams-rtt)** | 5.03 ms | 5.75 ms | 0.16 ms | | 2 | 🇩🇪 Frankfurt (FRA) | **[5.9 ms](/docs/network/latency/pairs/fra-ams-rtt)** | 5.62 ms | 6.84 ms | 0.25 ms | | 3 | 🇫🇷 Paris (PAR) | **[7 ms](/docs/network/latency/pairs/par-ams-rtt)** | 6.65 ms | 8.5 ms | 0.33 ms | | 4 | 🇩🇪 Berlin (BER) | **[7.2 ms](/docs/network/latency/pairs/ber-ams-rtt)** | 6.91 ms | 7.91 ms | 0.23 ms | | 5 | 🇫🇷 Marseille (MRS) | **[20.3 ms](/docs/network/latency/pairs/mrs-ams-rtt)** | 19.62 ms | 22.75 ms | 0.64 ms | | 6 | 🇷🇺 Moscow (MOW) | **[37.6 ms](/docs/network/latency/pairs/mow-ams-rtt)** | 37.09 ms | 39.9 ms | 0.57 ms | | 7 | 🇺🇸 New York (NYC) | **[68.7 ms](/docs/network/latency/pairs/nyc-ams-rtt)** | 66.97 ms | 74.13 ms | 1.38 ms | | 8 | 🇺🇸 Ashburn (IAD) | **[75.2 ms](/docs/network/latency/pairs/iad-ams-rtt)** | 72.76 ms | 84.12 ms | 2.32 ms | | 9 | 🇺🇸 Miami (MIA) | **[106 ms](/docs/network/latency/pairs/mia-ams-rtt)** | 103.48 ms | 113.5 ms | 2.2 ms | | 10 | 🇺🇸 Seattle (SEA) | **[128.6 ms](/docs/network/latency/pairs/sea-ams-rtt)** | 126.47 ms | 138.71 ms | 2.06 ms | ## Inbound Latency to Amsterdam (AMS) Round-trip time in milliseconds **from all other PoPs to Amsterdam**. The measurement origin matters: inbound and outbound RTT can differ for the same pair. | Source | RTT (ms) | Fiber Floor | Efficiency | Jitter | Loss | Tier | | ----------------------- | ---------------------------------------------------- | ----------- | ---------- | ------- | ----- | --------- | | 🇬🇧 London (LON) | **[5.2](/docs/network/latency/pairs/lon-ams-rtt)** | 3.51 ms | 67.4% | 0.12 ms | 0% | Ultra-Low | | 🇩🇪 Frankfurt (FRA) | **[5.9](/docs/network/latency/pairs/fra-ams-rtt)** | 3.57 ms | 60.6% | 0.12 ms | 0% | Ultra-Low | | 🇫🇷 Paris (PAR) | **[7](/docs/network/latency/pairs/par-ams-rtt)** | 4.22 ms | 60.3% | 0.12 ms | 0% | Ultra-Low | | 🇩🇪 Berlin (BER) | **[7.2](/docs/network/latency/pairs/ber-ams-rtt)** | 5.67 ms | 78.8% | 0.12 ms | 0% | Ultra-Low | | 🇫🇷 Marseille (MRS) | **[20.3](/docs/network/latency/pairs/mrs-ams-rtt)** | 9.89 ms | 48.7% | 0.17 ms | 0.16% | Ultra-Low | | 🇷🇺 Moscow (MOW) | **[37.6](/docs/network/latency/pairs/mow-ams-rtt)** | 21.1 ms | 56.1% | 0.28 ms | 0.17% | Excellent | | 🇺🇸 New York (NYC) | **[68.7](/docs/network/latency/pairs/nyc-ams-rtt)** | 57.56 ms | 83.8% | 0.33 ms | 0.03% | Excellent | | 🇺🇸 Ashburn (IAD) | **[75.2](/docs/network/latency/pairs/iad-ams-rtt)** | 60.94 ms | 81% | 0.34 ms | 0.02% | Excellent | | 🇺🇸 Miami (MIA) | **[106](/docs/network/latency/pairs/mia-ams-rtt)** | 73.03 ms | 68.9% | 0.87 ms | 0.08% | Good | | 🇺🇸 Seattle (SEA) | **[128.6](/docs/network/latency/pairs/sea-ams-rtt)** | 76.86 ms | 59.8% | 1.46 ms | 0.13% | Good | | 🇺🇸 Los Angeles (LAX) | **[132.1](/docs/network/latency/pairs/lax-ams-rtt)** | 87.74 ms | 66.4% | 0.78 ms | 0% | Good | | 🇭🇰 Hong Kong (HKG) | **[154.4](/docs/network/latency/pairs/hkg-ams-rtt)** | 91.05 ms | 59% | 0.87 ms | 0% | Fair | | 🇸🇬 Singapore (SIN) | **[155.8](/docs/network/latency/pairs/sin-ams-rtt)** | 102.86 ms | 66% | 1.04 ms | 0.19% | Fair | | 🇿🇦 Johannesburg (JNB) | **[163.2](/docs/network/latency/pairs/jnb-ams-rtt)** | 88.05 ms | 54% | 1.84 ms | 0% | Fair | | 🇹🇼 Taipei (TPE) | **[166.5](/docs/network/latency/pairs/tpe-ams-rtt)** | 92.72 ms | 55.7% | 0.78 ms | 0% | Fair | | 🇧🇷 São Paulo (GRU) | **[180](/docs/network/latency/pairs/gru-ams-rtt)** | 95.79 ms | 53.2% | 1.53 ms | 0% | Fair | | 🇯🇵 Tokyo (TYO) | **[197.4](/docs/network/latency/pairs/tyo-ams-rtt)** | 91.18 ms | 46.2% | 1.09 ms | 0% | Fair | | 🇦🇺 Melbourne (MEL) | **[252.8](/docs/network/latency/pairs/mel-ams-rtt)** | 161.98 ms | 64.1% | 1.86 ms | 0% | High | | 🇦🇺 Sydney (SYD) | **[258.9](/docs/network/latency/pairs/syd-ams-rtt)** | 162.93 ms | 62.9% | 1.18 ms | 0.16% | High | ## Europe Peers Amsterdam is one of 7 Hats Network PoPs in **Europe**. The table lists outbound measurements within the region. | PoP | Code | Country | RTT (ms) | Tier | | --------------------------------------- | ---- | ------- | -------- | --------- | | [🇩🇪 Berlin (BER)](./ber-berlin) | BER | Germany | **7.2** | Ultra-Low | | [🇩🇪 Frankfurt (FRA)](./fra-frankfurt) | FRA | Germany | **5.9** | Ultra-Low | | [🇬🇧 London (LON)](./lon-london) | LON | UK | **5.2** | Ultra-Low | | [🇫🇷 Marseille (MRS)](./mrs-marseille) | MRS | France | **19.6** | Ultra-Low | | [🇷🇺 Moscow (MOW)](./mow-moscow) | MOW | Russia | **38.1** | Excellent | | [🇫🇷 Paris (PAR)](./par-paris) | PAR | France | **7** | Ultra-Low | ## Questions ### What is the fastest route to Amsterdam? The fastest measured route to Amsterdam (AMS) is [London (LON) → Amsterdam (AMS)](/docs/network/latency/pairs/lon-ams-rtt), averaging **5.2 ms** RTT (Ultra-Low). ### What is the fastest route from Amsterdam? The fastest measured route from Amsterdam (AMS) is [Amsterdam (AMS) → London (LON)](/docs/network/latency/pairs/ams-lon-rtt), averaging **5.2 ms** RTT (Ultra-Low). ### How well connected is Amsterdam to the Hats Network backbone? Amsterdam (AMS) maintains measured routes to all 19 other PoPs across 5 continents. The slowest route, [Amsterdam (AMS) → Sydney (SYD)](/docs/network/latency/pairs/ams-syd-rtt), averages **258 ms** RTT. *** *Round published August 16, 2026. [All 20 PoPs](/docs/network/latency) · [Dataset and licence](/opendata/latency/).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/ams-amsterdam). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Berlin network latency (BER) Berlin (BER) is one of two German PoPs in this dataset. Compare its routes with Frankfurt when choosing a measurement origin. Measurement round: `2026-08-16T04:07:28Z`. The published round covers **19 outbound routes** from Berlin. Incoming routes are measured separately; neither direction describes last-mile performance. ## Location * **Location:** Berlin, Germany (Europe) * **Region:** Europe * **Role:** Backbone interconnection point for Hats Network (AS203314) * **Public city reference:** [GeoNames 2950159](https://www.geonames.org/2950159); 52.52437, 13.41053 * **Coverage:** RTT measurements to 19 other PoPs across 5 continents ### Interconnection * **Facility:** BER1 * **Upstream transit:** DE-CIX / RETN / NTT * **Peering / IX:** BCIX > Interactive content is available on the canonical HTML page. ## Round Summary | Metric | Value | | --------------------- | ----------------------------------------------------------------------------------------------------------- | | Fastest Route | [🇩🇪 Berlin (BER) → 🇩🇪 Frankfurt (FRA)](/docs/network/latency/pairs/ber-fra-rtt) (**6.1 ms**); Ultra-Low | | Slowest Route | [🇩🇪 Berlin (BER) → 🇦🇺 Sydney (SYD)](/docs/network/latency/pairs/ber-syd-rtt) (**262.1 ms**) | | Average RTT | **114.8 ms** | | Average Jitter | **1.11 ms** | | Average Packet Loss | **0.06%** | | Best Fiber Efficiency | **78.8%** | | Intra-Region Peers | 6 PoPs in Europe | | Total Routes | 19 outbound / 19 inbound (100% coverage) | > Jitter, packet-loss, percentile, and trend fields are estimated comparative indicators. Read the [theoretical fiber latency methodology](/docs/network/latency/theoretical-fiber-latency). ## Regional Routes Berlin's 6 other **Europe** PoPs have RTTs from **6.1 ms** (Frankfurt (FRA)) to **28.6 ms** (Moscow (MOW)). Berlin is in northeastern Germany, east of Frankfurt. ## Outbound Latency from Berlin (BER) Round-trip time in milliseconds **from Berlin** to all other backbone PoPs, sorted fastest to slowest. | Destination | RTT (ms) | Fiber Floor | Efficiency | Jitter | Loss | Tier | | ----------------------- | ---------------------------------------------------- | ----------- | ---------- | ------- | ----- | --------- | | 🇩🇪 Frankfurt (FRA) | **[6.1](/docs/network/latency/pairs/ber-fra-rtt)** | 4.16 ms | 68.2% | 0.12 ms | 0% | Ultra-Low | | 🇳🇱 Amsterdam (AMS) | **[7.2](/docs/network/latency/pairs/ber-ams-rtt)** | 5.67 ms | 78.8% | 0.12 ms | 0% | Ultra-Low | | 🇬🇧 London (LON) | **[15.2](/docs/network/latency/pairs/ber-lon-rtt)** | 9.15 ms | 60.2% | 0.15 ms | 0.14% | Ultra-Low | | 🇫🇷 Paris (PAR) | **[15.6](/docs/network/latency/pairs/ber-par-rtt)** | 8.62 ms | 55.3% | 0.16 ms | 0% | Ultra-Low | | 🇫🇷 Marseille (MRS) | **[20.2](/docs/network/latency/pairs/ber-mrs-rtt)** | 11.62 ms | 57.5% | 0.16 ms | 0.06% | Ultra-Low | | 🇷🇺 Moscow (MOW) | **[28.6](/docs/network/latency/pairs/ber-mow-rtt)** | 15.8 ms | 55.2% | 0.33 ms | 0% | Ultra-Low | | 🇺🇸 New York (NYC) | **[79.6](/docs/network/latency/pairs/ber-nyc-rtt)** | 62.7 ms | 78.8% | 0.5 ms | 0.08% | Excellent | | 🇺🇸 Ashburn (IAD) | **[85.7](/docs/network/latency/pairs/ber-iad-rtt)** | 66.05 ms | 77.1% | 0.79 ms | 0% | Good | | 🇺🇸 Miami (MIA) | **[118.3](/docs/network/latency/pairs/ber-mia-rtt)** | 78.44 ms | 66.3% | 0.76 ms | 0% | Good | | 🇺🇸 Seattle (SEA) | **[139.7](/docs/network/latency/pairs/ber-sea-rtt)** | 79.73 ms | 57.1% | 1.63 ms | 0.2% | Good | | 🇺🇸 Los Angeles (LAX) | **[140.9](/docs/network/latency/pairs/ber-lax-rtt)** | 91.39 ms | 64.9% | 1.62 ms | 0.1% | Good | | 🇭🇰 Hong Kong (HKG) | **[144.9](/docs/network/latency/pairs/ber-hkg-rtt)** | 85.85 ms | 59.3% | 1.29 ms | 0.08% | Good | | 🇹🇼 Taipei (TPE) | **[156.8](/docs/network/latency/pairs/ber-tpe-rtt)** | 87.84 ms | 56% | 1.45 ms | 0.17% | Fair | | 🇸🇬 Singapore (SIN) | **[158.3](/docs/network/latency/pairs/ber-sin-rtt)** | 97.19 ms | 61.4% | 0.87 ms | 0.19% | Fair | | 🇿🇦 Johannesburg (JNB) | **[172.1](/docs/network/latency/pairs/ber-jnb-rtt)** | 86.53 ms | 50.3% | 1.56 ms | 0% | Fair | | 🇧🇷 São Paulo (GRU) | **[186.5](/docs/network/latency/pairs/ber-gru-rtt)** | 100.2 ms | 53.7% | 2.2 ms | 0.01% | Fair | | 🇯🇵 Tokyo (TYO) | **[187.5](/docs/network/latency/pairs/ber-tyo-rtt)** | 87.53 ms | 46.7% | 1.21 ms | 0% | Fair | | 🇦🇺 Melbourne (MEL) | **[256.2](/docs/network/latency/pairs/ber-mel-rtt)** | 156.34 ms | 61% | 3.01 ms | 0.03% | High | | 🇦🇺 Sydney (SYD) | **[262.1](/docs/network/latency/pairs/ber-syd-rtt)** | 157.54 ms | 60.1% | 3.1 ms | 0.15% | High | ## Fastest Routes to Berlin (BER) The 10 fastest measured routes **to Berlin**, ranked by average RTT. Min / max / stdev are computed from the same measurement round (50 ICMP echo probes per route, 100 ms interval); routes without a published per-sample round show the measured average RTT only. | Rank | Source | Avg RTT | Min | Max | Stdev | | ---- | ---------------------- | ------------------------------------------------------- | --------- | --------- | ------- | | 1 | 🇩🇪 Frankfurt (FRA) | **[6.1 ms](/docs/network/latency/pairs/fra-ber-rtt)** | 5.91 ms | 6.79 ms | 0.18 ms | | 2 | 🇳🇱 Amsterdam (AMS) | **[7.2 ms](/docs/network/latency/pairs/ams-ber-rtt)** | 6.88 ms | 8.21 ms | 0.31 ms | | 3 | 🇫🇷 Paris (PAR) | **[15.5 ms](/docs/network/latency/pairs/par-ber-rtt)** | 14.87 ms | 17.28 ms | 0.57 ms | | 4 | 🇬🇧 London (LON) | **[16.2 ms](/docs/network/latency/pairs/lon-ber-rtt)** | 15.5 ms | 19.35 ms | 0.73 ms | | 5 | 🇫🇷 Marseille (MRS) | **[20.6 ms](/docs/network/latency/pairs/mrs-ber-rtt)** | 19.99 ms | 21.74 ms | 0.35 ms | | 6 | 🇷🇺 Moscow (MOW) | **[27.5 ms](/docs/network/latency/pairs/mow-ber-rtt)** | 26.72 ms | 29.96 ms | 0.74 ms | | 7 | 🇺🇸 New York (NYC) | **[78.8 ms](/docs/network/latency/pairs/nyc-ber-rtt)** | 77.15 ms | 84.9 ms | 1.59 ms | | 8 | 🇺🇸 Ashburn (IAD) | **[86.1 ms](/docs/network/latency/pairs/iad-ber-rtt)** | 84.07 ms | 94.45 ms | 2.05 ms | | 9 | 🇺🇸 Miami (MIA) | **[117.3 ms](/docs/network/latency/pairs/mia-ber-rtt)** | 113.83 ms | 127.02 ms | 2.73 ms | | 10 | 🇺🇸 Los Angeles (LAX) | **[139.3 ms](/docs/network/latency/pairs/lax-ber-rtt)** | 136.33 ms | 148.84 ms | 2.96 ms | ## Inbound Latency to Berlin (BER) Round-trip time in milliseconds **from all other PoPs to Berlin**. The measurement origin matters: inbound and outbound RTT can differ for the same pair. | Source | RTT (ms) | Fiber Floor | Efficiency | Jitter | Loss | Tier | | ----------------------- | ---------------------------------------------------- | ----------- | ---------- | ------- | ----- | --------- | | 🇩🇪 Frankfurt (FRA) | **[6.1](/docs/network/latency/pairs/fra-ber-rtt)** | 4.16 ms | 68.2% | 0.12 ms | 0% | Ultra-Low | | 🇳🇱 Amsterdam (AMS) | **[7.2](/docs/network/latency/pairs/ams-ber-rtt)** | 5.67 ms | 78.8% | 0.12 ms | 0.19% | Ultra-Low | | 🇫🇷 Paris (PAR) | **[15.5](/docs/network/latency/pairs/par-ber-rtt)** | 8.62 ms | 55.6% | 0.13 ms | 0% | Ultra-Low | | 🇬🇧 London (LON) | **[16.2](/docs/network/latency/pairs/lon-ber-rtt)** | 9.15 ms | 56.5% | 0.12 ms | 0.02% | Ultra-Low | | 🇫🇷 Marseille (MRS) | **[20.6](/docs/network/latency/pairs/mrs-ber-rtt)** | 11.62 ms | 56.4% | 0.24 ms | 0% | Ultra-Low | | 🇷🇺 Moscow (MOW) | **[27.5](/docs/network/latency/pairs/mow-ber-rtt)** | 15.8 ms | 57.5% | 0.3 ms | 0.13% | Ultra-Low | | 🇺🇸 New York (NYC) | **[78.8](/docs/network/latency/pairs/nyc-ber-rtt)** | 62.7 ms | 79.6% | 0.67 ms | 0% | Excellent | | 🇺🇸 Ashburn (IAD) | **[86.1](/docs/network/latency/pairs/iad-ber-rtt)** | 66.05 ms | 76.7% | 0.7 ms | 0% | Good | | 🇺🇸 Miami (MIA) | **[117.3](/docs/network/latency/pairs/mia-ber-rtt)** | 78.44 ms | 66.9% | 1.4 ms | 0% | Good | | 🇺🇸 Los Angeles (LAX) | **[139.3](/docs/network/latency/pairs/lax-ber-rtt)** | 91.39 ms | 65.6% | 1.58 ms | 0.12% | Good | | 🇺🇸 Seattle (SEA) | **[140.9](/docs/network/latency/pairs/sea-ber-rtt)** | 79.73 ms | 56.6% | 1.36 ms | 0.01% | Good | | 🇭🇰 Hong Kong (HKG) | **[145.1](/docs/network/latency/pairs/hkg-ber-rtt)** | 85.85 ms | 59.2% | 0.83 ms | 0.18% | Good | | 🇹🇼 Taipei (TPE) | **[158.6](/docs/network/latency/pairs/tpe-ber-rtt)** | 87.84 ms | 55.4% | 1.84 ms | 0.09% | Fair | | 🇸🇬 Singapore (SIN) | **[160.3](/docs/network/latency/pairs/sin-ber-rtt)** | 97.19 ms | 60.6% | 0.88 ms | 0.19% | Fair | | 🇿🇦 Johannesburg (JNB) | **[172.1](/docs/network/latency/pairs/jnb-ber-rtt)** | 86.53 ms | 50.3% | 1.75 ms | 0.11% | Fair | | 🇯🇵 Tokyo (TYO) | **[186.1](/docs/network/latency/pairs/tyo-ber-rtt)** | 87.53 ms | 47% | 1.64 ms | 0% | Fair | | 🇧🇷 São Paulo (GRU) | **[191.3](/docs/network/latency/pairs/gru-ber-rtt)** | 100.2 ms | 52.4% | 2.12 ms | 0% | Fair | | 🇦🇺 Melbourne (MEL) | **[254.7](/docs/network/latency/pairs/mel-ber-rtt)** | 156.34 ms | 61.4% | 1.95 ms | 0% | High | | 🇦🇺 Sydney (SYD) | **[260.5](/docs/network/latency/pairs/syd-ber-rtt)** | 157.54 ms | 60.5% | 2.14 ms | 0% | High | ## Europe Peers Berlin is one of 7 Hats Network PoPs in **Europe**. The table lists outbound measurements within the region. | PoP | Code | Country | RTT (ms) | Tier | | --------------------------------------- | ---- | ----------- | -------- | --------- | | [🇳🇱 Amsterdam (AMS)](./ams-amsterdam) | AMS | Netherlands | **7.2** | Ultra-Low | | [🇩🇪 Frankfurt (FRA)](./fra-frankfurt) | FRA | Germany | **6.1** | Ultra-Low | | [🇬🇧 London (LON)](./lon-london) | LON | UK | **15.2** | Ultra-Low | | [🇫🇷 Marseille (MRS)](./mrs-marseille) | MRS | France | **20.2** | Ultra-Low | | [🇷🇺 Moscow (MOW)](./mow-moscow) | MOW | Russia | **28.6** | Ultra-Low | | [🇫🇷 Paris (PAR)](./par-paris) | PAR | France | **15.6** | Ultra-Low | ## Questions ### What is the fastest route to Berlin? The fastest measured route to Berlin (BER) is [Frankfurt (FRA) → Berlin (BER)](/docs/network/latency/pairs/fra-ber-rtt), averaging **6.1 ms** RTT (Ultra-Low). ### What is the fastest route from Berlin? The fastest measured route from Berlin (BER) is [Berlin (BER) → Frankfurt (FRA)](/docs/network/latency/pairs/ber-fra-rtt), averaging **6.1 ms** RTT (Ultra-Low). ### How well connected is Berlin to the Hats Network backbone? Berlin (BER) maintains measured routes to all 19 other PoPs across 5 continents. The slowest route, [Berlin (BER) → Sydney (SYD)](/docs/network/latency/pairs/ber-syd-rtt), averages **262.1 ms** RTT. *** *Round published August 16, 2026. [All 20 PoPs](/docs/network/latency) · [Dataset and licence](/opendata/latency/).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/ber-berlin). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Frankfurt network latency (FRA) Frankfurt (FRA) is a Hats Network PoP in Germany. Its measurements cover European peers and destinations on four other continents. Measurement round: `2026-08-16T04:07:28Z`. The published round covers **19 outbound routes** from Frankfurt. Incoming routes are measured separately; neither direction describes last-mile performance. ## Location * **Location:** Frankfurt, Germany (Europe) * **Region:** Europe * **Role:** Backbone interconnection point for Hats Network (AS203314) * **Public city reference:** [GeoNames 2925533](https://www.geonames.org/2925533); 50.11552, 8.68417 * **Coverage:** RTT measurements to 19 other PoPs across 5 continents ### Interconnection * **Facility:** FRA1 > Interactive content is available on the canonical HTML page. ## Round Summary | Metric | Value | | --------------------- | -------------------------------------------------------------------------------------------------------------- | | Fastest Route | [🇩🇪 Frankfurt (FRA) → 🇳🇱 Amsterdam (AMS)](/docs/network/latency/pairs/fra-ams-rtt) (**5.9 ms**); Ultra-Low | | Slowest Route | [🇩🇪 Frankfurt (FRA) → 🇦🇺 Sydney (SYD)](/docs/network/latency/pairs/fra-syd-rtt) (**256 ms**) | | Average RTT | **114.0 ms** | | Average Jitter | **0.95 ms** | | Average Packet Loss | **0.05%** | | Best Fiber Efficiency | **81.9%** | | Intra-Region Peers | 6 PoPs in Europe | | Total Routes | 19 outbound / 19 inbound (100% coverage) | > Jitter, packet-loss, percentile, and trend fields are estimated comparative indicators. Read the [theoretical fiber latency methodology](/docs/network/latency/theoretical-fiber-latency). ## Regional Routes Frankfurt's 6 other **Europe** PoPs have RTTs from **5.9 ms** (Amsterdam (AMS)) to **34.7 ms** (Moscow (MOW)). Frankfurt is in western Germany, between the Amsterdam and Berlin measurement origins. ## Outbound Latency from Frankfurt (FRA) Round-trip time in milliseconds **from Frankfurt** to all other backbone PoPs, sorted fastest to slowest. | Destination | RTT (ms) | Fiber Floor | Efficiency | Jitter | Loss | Tier | | ----------------------- | ---------------------------------------------------- | ----------- | ---------- | ------- | ----- | --------- | | 🇳🇱 Amsterdam (AMS) | **[5.9](/docs/network/latency/pairs/fra-ams-rtt)** | 3.57 ms | 60.6% | 0.12 ms | 0% | Ultra-Low | | 🇩🇪 Berlin (BER) | **[6.1](/docs/network/latency/pairs/fra-ber-rtt)** | 4.16 ms | 68.2% | 0.12 ms | 0% | Ultra-Low | | 🇫🇷 Paris (PAR) | **[7.6](/docs/network/latency/pairs/fra-par-rtt)** | 4.7 ms | 61.8% | 0.12 ms | 0% | Ultra-Low | | 🇬🇧 London (LON) | **[12.9](/docs/network/latency/pairs/fra-lon-rtt)** | 6.26 ms | 48.5% | 0.12 ms | 0% | Ultra-Low | | 🇫🇷 Marseille (MRS) | **[16.4](/docs/network/latency/pairs/fra-mrs-rtt)** | 7.82 ms | 47.7% | 0.19 ms | 0% | Ultra-Low | | 🇷🇺 Moscow (MOW) | **[34.7](/docs/network/latency/pairs/fra-mow-rtt)** | 19.85 ms | 57.2% | 0.25 ms | 0.13% | Excellent | | 🇺🇸 New York (NYC) | **[74.4](/docs/network/latency/pairs/fra-nyc-rtt)** | 60.9 ms | 81.9% | 0.73 ms | 0% | Excellent | | 🇺🇸 Ashburn (IAD) | **[80.7](/docs/network/latency/pairs/fra-iad-rtt)** | 64.3 ms | 79.7% | 0.91 ms | 0.04% | Good | | 🇺🇸 Miami (MIA) | **[112](/docs/network/latency/pairs/fra-mia-rtt)** | 76.17 ms | 68% | 1.27 ms | 0.07% | Good | | 🇺🇸 Seattle (SEA) | **[135.2](/docs/network/latency/pairs/fra-sea-rtt)** | 80.34 ms | 59.4% | 0.98 ms | 0.16% | Good | | 🇺🇸 Los Angeles (LAX) | **[141.8](/docs/network/latency/pairs/fra-lax-rtt)** | 91.31 ms | 64.4% | 1.06 ms | 0% | Good | | 🇸🇬 Singapore (SIN) | **[151.4](/docs/network/latency/pairs/fra-sin-rtt)** | 100.56 ms | 66.4% | 0.7 ms | 0% | Fair | | 🇭🇰 Hong Kong (HKG) | **[152](/docs/network/latency/pairs/fra-hkg-rtt)** | 89.87 ms | 59.1% | 1.18 ms | 0% | Fair | | 🇹🇼 Taipei (TPE) | **[163.8](/docs/network/latency/pairs/fra-tpe-rtt)** | 91.96 ms | 56.1% | 1.28 ms | 0.04% | Fair | | 🇧🇷 São Paulo (GRU) | **[181.3](/docs/network/latency/pairs/fra-gru-rtt)** | 96.04 ms | 53% | 2.02 ms | 0.04% | Fair | | 🇿🇦 Johannesburg (JNB) | **[188.4](/docs/network/latency/pairs/fra-jnb-rtt)** | 84.87 ms | 45% | 1.61 ms | 0.1% | Fair | | 🇯🇵 Tokyo (TYO) | **[194.5](/docs/network/latency/pairs/fra-tyo-rtt)** | 91.61 ms | 47.1% | 1.29 ms | 0.12% | Fair | | 🇦🇺 Melbourne (MEL) | **[250.1](/docs/network/latency/pairs/fra-mel-rtt)** | 159.77 ms | 63.9% | 2.57 ms | 0% | High | | 🇦🇺 Sydney (SYD) | **[256](/docs/network/latency/pairs/fra-syd-rtt)** | 161.37 ms | 63% | 1.5 ms | 0.16% | High | ## Fastest Routes to Frankfurt (FRA) The 10 fastest measured routes **to Frankfurt**, ranked by average RTT. Min / max / stdev are computed from the same measurement round (50 ICMP echo probes per route, 100 ms interval); routes without a published per-sample round show the measured average RTT only. | Rank | Source | Avg RTT | Min | Max | Stdev | | ---- | -------------------- | ------------------------------------------------------- | --------- | --------- | ------- | | 1 | 🇳🇱 Amsterdam (AMS) | **[5.9 ms](/docs/network/latency/pairs/ams-fra-rtt)** | 5.67 ms | 7.15 ms | 0.23 ms | | 2 | 🇩🇪 Berlin (BER) | **[6.1 ms](/docs/network/latency/pairs/ber-fra-rtt)** | 5.9 ms | 6.73 ms | 0.18 ms | | 3 | 🇫🇷 Paris (PAR) | **[7.6 ms](/docs/network/latency/pairs/par-fra-rtt)** | 7.35 ms | 8.42 ms | 0.22 ms | | 4 | 🇬🇧 London (LON) | **[13.6 ms](/docs/network/latency/pairs/lon-fra-rtt)** | 13.18 ms | 15.43 ms | 0.42 ms | | 5 | 🇫🇷 Marseille (MRS) | **[16 ms](/docs/network/latency/pairs/mrs-fra-rtt)** | 15.12 ms | 18.91 ms | 0.79 ms | | 6 | 🇷🇺 Moscow (MOW) | **[35.9 ms](/docs/network/latency/pairs/mow-fra-rtt)** | 33.56 ms | 42.93 ms | 1.78 ms | | 7 | 🇺🇸 New York (NYC) | **[73.9 ms](/docs/network/latency/pairs/nyc-fra-rtt)** | 71.94 ms | 81.47 ms | 1.98 ms | | 8 | 🇺🇸 Ashburn (IAD) | **[81.2 ms](/docs/network/latency/pairs/iad-fra-rtt)** | 79.1 ms | 86.57 ms | 1.81 ms | | 9 | 🇺🇸 Miami (MIA) | **[112.2 ms](/docs/network/latency/pairs/mia-fra-rtt)** | 108.19 ms | 121.5 ms | 3.61 ms | | 10 | 🇺🇸 Seattle (SEA) | **[133.2 ms](/docs/network/latency/pairs/sea-fra-rtt)** | 130.2 ms | 141.75 ms | 2.72 ms | ## Inbound Latency to Frankfurt (FRA) Round-trip time in milliseconds **from all other PoPs to Frankfurt**. The measurement origin matters: inbound and outbound RTT can differ for the same pair. | Source | RTT (ms) | Fiber Floor | Efficiency | Jitter | Loss | Tier | | ----------------------- | ---------------------------------------------------- | ----------- | ---------- | ------- | ----- | --------- | | 🇳🇱 Amsterdam (AMS) | **[5.9](/docs/network/latency/pairs/ams-fra-rtt)** | 3.57 ms | 60.6% | 0.12 ms | 0.05% | Ultra-Low | | 🇩🇪 Berlin (BER) | **[6.1](/docs/network/latency/pairs/ber-fra-rtt)** | 4.16 ms | 68.2% | 0.12 ms | 0% | Ultra-Low | | 🇫🇷 Paris (PAR) | **[7.6](/docs/network/latency/pairs/par-fra-rtt)** | 4.7 ms | 61.8% | 0.12 ms | 0.1% | Ultra-Low | | 🇬🇧 London (LON) | **[13.6](/docs/network/latency/pairs/lon-fra-rtt)** | 6.26 ms | 46.1% | 0.12 ms | 0.03% | Ultra-Low | | 🇫🇷 Marseille (MRS) | **[16](/docs/network/latency/pairs/mrs-fra-rtt)** | 7.82 ms | 48.9% | 0.18 ms | 0.15% | Ultra-Low | | 🇷🇺 Moscow (MOW) | **[35.9](/docs/network/latency/pairs/mow-fra-rtt)** | 19.85 ms | 55.3% | 0.3 ms | 0% | Excellent | | 🇺🇸 New York (NYC) | **[73.9](/docs/network/latency/pairs/nyc-fra-rtt)** | 60.9 ms | 82.4% | 0.74 ms | 0.09% | Excellent | | 🇺🇸 Ashburn (IAD) | **[81.2](/docs/network/latency/pairs/iad-fra-rtt)** | 64.3 ms | 79.2% | 0.95 ms | 0.1% | Good | | 🇺🇸 Miami (MIA) | **[112.2](/docs/network/latency/pairs/mia-fra-rtt)** | 76.17 ms | 67.9% | 0.83 ms | 0.06% | Good | | 🇺🇸 Seattle (SEA) | **[133.2](/docs/network/latency/pairs/sea-fra-rtt)** | 80.34 ms | 60.3% | 0.66 ms | 0.11% | Good | | 🇺🇸 Los Angeles (LAX) | **[142.3](/docs/network/latency/pairs/lax-fra-rtt)** | 91.31 ms | 64.2% | 0.71 ms | 0% | Good | | 🇭🇰 Hong Kong (HKG) | **[150.7](/docs/network/latency/pairs/hkg-fra-rtt)** | 89.87 ms | 59.6% | 1.66 ms | 0% | Fair | | 🇸🇬 Singapore (SIN) | **[151.6](/docs/network/latency/pairs/sin-fra-rtt)** | 100.56 ms | 66.3% | 0.77 ms | 0.04% | Fair | | 🇹🇼 Taipei (TPE) | **[166](/docs/network/latency/pairs/tpe-fra-rtt)** | 91.96 ms | 55.4% | 1.67 ms | 0.13% | Fair | | 🇿🇦 Johannesburg (JNB) | **[166](/docs/network/latency/pairs/jnb-fra-rtt)** | 84.87 ms | 51.1% | 1.86 ms | 0.11% | Fair | | 🇧🇷 São Paulo (GRU) | **[186.2](/docs/network/latency/pairs/gru-fra-rtt)** | 96.04 ms | 51.6% | 2 ms | 0% | Fair | | 🇯🇵 Tokyo (TYO) | **[193.2](/docs/network/latency/pairs/tyo-fra-rtt)** | 91.61 ms | 47.4% | 2.16 ms | 0% | Fair | | 🇦🇺 Melbourne (MEL) | **[248.6](/docs/network/latency/pairs/mel-fra-rtt)** | 159.77 ms | 64.3% | 2.09 ms | 0.16% | Fair | | 🇦🇺 Sydney (SYD) | **[254.4](/docs/network/latency/pairs/syd-fra-rtt)** | 161.37 ms | 63.4% | 2.54 ms | 0% | High | ## Europe Peers Frankfurt is one of 7 Hats Network PoPs in **Europe**. The table lists outbound measurements within the region. | PoP | Code | Country | RTT (ms) | Tier | | --------------------------------------- | ---- | ----------- | -------- | --------- | | [🇳🇱 Amsterdam (AMS)](./ams-amsterdam) | AMS | Netherlands | **5.9** | Ultra-Low | | [🇩🇪 Berlin (BER)](./ber-berlin) | BER | Germany | **6.1** | Ultra-Low | | [🇬🇧 London (LON)](./lon-london) | LON | UK | **12.9** | Ultra-Low | | [🇫🇷 Marseille (MRS)](./mrs-marseille) | MRS | France | **16.4** | Ultra-Low | | [🇷🇺 Moscow (MOW)](./mow-moscow) | MOW | Russia | **34.7** | Excellent | | [🇫🇷 Paris (PAR)](./par-paris) | PAR | France | **7.6** | Ultra-Low | ## Questions ### What is the fastest route to Frankfurt? The fastest measured route to Frankfurt (FRA) is [Amsterdam (AMS) → Frankfurt (FRA)](/docs/network/latency/pairs/ams-fra-rtt), averaging **5.9 ms** RTT (Ultra-Low). ### What is the fastest route from Frankfurt? The fastest measured route from Frankfurt (FRA) is [Frankfurt (FRA) → Amsterdam (AMS)](/docs/network/latency/pairs/fra-ams-rtt), averaging **5.9 ms** RTT (Ultra-Low). ### How well connected is Frankfurt to the Hats Network backbone? Frankfurt (FRA) maintains measured routes to all 19 other PoPs across 5 continents. The slowest route, [Frankfurt (FRA) → Sydney (SYD)](/docs/network/latency/pairs/fra-syd-rtt), averages **256 ms** RTT. *** *Round published August 16, 2026. [All 20 PoPs](/docs/network/latency) · [Dataset and licence](/opendata/latency/).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/fra-frankfurt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # São Paulo network latency (GRU) São Paulo (GRU) is the South American PoP in this dataset. Every listed peer is in another region. Measurement round: `2026-08-16T04:07:28Z`. The published round covers **19 outbound routes** from São Paulo. Incoming routes are measured separately; neither direction describes last-mile performance. ## Location * **Location:** São Paulo, Brazil (South America) * **Region:** South America * **Role:** Backbone interconnection point for Hats Network (AS203314) * **Public city reference:** [GeoNames 3448439](https://www.geonames.org/3448439); -23.54750, -46.63611 * **Coverage:** RTT measurements to 19 other PoPs across 4 continents ### Interconnection * **Facility:** GRU1 * **Upstream transit:** NTT / Ascenty > Interactive content is available on the canonical HTML page. ## Round Summary | Metric | Value | | --------------------- | --------------------------------------------------------------------------------------------------------- | | Fastest Route | [🇧🇷 São Paulo (GRU) → 🇺🇸 Miami (MIA)](/docs/network/latency/pairs/gru-mia-rtt) (**74 ms**); Excellent | | Slowest Route | [🇧🇷 São Paulo (GRU) → 🇿🇦 Johannesburg (JNB)](/docs/network/latency/pairs/gru-jnb-rtt) (**333.8 ms**) | | Average RTT | **203.4 ms** | | Average Jitter | **1.75 ms** | | Average Packet Loss | **0.05%** | | Best Fiber Efficiency | **86.6%** | | Intra-Region Peers | 0 PoP in South America | | Total Routes | 19 outbound / 19 inbound (100% coverage) | > Jitter, packet-loss, percentile, and trend fields are estimated comparative indicators. Read the [theoretical fiber latency methodology](/docs/network/latency/theoretical-fiber-latency). ## Regional Routes São Paulo is Hats Network's only PoP in **South America**. The lowest RTT in this round is to Miami (MIA) at **74 ms** RTT. São Paulo is in southeastern Brazil and is the dataset's only South American city. ## Outbound Latency from São Paulo (GRU) Round-trip time in milliseconds **from São Paulo** to all other backbone PoPs, sorted fastest to slowest. | Destination | RTT (ms) | Fiber Floor | Efficiency | Jitter | Loss | Tier | | ----------------------- | ---------------------------------------------------- | ----------- | ---------- | ------- | ----- | --------- | | 🇺🇸 Miami (MIA) | **[74](/docs/network/latency/pairs/gru-mia-rtt)** | 64.11 ms | 86.6% | 0.62 ms | 0.16% | Excellent | | 🇺🇸 New York (NYC) | **[122](/docs/network/latency/pairs/gru-nyc-rtt)** | 74.99 ms | 61.5% | 1.29 ms | 0% | Good | | 🇺🇸 Ashburn (IAD) | **[128](/docs/network/latency/pairs/gru-iad-rtt)** | 74.79 ms | 58.4% | 0.6 ms | 0.09% | Good | | 🇺🇸 Los Angeles (LAX) | **[130.8](/docs/network/latency/pairs/gru-lax-rtt)** | 96.89 ms | 74.1% | 0.66 ms | 0% | Good | | 🇺🇸 Seattle (SEA) | **[155.9](/docs/network/latency/pairs/gru-sea-rtt)** | 106.76 ms | 68.5% | 1.54 ms | 0.19% | Fair | | 🇬🇧 London (LON) | **[175.8](/docs/network/latency/pairs/gru-lon-rtt)** | 92.76 ms | 52.8% | 1.62 ms | 0% | Fair | | 🇳🇱 Amsterdam (AMS) | **[180](/docs/network/latency/pairs/gru-ams-rtt)** | 95.79 ms | 53.2% | 1.53 ms | 0% | Fair | | 🇫🇷 Paris (PAR) | **[180.2](/docs/network/latency/pairs/gru-par-rtt)** | 91.83 ms | 51% | 1.95 ms | 0.14% | Fair | | 🇩🇪 Frankfurt (FRA) | **[186.2](/docs/network/latency/pairs/gru-fra-rtt)** | 96.04 ms | 51.6% | 2 ms | 0% | Fair | | 🇫🇷 Marseille (MRS) | **[187.7](/docs/network/latency/pairs/gru-mrs-rtt)** | 89.24 ms | 47.5% | 1.08 ms | 0.14% | Fair | | 🇩🇪 Berlin (BER) | **[191.3](/docs/network/latency/pairs/gru-ber-rtt)** | 100.2 ms | 52.4% | 2.12 ms | 0% | Fair | | 🇷🇺 Moscow (MOW) | **[219.5](/docs/network/latency/pairs/gru-mow-rtt)** | 115.47 ms | 52.6% | 1.57 ms | 0% | Fair | | 🇯🇵 Tokyo (TYO) | **[230.3](/docs/network/latency/pairs/gru-tyo-rtt)** | 181.46 ms | 78.8% | 2.42 ms | 0% | Fair | | 🇹🇼 Taipei (TPE) | **[259.1](/docs/network/latency/pairs/gru-tpe-rtt)** | 184.29 ms | 71.1% | 3.01 ms | 0% | High | | 🇦🇺 Sydney (SYD) | **[265.5](/docs/network/latency/pairs/gru-syd-rtt)** | 131 ms | 49.3% | 1.66 ms | 0% | High | | 🇭🇰 Hong Kong (HKG) | **[273.5](/docs/network/latency/pairs/gru-hkg-rtt)** | 176.86 ms | 64.7% | 2.57 ms | 0% | High | | 🇦🇺 Melbourne (MEL) | **[274.7](/docs/network/latency/pairs/gru-mel-rtt)** | 128.34 ms | 46.7% | 1.71 ms | 0.08% | High | | 🇸🇬 Singapore (SIN) | **[297.1](/docs/network/latency/pairs/gru-sin-rtt)** | 156.67 ms | 52.7% | 2.14 ms | 0.08% | High | | 🇿🇦 Johannesburg (JNB) | **[333.8](/docs/network/latency/pairs/gru-jnb-rtt)** | 72.88 ms | 21.8% | 3.1 ms | 0.07% | High | ## Fastest Routes to São Paulo (GRU) The 10 fastest measured routes **to São Paulo**, ranked by average RTT. Min / max / stdev are computed from the same measurement round (50 ICMP echo probes per route, 100 ms interval); routes without a published per-sample round show the measured average RTT only. | Rank | Source | Avg RTT | Min | Max | Stdev | | ---- | ---------------------- | ------------------------------------------------------- | --------- | --------- | ------- | | 1 | 🇺🇸 Ashburn (IAD) | **[101.4 ms](/docs/network/latency/pairs/iad-gru-rtt)** | 98.16 ms | 107.74 ms | 2.31 ms | | 2 | 🇺🇸 New York (NYC) | **[106.9 ms](/docs/network/latency/pairs/nyc-gru-rtt)** | 104.25 ms | 117.31 ms | 2.21 ms | | 3 | 🇺🇸 Miami (MIA) | **[128.7 ms](/docs/network/latency/pairs/mia-gru-rtt)** | 126.78 ms | 136.62 ms | 1.96 ms | | 4 | 🇺🇸 Los Angeles (LAX) | **[131.3 ms](/docs/network/latency/pairs/lax-gru-rtt)** | 125.74 ms | 147.12 ms | 4.94 ms | | 5 | 🇺🇸 Seattle (SEA) | **[155.7 ms](/docs/network/latency/pairs/sea-gru-rtt)** | 152.75 ms | 163 ms | 2.37 ms | | 6 | 🇬🇧 London (LON) | **[170.7 ms](/docs/network/latency/pairs/lon-gru-rtt)** | 165.28 ms | 185.16 ms | 4.35 ms | | 7 | 🇳🇱 Amsterdam (AMS) | **[175.5 ms](/docs/network/latency/pairs/ams-gru-rtt)** | 166.82 ms | 195.11 ms | 6.86 ms | | 8 | 🇫🇷 Paris (PAR) | **[175.8 ms](/docs/network/latency/pairs/par-gru-rtt)** | 173.09 ms | 186.7 ms | 2.78 ms | | 9 | 🇩🇪 Frankfurt (FRA) | **[181.3 ms](/docs/network/latency/pairs/fra-gru-rtt)** | 176.28 ms | 197.75 ms | 3.73 ms | | 10 | 🇩🇪 Berlin (BER) | **[186.5 ms](/docs/network/latency/pairs/ber-gru-rtt)** | 179.2 ms | 206.69 ms | 6.78 ms | ## Inbound Latency to São Paulo (GRU) Round-trip time in milliseconds **from all other PoPs to São Paulo**. The measurement origin matters: inbound and outbound RTT can differ for the same pair. | Source | RTT (ms) | Fiber Floor | Efficiency | Jitter | Loss | Tier | | ----------------------- | ---------------------------------------------------- | ----------- | ---------- | ------- | ----- | ---- | | 🇺🇸 Ashburn (IAD) | **[101.4](/docs/network/latency/pairs/iad-gru-rtt)** | 74.79 ms | 73.8% | 1.02 ms | 0% | Good | | 🇺🇸 New York (NYC) | **[106.9](/docs/network/latency/pairs/nyc-gru-rtt)** | 74.99 ms | 70.1% | 0.74 ms | 0% | Good | | 🇺🇸 Miami (MIA) | **[128.7](/docs/network/latency/pairs/mia-gru-rtt)** | 64.11 ms | 49.8% | 0.81 ms | 0.16% | Good | | 🇺🇸 Los Angeles (LAX) | **[131.3](/docs/network/latency/pairs/lax-gru-rtt)** | 96.89 ms | 73.8% | 0.87 ms | 0% | Good | | 🇺🇸 Seattle (SEA) | **[155.7](/docs/network/latency/pairs/sea-gru-rtt)** | 106.76 ms | 68.6% | 1.76 ms | 0% | Fair | | 🇬🇧 London (LON) | **[170.7](/docs/network/latency/pairs/lon-gru-rtt)** | 92.76 ms | 54.3% | 1.06 ms | 0.02% | Fair | | 🇳🇱 Amsterdam (AMS) | **[175.5](/docs/network/latency/pairs/ams-gru-rtt)** | 95.79 ms | 54.6% | 1.53 ms | 0% | Fair | | 🇫🇷 Paris (PAR) | **[175.8](/docs/network/latency/pairs/par-gru-rtt)** | 91.83 ms | 52.2% | 1.26 ms | 0% | Fair | | 🇩🇪 Frankfurt (FRA) | **[181.3](/docs/network/latency/pairs/fra-gru-rtt)** | 96.04 ms | 53% | 2.02 ms | 0.04% | Fair | | 🇩🇪 Berlin (BER) | **[186.5](/docs/network/latency/pairs/ber-gru-rtt)** | 100.2 ms | 53.7% | 2.2 ms | 0.01% | Fair | | 🇫🇷 Marseille (MRS) | **[186.7](/docs/network/latency/pairs/mrs-gru-rtt)** | 89.24 ms | 47.8% | 1.4 ms | 0.05% | Fair | | 🇷🇺 Moscow (MOW) | **[214.4](/docs/network/latency/pairs/mow-gru-rtt)** | 115.47 ms | 53.9% | 1.47 ms | 0% | Fair | | 🇯🇵 Tokyo (TYO) | **[230.5](/docs/network/latency/pairs/tyo-gru-rtt)** | 181.46 ms | 78.7% | 1.53 ms | 0% | Fair | | 🇹🇼 Taipei (TPE) | **[273](/docs/network/latency/pairs/tpe-gru-rtt)** | 184.29 ms | 67.5% | 2.95 ms | 0.03% | High | | 🇭🇰 Hong Kong (HKG) | **[287.6](/docs/network/latency/pairs/hkg-gru-rtt)** | 176.86 ms | 61.5% | 2.69 ms | 0.03% | High | | 🇦🇺 Sydney (SYD) | **[296.7](/docs/network/latency/pairs/syd-gru-rtt)** | 131 ms | 44.2% | 1.93 ms | 0% | High | | 🇦🇺 Melbourne (MEL) | **[305.3](/docs/network/latency/pairs/mel-gru-rtt)** | 128.34 ms | 42% | 1.69 ms | 0% | High | | 🇸🇬 Singapore (SIN) | **[311.4](/docs/network/latency/pairs/sin-gru-rtt)** | 156.67 ms | 50.3% | 2.4 ms | 0.1% | High | | 🇿🇦 Johannesburg (JNB) | **[328.7](/docs/network/latency/pairs/jnb-gru-rtt)** | 72.88 ms | 22.2% | 2.32 ms | 0.06% | High | ## Questions ### What is the fastest route to São Paulo? The fastest measured route to São Paulo (GRU) is [Ashburn (IAD) → São Paulo (GRU)](/docs/network/latency/pairs/iad-gru-rtt), averaging **101.4 ms** RTT (Good). ### What is the fastest route from São Paulo? The fastest measured route from São Paulo (GRU) is [São Paulo (GRU) → Miami (MIA)](/docs/network/latency/pairs/gru-mia-rtt), averaging **74 ms** RTT (Excellent). ### How well connected is São Paulo to the Hats Network backbone? São Paulo (GRU) maintains measured routes to all 19 other PoPs across 4 continents. The slowest route, [São Paulo (GRU) → Johannesburg (JNB)](/docs/network/latency/pairs/gru-jnb-rtt), averages **333.8 ms** RTT. *** *Round published August 16, 2026. [All 20 PoPs](/docs/network/latency) · [Dataset and licence](/opendata/latency/).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/gru-sao-paulo). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Hong Kong network latency (HKG) Hong Kong (HKG) is one of the network's East Asian PoPs. Compare the regional routes with Taipei, Tokyo and Singapore. Measurement round: `2026-08-16T04:07:28Z`. The published round covers **19 outbound routes** from Hong Kong. Incoming routes are measured separately; neither direction describes last-mile performance. ## Location * **Location:** Hong Kong, Hong Kong (Asia Pacific) * **Region:** Asia Pacific * **Role:** Backbone interconnection point for Hats Network (AS203314) * **Public city reference:** [GeoNames 1819729](https://www.geonames.org/1819729); 22.27832, 114.17469 * **Coverage:** RTT measurements to 19 other PoPs across 5 continents ### Interconnection * **Facility:** HKG3 * **Upstream transit:** HE / Cogent / CDN77 * **Peering / IX:** Equinix IX / HKIX > Interactive content is available on the canonical HTML page. ## Round Summary | Metric | Value | | --------------------- | ------------------------------------------------------------------------------------------------------------ | | Fastest Route | [🇭🇰 Hong Kong (HKG) → 🇹🇼 Taipei (TPE)](/docs/network/latency/pairs/hkg-tpe-rtt) (**14.7 ms**); Ultra-Low | | Slowest Route | [🇭🇰 Hong Kong (HKG) → 🇿🇦 Johannesburg (JNB)](/docs/network/latency/pairs/hkg-jnb-rtt) (**316.7 ms**) | | Average RTT | **151.1 ms** | | Average Jitter | **1.15 ms** | | Average Packet Loss | **0.06%** | | Best Fiber Efficiency | **80.1%** | | Intra-Region Peers | 5 PoPs in Asia Pacific | | Total Routes | 19 outbound / 19 inbound (100% coverage) | > Jitter, packet-loss, percentile, and trend fields are estimated comparative indicators. Read the [theoretical fiber latency methodology](/docs/network/latency/theoretical-fiber-latency). ## Regional Routes Hong Kong's 5 other **Asia Pacific** PoPs have RTTs from **14.7 ms** (Taipei (TPE)) to **138.4 ms** (Melbourne (MEL)). Hong Kong is on the coast of southern China, southwest of Taipei. ## Outbound Latency from Hong Kong (HKG) Round-trip time in milliseconds **from Hong Kong** to all other backbone PoPs, sorted fastest to slowest. | Destination | RTT (ms) | Fiber Floor | Efficiency | Jitter | Loss | Tier | | ----------------------- | ---------------------------------------------------- | ----------- | ---------- | ------- | ----- | --------- | | 🇹🇼 Taipei (TPE) | **[14.7](/docs/network/latency/pairs/hkg-tpe-rtt)** | 7.94 ms | 54% | 0.14 ms | 0.08% | Ultra-Low | | 🇸🇬 Singapore (SIN) | **[31.5](/docs/network/latency/pairs/hkg-sin-rtt)** | 25.24 ms | 80.1% | 0.33 ms | 0.05% | Excellent | | 🇯🇵 Tokyo (TYO) | **[44.9](/docs/network/latency/pairs/hkg-tyo-rtt)** | 28.24 ms | 62.9% | 0.23 ms | 0.08% | Excellent | | 🇷🇺 Moscow (MOW) | **[118.2](/docs/network/latency/pairs/hkg-mow-rtt)** | 70.05 ms | 59.3% | 1.17 ms | 0.11% | Good | | 🇺🇸 Seattle (SEA) | **[131.6](/docs/network/latency/pairs/hkg-sea-rtt)** | 102.2 ms | 77.7% | 0.82 ms | 0.05% | Good | | 🇦🇺 Sydney (SYD) | **[137.2](/docs/network/latency/pairs/hkg-syd-rtt)** | 71.97 ms | 52.5% | 0.86 ms | 0.17% | Good | | 🇦🇺 Melbourne (MEL) | **[138.4](/docs/network/latency/pairs/hkg-mel-rtt)** | 72.38 ms | 52.3% | 1.41 ms | 0% | Good | | 🇺🇸 Los Angeles (LAX) | **[145](/docs/network/latency/pairs/hkg-lax-rtt)** | 114.3 ms | 78.8% | 1.4 ms | 0.06% | Good | | 🇩🇪 Berlin (BER) | **[145.1](/docs/network/latency/pairs/hkg-ber-rtt)** | 85.85 ms | 59.2% | 0.83 ms | 0.18% | Good | | 🇩🇪 Frankfurt (FRA) | **[150.7](/docs/network/latency/pairs/hkg-fra-rtt)** | 89.87 ms | 59.6% | 1.66 ms | 0% | Fair | | 🇳🇱 Amsterdam (AMS) | **[154.4](/docs/network/latency/pairs/hkg-ams-rtt)** | 91.05 ms | 59% | 0.87 ms | 0% | Fair | | 🇬🇧 London (LON) | **[158.7](/docs/network/latency/pairs/hkg-lon-rtt)** | 94.44 ms | 59.5% | 0.8 ms | 0% | Fair | | 🇫🇷 Paris (PAR) | **[159.4](/docs/network/latency/pairs/hkg-par-rtt)** | 94.48 ms | 59.3% | 1.21 ms | 0.07% | Fair | | 🇫🇷 Marseille (MRS) | **[165.4](/docs/network/latency/pairs/hkg-mrs-rtt)** | 95.48 ms | 57.7% | 0.85 ms | 0.09% | Fair | | 🇺🇸 New York (NYC) | **[184.7](/docs/network/latency/pairs/hkg-nyc-rtt)** | 127.11 ms | 68.8% | 0.98 ms | 0% | Fair | | 🇺🇸 Ashburn (IAD) | **[186.2](/docs/network/latency/pairs/hkg-iad-rtt)** | 128.37 ms | 68.9% | 1.54 ms | 0.04% | Fair | | 🇺🇸 Miami (MIA) | **[201.1](/docs/network/latency/pairs/hkg-mia-rtt)** | 141.69 ms | 70.5% | 0.93 ms | 0.05% | Fair | | 🇧🇷 São Paulo (GRU) | **[287.6](/docs/network/latency/pairs/hkg-gru-rtt)** | 176.86 ms | 61.5% | 2.69 ms | 0.03% | High | | 🇿🇦 Johannesburg (JNB) | **[316.7](/docs/network/latency/pairs/hkg-jnb-rtt)** | 104.93 ms | 33.1% | 3.12 ms | 0% | High | ## Fastest Routes to Hong Kong (HKG) The 10 fastest measured routes **to Hong Kong**, ranked by average RTT. Min / max / stdev are computed from the same measurement round (50 ICMP echo probes per route, 100 ms interval); routes without a published per-sample round show the measured average RTT only. | Rank | Source | Avg RTT | Min | Max | Stdev | | ---- | ---------------------- | ------------------------------------------------------- | --------- | --------- | ------- | | 1 | 🇹🇼 Taipei (TPE) | **[15.5 ms](/docs/network/latency/pairs/tpe-hkg-rtt)** | 15.09 ms | 17.58 ms | 0.41 ms | | 2 | 🇸🇬 Singapore (SIN) | **[30.8 ms](/docs/network/latency/pairs/sin-hkg-rtt)** | 30.29 ms | 33.37 ms | 0.58 ms | | 3 | 🇯🇵 Tokyo (TYO) | **[44.8 ms](/docs/network/latency/pairs/tyo-hkg-rtt)** | 43.35 ms | 48.43 ms | 1.13 ms | | 4 | 🇷🇺 Moscow (MOW) | **[118.6 ms](/docs/network/latency/pairs/mow-hkg-rtt)** | 116.41 ms | 127.8 ms | 2.24 ms | | 5 | 🇺🇸 Seattle (SEA) | **[130.4 ms](/docs/network/latency/pairs/sea-hkg-rtt)** | 125.05 ms | 145.66 ms | 4.85 ms | | 6 | 🇦🇺 Sydney (SYD) | **[135.7 ms](/docs/network/latency/pairs/syd-hkg-rtt)** | 124.18 ms | 155.08 ms | 6.42 ms | | 7 | 🇦🇺 Melbourne (MEL) | **[138.5 ms](/docs/network/latency/pairs/mel-hkg-rtt)** | 133.73 ms | 150.99 ms | 4.36 ms | | 8 | 🇩🇪 Berlin (BER) | **[144.9 ms](/docs/network/latency/pairs/ber-hkg-rtt)** | 140.42 ms | 158.73 ms | 3.7 ms | | 9 | 🇺🇸 Los Angeles (LAX) | **[146.5 ms](/docs/network/latency/pairs/lax-hkg-rtt)** | 143.01 ms | 158.89 ms | 3.29 ms | | 10 | 🇩🇪 Frankfurt (FRA) | **[152 ms](/docs/network/latency/pairs/fra-hkg-rtt)** | 144.15 ms | 171.03 ms | 6.84 ms | ## Inbound Latency to Hong Kong (HKG) Round-trip time in milliseconds **from all other PoPs to Hong Kong**. The measurement origin matters: inbound and outbound RTT can differ for the same pair. | Source | RTT (ms) | Fiber Floor | Efficiency | Jitter | Loss | Tier | | ----------------------- | ---------------------------------------------------- | ----------- | ---------- | ------- | ----- | --------- | | 🇹🇼 Taipei (TPE) | **[15.5](/docs/network/latency/pairs/tpe-hkg-rtt)** | 7.94 ms | 51.2% | 0.16 ms | 0.02% | Ultra-Low | | 🇸🇬 Singapore (SIN) | **[30.8](/docs/network/latency/pairs/sin-hkg-rtt)** | 25.24 ms | 81.9% | 0.23 ms | 0.1% | Excellent | | 🇯🇵 Tokyo (TYO) | **[44.8](/docs/network/latency/pairs/tyo-hkg-rtt)** | 28.24 ms | 63% | 0.45 ms | 0.07% | Excellent | | 🇷🇺 Moscow (MOW) | **[118.6](/docs/network/latency/pairs/mow-hkg-rtt)** | 70.05 ms | 59.1% | 0.77 ms | 0.13% | Good | | 🇺🇸 Seattle (SEA) | **[130.4](/docs/network/latency/pairs/sea-hkg-rtt)** | 102.2 ms | 78.4% | 0.7 ms | 0% | Good | | 🇦🇺 Sydney (SYD) | **[135.7](/docs/network/latency/pairs/syd-hkg-rtt)** | 71.97 ms | 53% | 0.65 ms | 0.19% | Good | | 🇦🇺 Melbourne (MEL) | **[138.5](/docs/network/latency/pairs/mel-hkg-rtt)** | 72.38 ms | 52.3% | 1.39 ms | 0.19% | Good | | 🇩🇪 Berlin (BER) | **[144.9](/docs/network/latency/pairs/ber-hkg-rtt)** | 85.85 ms | 59.3% | 1.29 ms | 0.08% | Good | | 🇺🇸 Los Angeles (LAX) | **[146.5](/docs/network/latency/pairs/lax-hkg-rtt)** | 114.3 ms | 78% | 1.42 ms | 0% | Good | | 🇩🇪 Frankfurt (FRA) | **[152](/docs/network/latency/pairs/fra-hkg-rtt)** | 89.87 ms | 59.1% | 1.18 ms | 0% | Fair | | 🇳🇱 Amsterdam (AMS) | **[155.9](/docs/network/latency/pairs/ams-hkg-rtt)** | 91.05 ms | 58.4% | 1.21 ms | 0.05% | Fair | | 🇫🇷 Paris (PAR) | **[160.1](/docs/network/latency/pairs/par-hkg-rtt)** | 94.48 ms | 59% | 1.29 ms | 0.05% | Fair | | 🇬🇧 London (LON) | **[160.6](/docs/network/latency/pairs/lon-hkg-rtt)** | 94.44 ms | 58.8% | 0.84 ms | 0.07% | Fair | | 🇫🇷 Marseille (MRS) | **[166.4](/docs/network/latency/pairs/mrs-hkg-rtt)** | 95.48 ms | 57.4% | 1.49 ms | 0% | Fair | | 🇺🇸 New York (NYC) | **[187.1](/docs/network/latency/pairs/nyc-hkg-rtt)** | 127.11 ms | 67.9% | 1.76 ms | 0% | Fair | | 🇺🇸 Ashburn (IAD) | **[187.2](/docs/network/latency/pairs/iad-hkg-rtt)** | 128.37 ms | 68.6% | 1 ms | 0.05% | Fair | | 🇺🇸 Miami (MIA) | **[199.5](/docs/network/latency/pairs/mia-hkg-rtt)** | 141.69 ms | 71% | 2.35 ms | 0.06% | Fair | | 🇧🇷 São Paulo (GRU) | **[273.5](/docs/network/latency/pairs/gru-hkg-rtt)** | 176.86 ms | 64.7% | 2.57 ms | 0% | High | | 🇿🇦 Johannesburg (JNB) | **[318](/docs/network/latency/pairs/jnb-hkg-rtt)** | 104.93 ms | 33% | 3.32 ms | 0.19% | High | ## Asia Pacific Peers Hong Kong is one of 6 Hats Network PoPs in **Asia Pacific**. The table lists outbound measurements within the region. | PoP | Code | Country | RTT (ms) | Tier | | --------------------------------------- | ---- | --------- | --------- | --------- | | [🇦🇺 Melbourne (MEL)](./mel-melbourne) | MEL | Australia | **138.4** | Good | | [🇸🇬 Singapore (SIN)](./sin-singapore) | SIN | Singapore | **31.5** | Excellent | | [🇦🇺 Sydney (SYD)](./syd-sydney) | SYD | Australia | **137.2** | Good | | [🇹🇼 Taipei (TPE)](./tpe-taipei) | TPE | Taiwan | **14.7** | Ultra-Low | | [🇯🇵 Tokyo (TYO)](./tyo-tokyo) | TYO | Japan | **44.9** | Excellent | ## Questions ### What is the fastest route to Hong Kong? The fastest measured route to Hong Kong (HKG) is [Taipei (TPE) → Hong Kong (HKG)](/docs/network/latency/pairs/tpe-hkg-rtt), averaging **15.5 ms** RTT (Ultra-Low). ### What is the fastest route from Hong Kong? The fastest measured route from Hong Kong (HKG) is [Hong Kong (HKG) → Taipei (TPE)](/docs/network/latency/pairs/hkg-tpe-rtt), averaging **14.7 ms** RTT (Ultra-Low). ### How well connected is Hong Kong to the Hats Network backbone? Hong Kong (HKG) maintains measured routes to all 19 other PoPs across 5 continents. The slowest route, [Hong Kong (HKG) → Johannesburg (JNB)](/docs/network/latency/pairs/hkg-jnb-rtt), averages **316.7 ms** RTT. *** *Round published August 16, 2026. [All 20 PoPs](/docs/network/latency) · [Dataset and licence](/opendata/latency/).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/hkg-hong-kong). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Ashburn network latency (IAD) Ashburn (IAD) is the network's Northern Virginia PoP. Compare its East Coast measurements with New York and Miami. Measurement round: `2026-08-16T04:07:28Z`. The published round covers **19 outbound routes** from Ashburn. Incoming routes are measured separately; neither direction describes last-mile performance. ## Location * **Location:** Ashburn, USA (North America) * **Region:** North America * **Role:** Backbone interconnection point for Hats Network (AS203314) * **Public city reference:** [GeoNames 4744870](https://www.geonames.org/4744870); 39.04372, -77.48749 * **Coverage:** RTT measurements to 19 other PoPs across 5 continents ### Interconnection * **Facility:** IAD1 * **Upstream transit:** HE / Telecom Italia / Comcast / GTT / Cogent > Interactive content is available on the canonical HTML page. ## Round Summary | Metric | Value | | --------------------- | --------------------------------------------------------------------------------------------------------- | | Fastest Route | [🇺🇸 Ashburn (IAD) → 🇺🇸 New York (NYC)](/docs/network/latency/pairs/iad-nyc-rtt) (**6 ms**); Ultra-Low | | Slowest Route | [🇺🇸 Ashburn (IAD) → 🇿🇦 Johannesburg (JNB)](/docs/network/latency/pairs/iad-jnb-rtt) (**228.3 ms**) | | Average RTT | **115.0 ms** | | Average Jitter | **0.87 ms** | | Average Packet Loss | **0.06%** | | Best Fiber Efficiency | **82.6%** | | Intra-Region Peers | 4 PoPs in North America | | Total Routes | 19 outbound / 19 inbound (100% coverage) | > Jitter, packet-loss, percentile, and trend fields are estimated comparative indicators. Read the [theoretical fiber latency methodology](/docs/network/latency/theoretical-fiber-latency). ## Regional Routes Ashburn's 4 other **North America** PoPs have RTTs from **6 ms** (New York (NYC)) to **62.3 ms** (Seattle (SEA)). Ashburn is in Northern Virginia, inland from the US East Coast. ## Outbound Latency from Ashburn (IAD) Round-trip time in milliseconds **from Ashburn** to all other backbone PoPs, sorted fastest to slowest. | Destination | RTT (ms) | Fiber Floor | Efficiency | Jitter | Loss | Tier | | ----------------------- | ---------------------------------------------------- | ----------- | ---------- | ------- | ----- | --------- | | 🇺🇸 New York (NYC) | **[6](/docs/network/latency/pairs/iad-nyc-rtt)** | 3.44 ms | 57.3% | 0.12 ms | 0.17% | Ultra-Low | | 🇺🇸 Miami (MIA) | **[27.4](/docs/network/latency/pairs/iad-mia-rtt)** | 14.62 ms | 53.4% | 0.15 ms | 0% | Ultra-Low | | 🇺🇸 Los Angeles (LAX) | **[60.7](/docs/network/latency/pairs/iad-lax-rtt)** | 35.84 ms | 59% | 0.32 ms | 0% | Excellent | | 🇺🇸 Seattle (SEA) | **[62.3](/docs/network/latency/pairs/iad-sea-rtt)** | 36.27 ms | 58.2% | 0.44 ms | 0.12% | Excellent | | 🇬🇧 London (LON) | **[70.3](/docs/network/latency/pairs/iad-lon-rtt)** | 58.1 ms | 82.6% | 0.81 ms | 0.19% | Excellent | | 🇳🇱 Amsterdam (AMS) | **[75.2](/docs/network/latency/pairs/iad-ams-rtt)** | 60.94 ms | 81% | 0.34 ms | 0.02% | Excellent | | 🇫🇷 Paris (PAR) | **[76.1](/docs/network/latency/pairs/iad-par-rtt)** | 60.73 ms | 79.8% | 0.7 ms | 0.03% | Excellent | | 🇩🇪 Frankfurt (FRA) | **[81.2](/docs/network/latency/pairs/iad-fra-rtt)** | 64.3 ms | 79.2% | 0.95 ms | 0.1% | Good | | 🇩🇪 Berlin (BER) | **[86.1](/docs/network/latency/pairs/iad-ber-rtt)** | 66.05 ms | 76.7% | 0.7 ms | 0% | Good | | 🇫🇷 Marseille (MRS) | **[86.5](/docs/network/latency/pairs/iad-mrs-rtt)** | 65.34 ms | 75.5% | 0.72 ms | 0% | Good | | 🇧🇷 São Paulo (GRU) | **[101.4](/docs/network/latency/pairs/iad-gru-rtt)** | 74.79 ms | 73.8% | 1.02 ms | 0% | Good | | 🇷🇺 Moscow (MOW) | **[112.8](/docs/network/latency/pairs/iad-mow-rtt)** | 76.88 ms | 68.2% | 0.79 ms | 0.13% | Good | | 🇯🇵 Tokyo (TYO) | **[141.9](/docs/network/latency/pairs/iad-tyo-rtt)** | 106.69 ms | 75.2% | 1.23 ms | 0.02% | Good | | 🇹🇼 Taipei (TPE) | **[172.3](/docs/network/latency/pairs/iad-tpe-rtt)** | 123.78 ms | 71.8% | 1.59 ms | 0.04% | Fair | | 🇭🇰 Hong Kong (HKG) | **[187.2](/docs/network/latency/pairs/iad-hkg-rtt)** | 128.37 ms | 68.6% | 1 ms | 0.05% | Fair | | 🇦🇺 Sydney (SYD) | **[193.6](/docs/network/latency/pairs/iad-syd-rtt)** | 153.48 ms | 79.3% | 1.68 ms | 0% | Fair | | 🇦🇺 Melbourne (MEL) | **[203.9](/docs/network/latency/pairs/iad-mel-rtt)** | 160.09 ms | 78.5% | 1.06 ms | 0.15% | Fair | | 🇸🇬 Singapore (SIN) | **[211.5](/docs/network/latency/pairs/iad-sin-rtt)** | 152.14 ms | 71.9% | 1.77 ms | 0% | Fair | | 🇿🇦 Johannesburg (JNB) | **[228.3](/docs/network/latency/pairs/iad-jnb-rtt)** | 128.11 ms | 56.1% | 1.05 ms | 0.04% | Fair | ## Fastest Routes to Ashburn (IAD) The 10 fastest measured routes **to Ashburn**, ranked by average RTT. Min / max / stdev are computed from the same measurement round (50 ICMP echo probes per route, 100 ms interval); routes without a published per-sample round show the measured average RTT only. | Rank | Source | Avg RTT | Min | Max | Stdev | | ---- | ---------------------- | ------------------------------------------------------ | -------- | -------- | ------- | | 1 | 🇺🇸 New York (NYC) | **[6 ms](/docs/network/latency/pairs/nyc-iad-rtt)** | 5.58 ms | 7.18 ms | 0.3 ms | | 2 | 🇺🇸 Miami (MIA) | **[27.3 ms](/docs/network/latency/pairs/mia-iad-rtt)** | 26.26 ms | 30.45 ms | 1.02 ms | | 3 | 🇺🇸 Los Angeles (LAX) | **[60.4 ms](/docs/network/latency/pairs/lax-iad-rtt)** | 57.19 ms | 70.55 ms | 2.61 ms | | 4 | 🇺🇸 Seattle (SEA) | **[61.5 ms](/docs/network/latency/pairs/sea-iad-rtt)** | 57.92 ms | 70.85 ms | 2.96 ms | | 5 | 🇬🇧 London (LON) | **[71 ms](/docs/network/latency/pairs/lon-iad-rtt)** | 68.31 ms | 80.67 ms | 2.3 ms | | 6 | 🇫🇷 Paris (PAR) | **[75.5 ms](/docs/network/latency/pairs/par-iad-rtt)** | 72.75 ms | 85.2 ms | 2.46 ms | | 7 | 🇳🇱 Amsterdam (AMS) | **[75.7 ms](/docs/network/latency/pairs/ams-iad-rtt)** | 72.76 ms | 82.67 ms | 2.28 ms | | 8 | 🇩🇪 Frankfurt (FRA) | **[80.7 ms](/docs/network/latency/pairs/fra-iad-rtt)** | 79.19 ms | 87.49 ms | 1.42 ms | | 9 | 🇩🇪 Berlin (BER) | **[85.7 ms](/docs/network/latency/pairs/ber-iad-rtt)** | 83.7 ms | 92.94 ms | 2.12 ms | | 10 | 🇫🇷 Marseille (MRS) | **[85.7 ms](/docs/network/latency/pairs/mrs-iad-rtt)** | 82.1 ms | 94.47 ms | 2.63 ms | ## Inbound Latency to Ashburn (IAD) Round-trip time in milliseconds **from all other PoPs to Ashburn**. The measurement origin matters: inbound and outbound RTT can differ for the same pair. | Source | RTT (ms) | Fiber Floor | Efficiency | Jitter | Loss | Tier | | ----------------------- | ---------------------------------------------------- | ----------- | ---------- | ------- | ----- | --------- | | 🇺🇸 New York (NYC) | **[6](/docs/network/latency/pairs/nyc-iad-rtt)** | 3.44 ms | 57.3% | 0.12 ms | 0.01% | Ultra-Low | | 🇺🇸 Miami (MIA) | **[27.3](/docs/network/latency/pairs/mia-iad-rtt)** | 14.62 ms | 53.6% | 0.15 ms | 0% | Ultra-Low | | 🇺🇸 Los Angeles (LAX) | **[60.4](/docs/network/latency/pairs/lax-iad-rtt)** | 35.84 ms | 59.3% | 0.64 ms | 0.06% | Excellent | | 🇺🇸 Seattle (SEA) | **[61.5](/docs/network/latency/pairs/sea-iad-rtt)** | 36.27 ms | 59% | 0.55 ms | 0% | Excellent | | 🇬🇧 London (LON) | **[71](/docs/network/latency/pairs/lon-iad-rtt)** | 58.1 ms | 81.8% | 0.52 ms | 0.13% | Excellent | | 🇫🇷 Paris (PAR) | **[75.5](/docs/network/latency/pairs/par-iad-rtt)** | 60.73 ms | 80.4% | 0.86 ms | 0.02% | Excellent | | 🇳🇱 Amsterdam (AMS) | **[75.7](/docs/network/latency/pairs/ams-iad-rtt)** | 60.94 ms | 80.5% | 0.75 ms | 0.05% | Excellent | | 🇩🇪 Frankfurt (FRA) | **[80.7](/docs/network/latency/pairs/fra-iad-rtt)** | 64.3 ms | 79.7% | 0.91 ms | 0.04% | Good | | 🇩🇪 Berlin (BER) | **[85.7](/docs/network/latency/pairs/ber-iad-rtt)** | 66.05 ms | 77.1% | 0.79 ms | 0% | Good | | 🇫🇷 Marseille (MRS) | **[85.7](/docs/network/latency/pairs/mrs-iad-rtt)** | 65.34 ms | 76.2% | 0.46 ms | 0.05% | Good | | 🇷🇺 Moscow (MOW) | **[113.9](/docs/network/latency/pairs/mow-iad-rtt)** | 76.88 ms | 67.5% | 1 ms | 0.08% | Good | | 🇧🇷 São Paulo (GRU) | **[128](/docs/network/latency/pairs/gru-iad-rtt)** | 74.79 ms | 58.4% | 0.6 ms | 0.09% | Good | | 🇯🇵 Tokyo (TYO) | **[142.9](/docs/network/latency/pairs/tyo-iad-rtt)** | 106.69 ms | 74.7% | 0.69 ms | 0% | Good | | 🇹🇼 Taipei (TPE) | **[171.6](/docs/network/latency/pairs/tpe-iad-rtt)** | 123.78 ms | 72.1% | 1.7 ms | 0% | Fair | | 🇭🇰 Hong Kong (HKG) | **[186.2](/docs/network/latency/pairs/hkg-iad-rtt)** | 128.37 ms | 68.9% | 1.54 ms | 0.04% | Fair | | 🇦🇺 Sydney (SYD) | **[195.3](/docs/network/latency/pairs/syd-iad-rtt)** | 153.48 ms | 78.6% | 1.26 ms | 0.15% | Fair | | 🇦🇺 Melbourne (MEL) | **[203.9](/docs/network/latency/pairs/mel-iad-rtt)** | 160.09 ms | 78.5% | 2.45 ms | 0.02% | Fair | | 🇸🇬 Singapore (SIN) | **[210](/docs/network/latency/pairs/sin-iad-rtt)** | 152.14 ms | 72.4% | 2.32 ms | 0% | Fair | | 🇿🇦 Johannesburg (JNB) | **[229](/docs/network/latency/pairs/jnb-iad-rtt)** | 128.11 ms | 55.9% | 1.29 ms | 0% | Fair | ## North America Peers Ashburn is one of 5 Hats Network PoPs in **North America**. The table lists outbound measurements within the region. | PoP | Code | Country | RTT (ms) | Tier | | ------------------------------------------- | ---- | ------- | -------- | --------- | | [🇺🇸 Los Angeles (LAX)](./lax-los-angeles) | LAX | USA | **60.7** | Excellent | | [🇺🇸 Miami (MIA)](./mia-miami) | MIA | USA | **27.4** | Ultra-Low | | [🇺🇸 New York (NYC)](./nyc-new-york) | NYC | USA | **6** | Ultra-Low | | [🇺🇸 Seattle (SEA)](./sea-seattle) | SEA | USA | **62.3** | Excellent | ## Questions ### What is the fastest route to Ashburn? The fastest measured route to Ashburn (IAD) is [New York (NYC) → Ashburn (IAD)](/docs/network/latency/pairs/nyc-iad-rtt), averaging **6 ms** RTT (Ultra-Low). ### What is the fastest route from Ashburn? The fastest measured route from Ashburn (IAD) is [Ashburn (IAD) → New York (NYC)](/docs/network/latency/pairs/iad-nyc-rtt), averaging **6 ms** RTT (Ultra-Low). ### How well connected is Ashburn to the Hats Network backbone? Ashburn (IAD) maintains measured routes to all 19 other PoPs across 5 continents. The slowest route, [Ashburn (IAD) → Johannesburg (JNB)](/docs/network/latency/pairs/iad-jnb-rtt), averages **228.3 ms** RTT. *** *Round published August 16, 2026. [All 20 PoPs](/docs/network/latency) · [Dataset and licence](/opendata/latency/).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/iad-ashburn). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Global Ping & RTT Latency Matrix | AS203314 > **Published round** > > Snapshot: **August 16, 2026**. The pipeline runs daily, but this page reflects the published round > below. See [measurement limits](#methodology--data-sources) before using the results. Measurement round: `2026-08-16T04:07:28Z`. ## Overview The matrix records **round-trip time (RTT)** between Hats Network PoPs, in milliseconds. Each row is a measurement origin. The reverse direction may produce a different result. ### Dataset Coverage * **Nodes:** 20 * **Directed routes:** 380 / 380 (100.0% coverage) > **Important:** RTT is *not* one-way latency. Values represent backbone PoP-to-PoP paths, not last-mile end-user performance. ## Methodology & Data Sources ### Data sources The publishing pipeline combines the available backbone measurements into one dated snapshot. Physical references use WGS-84 city-centre distances and optical propagation assumptions. They do not identify the actual cable path. Review the [coordinate and model inputs](/docs/network/latency/theoretical-fiber-latency), then reproduce the calculation on the [formula verification page](/docs/network/latency/theoretical-latency-verification). ### Limitations * This dataset is **not an SLA** and should be treated as *indicative best-case backbone RTT*. * End-user latency depends on local access networks, congestion, routing policy, and traffic engineering. > Interactive content is available on the canonical HTML page. ## Full RTT Matrix Rows are origins; columns are destinations. Values are RTT in milliseconds. Bold entries mark the lowest value in each destination column. The static table below contains the same measurements. > Interactive content is available on the canonical HTML page. View raw RTT table (accessible / printable) | Source | AMS | BER | FRA | LON | MRS | MOW | PAR | IAD | LAX | MIA | NYC | SEA | GRU | HKG | MEL | SIN | SYD | TPE | TYO | JNB | | ------- | ------- | ------- | ------- | ------- | ------- | -------- | ------- | ----- | -------- | -------- | ----- | -------- | --------- | -------- | ------- | -------- | ------- | -------- | -------- | --------- | | **AMS** | N/A | 7.2 | **5.9** | **5.2** | 19.6 | 38.1 | 7 | 75.7 | 133.7 | 106 | 68.6 | 130.2 | 175.5 | 155.9 | 252.5 | 157.6 | 258 | 168.7 | 198.2 | **171.9** | | **BER** | 7.2 | N/A | 6.1 | 15.2 | 20.2 | **28.6** | 15.6 | 85.7 | 140.9 | 118.3 | 79.6 | 139.7 | 186.5 | 144.9 | 256.2 | 158.3 | 262.1 | 156.8 | 187.5 | 172.1 | | **FRA** | 5.9 | **6.1** | N/A | 12.9 | 16.4 | 34.7 | 7.6 | 80.7 | 141.8 | 112 | 74.4 | 135.2 | 181.3 | 152 | 250.1 | 151.4 | 256 | 163.8 | 194.5 | 188.4 | | **LON** | **5.2** | 16.2 | 13.6 | N/A | 17.9 | 43.5 | **6.4** | 71 | 127.5 | 101.9 | 63.8 | 123.2 | 170.7 | 160.6 | 249.6 | 156.5 | 258.1 | 173.8 | 202.9 | 179.6 | | **MRS** | 20.3 | 20.6 | 16 | 18.9 | N/A | 49.6 | 8.9 | 85.7 | 144.5 | 112.4 | 79.8 | 141.7 | 186.7 | 166.4 | 234.2 | 139.6 | 240.6 | 180.9 | 203.8 | 198.5 | | **MOW** | 37.6 | 27.5 | 35.9 | 42.4 | 49.9 | N/A | 44 | 113.9 | 177.7 | 144.5 | 107.5 | 164.8 | 214.4 | 118.6 | 235.9 | 147.3 | 241.8 | 132.1 | 162.2 | 201.9 | | **PAR** | 7 | 15.5 | 7.6 | 6.4 | **8.9** | 44.7 | N/A | 75.5 | 136.5 | 106.8 | 68.9 | 128 | 175.8 | 160.1 | 246.4 | 150.3 | 249.7 | 171.9 | 204.7 | 180.9 | | **IAD** | 75.2 | 86.1 | 81.2 | 70.3 | 86.5 | 112.8 | 76.1 | N/A | 60.7 | **27.4** | **6** | 62.3 | **101.4** | 187.2 | 203.9 | 211.5 | 193.6 | 172.3 | 141.9 | 228.3 | | **LAX** | 132.1 | 139.3 | 142.3 | 129.2 | 143.9 | 177 | 135.9 | 60.4 | N/A | 57.3 | 57.9 | **26.9** | 131.3 | 146.5 | 148.2 | 167.7 | 137.6 | 132.9 | 101.2 | 287.2 | | **MIA** | 106 | 117.3 | 112.2 | 101.8 | 113.7 | 145.5 | 106.2 | 27.3 | 56.8 | N/A | 33.3 | 81.9 | 128.7 | 199.5 | 200.7 | 223.1 | 191.5 | 185.1 | 156.3 | 259.8 | | **NYC** | 68.7 | 78.8 | 73.9 | 63.5 | 80.2 | 107 | 68.9 | **6** | 58.9 | 32.9 | N/A | 58.7 | 106.9 | 187.1 | 205.1 | 208.7 | 192.6 | 173.4 | 143.1 | 221.5 | | **SEA** | 128.6 | 140.9 | 133.2 | 124.1 | 141.2 | 164.6 | 129.2 | 61.5 | **25.9** | 81.7 | 59.5 | N/A | 155.7 | 130.4 | 170.1 | 151.1 | 161.3 | 114.7 | 85 | 282.1 | | **GRU** | 180 | 191.3 | 186.2 | 175.8 | 187.7 | 219.5 | 180.2 | 128 | 130.8 | 74 | 122 | 155.9 | N/A | 273.5 | 274.7 | 297.1 | 265.5 | 259.1 | 230.3 | 333.8 | | **HKG** | 154.4 | 145.1 | 150.7 | 158.7 | 165.4 | 118.2 | 159.4 | 186.2 | 145 | 201.1 | 184.7 | 131.6 | 287.6 | N/A | 138.4 | **31.5** | 137.2 | **14.7** | 44.9 | 316.7 | | **MEL** | 252.8 | 254.7 | 248.6 | 252 | 233.5 | 236.1 | 246.5 | 203.9 | 148.4 | 201 | 204.2 | 172.1 | 305.3 | 138.5 | N/A | 88.4 | **9.8** | 143.2 | 112 | 405.5 | | **SIN** | 155.8 | 160.3 | 151.6 | 155.4 | 140.6 | 147.7 | 151.4 | 210 | 168.2 | 221.7 | 209.3 | 151.5 | 311.4 | 30.8 | 88.6 | N/A | 94.5 | 45.8 | 68.9 | 312.6 | | **SYD** | 258.9 | 260.5 | 254.4 | 256.5 | 240.5 | 242.2 | 250.3 | 195.3 | 136.5 | 191.5 | 191.9 | 161.9 | 296.7 | 135.7 | **9.8** | 94.5 | N/A | 131.7 | 101.3 | 412.5 | | **TPE** | 166.5 | 158.6 | 166 | 173 | 179.5 | 132.7 | 173.8 | 171.6 | 132 | 185.7 | 174.2 | 115.8 | 273 | **15.5** | 142 | 44.3 | 133.2 | N/A | **31.9** | 331 | | **TYO** | 197.4 | 186.1 | 193.2 | 202.7 | 205.6 | 161.7 | 204.8 | 142.9 | 101.2 | 156.5 | 144.6 | 84.8 | 230.5 | 44.8 | 113.9 | 69.4 | 101.9 | 32.3 | N/A | 359.2 | | **JNB** | 163.2 | 172.1 | 166 | 158 | 172 | 200.7 | 164.4 | 229 | 285.5 | 259.9 | 221.8 | 281.2 | 328.7 | 318 | 406.2 | 311.6 | 412.6 | 329.8 | 360.5 | N/A | ## Regional Breakdown ### Europe #### Intra-Region Routes | Route | RTT (ms) | | --------------------------------- | -------- | | Amsterdam (AMS) → London (LON) | 5.2 | | Amsterdam (AMS) → Frankfurt (FRA) | 5.9 | | Berlin (BER) → Frankfurt (FRA) | 6.1 | | London (LON) → Paris (PAR) | 6.4 | | Amsterdam (AMS) → Paris (PAR) | 7 | | Amsterdam (AMS) → Berlin (BER) | 7.2 | | Frankfurt (FRA) → Paris (PAR) | 7.6 | | Marseille (MRS) → Paris (PAR) | 8.9 | | Frankfurt (FRA) → London (LON) | 12.9 | | Berlin (BER) → London (LON) | 15.2 | | Berlin (BER) → Paris (PAR) | 15.6 | | Frankfurt (FRA) → Marseille (MRS) | 16.4 | | London (LON) → Marseille (MRS) | 17.9 | | Amsterdam (AMS) → Marseille (MRS) | 19.6 | | Berlin (BER) → Marseille (MRS) | 20.2 | | Berlin (BER) → Moscow (MOW) | 28.6 | | Frankfurt (FRA) → Moscow (MOW) | 34.7 | | Amsterdam (AMS) → Moscow (MOW) | 38.1 | | London (LON) → Moscow (MOW) | 43.5 | | Moscow (MOW) → Paris (PAR) | 44 | | Marseille (MRS) → Moscow (MOW) | 49.6 | #### Key Cross-Region Routes | Route | RTT (ms) | Far Region | | -------------------------------- | -------- | ------------- | | London (LON) → New York (NYC) | 63.8 | North America | | Amsterdam (AMS) → New York (NYC) | 68.6 | North America | | Paris (PAR) → New York (NYC) | 68.9 | North America | | London (LON) → Ashburn (IAD) | 71 | North America | | Frankfurt (FRA) → New York (NYC) | 74.4 | North America | | Paris (PAR) → Ashburn (IAD) | 75.5 | North America | | Amsterdam (AMS) → Ashburn (IAD) | 75.7 | North America | | Berlin (BER) → New York (NYC) | 79.6 | North America | ### North America #### Intra-Region Routes | Route | RTT (ms) | | ---------------------------------- | -------- | | Ashburn (IAD) → New York (NYC) | 6 | | Los Angeles (LAX) → Seattle (SEA) | 26.9 | | Ashburn (IAD) → Miami (MIA) | 27.4 | | Miami (MIA) → New York (NYC) | 33.3 | | Los Angeles (LAX) → Miami (MIA) | 57.3 | | Los Angeles (LAX) → New York (NYC) | 57.9 | | New York (NYC) → Seattle (SEA) | 58.7 | | Ashburn (IAD) → Los Angeles (LAX) | 60.7 | | Ashburn (IAD) → Seattle (SEA) | 62.3 | | Miami (MIA) → Seattle (SEA) | 81.9 | #### Key Cross-Region Routes | Route | RTT (ms) | Far Region | | -------------------------------- | -------- | ---------- | | New York (NYC) → London (LON) | 63.5 | Europe | | New York (NYC) → Amsterdam (AMS) | 68.7 | Europe | | New York (NYC) → Paris (PAR) | 68.9 | Europe | | Ashburn (IAD) → London (LON) | 70.3 | Europe | | New York (NYC) → Frankfurt (FRA) | 73.9 | Europe | | Ashburn (IAD) → Amsterdam (AMS) | 75.2 | Europe | | Ashburn (IAD) → Paris (PAR) | 76.1 | Europe | | New York (NYC) → Berlin (BER) | 78.8 | Europe | ### South America #### Key Cross-Region Routes | Route | RTT (ms) | Far Region | | ----------------------------------- | -------- | ------------- | | São Paulo (GRU) → Miami (MIA) | 74 | North America | | São Paulo (GRU) → New York (NYC) | 122 | North America | | São Paulo (GRU) → Ashburn (IAD) | 128 | North America | | São Paulo (GRU) → Los Angeles (LAX) | 130.8 | North America | | São Paulo (GRU) → Seattle (SEA) | 155.9 | North America | | São Paulo (GRU) → London (LON) | 175.8 | Europe | | São Paulo (GRU) → Amsterdam (AMS) | 180 | Europe | | São Paulo (GRU) → Paris (PAR) | 180.2 | Europe | ### Asia Pacific #### Intra-Region Routes | Route | RTT (ms) | | --------------------------------- | -------- | | Melbourne (MEL) → Sydney (SYD) | 9.8 | | Hong Kong (HKG) → Taipei (TPE) | 14.7 | | Hong Kong (HKG) → Singapore (SIN) | 31.5 | | Taipei (TPE) → Tokyo (TYO) | 31.9 | | Hong Kong (HKG) → Tokyo (TYO) | 44.9 | | Singapore (SIN) → Taipei (TPE) | 45.8 | | Singapore (SIN) → Tokyo (TYO) | 68.9 | | Melbourne (MEL) → Singapore (SIN) | 88.4 | | Singapore (SIN) → Sydney (SYD) | 94.5 | | Sydney (SYD) → Tokyo (TYO) | 101.3 | | Melbourne (MEL) → Tokyo (TYO) | 112 | | Sydney (SYD) → Taipei (TPE) | 131.7 | | Hong Kong (HKG) → Sydney (SYD) | 137.2 | | Hong Kong (HKG) → Melbourne (MEL) | 138.4 | | Melbourne (MEL) → Taipei (TPE) | 143.2 | #### Key Cross-Region Routes | Route | RTT (ms) | Far Region | | -------------------------------- | -------- | ------------- | | Tokyo (TYO) → Seattle (SEA) | 84.8 | North America | | Tokyo (TYO) → Los Angeles (LAX) | 101.2 | North America | | Taipei (TPE) → Seattle (SEA) | 115.8 | North America | | Hong Kong (HKG) → Moscow (MOW) | 118.2 | Europe | | Hong Kong (HKG) → Seattle (SEA) | 131.6 | North America | | Taipei (TPE) → Los Angeles (LAX) | 132 | North America | | Taipei (TPE) → Moscow (MOW) | 132.7 | Europe | | Sydney (SYD) → Los Angeles (LAX) | 136.5 | North America | ### Africa #### Key Cross-Region Routes | Route | RTT (ms) | Far Region | | ------------------------------------ | -------- | ------------- | | Johannesburg (JNB) → London (LON) | 158 | Europe | | Johannesburg (JNB) → Amsterdam (AMS) | 163.2 | Europe | | Johannesburg (JNB) → Paris (PAR) | 164.4 | Europe | | Johannesburg (JNB) → Frankfurt (FRA) | 166 | Europe | | Johannesburg (JNB) → Marseille (MRS) | 172 | Europe | | Johannesburg (JNB) → Berlin (BER) | 172.1 | Europe | | Johannesburg (JNB) → Moscow (MOW) | 200.7 | Europe | | Johannesburg (JNB) → New York (NYC) | 221.8 | North America | ## Node Reference | Code | City | Country/Region | Continent | | ------- | ---------------------------------- | -------------- | ------------- | | **AMS** | [Amsterdam](./ams-amsterdam) | Netherlands | Europe | | **BER** | [Berlin](./ber-berlin) | Germany | Europe | | **FRA** | [Frankfurt](./fra-frankfurt) | Germany | Europe | | **LON** | [London](./lon-london) | UK | Europe | | **MRS** | [Marseille](./mrs-marseille) | France | Europe | | **MOW** | [Moscow](./mow-moscow) | Russia | Europe | | **PAR** | [Paris](./par-paris) | France | Europe | | **IAD** | [Ashburn](./iad-ashburn) | USA | North America | | **LAX** | [Los Angeles](./lax-los-angeles) | USA | North America | | **MIA** | [Miami](./mia-miami) | USA | North America | | **NYC** | [New York](./nyc-new-york) | USA | North America | | **SEA** | [Seattle](./sea-seattle) | USA | North America | | **GRU** | [São Paulo](./gru-sao-paulo) | Brazil | South America | | **HKG** | [Hong Kong](./hkg-hong-kong) | Hong Kong | Asia Pacific | | **MEL** | [Melbourne](./mel-melbourne) | Australia | Asia Pacific | | **SIN** | [Singapore](./sin-singapore) | Singapore | Asia Pacific | | **SYD** | [Sydney](./syd-sydney) | Australia | Asia Pacific | | **TPE** | [Taipei](./tpe-taipei) | Taiwan | Asia Pacific | | **TYO** | [Tokyo](./tyo-tokyo) | Japan | Asia Pacific | | **JNB** | [Johannesburg](./jnb-johannesburg) | South Africa | Africa | ## Browse Latency Data * [Latency Coordinates & Fiber Model Inputs](./theoretical-fiber-latency): GeoNames resolution, physical constants, engineering assumptions, and source boundaries * [Theoretical Latency Formula Verification](./theoretical-latency-verification): rendered equations, unit checks, boundary conditions, and a complete Tokyo-to-Sydney reconciliation * [By PoP](#node-reference): per-node outbound/inbound RTT summaries * [By City Pair](./pairs): individual source-to-destination RTT pages * [By Region & Corridor](./regions): regional and inter-regional overviews *** *Published snapshot: August 16, 2026. [Download the dataset](/opendata/latency/).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Johannesburg network latency (JNB) Johannesburg (JNB) is the African PoP in this dataset. Its published routes connect it with every other measurement origin. Measurement round: `2026-08-16T04:07:28Z`. The published round covers **19 outbound routes** from Johannesburg. Incoming routes are measured separately; neither direction describes last-mile performance. ## Location * **Location:** Johannesburg, South Africa (Africa) * **Region:** Africa * **Role:** Backbone interconnection point for Hats Network (AS203314) * **Public city reference:** [GeoNames 993800](https://www.geonames.org/993800); -26.20227, 28.04363 * **Coverage:** RTT measurements to 19 other PoPs across 4 continents ### Interconnection * **Facility:** JNB1 * **Upstream transit:** HE Only > Interactive content is available on the canonical HTML page. ## Round Summary | Metric | Value | | --------------------- | --------------------------------------------------------------------------------------------------------- | | Fastest Route | [🇿🇦 Johannesburg (JNB) → 🇬🇧 London (LON)](/docs/network/latency/pairs/jnb-lon-rtt) (**158 ms**); Fair | | Slowest Route | [🇿🇦 Johannesburg (JNB) → 🇦🇺 Sydney (SYD)](/docs/network/latency/pairs/jnb-syd-rtt) (**412.6 ms**) | | Average RTT | **260.1 ms** | | Average Jitter | **2.18 ms** | | Average Packet Loss | **0.06%** | | Best Fiber Efficiency | **57.5%** | | Intra-Region Peers | 0 PoP in Africa | | Total Routes | 19 outbound / 19 inbound (100% coverage) | > Jitter, packet-loss, percentile, and trend fields are estimated comparative indicators. Read the [theoretical fiber latency methodology](/docs/network/latency/theoretical-fiber-latency). ## Regional Routes Johannesburg is Hats Network's only PoP in **Africa**. The lowest RTT in this round is to London (LON) at **158 ms** RTT. Johannesburg is inland in northeastern South Africa. ## Outbound Latency from Johannesburg (JNB) Round-trip time in milliseconds **from Johannesburg** to all other backbone PoPs, sorted fastest to slowest. | Destination | RTT (ms) | Fiber Floor | Efficiency | Jitter | Loss | Tier | | ---------------------- | ---------------------------------------------------- | ----------- | ---------- | ------- | ----- | ---- | | 🇬🇧 London (LON) | **[158](/docs/network/latency/pairs/jnb-lon-rtt)** | 88.52 ms | 56% | 0.77 ms | 0.12% | Fair | | 🇳🇱 Amsterdam (AMS) | **[163.2](/docs/network/latency/pairs/jnb-ams-rtt)** | 88.05 ms | 54% | 1.84 ms | 0% | Fair | | 🇫🇷 Paris (PAR) | **[164.4](/docs/network/latency/pairs/jnb-par-rtt)** | 85.17 ms | 51.8% | 1.05 ms | 0% | Fair | | 🇩🇪 Frankfurt (FRA) | **[166](/docs/network/latency/pairs/jnb-fra-rtt)** | 84.87 ms | 51.1% | 1.86 ms | 0.11% | Fair | | 🇫🇷 Marseille (MRS) | **[172](/docs/network/latency/pairs/jnb-mrs-rtt)** | 78.71 ms | 45.8% | 1.53 ms | 0% | Fair | | 🇩🇪 Berlin (BER) | **[172.1](/docs/network/latency/pairs/jnb-ber-rtt)** | 86.53 ms | 50.3% | 1.75 ms | 0.11% | Fair | | 🇷🇺 Moscow (MOW) | **[200.7](/docs/network/latency/pairs/jnb-mow-rtt)** | 89.37 ms | 44.5% | 2.17 ms | 0.16% | Fair | | 🇺🇸 New York (NYC) | **[221.8](/docs/network/latency/pairs/jnb-nyc-rtt)** | 125.67 ms | 56.7% | 2.01 ms | 0% | Fair | | 🇺🇸 Ashburn (IAD) | **[229](/docs/network/latency/pairs/jnb-iad-rtt)** | 128.11 ms | 55.9% | 1.29 ms | 0% | Fair | | 🇺🇸 Miami (MIA) | **[259.9](/docs/network/latency/pairs/jnb-mia-rtt)** | 126.78 ms | 48.8% | 3.06 ms | 0.1% | High | | 🇺🇸 Seattle (SEA) | **[281.2](/docs/network/latency/pairs/jnb-sea-rtt)** | 161.61 ms | 57.5% | 1.91 ms | 0.11% | High | | 🇺🇸 Los Angeles (LAX) | **[285.5](/docs/network/latency/pairs/jnb-lax-rtt)** | 163.35 ms | 57.2% | 2.14 ms | 0.01% | High | | 🇸🇬 Singapore (SIN) | **[311.6](/docs/network/latency/pairs/jnb-sin-rtt)** | 84.85 ms | 27.2% | 3.35 ms | 0% | High | | 🇭🇰 Hong Kong (HKG) | **[318](/docs/network/latency/pairs/jnb-hkg-rtt)** | 104.93 ms | 33% | 3.32 ms | 0.19% | High | | 🇧🇷 São Paulo (GRU) | **[328.7](/docs/network/latency/pairs/jnb-gru-rtt)** | 72.88 ms | 22.2% | 2.32 ms | 0.06% | High | | 🇹🇼 Taipei (TPE) | **[329.8](/docs/network/latency/pairs/jnb-tpe-rtt)** | 112.87 ms | 34.2% | 3.33 ms | 0% | High | | 🇯🇵 Tokyo (TYO) | **[360.5](/docs/network/latency/pairs/jnb-tyo-rtt)** | 132.57 ms | 36.8% | 2.79 ms | 0.06% | High | | 🇦🇺 Melbourne (MEL) | **[406.2](/docs/network/latency/pairs/jnb-mel-rtt)** | 101.33 ms | 24.9% | 2.56 ms | 0% | High | | 🇦🇺 Sydney (SYD) | **[412.6](/docs/network/latency/pairs/jnb-syd-rtt)** | 108.32 ms | 26.3% | 2.3 ms | 0.07% | High | ## Fastest Routes to Johannesburg (JNB) The 10 fastest measured routes **to Johannesburg**, ranked by average RTT. Min / max / stdev are computed from the same measurement round (50 ICMP echo probes per route, 100 ms interval); routes without a published per-sample round show the measured average RTT only. | Rank | Source | Avg RTT | Min | Max | Stdev | | ---- | -------------------- | ------------------------------------------------------- | --------- | --------- | -------- | | 1 | 🇳🇱 Amsterdam (AMS) | **[171.9 ms](/docs/network/latency/pairs/ams-jnb-rtt)** | 163.56 ms | 197.47 ms | 6.53 ms | | 2 | 🇩🇪 Berlin (BER) | **[172.1 ms](/docs/network/latency/pairs/ber-jnb-rtt)** | 163.72 ms | 199.65 ms | 7.34 ms | | 3 | 🇬🇧 London (LON) | **[179.6 ms](/docs/network/latency/pairs/lon-jnb-rtt)** | 171.6 ms | 203.66 ms | 5.96 ms | | 4 | 🇫🇷 Paris (PAR) | **[180.9 ms](/docs/network/latency/pairs/par-jnb-rtt)** | 176.39 ms | 195.6 ms | 3.41 ms | | 5 | 🇩🇪 Frankfurt (FRA) | **[188.4 ms](/docs/network/latency/pairs/fra-jnb-rtt)** | 182.33 ms | 218.93 ms | 6.75 ms | | 6 | 🇫🇷 Marseille (MRS) | **[198.5 ms](/docs/network/latency/pairs/mrs-jnb-rtt)** | 194.33 ms | 207.86 ms | 3.18 ms | | 7 | 🇷🇺 Moscow (MOW) | **[201.9 ms](/docs/network/latency/pairs/mow-jnb-rtt)** | 192.01 ms | 232.17 ms | 7.94 ms | | 8 | 🇺🇸 New York (NYC) | **[221.5 ms](/docs/network/latency/pairs/nyc-jnb-rtt)** | 209.09 ms | 252.44 ms | 9.18 ms | | 9 | 🇺🇸 Ashburn (IAD) | **[228.3 ms](/docs/network/latency/pairs/iad-jnb-rtt)** | 216.38 ms | 264.39 ms | 10.74 ms | | 10 | 🇺🇸 Miami (MIA) | **[259.8 ms](/docs/network/latency/pairs/mia-jnb-rtt)** | 252.44 ms | 277.9 ms | 6.11 ms | ## Inbound Latency to Johannesburg (JNB) Round-trip time in milliseconds **from all other PoPs to Johannesburg**. The measurement origin matters: inbound and outbound RTT can differ for the same pair. | Source | RTT (ms) | Fiber Floor | Efficiency | Jitter | Loss | Tier | | ---------------------- | ---------------------------------------------------- | ----------- | ---------- | ------- | ----- | ---- | | 🇳🇱 Amsterdam (AMS) | **[171.9](/docs/network/latency/pairs/ams-jnb-rtt)** | 88.05 ms | 51.2% | 1.1 ms | 0.1% | Fair | | 🇩🇪 Berlin (BER) | **[172.1](/docs/network/latency/pairs/ber-jnb-rtt)** | 86.53 ms | 50.3% | 1.56 ms | 0% | Fair | | 🇬🇧 London (LON) | **[179.6](/docs/network/latency/pairs/lon-jnb-rtt)** | 88.52 ms | 49.3% | 0.86 ms | 0% | Fair | | 🇫🇷 Paris (PAR) | **[180.9](/docs/network/latency/pairs/par-jnb-rtt)** | 85.17 ms | 47.1% | 0.94 ms | 0.17% | Fair | | 🇩🇪 Frankfurt (FRA) | **[188.4](/docs/network/latency/pairs/fra-jnb-rtt)** | 84.87 ms | 45% | 1.61 ms | 0.1% | Fair | | 🇫🇷 Marseille (MRS) | **[198.5](/docs/network/latency/pairs/mrs-jnb-rtt)** | 78.71 ms | 39.7% | 1.78 ms | 0% | Fair | | 🇷🇺 Moscow (MOW) | **[201.9](/docs/network/latency/pairs/mow-jnb-rtt)** | 89.37 ms | 44.3% | 1.29 ms | 0% | Fair | | 🇺🇸 New York (NYC) | **[221.5](/docs/network/latency/pairs/nyc-jnb-rtt)** | 125.67 ms | 56.7% | 2.09 ms | 0% | Fair | | 🇺🇸 Ashburn (IAD) | **[228.3](/docs/network/latency/pairs/iad-jnb-rtt)** | 128.11 ms | 56.1% | 1.05 ms | 0.04% | Fair | | 🇺🇸 Miami (MIA) | **[259.8](/docs/network/latency/pairs/mia-jnb-rtt)** | 126.78 ms | 48.8% | 1.51 ms | 0.19% | High | | 🇺🇸 Seattle (SEA) | **[282.1](/docs/network/latency/pairs/sea-jnb-rtt)** | 161.61 ms | 57.3% | 1.53 ms | 0.15% | High | | 🇺🇸 Los Angeles (LAX) | **[287.2](/docs/network/latency/pairs/lax-jnb-rtt)** | 163.35 ms | 56.9% | 1.6 ms | 0.08% | High | | 🇸🇬 Singapore (SIN) | **[312.6](/docs/network/latency/pairs/sin-jnb-rtt)** | 84.85 ms | 27.1% | 1.51 ms | 0.19% | High | | 🇭🇰 Hong Kong (HKG) | **[316.7](/docs/network/latency/pairs/hkg-jnb-rtt)** | 104.93 ms | 33.1% | 3.12 ms | 0% | High | | 🇹🇼 Taipei (TPE) | **[331](/docs/network/latency/pairs/tpe-jnb-rtt)** | 112.87 ms | 34.1% | 1.92 ms | 0% | High | | 🇧🇷 São Paulo (GRU) | **[333.8](/docs/network/latency/pairs/gru-jnb-rtt)** | 72.88 ms | 21.8% | 3.1 ms | 0.07% | High | | 🇯🇵 Tokyo (TYO) | **[359.2](/docs/network/latency/pairs/tyo-jnb-rtt)** | 132.57 ms | 36.9% | 2.19 ms | 0.05% | High | | 🇦🇺 Melbourne (MEL) | **[405.5](/docs/network/latency/pairs/mel-jnb-rtt)** | 101.33 ms | 25% | 3.75 ms | 0.06% | High | | 🇦🇺 Sydney (SYD) | **[412.5](/docs/network/latency/pairs/syd-jnb-rtt)** | 108.32 ms | 26.3% | 3.8 ms | 0% | High | ## Questions ### What is the fastest route to Johannesburg? The fastest measured route to Johannesburg (JNB) is [Amsterdam (AMS) → Johannesburg (JNB)](/docs/network/latency/pairs/ams-jnb-rtt), averaging **171.9 ms** RTT (Fair). ### What is the fastest route from Johannesburg? The fastest measured route from Johannesburg (JNB) is [Johannesburg (JNB) → London (LON)](/docs/network/latency/pairs/jnb-lon-rtt), averaging **158 ms** RTT (Fair). ### How well connected is Johannesburg to the Hats Network backbone? Johannesburg (JNB) maintains measured routes to all 19 other PoPs across 4 continents. The slowest route, [Johannesburg (JNB) → Sydney (SYD)](/docs/network/latency/pairs/jnb-syd-rtt), averages **412.6 ms** RTT. *** *Round published August 16, 2026. [All 20 PoPs](/docs/network/latency) · [Dataset and licence](/opendata/latency/).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/jnb-johannesburg). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Los Angeles network latency (LAX) Los Angeles (LAX) is a US West Coast PoP. The round includes routes to East Asia as well as US and European destinations. Measurement round: `2026-08-16T04:07:28Z`. The published round covers **19 outbound routes** from Los Angeles. Incoming routes are measured separately; neither direction describes last-mile performance. ## Location * **Location:** Los Angeles, USA (North America) * **Region:** North America * **Role:** Backbone interconnection point for Hats Network (AS203314) * **Public city reference:** [GeoNames 5368361](https://www.geonames.org/5368361); 34.05223, -118.24368 * **Coverage:** RTT measurements to 19 other PoPs across 5 continents ### Interconnection * **Facility:** LAX1 * **Upstream transit:** HE / Cogent / Comcast > Interactive content is available on the canonical HTML page. ## Round Summary | Metric | Value | | --------------------- | --------------------------------------------------------------------------------------------------------------- | | Fastest Route | [🇺🇸 Los Angeles (LAX) → 🇺🇸 Seattle (SEA)](/docs/network/latency/pairs/lax-sea-rtt) (**26.9 ms**); Ultra-Low | | Slowest Route | [🇺🇸 Los Angeles (LAX) → 🇿🇦 Johannesburg (JNB)](/docs/network/latency/pairs/lax-jnb-rtt) (**287.2 ms**) | | Average RTT | **129.2 ms** | | Average Jitter | **1.03 ms** | | Average Packet Loss | **0.04%** | | Best Fiber Efficiency | **85.9%** | | Intra-Region Peers | 4 PoPs in North America | | Total Routes | 19 outbound / 19 inbound (100% coverage) | > Jitter, packet-loss, percentile, and trend fields are estimated comparative indicators. Read the [theoretical fiber latency methodology](/docs/network/latency/theoretical-fiber-latency). ## Regional Routes Los Angeles's 4 other **North America** PoPs have RTTs from **26.9 ms** (Seattle (SEA)) to **60.4 ms** (Ashburn (IAD)). Los Angeles is in southern California on the US Pacific coast. ## Outbound Latency from Los Angeles (LAX) Round-trip time in milliseconds **from Los Angeles** to all other backbone PoPs, sorted fastest to slowest. | Destination | RTT (ms) | Fiber Floor | Efficiency | Jitter | Loss | Tier | | ----------------------- | ---------------------------------------------------- | ----------- | ---------- | ------- | ----- | --------- | | 🇺🇸 Seattle (SEA) | **[26.9](/docs/network/latency/pairs/lax-sea-rtt)** | 15.12 ms | 56.2% | 0.3 ms | 0.13% | Ultra-Low | | 🇺🇸 Miami (MIA) | **[57.3](/docs/network/latency/pairs/lax-mia-rtt)** | 36.86 ms | 64.3% | 0.34 ms | 0% | Excellent | | 🇺🇸 New York (NYC) | **[57.9](/docs/network/latency/pairs/lax-nyc-rtt)** | 38.63 ms | 66.7% | 0.34 ms | 0.09% | Excellent | | 🇺🇸 Ashburn (IAD) | **[60.4](/docs/network/latency/pairs/lax-iad-rtt)** | 35.84 ms | 59.3% | 0.64 ms | 0.06% | Excellent | | 🇯🇵 Tokyo (TYO) | **[101.2](/docs/network/latency/pairs/lax-tyo-rtt)** | 86.51 ms | 85.5% | 0.82 ms | 0% | Good | | 🇬🇧 London (LON) | **[129.2](/docs/network/latency/pairs/lax-lon-rtt)** | 85.95 ms | 66.5% | 1.17 ms | 0% | Good | | 🇧🇷 São Paulo (GRU) | **[131.3](/docs/network/latency/pairs/lax-gru-rtt)** | 96.89 ms | 73.8% | 0.87 ms | 0% | Good | | 🇳🇱 Amsterdam (AMS) | **[132.1](/docs/network/latency/pairs/lax-ams-rtt)** | 87.74 ms | 66.4% | 0.78 ms | 0% | Good | | 🇹🇼 Taipei (TPE) | **[132.9](/docs/network/latency/pairs/lax-tpe-rtt)** | 107 ms | 80.5% | 0.87 ms | 0.08% | Good | | 🇫🇷 Paris (PAR) | **[135.9](/docs/network/latency/pairs/lax-par-rtt)** | 89.18 ms | 65.6% | 0.62 ms | 0% | Good | | 🇦🇺 Sydney (SYD) | **[137.6](/docs/network/latency/pairs/lax-syd-rtt)** | 118.13 ms | 85.9% | 1.65 ms | 0% | Good | | 🇩🇪 Berlin (BER) | **[139.3](/docs/network/latency/pairs/lax-ber-rtt)** | 91.39 ms | 65.6% | 1.58 ms | 0.12% | Good | | 🇩🇪 Frankfurt (FRA) | **[142.3](/docs/network/latency/pairs/lax-fra-rtt)** | 91.31 ms | 64.2% | 0.71 ms | 0% | Good | | 🇫🇷 Marseille (MRS) | **[143.9](/docs/network/latency/pairs/lax-mrs-rtt)** | 95.09 ms | 66.1% | 1.05 ms | 0.16% | Good | | 🇭🇰 Hong Kong (HKG) | **[146.5](/docs/network/latency/pairs/lax-hkg-rtt)** | 114.3 ms | 78% | 1.42 ms | 0% | Good | | 🇦🇺 Melbourne (MEL) | **[148.2](/docs/network/latency/pairs/lax-mel-rtt)** | 125 ms | 84.3% | 1.74 ms | 0.06% | Good | | 🇸🇬 Singapore (SIN) | **[167.7](/docs/network/latency/pairs/lax-sin-rtt)** | 138.43 ms | 82.5% | 1.33 ms | 0% | Fair | | 🇷🇺 Moscow (MOW) | **[177](/docs/network/latency/pairs/lax-mow-rtt)** | 95.9 ms | 54.2% | 1.7 ms | 0% | Fair | | 🇿🇦 Johannesburg (JNB) | **[287.2](/docs/network/latency/pairs/lax-jnb-rtt)** | 163.35 ms | 56.9% | 1.6 ms | 0.08% | High | ## Fastest Routes to Los Angeles (LAX) The 10 fastest measured routes **to Los Angeles**, ranked by average RTT. Min / max / stdev are computed from the same measurement round (50 ICMP echo probes per route, 100 ms interval); routes without a published per-sample round show the measured average RTT only. | Rank | Source | Avg RTT | Min | Max | Stdev | | ---- | -------------------- | ------------------------------------------------------- | --------- | --------- | ------- | | 1 | 🇺🇸 Seattle (SEA) | **[25.9 ms](/docs/network/latency/pairs/sea-lax-rtt)** | 25.44 ms | 27.56 ms | 0.42 ms | | 2 | 🇺🇸 Miami (MIA) | **[56.8 ms](/docs/network/latency/pairs/mia-lax-rtt)** | 54.44 ms | 64.31 ms | 2.11 ms | | 3 | 🇺🇸 New York (NYC) | **[58.9 ms](/docs/network/latency/pairs/nyc-lax-rtt)** | 56.67 ms | 64.24 ms | 2.02 ms | | 4 | 🇺🇸 Ashburn (IAD) | **[60.7 ms](/docs/network/latency/pairs/iad-lax-rtt)** | 58.56 ms | 67.73 ms | 1.84 ms | | 5 | 🇯🇵 Tokyo (TYO) | **[101.2 ms](/docs/network/latency/pairs/tyo-lax-rtt)** | 98.89 ms | 105.47 ms | 1.62 ms | | 6 | 🇬🇧 London (LON) | **[127.5 ms](/docs/network/latency/pairs/lon-lax-rtt)** | 123.84 ms | 138.75 ms | 3.03 ms | | 7 | 🇧🇷 São Paulo (GRU) | **[130.8 ms](/docs/network/latency/pairs/gru-lax-rtt)** | 125.52 ms | 141.14 ms | 3.75 ms | | 8 | 🇹🇼 Taipei (TPE) | **[132 ms](/docs/network/latency/pairs/tpe-lax-rtt)** | 129.33 ms | 144.07 ms | 2.56 ms | | 9 | 🇳🇱 Amsterdam (AMS) | **[133.7 ms](/docs/network/latency/pairs/ams-lax-rtt)** | 130.58 ms | 141.4 ms | 2.49 ms | | 10 | 🇫🇷 Paris (PAR) | **[136.5 ms](/docs/network/latency/pairs/par-lax-rtt)** | 130.92 ms | 152.23 ms | 5.13 ms | ## Inbound Latency to Los Angeles (LAX) Round-trip time in milliseconds **from all other PoPs to Los Angeles**. The measurement origin matters: inbound and outbound RTT can differ for the same pair. | Source | RTT (ms) | Fiber Floor | Efficiency | Jitter | Loss | Tier | | ----------------------- | ---------------------------------------------------- | ----------- | ---------- | ------- | ----- | --------- | | 🇺🇸 Seattle (SEA) | **[25.9](/docs/network/latency/pairs/sea-lax-rtt)** | 15.12 ms | 58.4% | 0.24 ms | 0% | Ultra-Low | | 🇺🇸 Miami (MIA) | **[56.8](/docs/network/latency/pairs/mia-lax-rtt)** | 36.86 ms | 64.9% | 0.59 ms | 0% | Excellent | | 🇺🇸 New York (NYC) | **[58.9](/docs/network/latency/pairs/nyc-lax-rtt)** | 38.63 ms | 65.6% | 0.46 ms | 0% | Excellent | | 🇺🇸 Ashburn (IAD) | **[60.7](/docs/network/latency/pairs/iad-lax-rtt)** | 35.84 ms | 59% | 0.32 ms | 0% | Excellent | | 🇯🇵 Tokyo (TYO) | **[101.2](/docs/network/latency/pairs/tyo-lax-rtt)** | 86.51 ms | 85.5% | 0.97 ms | 0.02% | Good | | 🇬🇧 London (LON) | **[127.5](/docs/network/latency/pairs/lon-lax-rtt)** | 85.95 ms | 67.4% | 0.86 ms | 0.17% | Good | | 🇧🇷 São Paulo (GRU) | **[130.8](/docs/network/latency/pairs/gru-lax-rtt)** | 96.89 ms | 74.1% | 0.66 ms | 0% | Good | | 🇹🇼 Taipei (TPE) | **[132](/docs/network/latency/pairs/tpe-lax-rtt)** | 107 ms | 81.1% | 1.31 ms | 0% | Good | | 🇳🇱 Amsterdam (AMS) | **[133.7](/docs/network/latency/pairs/ams-lax-rtt)** | 87.74 ms | 65.6% | 0.65 ms | 0% | Good | | 🇫🇷 Paris (PAR) | **[136.5](/docs/network/latency/pairs/par-lax-rtt)** | 89.18 ms | 65.3% | 1.58 ms | 0.11% | Good | | 🇦🇺 Sydney (SYD) | **[136.5](/docs/network/latency/pairs/syd-lax-rtt)** | 118.13 ms | 86.5% | 1.54 ms | 0.12% | Good | | 🇩🇪 Berlin (BER) | **[140.9](/docs/network/latency/pairs/ber-lax-rtt)** | 91.39 ms | 64.9% | 1.62 ms | 0.1% | Good | | 🇩🇪 Frankfurt (FRA) | **[141.8](/docs/network/latency/pairs/fra-lax-rtt)** | 91.31 ms | 64.4% | 1.06 ms | 0% | Good | | 🇫🇷 Marseille (MRS) | **[144.5](/docs/network/latency/pairs/mrs-lax-rtt)** | 95.09 ms | 65.8% | 0.78 ms | 0% | Good | | 🇭🇰 Hong Kong (HKG) | **[145](/docs/network/latency/pairs/hkg-lax-rtt)** | 114.3 ms | 78.8% | 1.4 ms | 0.06% | Good | | 🇦🇺 Melbourne (MEL) | **[148.4](/docs/network/latency/pairs/mel-lax-rtt)** | 125 ms | 84.2% | 1.28 ms | 0.07% | Good | | 🇸🇬 Singapore (SIN) | **[168.2](/docs/network/latency/pairs/sin-lax-rtt)** | 138.43 ms | 82.3% | 1.78 ms | 0.03% | Fair | | 🇷🇺 Moscow (MOW) | **[177.7](/docs/network/latency/pairs/mow-lax-rtt)** | 95.9 ms | 54% | 1.55 ms | 0% | Fair | | 🇿🇦 Johannesburg (JNB) | **[285.5](/docs/network/latency/pairs/jnb-lax-rtt)** | 163.35 ms | 57.2% | 2.14 ms | 0.01% | High | ## North America Peers Los Angeles is one of 5 Hats Network PoPs in **North America**. The table lists outbound measurements within the region. | PoP | Code | Country | RTT (ms) | Tier | | ------------------------------------- | ---- | ------- | -------- | --------- | | [🇺🇸 Ashburn (IAD)](./iad-ashburn) | IAD | USA | **60.4** | Excellent | | [🇺🇸 Miami (MIA)](./mia-miami) | MIA | USA | **57.3** | Excellent | | [🇺🇸 New York (NYC)](./nyc-new-york) | NYC | USA | **57.9** | Excellent | | [🇺🇸 Seattle (SEA)](./sea-seattle) | SEA | USA | **26.9** | Ultra-Low | ## Questions ### What is the fastest route to Los Angeles? The fastest measured route to Los Angeles (LAX) is [Seattle (SEA) → Los Angeles (LAX)](/docs/network/latency/pairs/sea-lax-rtt), averaging **25.9 ms** RTT (Ultra-Low). ### What is the fastest route from Los Angeles? The fastest measured route from Los Angeles (LAX) is [Los Angeles (LAX) → Seattle (SEA)](/docs/network/latency/pairs/lax-sea-rtt), averaging **26.9 ms** RTT (Ultra-Low). ### How well connected is Los Angeles to the Hats Network backbone? Los Angeles (LAX) maintains measured routes to all 19 other PoPs across 5 continents. The slowest route, [Los Angeles (LAX) → Johannesburg (JNB)](/docs/network/latency/pairs/lax-jnb-rtt), averages **287.2 ms** RTT. *** *Round published August 16, 2026. [All 20 PoPs](/docs/network/latency) · [Dataset and licence](/opendata/latency/).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/lax-los-angeles). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # London network latency (LON) ## Routes Published ICMP round-trip measurements between London and the other backbone PoPs. These are inter-PoP measurements, not last-mile speed tests. | Destination | Outbound RTT (ms) | Inbound RTT (ms) | | --- | ---: | ---: | | Amsterdam (AMS) | [5.2](/docs/network/latency/pairs/lon-ams-rtt) | [5.2](/docs/network/latency/pairs/ams-lon-rtt) | | Paris (PAR) | [6.4](/docs/network/latency/pairs/lon-par-rtt) | [6.4](/docs/network/latency/pairs/par-lon-rtt) | | Frankfurt (FRA) | [13.6](/docs/network/latency/pairs/lon-fra-rtt) | [12.9](/docs/network/latency/pairs/fra-lon-rtt) | | Berlin (BER) | [16.2](/docs/network/latency/pairs/lon-ber-rtt) | [15.2](/docs/network/latency/pairs/ber-lon-rtt) | | Marseille (MRS) | [17.9](/docs/network/latency/pairs/lon-mrs-rtt) | [18.9](/docs/network/latency/pairs/mrs-lon-rtt) | | Moscow (MOW) | [43.5](/docs/network/latency/pairs/lon-mow-rtt) | [42.4](/docs/network/latency/pairs/mow-lon-rtt) | | New York (NYC) | [63.8](/docs/network/latency/pairs/lon-nyc-rtt) | [63.5](/docs/network/latency/pairs/nyc-lon-rtt) | | Ashburn (IAD) | [71.0](/docs/network/latency/pairs/lon-iad-rtt) | [70.3](/docs/network/latency/pairs/iad-lon-rtt) | | Miami (MIA) | [101.9](/docs/network/latency/pairs/lon-mia-rtt) | [101.8](/docs/network/latency/pairs/mia-lon-rtt) | | Seattle (SEA) | [123.2](/docs/network/latency/pairs/lon-sea-rtt) | [124.1](/docs/network/latency/pairs/sea-lon-rtt) | | Los Angeles (LAX) | [127.5](/docs/network/latency/pairs/lon-lax-rtt) | [129.2](/docs/network/latency/pairs/lax-lon-rtt) | | Singapore (SIN) | [156.5](/docs/network/latency/pairs/lon-sin-rtt) | [155.4](/docs/network/latency/pairs/sin-lon-rtt) | | Hong Kong (HKG) | [160.6](/docs/network/latency/pairs/lon-hkg-rtt) | [158.7](/docs/network/latency/pairs/hkg-lon-rtt) | | São Paulo (GRU) | [170.7](/docs/network/latency/pairs/lon-gru-rtt) | [175.8](/docs/network/latency/pairs/gru-lon-rtt) | | Taipei (TPE) | [173.8](/docs/network/latency/pairs/lon-tpe-rtt) | [173.0](/docs/network/latency/pairs/tpe-lon-rtt) | | Johannesburg (JNB) | [179.6](/docs/network/latency/pairs/lon-jnb-rtt) | [158.0](/docs/network/latency/pairs/jnb-lon-rtt) | | Tokyo (TYO) | [202.9](/docs/network/latency/pairs/lon-tyo-rtt) | [202.7](/docs/network/latency/pairs/tyo-lon-rtt) | | Melbourne (MEL) | [249.6](/docs/network/latency/pairs/lon-mel-rtt) | [252.0](/docs/network/latency/pairs/mel-lon-rtt) | | Sydney (SYD) | [258.1](/docs/network/latency/pairs/lon-syd-rtt) | [256.5](/docs/network/latency/pairs/syd-lon-rtt) | ## Node & interconnection Network Hats Network / AS203314 Location London, UK Facility LON1 Transit HE / Cogent / LINX Peering / IX LINX / DE-CIX London Region Europe London, UK City reference image City reference: [GeoNames 2643743](https://www.geonames.org/2643743); 51.50853, -0.12574. Coordinates refer to the city, not the facility. ## Measurement notes ### Why are outbound and inbound values different? Outbound probes start at London; inbound probes start at the remote PoP. Each value is a complete round trip. Routing, congestion and the measurement origin can produce different results in the two directions. ### What do the probe statistics describe? Minimum, maximum and standard deviation describe the published samples in each route's probe round. Each route includes its sample count, interval and timestamp. A short round does not establish long-term reliability or an SLA. ### What is measured, and what is modeled? The comparison table shows measured RTT. Fiber floors, efficiency, jitter estimates and trend comparisons on the individual route pages are calculated indicators. They do not identify a physical cable path or replace loss and jitter measurements. Read the [methodology](/docs/network/latency/theoretical-fiber-latency). Measurement round: `2026-08-16T04:07:28Z`. [All PoPs](/docs/network/latency) / [Dataset and licence](/opendata/latency/). --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/lon-london). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Melbourne network latency (MEL) Melbourne (MEL) is one of two Australian PoPs in the dataset. Sydney is the other Australian measurement origin. Measurement round: `2026-08-16T04:07:28Z`. The published round covers **19 outbound routes** from Melbourne. Incoming routes are measured separately; neither direction describes last-mile performance. ## Location * **Location:** Melbourne, Australia (Asia Pacific) * **Region:** Asia Pacific * **Role:** Backbone interconnection point for Hats Network (AS203314) * **Public city reference:** [GeoNames 2158177](https://www.geonames.org/2158177); -37.81400, 144.96332 * **Coverage:** RTT measurements to 19 other PoPs across 5 continents ### Interconnection * **Facility:** MEL1 * **Upstream transit:** HE / Superloop > Interactive content is available on the canonical HTML page. ## Round Summary | Metric | Value | | --------------------- | ----------------------------------------------------------------------------------------------------------- | | Fastest Route | [🇦🇺 Melbourne (MEL) → 🇦🇺 Sydney (SYD)](/docs/network/latency/pairs/mel-syd-rtt) (**9.8 ms**); Ultra-Low | | Slowest Route | [🇦🇺 Melbourne (MEL) → 🇿🇦 Johannesburg (JNB)](/docs/network/latency/pairs/mel-jnb-rtt) (**405.5 ms**) | | Average RTT | **203.0 ms** | | Average Jitter | **1.75 ms** | | Average Packet Loss | **0.04%** | | Best Fiber Efficiency | **84.2%** | | Intra-Region Peers | 5 PoPs in Asia Pacific | | Total Routes | 19 outbound / 19 inbound (100% coverage) | > Jitter, packet-loss, percentile, and trend fields are estimated comparative indicators. Read the [theoretical fiber latency methodology](/docs/network/latency/theoretical-fiber-latency). ## Regional Routes Melbourne's 5 other **Asia Pacific** PoPs have RTTs from **9.8 ms** (Sydney (SYD)) to **143.2 ms** (Taipei (TPE)). Melbourne is in southeastern Australia, southwest of Sydney. ## Outbound Latency from Melbourne (MEL) Round-trip time in milliseconds **from Melbourne** to all other backbone PoPs, sorted fastest to slowest. | Destination | RTT (ms) | Fiber Floor | Efficiency | Jitter | Loss | Tier | | ----------------------- | ---------------------------------------------------- | ----------- | ---------- | ------- | ----- | --------- | | 🇦🇺 Sydney (SYD) | **[9.8](/docs/network/latency/pairs/mel-syd-rtt)** | 6.99 ms | 71.3% | 0.12 ms | 0% | Ultra-Low | | 🇸🇬 Singapore (SIN) | **[88.4](/docs/network/latency/pairs/mel-sin-rtt)** | 59.22 ms | 67% | 0.89 ms | 0.02% | Good | | 🇯🇵 Tokyo (TYO) | **[112](/docs/network/latency/pairs/mel-tyo-rtt)** | 79.87 ms | 71.3% | 0.86 ms | 0% | Good | | 🇭🇰 Hong Kong (HKG) | **[138.5](/docs/network/latency/pairs/mel-hkg-rtt)** | 72.38 ms | 52.3% | 1.39 ms | 0.19% | Good | | 🇹🇼 Taipei (TPE) | **[143.2](/docs/network/latency/pairs/mel-tpe-rtt)** | 72.25 ms | 50.5% | 0.65 ms | 0% | Good | | 🇺🇸 Los Angeles (LAX) | **[148.4](/docs/network/latency/pairs/mel-lax-rtt)** | 125 ms | 84.2% | 1.28 ms | 0.07% | Good | | 🇺🇸 Seattle (SEA) | **[172.1](/docs/network/latency/pairs/mel-sea-rtt)** | 128.93 ms | 74.9% | 1.49 ms | 0% | Fair | | 🇺🇸 Miami (MIA) | **[201](/docs/network/latency/pairs/mel-mia-rtt)** | 152.69 ms | 76% | 1.83 ms | 0% | Fair | | 🇺🇸 Ashburn (IAD) | **[203.9](/docs/network/latency/pairs/mel-iad-rtt)** | 160.09 ms | 78.5% | 2.45 ms | 0.02% | Fair | | 🇺🇸 New York (NYC) | **[204.2](/docs/network/latency/pairs/mel-nyc-rtt)** | 163.27 ms | 80% | 2.1 ms | 0% | Fair | | 🇫🇷 Marseille (MRS) | **[233.5](/docs/network/latency/pairs/mel-mrs-rtt)** | 162.36 ms | 69.5% | 2.54 ms | 0% | Fair | | 🇷🇺 Moscow (MOW) | **[236.1](/docs/network/latency/pairs/mel-mow-rtt)** | 141.13 ms | 59.8% | 2.48 ms | 0.05% | Fair | | 🇫🇷 Paris (PAR) | **[246.5](/docs/network/latency/pairs/mel-par-rtt)** | 164.4 ms | 66.7% | 1.53 ms | 0.13% | Fair | | 🇩🇪 Frankfurt (FRA) | **[248.6](/docs/network/latency/pairs/mel-fra-rtt)** | 159.77 ms | 64.3% | 2.09 ms | 0.16% | Fair | | 🇬🇧 London (LON) | **[252](/docs/network/latency/pairs/mel-lon-rtt)** | 165.48 ms | 65.7% | 2.3 ms | 0.04% | High | | 🇳🇱 Amsterdam (AMS) | **[252.8](/docs/network/latency/pairs/mel-ams-rtt)** | 161.98 ms | 64.1% | 1.86 ms | 0% | High | | 🇩🇪 Berlin (BER) | **[254.7](/docs/network/latency/pairs/mel-ber-rtt)** | 156.34 ms | 61.4% | 1.95 ms | 0% | High | | 🇧🇷 São Paulo (GRU) | **[305.3](/docs/network/latency/pairs/mel-gru-rtt)** | 128.34 ms | 42% | 1.69 ms | 0% | High | | 🇿🇦 Johannesburg (JNB) | **[405.5](/docs/network/latency/pairs/mel-jnb-rtt)** | 101.33 ms | 25% | 3.75 ms | 0.06% | High | ## Fastest Routes to Melbourne (MEL) The 10 fastest measured routes **to Melbourne**, ranked by average RTT. Min / max / stdev are computed from the same measurement round (50 ICMP echo probes per route, 100 ms interval); routes without a published per-sample round show the measured average RTT only. | Rank | Source | Avg RTT | Min | Max | Stdev | | ---- | ---------------------- | ------------------------------------------------------- | --------- | --------- | ------- | | 1 | 🇦🇺 Sydney (SYD) | **[9.8 ms](/docs/network/latency/pairs/syd-mel-rtt)** | 9.38 ms | 11.84 ms | 0.46 ms | | 2 | 🇸🇬 Singapore (SIN) | **[88.6 ms](/docs/network/latency/pairs/sin-mel-rtt)** | 86.6 ms | 94.44 ms | 1.9 ms | | 3 | 🇯🇵 Tokyo (TYO) | **[113.9 ms](/docs/network/latency/pairs/tyo-mel-rtt)** | 107.58 ms | 128.46 ms | 4.6 ms | | 4 | 🇭🇰 Hong Kong (HKG) | **[138.4 ms](/docs/network/latency/pairs/hkg-mel-rtt)** | 132.66 ms | 153.75 ms | 4.9 ms | | 5 | 🇹🇼 Taipei (TPE) | **[142 ms](/docs/network/latency/pairs/tpe-mel-rtt)** | 135.86 ms | 164.91 ms | 6.17 ms | | 6 | 🇺🇸 Los Angeles (LAX) | **[148.2 ms](/docs/network/latency/pairs/lax-mel-rtt)** | 142.14 ms | 162.19 ms | 5.38 ms | | 7 | 🇺🇸 Seattle (SEA) | **[170.1 ms](/docs/network/latency/pairs/sea-mel-rtt)** | 164.93 ms | 189.47 ms | 4.46 ms | | 8 | 🇺🇸 Miami (MIA) | **[200.7 ms](/docs/network/latency/pairs/mia-mel-rtt)** | 190.44 ms | 231.45 ms | 9.1 ms | | 9 | 🇺🇸 Ashburn (IAD) | **[203.9 ms](/docs/network/latency/pairs/iad-mel-rtt)** | 200.88 ms | 217.16 ms | 3.32 ms | | 10 | 🇺🇸 New York (NYC) | **[205.1 ms](/docs/network/latency/pairs/nyc-mel-rtt)** | 196.41 ms | 229.31 ms | 6.74 ms | ## Inbound Latency to Melbourne (MEL) Round-trip time in milliseconds **from all other PoPs to Melbourne**. The measurement origin matters: inbound and outbound RTT can differ for the same pair. | Source | RTT (ms) | Fiber Floor | Efficiency | Jitter | Loss | Tier | | ----------------------- | ---------------------------------------------------- | ----------- | ---------- | ------- | ----- | --------- | | 🇦🇺 Sydney (SYD) | **[9.8](/docs/network/latency/pairs/syd-mel-rtt)** | 6.99 ms | 71.3% | 0.12 ms | 0.19% | Ultra-Low | | 🇸🇬 Singapore (SIN) | **[88.6](/docs/network/latency/pairs/sin-mel-rtt)** | 59.22 ms | 66.8% | 0.7 ms | 0% | Good | | 🇯🇵 Tokyo (TYO) | **[113.9](/docs/network/latency/pairs/tyo-mel-rtt)** | 79.87 ms | 70.1% | 1.25 ms | 0.05% | Good | | 🇭🇰 Hong Kong (HKG) | **[138.4](/docs/network/latency/pairs/hkg-mel-rtt)** | 72.38 ms | 52.3% | 1.41 ms | 0% | Good | | 🇹🇼 Taipei (TPE) | **[142](/docs/network/latency/pairs/tpe-mel-rtt)** | 72.25 ms | 50.9% | 1.12 ms | 0.08% | Good | | 🇺🇸 Los Angeles (LAX) | **[148.2](/docs/network/latency/pairs/lax-mel-rtt)** | 125 ms | 84.3% | 1.74 ms | 0.06% | Good | | 🇺🇸 Seattle (SEA) | **[170.1](/docs/network/latency/pairs/sea-mel-rtt)** | 128.93 ms | 75.8% | 0.98 ms | 0% | Fair | | 🇺🇸 Miami (MIA) | **[200.7](/docs/network/latency/pairs/mia-mel-rtt)** | 152.69 ms | 76.1% | 1.72 ms | 0% | Fair | | 🇺🇸 Ashburn (IAD) | **[203.9](/docs/network/latency/pairs/iad-mel-rtt)** | 160.09 ms | 78.5% | 1.06 ms | 0.15% | Fair | | 🇺🇸 New York (NYC) | **[205.1](/docs/network/latency/pairs/nyc-mel-rtt)** | 163.27 ms | 79.6% | 1.79 ms | 0.04% | Fair | | 🇫🇷 Marseille (MRS) | **[234.2](/docs/network/latency/pairs/mrs-mel-rtt)** | 162.36 ms | 69.3% | 1.1 ms | 0% | Fair | | 🇷🇺 Moscow (MOW) | **[235.9](/docs/network/latency/pairs/mow-mel-rtt)** | 141.13 ms | 59.8% | 1.11 ms | 0% | Fair | | 🇫🇷 Paris (PAR) | **[246.4](/docs/network/latency/pairs/par-mel-rtt)** | 164.4 ms | 66.7% | 1.59 ms | 0.17% | Fair | | 🇬🇧 London (LON) | **[249.6](/docs/network/latency/pairs/lon-mel-rtt)** | 165.48 ms | 66.3% | 2.17 ms | 0% | Fair | | 🇩🇪 Frankfurt (FRA) | **[250.1](/docs/network/latency/pairs/fra-mel-rtt)** | 159.77 ms | 63.9% | 2.57 ms | 0% | High | | 🇳🇱 Amsterdam (AMS) | **[252.5](/docs/network/latency/pairs/ams-mel-rtt)** | 161.98 ms | 64.1% | 1.77 ms | 0.03% | High | | 🇩🇪 Berlin (BER) | **[256.2](/docs/network/latency/pairs/ber-mel-rtt)** | 156.34 ms | 61% | 3.01 ms | 0.03% | High | | 🇧🇷 São Paulo (GRU) | **[274.7](/docs/network/latency/pairs/gru-mel-rtt)** | 128.34 ms | 46.7% | 1.71 ms | 0.08% | High | | 🇿🇦 Johannesburg (JNB) | **[406.2](/docs/network/latency/pairs/jnb-mel-rtt)** | 101.33 ms | 24.9% | 2.56 ms | 0% | High | ## Asia Pacific Peers Melbourne is one of 6 Hats Network PoPs in **Asia Pacific**. The table lists outbound measurements within the region. | PoP | Code | Country | RTT (ms) | Tier | | --------------------------------------- | ---- | --------- | --------- | --------- | | [🇭🇰 Hong Kong (HKG)](./hkg-hong-kong) | HKG | Hong Kong | **138.5** | Good | | [🇸🇬 Singapore (SIN)](./sin-singapore) | SIN | Singapore | **88.4** | Good | | [🇦🇺 Sydney (SYD)](./syd-sydney) | SYD | Australia | **9.8** | Ultra-Low | | [🇹🇼 Taipei (TPE)](./tpe-taipei) | TPE | Taiwan | **143.2** | Good | | [🇯🇵 Tokyo (TYO)](./tyo-tokyo) | TYO | Japan | **112** | Good | ## Questions ### What is the fastest route to Melbourne? The fastest measured route to Melbourne (MEL) is [Sydney (SYD) → Melbourne (MEL)](/docs/network/latency/pairs/syd-mel-rtt), averaging **9.8 ms** RTT (Ultra-Low). ### What is the fastest route from Melbourne? The fastest measured route from Melbourne (MEL) is [Melbourne (MEL) → Sydney (SYD)](/docs/network/latency/pairs/mel-syd-rtt), averaging **9.8 ms** RTT (Ultra-Low). ### How well connected is Melbourne to the Hats Network backbone? Melbourne (MEL) maintains measured routes to all 19 other PoPs across 5 continents. The slowest route, [Melbourne (MEL) → Johannesburg (JNB)](/docs/network/latency/pairs/mel-jnb-rtt), averages **405.5 ms** RTT. *** *Round published August 16, 2026. [All 20 PoPs](/docs/network/latency) · [Dataset and licence](/opendata/latency/).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/mel-melbourne). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Miami network latency (MIA) Miami (MIA) is a Hats Network PoP in Florida. Its measurements include São Paulo and the other North American nodes. Measurement round: `2026-08-16T04:07:28Z`. The published round covers **19 outbound routes** from Miami. Incoming routes are measured separately; neither direction describes last-mile performance. ## Location * **Location:** Miami, USA (North America) * **Region:** North America * **Role:** Backbone interconnection point for Hats Network (AS203314) * **Public city reference:** [GeoNames 4164138](https://www.geonames.org/4164138); 25.77427, -80.19366 * **Coverage:** RTT measurements to 19 other PoPs across 5 continents ### Interconnection * **Facility:** MIA1 * **Upstream transit:** Cogent / GTT * **Peering / IX:** NAP of the Americas / Equinix MI > Interactive content is available on the canonical HTML page. ## Round Summary | Metric | Value | | --------------------- | --------------------------------------------------------------------------------------------------------- | | Fastest Route | [🇺🇸 Miami (MIA) → 🇺🇸 Ashburn (IAD)](/docs/network/latency/pairs/mia-iad-rtt) (**27.3 ms**); Ultra-Low | | Slowest Route | [🇺🇸 Miami (MIA) → 🇿🇦 Johannesburg (JNB)](/docs/network/latency/pairs/mia-jnb-rtt) (**259.8 ms**) | | Average RTT | **134.0 ms** | | Average Jitter | **1.06 ms** | | Average Packet Loss | **0.06%** | | Best Fiber Efficiency | **76.9%** | | Intra-Region Peers | 4 PoPs in North America | | Total Routes | 19 outbound / 19 inbound (100% coverage) | > Jitter, packet-loss, percentile, and trend fields are estimated comparative indicators. Read the [theoretical fiber latency methodology](/docs/network/latency/theoretical-fiber-latency). ## Regional Routes Miami's 4 other **North America** PoPs have RTTs from **27.3 ms** (Ashburn (IAD)) to **81.9 ms** (Seattle (SEA)). Miami is in southern Florida, southeast of the other US cities in this dataset. ## Outbound Latency from Miami (MIA) Round-trip time in milliseconds **from Miami** to all other backbone PoPs, sorted fastest to slowest. | Destination | RTT (ms) | Fiber Floor | Efficiency | Jitter | Loss | Tier | | ----------------------- | ---------------------------------------------------- | ----------- | ---------- | ------- | ----- | --------- | | 🇺🇸 Ashburn (IAD) | **[27.3](/docs/network/latency/pairs/mia-iad-rtt)** | 14.62 ms | 53.6% | 0.15 ms | 0% | Ultra-Low | | 🇺🇸 New York (NYC) | **[33.3](/docs/network/latency/pairs/mia-nyc-rtt)** | 17.17 ms | 51.6% | 0.33 ms | 0.08% | Excellent | | 🇺🇸 Los Angeles (LAX) | **[56.8](/docs/network/latency/pairs/mia-lax-rtt)** | 36.86 ms | 64.9% | 0.59 ms | 0% | Excellent | | 🇺🇸 Seattle (SEA) | **[81.9](/docs/network/latency/pairs/mia-sea-rtt)** | 43.08 ms | 52.6% | 0.47 ms | 0.1% | Good | | 🇬🇧 London (LON) | **[101.8](/docs/network/latency/pairs/mia-lon-rtt)** | 69.9 ms | 68.7% | 0.73 ms | 0% | Good | | 🇳🇱 Amsterdam (AMS) | **[106](/docs/network/latency/pairs/mia-ams-rtt)** | 73.03 ms | 68.9% | 0.87 ms | 0.08% | Good | | 🇫🇷 Paris (PAR) | **[106.2](/docs/network/latency/pairs/mia-par-rtt)** | 72.16 ms | 67.9% | 1.07 ms | 0% | Good | | 🇩🇪 Frankfurt (FRA) | **[112.2](/docs/network/latency/pairs/mia-fra-rtt)** | 76.17 ms | 67.9% | 0.83 ms | 0.06% | Good | | 🇫🇷 Marseille (MRS) | **[113.7](/docs/network/latency/pairs/mia-mrs-rtt)** | 75.92 ms | 66.8% | 1.23 ms | 0.14% | Good | | 🇩🇪 Berlin (BER) | **[117.3](/docs/network/latency/pairs/mia-ber-rtt)** | 78.44 ms | 66.9% | 1.4 ms | 0% | Good | | 🇧🇷 São Paulo (GRU) | **[128.7](/docs/network/latency/pairs/mia-gru-rtt)** | 64.11 ms | 49.8% | 0.81 ms | 0.16% | Good | | 🇷🇺 Moscow (MOW) | **[145.5](/docs/network/latency/pairs/mia-mow-rtt)** | 90.48 ms | 62.2% | 1.67 ms | 0.13% | Good | | 🇯🇵 Tokyo (TYO) | **[156.3](/docs/network/latency/pairs/mia-tyo-rtt)** | 117.71 ms | 75.3% | 0.95 ms | 0.03% | Fair | | 🇹🇼 Taipei (TPE) | **[185.1](/docs/network/latency/pairs/mia-tpe-rtt)** | 136.3 ms | 73.6% | 1.17 ms | 0.09% | Fair | | 🇦🇺 Sydney (SYD) | **[191.5](/docs/network/latency/pairs/mia-syd-rtt)** | 147.18 ms | 76.9% | 0.88 ms | 0% | Fair | | 🇭🇰 Hong Kong (HKG) | **[199.5](/docs/network/latency/pairs/mia-hkg-rtt)** | 141.69 ms | 71% | 2.35 ms | 0.06% | Fair | | 🇦🇺 Melbourne (MEL) | **[200.7](/docs/network/latency/pairs/mia-mel-rtt)** | 152.69 ms | 76.1% | 1.72 ms | 0% | Fair | | 🇸🇬 Singapore (SIN) | **[223.1](/docs/network/latency/pairs/mia-sin-rtt)** | 166.27 ms | 74.5% | 1.42 ms | 0.08% | Fair | | 🇿🇦 Johannesburg (JNB) | **[259.8](/docs/network/latency/pairs/mia-jnb-rtt)** | 126.78 ms | 48.8% | 1.51 ms | 0.19% | High | ## Fastest Routes to Miami (MIA) The 10 fastest measured routes **to Miami**, ranked by average RTT. Min / max / stdev are computed from the same measurement round (50 ICMP echo probes per route, 100 ms interval); routes without a published per-sample round show the measured average RTT only. | Rank | Source | Avg RTT | Min | Max | Stdev | | ---- | ---------------------- | ------------------------------------------------------- | --------- | --------- | ------- | | 1 | 🇺🇸 Ashburn (IAD) | **[27.4 ms](/docs/network/latency/pairs/iad-mia-rtt)** | 26.09 ms | 30.01 ms | 1.06 ms | | 2 | 🇺🇸 New York (NYC) | **[32.9 ms](/docs/network/latency/pairs/nyc-mia-rtt)** | 31.54 ms | 39.09 ms | 1.24 ms | | 3 | 🇺🇸 Los Angeles (LAX) | **[57.3 ms](/docs/network/latency/pairs/lax-mia-rtt)** | 55.45 ms | 63.93 ms | 1.73 ms | | 4 | 🇧🇷 São Paulo (GRU) | **[74 ms](/docs/network/latency/pairs/gru-mia-rtt)** | 70.55 ms | 83.79 ms | 3.07 ms | | 5 | 🇺🇸 Seattle (SEA) | **[81.7 ms](/docs/network/latency/pairs/sea-mia-rtt)** | 79.89 ms | 89.57 ms | 1.94 ms | | 6 | 🇬🇧 London (LON) | **[101.9 ms](/docs/network/latency/pairs/lon-mia-rtt)** | 97.83 ms | 115.14 ms | 3.85 ms | | 7 | 🇳🇱 Amsterdam (AMS) | **[106 ms](/docs/network/latency/pairs/ams-mia-rtt)** | 99.94 ms | 119.16 ms | 4.46 ms | | 8 | 🇫🇷 Paris (PAR) | **[106.8 ms](/docs/network/latency/pairs/par-mia-rtt)** | 101.15 ms | 125.38 ms | 5.28 ms | | 9 | 🇩🇪 Frankfurt (FRA) | **[112 ms](/docs/network/latency/pairs/fra-mia-rtt)** | 110.01 ms | 117.35 ms | 1.8 ms | | 10 | 🇫🇷 Marseille (MRS) | **[112.4 ms](/docs/network/latency/pairs/mrs-mia-rtt)** | 105.38 ms | 129.43 ms | 5.13 ms | ## Inbound Latency to Miami (MIA) Round-trip time in milliseconds **from all other PoPs to Miami**. The measurement origin matters: inbound and outbound RTT can differ for the same pair. | Source | RTT (ms) | Fiber Floor | Efficiency | Jitter | Loss | Tier | | ----------------------- | ---------------------------------------------------- | ----------- | ---------- | ------- | ----- | --------- | | 🇺🇸 Ashburn (IAD) | **[27.4](/docs/network/latency/pairs/iad-mia-rtt)** | 14.62 ms | 53.4% | 0.15 ms | 0% | Ultra-Low | | 🇺🇸 New York (NYC) | **[32.9](/docs/network/latency/pairs/nyc-mia-rtt)** | 17.17 ms | 52.2% | 0.39 ms | 0.15% | Excellent | | 🇺🇸 Los Angeles (LAX) | **[57.3](/docs/network/latency/pairs/lax-mia-rtt)** | 36.86 ms | 64.3% | 0.34 ms | 0% | Excellent | | 🇧🇷 São Paulo (GRU) | **[74](/docs/network/latency/pairs/gru-mia-rtt)** | 64.11 ms | 86.6% | 0.62 ms | 0.16% | Excellent | | 🇺🇸 Seattle (SEA) | **[81.7](/docs/network/latency/pairs/sea-mia-rtt)** | 43.08 ms | 52.7% | 0.84 ms | 0% | Good | | 🇬🇧 London (LON) | **[101.9](/docs/network/latency/pairs/lon-mia-rtt)** | 69.9 ms | 68.6% | 0.83 ms | 0.1% | Good | | 🇳🇱 Amsterdam (AMS) | **[106](/docs/network/latency/pairs/ams-mia-rtt)** | 73.03 ms | 68.9% | 1.03 ms | 0.01% | Good | | 🇫🇷 Paris (PAR) | **[106.8](/docs/network/latency/pairs/par-mia-rtt)** | 72.16 ms | 67.6% | 0.65 ms | 0.05% | Good | | 🇩🇪 Frankfurt (FRA) | **[112](/docs/network/latency/pairs/fra-mia-rtt)** | 76.17 ms | 68% | 1.27 ms | 0.07% | Good | | 🇫🇷 Marseille (MRS) | **[112.4](/docs/network/latency/pairs/mrs-mia-rtt)** | 75.92 ms | 67.5% | 0.96 ms | 0.1% | Good | | 🇩🇪 Berlin (BER) | **[118.3](/docs/network/latency/pairs/ber-mia-rtt)** | 78.44 ms | 66.3% | 0.76 ms | 0% | Good | | 🇷🇺 Moscow (MOW) | **[144.5](/docs/network/latency/pairs/mow-mia-rtt)** | 90.48 ms | 62.6% | 1.25 ms | 0.07% | Good | | 🇯🇵 Tokyo (TYO) | **[156.5](/docs/network/latency/pairs/tyo-mia-rtt)** | 117.71 ms | 75.2% | 1.36 ms | 0.14% | Fair | | 🇹🇼 Taipei (TPE) | **[185.7](/docs/network/latency/pairs/tpe-mia-rtt)** | 136.3 ms | 73.4% | 1.27 ms | 0% | Fair | | 🇦🇺 Sydney (SYD) | **[191.5](/docs/network/latency/pairs/syd-mia-rtt)** | 147.18 ms | 76.9% | 0.95 ms | 0.12% | Fair | | 🇦🇺 Melbourne (MEL) | **[201](/docs/network/latency/pairs/mel-mia-rtt)** | 152.69 ms | 76% | 1.83 ms | 0% | Fair | | 🇭🇰 Hong Kong (HKG) | **[201.1](/docs/network/latency/pairs/hkg-mia-rtt)** | 141.69 ms | 70.5% | 0.93 ms | 0.05% | Fair | | 🇸🇬 Singapore (SIN) | **[221.7](/docs/network/latency/pairs/sin-mia-rtt)** | 166.27 ms | 75% | 1.1 ms | 0.19% | Fair | | 🇿🇦 Johannesburg (JNB) | **[259.9](/docs/network/latency/pairs/jnb-mia-rtt)** | 126.78 ms | 48.8% | 3.06 ms | 0.1% | High | ## North America Peers Miami is one of 5 Hats Network PoPs in **North America**. The table lists outbound measurements within the region. | PoP | Code | Country | RTT (ms) | Tier | | ------------------------------------------- | ---- | ------- | -------- | --------- | | [🇺🇸 Ashburn (IAD)](./iad-ashburn) | IAD | USA | **27.3** | Ultra-Low | | [🇺🇸 Los Angeles (LAX)](./lax-los-angeles) | LAX | USA | **56.8** | Excellent | | [🇺🇸 New York (NYC)](./nyc-new-york) | NYC | USA | **33.3** | Excellent | | [🇺🇸 Seattle (SEA)](./sea-seattle) | SEA | USA | **81.9** | Good | ## Questions ### What is the fastest route to Miami? The fastest measured route to Miami (MIA) is [Ashburn (IAD) → Miami (MIA)](/docs/network/latency/pairs/iad-mia-rtt), averaging **27.4 ms** RTT (Ultra-Low). ### What is the fastest route from Miami? The fastest measured route from Miami (MIA) is [Miami (MIA) → Ashburn (IAD)](/docs/network/latency/pairs/mia-iad-rtt), averaging **27.3 ms** RTT (Ultra-Low). ### How well connected is Miami to the Hats Network backbone? Miami (MIA) maintains measured routes to all 19 other PoPs across 5 continents. The slowest route, [Miami (MIA) → Johannesburg (JNB)](/docs/network/latency/pairs/mia-jnb-rtt), averages **259.8 ms** RTT. *** *Round published August 16, 2026. [All 20 PoPs](/docs/network/latency) · [Dataset and licence](/opendata/latency/).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/mia-miami). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Moscow network latency (MOW) Moscow (MOW) is the easternmost European PoP in this dataset. The tables compare its European routes with longer intercontinental paths. Measurement round: `2026-08-16T04:07:28Z`. The published round covers **19 outbound routes** from Moscow. Incoming routes are measured separately; neither direction describes last-mile performance. ## Location * **Location:** Moscow, Russia (Europe) * **Region:** Europe * **Role:** Backbone interconnection point for Hats Network (AS203314) * **Public city reference:** [GeoNames 524901](https://www.geonames.org/524901); 55.75204, 37.61781 * **Coverage:** RTT measurements to 19 other PoPs across 5 continents ### Interconnection * **Facility:** MOW1 * **Upstream transit:** RETN / Melbicom > Interactive content is available on the canonical HTML page. ## Round Summary | Metric | Value | | --------------------- | --------------------------------------------------------------------------------------------------------- | | Fastest Route | [🇷🇺 Moscow (MOW) → 🇩🇪 Berlin (BER)](/docs/network/latency/pairs/mow-ber-rtt) (**27.5 ms**); Ultra-Low | | Slowest Route | [🇷🇺 Moscow (MOW) → 🇦🇺 Sydney (SYD)](/docs/network/latency/pairs/mow-syd-rtt) (**241.8 ms**) | | Average RTT | **126.3 ms** | | Average Jitter | **1.03 ms** | | Average Packet Loss | **0.06%** | | Best Fiber Efficiency | **68.6%** | | Intra-Region Peers | 6 PoPs in Europe | | Total Routes | 19 outbound / 19 inbound (100% coverage) | > Jitter, packet-loss, percentile, and trend fields are estimated comparative indicators. Read the [theoretical fiber latency methodology](/docs/network/latency/theoretical-fiber-latency). ## Regional Routes Moscow's 6 other **Europe** PoPs have RTTs from **27.5 ms** (Berlin (BER)) to **49.9 ms** (Marseille (MRS)). Moscow is in western Russia and is the easternmost European city in this dataset. ## Outbound Latency from Moscow (MOW) Round-trip time in milliseconds **from Moscow** to all other backbone PoPs, sorted fastest to slowest. | Destination | RTT (ms) | Fiber Floor | Efficiency | Jitter | Loss | Tier | | ----------------------- | ---------------------------------------------------- | ----------- | ---------- | ------- | ----- | --------- | | 🇩🇪 Berlin (BER) | **[27.5](/docs/network/latency/pairs/mow-ber-rtt)** | 15.8 ms | 57.5% | 0.3 ms | 0.13% | Ultra-Low | | 🇩🇪 Frankfurt (FRA) | **[35.9](/docs/network/latency/pairs/mow-fra-rtt)** | 19.85 ms | 55.3% | 0.3 ms | 0% | Excellent | | 🇳🇱 Amsterdam (AMS) | **[37.6](/docs/network/latency/pairs/mow-ams-rtt)** | 21.1 ms | 56.1% | 0.28 ms | 0.17% | Excellent | | 🇬🇧 London (LON) | **[42.4](/docs/network/latency/pairs/mow-lon-rtt)** | 24.56 ms | 57.9% | 0.46 ms | 0.18% | Excellent | | 🇫🇷 Paris (PAR) | **[44](/docs/network/latency/pairs/mow-par-rtt)** | 24.42 ms | 55.5% | 0.29 ms | 0% | Excellent | | 🇫🇷 Marseille (MRS) | **[49.9](/docs/network/latency/pairs/mow-mrs-rtt)** | 26.23 ms | 52.6% | 0.4 ms | 0.05% | Excellent | | 🇺🇸 New York (NYC) | **[107.5](/docs/network/latency/pairs/mow-nyc-rtt)** | 73.75 ms | 68.6% | 1 ms | 0% | Good | | 🇺🇸 Ashburn (IAD) | **[113.9](/docs/network/latency/pairs/mow-iad-rtt)** | 76.88 ms | 67.5% | 1 ms | 0.08% | Good | | 🇭🇰 Hong Kong (HKG) | **[118.6](/docs/network/latency/pairs/mow-hkg-rtt)** | 70.05 ms | 59.1% | 0.77 ms | 0.13% | Good | | 🇹🇼 Taipei (TPE) | **[132.1](/docs/network/latency/pairs/mow-tpe-rtt)** | 72.16 ms | 54.6% | 1.49 ms | 0% | Good | | 🇺🇸 Miami (MIA) | **[144.5](/docs/network/latency/pairs/mow-mia-rtt)** | 90.48 ms | 62.6% | 1.25 ms | 0.07% | Good | | 🇸🇬 Singapore (SIN) | **[147.3](/docs/network/latency/pairs/mow-sin-rtt)** | 82.49 ms | 56% | 1.11 ms | 0.17% | Good | | 🇯🇵 Tokyo (TYO) | **[162.2](/docs/network/latency/pairs/mow-tyo-rtt)** | 73.42 ms | 45.3% | 1.92 ms | 0.08% | Fair | | 🇺🇸 Seattle (SEA) | **[164.8](/docs/network/latency/pairs/mow-sea-rtt)** | 82.23 ms | 49.9% | 0.99 ms | 0% | Fair | | 🇺🇸 Los Angeles (LAX) | **[177.7](/docs/network/latency/pairs/mow-lax-rtt)** | 95.9 ms | 54% | 1.55 ms | 0% | Fair | | 🇿🇦 Johannesburg (JNB) | **[201.9](/docs/network/latency/pairs/mow-jnb-rtt)** | 89.37 ms | 44.3% | 1.29 ms | 0% | Fair | | 🇧🇷 São Paulo (GRU) | **[214.4](/docs/network/latency/pairs/mow-gru-rtt)** | 115.47 ms | 53.9% | 1.47 ms | 0% | Fair | | 🇦🇺 Melbourne (MEL) | **[235.9](/docs/network/latency/pairs/mow-mel-rtt)** | 141.13 ms | 59.8% | 1.11 ms | 0% | Fair | | 🇦🇺 Sydney (SYD) | **[241.8](/docs/network/latency/pairs/mow-syd-rtt)** | 141.84 ms | 58.7% | 2.66 ms | 0% | Fair | ## Fastest Routes to Moscow (MOW) The 10 fastest measured routes **to Moscow**, ranked by average RTT. Min / max / stdev are computed from the same measurement round (50 ICMP echo probes per route, 100 ms interval); routes without a published per-sample round show the measured average RTT only. | Rank | Source | Avg RTT | Min | Max | Stdev | | ---- | -------------------- | ------------------------------------------------------- | --------- | --------- | ------- | | 1 | 🇩🇪 Berlin (BER) | **[28.6 ms](/docs/network/latency/pairs/ber-mow-rtt)** | 27.35 ms | 31.58 ms | 1.1 ms | | 2 | 🇩🇪 Frankfurt (FRA) | **[34.7 ms](/docs/network/latency/pairs/fra-mow-rtt)** | 32.66 ms | 40.02 ms | 1.68 ms | | 3 | 🇳🇱 Amsterdam (AMS) | **[38.1 ms](/docs/network/latency/pairs/ams-mow-rtt)** | 35.69 ms | 44.57 ms | 1.9 ms | | 4 | 🇬🇧 London (LON) | **[43.5 ms](/docs/network/latency/pairs/lon-mow-rtt)** | 41.81 ms | 48.52 ms | 1.38 ms | | 5 | 🇫🇷 Paris (PAR) | **[44.7 ms](/docs/network/latency/pairs/par-mow-rtt)** | 43.04 ms | 51.46 ms | 1.56 ms | | 6 | 🇫🇷 Marseille (MRS) | **[49.6 ms](/docs/network/latency/pairs/mrs-mow-rtt)** | 47.26 ms | 58 ms | 2.18 ms | | 7 | 🇺🇸 New York (NYC) | **[107 ms](/docs/network/latency/pairs/nyc-mow-rtt)** | 101.46 ms | 119.28 ms | 4.26 ms | | 8 | 🇺🇸 Ashburn (IAD) | **[112.8 ms](/docs/network/latency/pairs/iad-mow-rtt)** | 109.18 ms | 122.64 ms | 3.63 ms | | 9 | 🇭🇰 Hong Kong (HKG) | **[118.2 ms](/docs/network/latency/pairs/hkg-mow-rtt)** | 115.72 ms | 128.25 ms | 2.23 ms | | 10 | 🇹🇼 Taipei (TPE) | **[132.7 ms](/docs/network/latency/pairs/tpe-mow-rtt)** | 127.77 ms | 145.28 ms | 3.91 ms | ## Inbound Latency to Moscow (MOW) Round-trip time in milliseconds **from all other PoPs to Moscow**. The measurement origin matters: inbound and outbound RTT can differ for the same pair. | Source | RTT (ms) | Fiber Floor | Efficiency | Jitter | Loss | Tier | | ----------------------- | ---------------------------------------------------- | ----------- | ---------- | ------- | ----- | --------- | | 🇩🇪 Berlin (BER) | **[28.6](/docs/network/latency/pairs/ber-mow-rtt)** | 15.8 ms | 55.2% | 0.33 ms | 0% | Ultra-Low | | 🇩🇪 Frankfurt (FRA) | **[34.7](/docs/network/latency/pairs/fra-mow-rtt)** | 19.85 ms | 57.2% | 0.25 ms | 0.13% | Excellent | | 🇳🇱 Amsterdam (AMS) | **[38.1](/docs/network/latency/pairs/ams-mow-rtt)** | 21.1 ms | 55.4% | 0.38 ms | 0% | Excellent | | 🇬🇧 London (LON) | **[43.5](/docs/network/latency/pairs/lon-mow-rtt)** | 24.56 ms | 56.5% | 0.2 ms | 0.06% | Excellent | | 🇫🇷 Paris (PAR) | **[44.7](/docs/network/latency/pairs/par-mow-rtt)** | 24.42 ms | 54.6% | 0.23 ms | 0.2% | Excellent | | 🇫🇷 Marseille (MRS) | **[49.6](/docs/network/latency/pairs/mrs-mow-rtt)** | 26.23 ms | 52.9% | 0.45 ms | 0.07% | Excellent | | 🇺🇸 New York (NYC) | **[107](/docs/network/latency/pairs/nyc-mow-rtt)** | 73.75 ms | 68.9% | 1.08 ms | 0.06% | Good | | 🇺🇸 Ashburn (IAD) | **[112.8](/docs/network/latency/pairs/iad-mow-rtt)** | 76.88 ms | 68.2% | 0.79 ms | 0.13% | Good | | 🇭🇰 Hong Kong (HKG) | **[118.2](/docs/network/latency/pairs/hkg-mow-rtt)** | 70.05 ms | 59.3% | 1.17 ms | 0.11% | Good | | 🇹🇼 Taipei (TPE) | **[132.7](/docs/network/latency/pairs/tpe-mow-rtt)** | 72.16 ms | 54.4% | 0.68 ms | 0.09% | Good | | 🇺🇸 Miami (MIA) | **[145.5](/docs/network/latency/pairs/mia-mow-rtt)** | 90.48 ms | 62.2% | 1.67 ms | 0.13% | Good | | 🇸🇬 Singapore (SIN) | **[147.7](/docs/network/latency/pairs/sin-mow-rtt)** | 82.49 ms | 55.8% | 0.99 ms | 0.11% | Good | | 🇯🇵 Tokyo (TYO) | **[161.7](/docs/network/latency/pairs/tyo-mow-rtt)** | 73.42 ms | 45.4% | 1.62 ms | 0.09% | Fair | | 🇺🇸 Seattle (SEA) | **[164.6](/docs/network/latency/pairs/sea-mow-rtt)** | 82.23 ms | 50% | 1.42 ms | 0% | Fair | | 🇺🇸 Los Angeles (LAX) | **[177](/docs/network/latency/pairs/lax-mow-rtt)** | 95.9 ms | 54.2% | 1.7 ms | 0% | Fair | | 🇿🇦 Johannesburg (JNB) | **[200.7](/docs/network/latency/pairs/jnb-mow-rtt)** | 89.37 ms | 44.5% | 2.17 ms | 0.16% | Fair | | 🇧🇷 São Paulo (GRU) | **[219.5](/docs/network/latency/pairs/gru-mow-rtt)** | 115.47 ms | 52.6% | 1.57 ms | 0% | Fair | | 🇦🇺 Melbourne (MEL) | **[236.1](/docs/network/latency/pairs/mel-mow-rtt)** | 141.13 ms | 59.8% | 2.48 ms | 0.05% | Fair | | 🇦🇺 Sydney (SYD) | **[242.2](/docs/network/latency/pairs/syd-mow-rtt)** | 141.84 ms | 58.6% | 1.9 ms | 0.11% | Fair | ## Europe Peers Moscow is one of 7 Hats Network PoPs in **Europe**. The table lists outbound measurements within the region. | PoP | Code | Country | RTT (ms) | Tier | | --------------------------------------- | ---- | ----------- | -------- | --------- | | [🇳🇱 Amsterdam (AMS)](./ams-amsterdam) | AMS | Netherlands | **37.6** | Excellent | | [🇩🇪 Berlin (BER)](./ber-berlin) | BER | Germany | **27.5** | Ultra-Low | | [🇩🇪 Frankfurt (FRA)](./fra-frankfurt) | FRA | Germany | **35.9** | Excellent | | [🇬🇧 London (LON)](./lon-london) | LON | UK | **42.4** | Excellent | | [🇫🇷 Marseille (MRS)](./mrs-marseille) | MRS | France | **49.9** | Excellent | | [🇫🇷 Paris (PAR)](./par-paris) | PAR | France | **44** | Excellent | ## Questions ### What is the fastest route to Moscow? The fastest measured route to Moscow (MOW) is [Berlin (BER) → Moscow (MOW)](/docs/network/latency/pairs/ber-mow-rtt), averaging **28.6 ms** RTT (Ultra-Low). ### What is the fastest route from Moscow? The fastest measured route from Moscow (MOW) is [Moscow (MOW) → Berlin (BER)](/docs/network/latency/pairs/mow-ber-rtt), averaging **27.5 ms** RTT (Ultra-Low). ### How well connected is Moscow to the Hats Network backbone? Moscow (MOW) maintains measured routes to all 19 other PoPs across 5 continents. The slowest route, [Moscow (MOW) → Sydney (SYD)](/docs/network/latency/pairs/mow-syd-rtt), averages **241.8 ms** RTT. *** *Round published August 16, 2026. [All 20 PoPs](/docs/network/latency) · [Dataset and licence](/opendata/latency/).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/mow-moscow). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Marseille network latency (MRS) Marseille (MRS) is one of the network's two French PoPs. Its route measurements can be compared with those from Paris. Measurement round: `2026-08-16T04:07:28Z`. The published round covers **19 outbound routes** from Marseille. Incoming routes are measured separately; neither direction describes last-mile performance. ## Location * **Location:** Marseille, France (Europe) * **Region:** Europe * **Role:** Backbone interconnection point for Hats Network (AS203314) * **Public city reference:** [GeoNames 2995469](https://www.geonames.org/2995469); 43.29695, 5.38107 * **Coverage:** RTT measurements to 19 other PoPs across 5 continents ### Interconnection * **Facility:** MRS1 > Interactive content is available on the canonical HTML page. ## Round Summary | Metric | Value | | --------------------- | ---------------------------------------------------------------------------------------------------------- | | Fastest Route | [🇫🇷 Marseille (MRS) → 🇫🇷 Paris (PAR)](/docs/network/latency/pairs/mrs-par-rtt) (**8.9 ms**); Ultra-Low | | Slowest Route | [🇫🇷 Marseille (MRS) → 🇦🇺 Sydney (SYD)](/docs/network/latency/pairs/mrs-syd-rtt) (**240.6 ms**) | | Average RTT | **118.4 ms** | | Average Jitter | **0.97 ms** | | Average Packet Loss | **0.05%** | | Best Fiber Efficiency | **77.6%** | | Intra-Region Peers | 6 PoPs in Europe | | Total Routes | 19 outbound / 19 inbound (100% coverage) | > Jitter, packet-loss, percentile, and trend fields are estimated comparative indicators. Read the [theoretical fiber latency methodology](/docs/network/latency/theoretical-fiber-latency). ## Regional Routes Marseille's 6 other **Europe** PoPs have RTTs from **8.9 ms** (Paris (PAR)) to **49.6 ms** (Moscow (MOW)). Marseille is on France's Mediterranean coast, south of Paris. ## Outbound Latency from Marseille (MRS) Round-trip time in milliseconds **from Marseille** to all other backbone PoPs, sorted fastest to slowest. | Destination | RTT (ms) | Fiber Floor | Efficiency | Jitter | Loss | Tier | | ----------------------- | ---------------------------------------------------- | ----------- | ---------- | ------- | ----- | --------- | | 🇫🇷 Paris (PAR) | **[8.9](/docs/network/latency/pairs/mrs-par-rtt)** | 6.47 ms | 72.7% | 0.12 ms | 0% | Ultra-Low | | 🇩🇪 Frankfurt (FRA) | **[16](/docs/network/latency/pairs/mrs-fra-rtt)** | 7.82 ms | 48.9% | 0.18 ms | 0.15% | Ultra-Low | | 🇬🇧 London (LON) | **[18.9](/docs/network/latency/pairs/mrs-lon-rtt)** | 9.82 ms | 51.9% | 0.15 ms | 0.09% | Ultra-Low | | 🇳🇱 Amsterdam (AMS) | **[20.3](/docs/network/latency/pairs/mrs-ams-rtt)** | 9.89 ms | 48.7% | 0.17 ms | 0.16% | Ultra-Low | | 🇩🇪 Berlin (BER) | **[20.6](/docs/network/latency/pairs/mrs-ber-rtt)** | 11.62 ms | 56.4% | 0.24 ms | 0% | Ultra-Low | | 🇷🇺 Moscow (MOW) | **[49.6](/docs/network/latency/pairs/mrs-mow-rtt)** | 26.23 ms | 52.9% | 0.45 ms | 0.07% | Excellent | | 🇺🇸 New York (NYC) | **[79.8](/docs/network/latency/pairs/mrs-nyc-rtt)** | 61.91 ms | 77.6% | 0.78 ms | 0% | Excellent | | 🇺🇸 Ashburn (IAD) | **[85.7](/docs/network/latency/pairs/mrs-iad-rtt)** | 65.34 ms | 76.2% | 0.46 ms | 0.05% | Good | | 🇺🇸 Miami (MIA) | **[112.4](/docs/network/latency/pairs/mrs-mia-rtt)** | 75.92 ms | 67.5% | 0.96 ms | 0.1% | Good | | 🇸🇬 Singapore (SIN) | **[139.6](/docs/network/latency/pairs/mrs-sin-rtt)** | 103.79 ms | 74.3% | 1.19 ms | 0% | Good | | 🇺🇸 Seattle (SEA) | **[141.7](/docs/network/latency/pairs/mrs-sea-rtt)** | 85.27 ms | 60.2% | 1.19 ms | 0.12% | Good | | 🇺🇸 Los Angeles (LAX) | **[144.5](/docs/network/latency/pairs/mrs-lax-rtt)** | 95.09 ms | 65.8% | 0.78 ms | 0% | Good | | 🇭🇰 Hong Kong (HKG) | **[166.4](/docs/network/latency/pairs/mrs-hkg-rtt)** | 95.48 ms | 57.4% | 1.49 ms | 0% | Fair | | 🇹🇼 Taipei (TPE) | **[180.9](/docs/network/latency/pairs/mrs-tpe-rtt)** | 98.2 ms | 54.3% | 1.86 ms | 0% | Fair | | 🇧🇷 São Paulo (GRU) | **[186.7](/docs/network/latency/pairs/mrs-gru-rtt)** | 89.24 ms | 47.8% | 1.4 ms | 0.05% | Fair | | 🇿🇦 Johannesburg (JNB) | **[198.5](/docs/network/latency/pairs/mrs-jnb-rtt)** | 78.71 ms | 39.7% | 1.78 ms | 0% | Fair | | 🇯🇵 Tokyo (TYO) | **[203.8](/docs/network/latency/pairs/mrs-tyo-rtt)** | 99.03 ms | 48.6% | 2.03 ms | 0.18% | Fair | | 🇦🇺 Melbourne (MEL) | **[234.2](/docs/network/latency/pairs/mrs-mel-rtt)** | 162.36 ms | 69.3% | 1.1 ms | 0% | Fair | | 🇦🇺 Sydney (SYD) | **[240.6](/docs/network/latency/pairs/mrs-syd-rtt)** | 165.41 ms | 68.7% | 2.14 ms | 0% | Fair | ## Fastest Routes to Marseille (MRS) The 10 fastest measured routes **to Marseille**, ranked by average RTT. Min / max / stdev are computed from the same measurement round (50 ICMP echo probes per route, 100 ms interval); routes without a published per-sample round show the measured average RTT only. | Rank | Source | Avg RTT | Min | Max | Stdev | | ---- | -------------------- | ------------------------------------------------------- | --------- | --------- | ------- | | 1 | 🇫🇷 Paris (PAR) | **[8.9 ms](/docs/network/latency/pairs/par-mrs-rtt)** | 8.59 ms | 9.69 ms | 0.26 ms | | 2 | 🇩🇪 Frankfurt (FRA) | **[16.4 ms](/docs/network/latency/pairs/fra-mrs-rtt)** | 15.96 ms | 17.74 ms | 0.37 ms | | 3 | 🇬🇧 London (LON) | **[17.9 ms](/docs/network/latency/pairs/lon-mrs-rtt)** | 17.02 ms | 21.69 ms | 0.84 ms | | 4 | 🇳🇱 Amsterdam (AMS) | **[19.6 ms](/docs/network/latency/pairs/ams-mrs-rtt)** | 19.03 ms | 20.95 ms | 0.5 ms | | 5 | 🇩🇪 Berlin (BER) | **[20.2 ms](/docs/network/latency/pairs/ber-mrs-rtt)** | 19.92 ms | 21.65 ms | 0.31 ms | | 6 | 🇷🇺 Moscow (MOW) | **[49.9 ms](/docs/network/latency/pairs/mow-mrs-rtt)** | 47.94 ms | 53.35 ms | 1.46 ms | | 7 | 🇺🇸 New York (NYC) | **[80.2 ms](/docs/network/latency/pairs/nyc-mrs-rtt)** | 78.94 ms | 85.06 ms | 1.32 ms | | 8 | 🇺🇸 Ashburn (IAD) | **[86.5 ms](/docs/network/latency/pairs/iad-mrs-rtt)** | 81.69 ms | 97.62 ms | 3.63 ms | | 9 | 🇺🇸 Miami (MIA) | **[113.7 ms](/docs/network/latency/pairs/mia-mrs-rtt)** | 110.36 ms | 126.2 ms | 3.72 ms | | 10 | 🇸🇬 Singapore (SIN) | **[140.6 ms](/docs/network/latency/pairs/sin-mrs-rtt)** | 138.21 ms | 150.15 ms | 2.29 ms | ## Inbound Latency to Marseille (MRS) Round-trip time in milliseconds **from all other PoPs to Marseille**. The measurement origin matters: inbound and outbound RTT can differ for the same pair. | Source | RTT (ms) | Fiber Floor | Efficiency | Jitter | Loss | Tier | | ----------------------- | ---------------------------------------------------- | ----------- | ---------- | ------- | ----- | --------- | | 🇫🇷 Paris (PAR) | **[8.9](/docs/network/latency/pairs/par-mrs-rtt)** | 6.47 ms | 72.7% | 0.12 ms | 0% | Ultra-Low | | 🇩🇪 Frankfurt (FRA) | **[16.4](/docs/network/latency/pairs/fra-mrs-rtt)** | 7.82 ms | 47.7% | 0.19 ms | 0% | Ultra-Low | | 🇬🇧 London (LON) | **[17.9](/docs/network/latency/pairs/lon-mrs-rtt)** | 9.82 ms | 54.8% | 0.16 ms | 0.14% | Ultra-Low | | 🇳🇱 Amsterdam (AMS) | **[19.6](/docs/network/latency/pairs/ams-mrs-rtt)** | 9.89 ms | 50.5% | 0.12 ms | 0.03% | Ultra-Low | | 🇩🇪 Berlin (BER) | **[20.2](/docs/network/latency/pairs/ber-mrs-rtt)** | 11.62 ms | 57.5% | 0.16 ms | 0.06% | Ultra-Low | | 🇷🇺 Moscow (MOW) | **[49.9](/docs/network/latency/pairs/mow-mrs-rtt)** | 26.23 ms | 52.6% | 0.4 ms | 0.05% | Excellent | | 🇺🇸 New York (NYC) | **[80.2](/docs/network/latency/pairs/nyc-mrs-rtt)** | 61.91 ms | 77.2% | 0.53 ms | 0% | Good | | 🇺🇸 Ashburn (IAD) | **[86.5](/docs/network/latency/pairs/iad-mrs-rtt)** | 65.34 ms | 75.5% | 0.72 ms | 0% | Good | | 🇺🇸 Miami (MIA) | **[113.7](/docs/network/latency/pairs/mia-mrs-rtt)** | 75.92 ms | 66.8% | 1.23 ms | 0.14% | Good | | 🇸🇬 Singapore (SIN) | **[140.6](/docs/network/latency/pairs/sin-mrs-rtt)** | 103.79 ms | 73.8% | 1.48 ms | 0.2% | Good | | 🇺🇸 Seattle (SEA) | **[141.2](/docs/network/latency/pairs/sea-mrs-rtt)** | 85.27 ms | 60.4% | 1.25 ms | 0.04% | Good | | 🇺🇸 Los Angeles (LAX) | **[143.9](/docs/network/latency/pairs/lax-mrs-rtt)** | 95.09 ms | 66.1% | 1.05 ms | 0.16% | Good | | 🇭🇰 Hong Kong (HKG) | **[165.4](/docs/network/latency/pairs/hkg-mrs-rtt)** | 95.48 ms | 57.7% | 0.85 ms | 0.09% | Fair | | 🇿🇦 Johannesburg (JNB) | **[172](/docs/network/latency/pairs/jnb-mrs-rtt)** | 78.71 ms | 45.8% | 1.53 ms | 0% | Fair | | 🇹🇼 Taipei (TPE) | **[179.5](/docs/network/latency/pairs/tpe-mrs-rtt)** | 98.2 ms | 54.7% | 1.4 ms | 0% | Fair | | 🇧🇷 São Paulo (GRU) | **[187.7](/docs/network/latency/pairs/gru-mrs-rtt)** | 89.24 ms | 47.5% | 1.08 ms | 0.14% | Fair | | 🇯🇵 Tokyo (TYO) | **[205.6](/docs/network/latency/pairs/tyo-mrs-rtt)** | 99.03 ms | 48.2% | 1.86 ms | 0.14% | Fair | | 🇦🇺 Melbourne (MEL) | **[233.5](/docs/network/latency/pairs/mel-mrs-rtt)** | 162.36 ms | 69.5% | 2.54 ms | 0% | Fair | | 🇦🇺 Sydney (SYD) | **[240.5](/docs/network/latency/pairs/syd-mrs-rtt)** | 165.41 ms | 68.8% | 2.18 ms | 0% | Fair | ## Europe Peers Marseille is one of 7 Hats Network PoPs in **Europe**. The table lists outbound measurements within the region. | PoP | Code | Country | RTT (ms) | Tier | | --------------------------------------- | ---- | ----------- | -------- | --------- | | [🇳🇱 Amsterdam (AMS)](./ams-amsterdam) | AMS | Netherlands | **20.3** | Ultra-Low | | [🇩🇪 Berlin (BER)](./ber-berlin) | BER | Germany | **20.6** | Ultra-Low | | [🇩🇪 Frankfurt (FRA)](./fra-frankfurt) | FRA | Germany | **16** | Ultra-Low | | [🇬🇧 London (LON)](./lon-london) | LON | UK | **18.9** | Ultra-Low | | [🇷🇺 Moscow (MOW)](./mow-moscow) | MOW | Russia | **49.6** | Excellent | | [🇫🇷 Paris (PAR)](./par-paris) | PAR | France | **8.9** | Ultra-Low | ## Questions ### What is the fastest route to Marseille? The fastest measured route to Marseille (MRS) is [Paris (PAR) → Marseille (MRS)](/docs/network/latency/pairs/par-mrs-rtt), averaging **8.9 ms** RTT (Ultra-Low). ### What is the fastest route from Marseille? The fastest measured route from Marseille (MRS) is [Marseille (MRS) → Paris (PAR)](/docs/network/latency/pairs/mrs-par-rtt), averaging **8.9 ms** RTT (Ultra-Low). ### How well connected is Marseille to the Hats Network backbone? Marseille (MRS) maintains measured routes to all 19 other PoPs across 5 continents. The slowest route, [Marseille (MRS) → Sydney (SYD)](/docs/network/latency/pairs/mrs-syd-rtt), averages **240.6 ms** RTT. *** *Round published August 16, 2026. [All 20 PoPs](/docs/network/latency) · [Dataset and licence](/opendata/latency/).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/mrs-marseille). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # New York network latency (NYC) New York (NYC) is a US East Coast PoP. The published routes include European, domestic and Asia-Pacific destinations. Measurement round: `2026-08-16T04:07:28Z`. The published round covers **19 outbound routes** from New York. Incoming routes are measured separately; neither direction describes last-mile performance. ## Location * **Location:** New York, USA (North America) * **Region:** North America * **Role:** Backbone interconnection point for Hats Network (AS203314) * **Public city reference:** [GeoNames 5128581](https://www.geonames.org/5128581); 40.71427, -74.00597 * **Coverage:** RTT measurements to 19 other PoPs across 5 continents ### Interconnection * **Facility:** NYC1 > Interactive content is available on the canonical HTML page. ## Round Summary | Metric | Value | | --------------------- | --------------------------------------------------------------------------------------------------------- | | Fastest Route | [🇺🇸 New York (NYC) → 🇺🇸 Ashburn (IAD)](/docs/network/latency/pairs/nyc-iad-rtt) (**6 ms**); Ultra-Low | | Slowest Route | [🇺🇸 New York (NYC) → 🇿🇦 Johannesburg (JNB)](/docs/network/latency/pairs/nyc-jnb-rtt) (**221.5 ms**) | | Average RTT | **112.4 ms** | | Average Jitter | **0.89 ms** | | Average Packet Loss | **0.03%** | | Best Fiber Efficiency | **86.1%** | | Intra-Region Peers | 4 PoPs in North America | | Total Routes | 19 outbound / 19 inbound (100% coverage) | > Jitter, packet-loss, percentile, and trend fields are estimated comparative indicators. Read the [theoretical fiber latency methodology](/docs/network/latency/theoretical-fiber-latency). ## Regional Routes New York's 4 other **North America** PoPs have RTTs from **6 ms** (Ashburn (IAD)) to **58.9 ms** (Los Angeles (LAX)). New York is on the US Atlantic coast, northeast of Ashburn. ## Outbound Latency from New York (NYC) Round-trip time in milliseconds **from New York** to all other backbone PoPs, sorted fastest to slowest. | Destination | RTT (ms) | Fiber Floor | Efficiency | Jitter | Loss | Tier | | ----------------------- | ---------------------------------------------------- | ----------- | ---------- | ------- | ----- | --------- | | 🇺🇸 Ashburn (IAD) | **[6](/docs/network/latency/pairs/nyc-iad-rtt)** | 3.44 ms | 57.3% | 0.12 ms | 0.01% | Ultra-Low | | 🇺🇸 Miami (MIA) | **[32.9](/docs/network/latency/pairs/nyc-mia-rtt)** | 17.17 ms | 52.2% | 0.39 ms | 0.15% | Excellent | | 🇺🇸 Seattle (SEA) | **[58.7](/docs/network/latency/pairs/nyc-sea-rtt)** | 37.95 ms | 64.7% | 0.35 ms | 0.11% | Excellent | | 🇺🇸 Los Angeles (LAX) | **[58.9](/docs/network/latency/pairs/nyc-lax-rtt)** | 38.63 ms | 65.6% | 0.46 ms | 0% | Excellent | | 🇬🇧 London (LON) | **[63.5](/docs/network/latency/pairs/nyc-lon-rtt)** | 54.69 ms | 86.1% | 0.7 ms | 0.01% | Excellent | | 🇳🇱 Amsterdam (AMS) | **[68.7](/docs/network/latency/pairs/nyc-ams-rtt)** | 57.56 ms | 83.8% | 0.33 ms | 0.03% | Excellent | | 🇫🇷 Paris (PAR) | **[68.9](/docs/network/latency/pairs/nyc-par-rtt)** | 57.31 ms | 83.2% | 0.43 ms | 0.15% | Excellent | | 🇩🇪 Frankfurt (FRA) | **[73.9](/docs/network/latency/pairs/nyc-fra-rtt)** | 60.9 ms | 82.4% | 0.74 ms | 0.09% | Excellent | | 🇩🇪 Berlin (BER) | **[78.8](/docs/network/latency/pairs/nyc-ber-rtt)** | 62.7 ms | 79.6% | 0.67 ms | 0% | Excellent | | 🇫🇷 Marseille (MRS) | **[80.2](/docs/network/latency/pairs/nyc-mrs-rtt)** | 61.91 ms | 77.2% | 0.53 ms | 0% | Good | | 🇧🇷 São Paulo (GRU) | **[106.9](/docs/network/latency/pairs/nyc-gru-rtt)** | 74.99 ms | 70.1% | 0.74 ms | 0% | Good | | 🇷🇺 Moscow (MOW) | **[107](/docs/network/latency/pairs/nyc-mow-rtt)** | 73.75 ms | 68.9% | 1.08 ms | 0.06% | Good | | 🇯🇵 Tokyo (TYO) | **[143.1](/docs/network/latency/pairs/nyc-tyo-rtt)** | 106.47 ms | 74.4% | 0.8 ms | 0% | Good | | 🇹🇼 Taipei (TPE) | **[173.4](/docs/network/latency/pairs/nyc-tpe-rtt)** | 122.89 ms | 70.9% | 1.21 ms | 0% | Fair | | 🇭🇰 Hong Kong (HKG) | **[187.1](/docs/network/latency/pairs/nyc-hkg-rtt)** | 127.11 ms | 67.9% | 1.76 ms | 0% | Fair | | 🇦🇺 Sydney (SYD) | **[192.6](/docs/network/latency/pairs/nyc-syd-rtt)** | 156.57 ms | 81.3% | 0.91 ms | 0% | Fair | | 🇦🇺 Melbourne (MEL) | **[205.1](/docs/network/latency/pairs/nyc-mel-rtt)** | 163.27 ms | 79.6% | 1.79 ms | 0.04% | Fair | | 🇸🇬 Singapore (SIN) | **[208.7](/docs/network/latency/pairs/nyc-sin-rtt)** | 150.29 ms | 72% | 1.83 ms | 0% | Fair | | 🇿🇦 Johannesburg (JNB) | **[221.5](/docs/network/latency/pairs/nyc-jnb-rtt)** | 125.67 ms | 56.7% | 2.09 ms | 0% | Fair | ## Fastest Routes to New York (NYC) The 10 fastest measured routes **to New York**, ranked by average RTT. Min / max / stdev are computed from the same measurement round (50 ICMP echo probes per route, 100 ms interval); routes without a published per-sample round show the measured average RTT only. | Rank | Source | Avg RTT | Min | Max | Stdev | | ---- | ---------------------- | ------------------------------------------------------ | -------- | -------- | ------- | | 1 | 🇺🇸 Ashburn (IAD) | **[6 ms](/docs/network/latency/pairs/iad-nyc-rtt)** | 5.77 ms | 6.79 ms | 0.2 ms | | 2 | 🇺🇸 Miami (MIA) | **[33.3 ms](/docs/network/latency/pairs/mia-nyc-rtt)** | 31.68 ms | 39.04 ms | 1.51 ms | | 3 | 🇺🇸 Los Angeles (LAX) | **[57.9 ms](/docs/network/latency/pairs/lax-nyc-rtt)** | 55.78 ms | 62.74 ms | 1.33 ms | | 4 | 🇺🇸 Seattle (SEA) | **[59.5 ms](/docs/network/latency/pairs/sea-nyc-rtt)** | 57.58 ms | 64.42 ms | 1.48 ms | | 5 | 🇬🇧 London (LON) | **[63.8 ms](/docs/network/latency/pairs/lon-nyc-rtt)** | 60.82 ms | 74.75 ms | 2.78 ms | | 6 | 🇳🇱 Amsterdam (AMS) | **[68.6 ms](/docs/network/latency/pairs/ams-nyc-rtt)** | 67.28 ms | 73 ms | 1.16 ms | | 7 | 🇫🇷 Paris (PAR) | **[68.9 ms](/docs/network/latency/pairs/par-nyc-rtt)** | 66.76 ms | 73.07 ms | 1.55 ms | | 8 | 🇩🇪 Frankfurt (FRA) | **[74.4 ms](/docs/network/latency/pairs/fra-nyc-rtt)** | 72.04 ms | 80.52 ms | 2.09 ms | | 9 | 🇩🇪 Berlin (BER) | **[79.6 ms](/docs/network/latency/pairs/ber-nyc-rtt)** | 78.03 ms | 83.78 ms | 1.28 ms | | 10 | 🇫🇷 Marseille (MRS) | **[79.8 ms](/docs/network/latency/pairs/mrs-nyc-rtt)** | 75.53 ms | 89.73 ms | 3.65 ms | ## Inbound Latency to New York (NYC) Round-trip time in milliseconds **from all other PoPs to New York**. The measurement origin matters: inbound and outbound RTT can differ for the same pair. | Source | RTT (ms) | Fiber Floor | Efficiency | Jitter | Loss | Tier | | ----------------------- | ---------------------------------------------------- | ----------- | ---------- | ------- | ----- | --------- | | 🇺🇸 Ashburn (IAD) | **[6](/docs/network/latency/pairs/iad-nyc-rtt)** | 3.44 ms | 57.3% | 0.12 ms | 0.17% | Ultra-Low | | 🇺🇸 Miami (MIA) | **[33.3](/docs/network/latency/pairs/mia-nyc-rtt)** | 17.17 ms | 51.6% | 0.33 ms | 0.08% | Excellent | | 🇺🇸 Los Angeles (LAX) | **[57.9](/docs/network/latency/pairs/lax-nyc-rtt)** | 38.63 ms | 66.7% | 0.34 ms | 0.09% | Excellent | | 🇺🇸 Seattle (SEA) | **[59.5](/docs/network/latency/pairs/sea-nyc-rtt)** | 37.95 ms | 63.8% | 0.66 ms | 0.16% | Excellent | | 🇬🇧 London (LON) | **[63.8](/docs/network/latency/pairs/lon-nyc-rtt)** | 54.69 ms | 85.7% | 0.42 ms | 0.12% | Excellent | | 🇳🇱 Amsterdam (AMS) | **[68.6](/docs/network/latency/pairs/ams-nyc-rtt)** | 57.56 ms | 83.9% | 0.31 ms | 0% | Excellent | | 🇫🇷 Paris (PAR) | **[68.9](/docs/network/latency/pairs/par-nyc-rtt)** | 57.31 ms | 83.2% | 0.71 ms | 0.05% | Excellent | | 🇩🇪 Frankfurt (FRA) | **[74.4](/docs/network/latency/pairs/fra-nyc-rtt)** | 60.9 ms | 81.9% | 0.73 ms | 0% | Excellent | | 🇩🇪 Berlin (BER) | **[79.6](/docs/network/latency/pairs/ber-nyc-rtt)** | 62.7 ms | 78.8% | 0.5 ms | 0.08% | Excellent | | 🇫🇷 Marseille (MRS) | **[79.8](/docs/network/latency/pairs/mrs-nyc-rtt)** | 61.91 ms | 77.6% | 0.78 ms | 0% | Excellent | | 🇷🇺 Moscow (MOW) | **[107.5](/docs/network/latency/pairs/mow-nyc-rtt)** | 73.75 ms | 68.6% | 1 ms | 0% | Good | | 🇧🇷 São Paulo (GRU) | **[122](/docs/network/latency/pairs/gru-nyc-rtt)** | 74.99 ms | 61.5% | 1.29 ms | 0% | Good | | 🇯🇵 Tokyo (TYO) | **[144.6](/docs/network/latency/pairs/tyo-nyc-rtt)** | 106.47 ms | 73.6% | 1.33 ms | 0.17% | Good | | 🇹🇼 Taipei (TPE) | **[174.2](/docs/network/latency/pairs/tpe-nyc-rtt)** | 122.89 ms | 70.5% | 1.24 ms | 0% | Fair | | 🇭🇰 Hong Kong (HKG) | **[184.7](/docs/network/latency/pairs/hkg-nyc-rtt)** | 127.11 ms | 68.8% | 0.98 ms | 0% | Fair | | 🇦🇺 Sydney (SYD) | **[191.9](/docs/network/latency/pairs/syd-nyc-rtt)** | 156.57 ms | 81.6% | 2.06 ms | 0.07% | Fair | | 🇦🇺 Melbourne (MEL) | **[204.2](/docs/network/latency/pairs/mel-nyc-rtt)** | 163.27 ms | 80% | 2.1 ms | 0% | Fair | | 🇸🇬 Singapore (SIN) | **[209.3](/docs/network/latency/pairs/sin-nyc-rtt)** | 150.29 ms | 71.8% | 1.38 ms | 0% | Fair | | 🇿🇦 Johannesburg (JNB) | **[221.8](/docs/network/latency/pairs/jnb-nyc-rtt)** | 125.67 ms | 56.7% | 2.01 ms | 0% | Fair | ## North America Peers New York is one of 5 Hats Network PoPs in **North America**. The table lists outbound measurements within the region. | PoP | Code | Country | RTT (ms) | Tier | | ------------------------------------------- | ---- | ------- | -------- | --------- | | [🇺🇸 Ashburn (IAD)](./iad-ashburn) | IAD | USA | **6** | Ultra-Low | | [🇺🇸 Los Angeles (LAX)](./lax-los-angeles) | LAX | USA | **58.9** | Excellent | | [🇺🇸 Miami (MIA)](./mia-miami) | MIA | USA | **32.9** | Excellent | | [🇺🇸 Seattle (SEA)](./sea-seattle) | SEA | USA | **58.7** | Excellent | ## Questions ### What is the fastest route to New York? The fastest measured route to New York (NYC) is [Ashburn (IAD) → New York (NYC)](/docs/network/latency/pairs/iad-nyc-rtt), averaging **6 ms** RTT (Ultra-Low). ### What is the fastest route from New York? The fastest measured route from New York (NYC) is [New York (NYC) → Ashburn (IAD)](/docs/network/latency/pairs/nyc-iad-rtt), averaging **6 ms** RTT (Ultra-Low). ### How well connected is New York to the Hats Network backbone? New York (NYC) maintains measured routes to all 19 other PoPs across 5 continents. The slowest route, [New York (NYC) → Johannesburg (JNB)](/docs/network/latency/pairs/nyc-jnb-rtt), averages **221.5 ms** RTT. *** *Round published August 16, 2026. [All 20 PoPs](/docs/network/latency) · [Dataset and licence](/opendata/latency/).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/nyc-new-york). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Paris network latency (PAR) Paris (PAR) is a Hats Network PoP in France. Outbound and inbound measurements are listed separately. Measurement round: `2026-08-16T04:07:28Z`. The published round covers **19 outbound routes** from Paris. Incoming routes are measured separately; neither direction describes last-mile performance. ## Location * **Location:** Paris, France (Europe) * **Region:** Europe * **Role:** Backbone interconnection point for Hats Network (AS203314) * **Public city reference:** [GeoNames 2988507](https://www.geonames.org/2988507); 48.85341, 2.34880 * **Coverage:** RTT measurements to 19 other PoPs across 5 continents ### Interconnection * **Facility:** PAR1 > Interactive content is available on the canonical HTML page. ## Round Summary | Metric | Value | | --------------------- | ------------------------------------------------------------------------------------------------------- | | Fastest Route | [🇫🇷 Paris (PAR) → 🇬🇧 London (LON)](/docs/network/latency/pairs/par-lon-rtt) (**6.4 ms**); Ultra-Low | | Slowest Route | [🇫🇷 Paris (PAR) → 🇦🇺 Sydney (SYD)](/docs/network/latency/pairs/par-syd-rtt) (**249.7 ms**) | | Average RTT | **112.9 ms** | | Average Jitter | **0.85 ms** | | Average Packet Loss | **0.05%** | | Best Fiber Efficiency | **83.2%** | | Intra-Region Peers | 6 PoPs in Europe | | Total Routes | 19 outbound / 19 inbound (100% coverage) | > Jitter, packet-loss, percentile, and trend fields are estimated comparative indicators. Read the [theoretical fiber latency methodology](/docs/network/latency/theoretical-fiber-latency). ## Regional Routes Paris's 6 other **Europe** PoPs have RTTs from **6.4 ms** (London (LON)) to **44.7 ms** (Moscow (MOW)). Paris is in northern France, between London and Marseille. ## Outbound Latency from Paris (PAR) Round-trip time in milliseconds **from Paris** to all other backbone PoPs, sorted fastest to slowest. | Destination | RTT (ms) | Fiber Floor | Efficiency | Jitter | Loss | Tier | | ----------------------- | ---------------------------------------------------- | ----------- | ---------- | ------- | ----- | --------- | | 🇬🇧 London (LON) | **[6.4](/docs/network/latency/pairs/par-lon-rtt)** | 3.37 ms | 52.7% | 0.12 ms | 0% | Ultra-Low | | 🇳🇱 Amsterdam (AMS) | **[7](/docs/network/latency/pairs/par-ams-rtt)** | 4.22 ms | 60.3% | 0.12 ms | 0% | Ultra-Low | | 🇩🇪 Frankfurt (FRA) | **[7.6](/docs/network/latency/pairs/par-fra-rtt)** | 4.7 ms | 61.8% | 0.12 ms | 0.1% | Ultra-Low | | 🇫🇷 Marseille (MRS) | **[8.9](/docs/network/latency/pairs/par-mrs-rtt)** | 6.47 ms | 72.7% | 0.12 ms | 0% | Ultra-Low | | 🇩🇪 Berlin (BER) | **[15.5](/docs/network/latency/pairs/par-ber-rtt)** | 8.62 ms | 55.6% | 0.13 ms | 0% | Ultra-Low | | 🇷🇺 Moscow (MOW) | **[44.7](/docs/network/latency/pairs/par-mow-rtt)** | 24.42 ms | 54.6% | 0.23 ms | 0.2% | Excellent | | 🇺🇸 New York (NYC) | **[68.9](/docs/network/latency/pairs/par-nyc-rtt)** | 57.31 ms | 83.2% | 0.71 ms | 0.05% | Excellent | | 🇺🇸 Ashburn (IAD) | **[75.5](/docs/network/latency/pairs/par-iad-rtt)** | 60.73 ms | 80.4% | 0.86 ms | 0.02% | Excellent | | 🇺🇸 Miami (MIA) | **[106.8](/docs/network/latency/pairs/par-mia-rtt)** | 72.16 ms | 67.6% | 0.65 ms | 0.05% | Good | | 🇺🇸 Seattle (SEA) | **[128](/docs/network/latency/pairs/par-sea-rtt)** | 78.98 ms | 61.7% | 0.85 ms | 0.07% | Good | | 🇺🇸 Los Angeles (LAX) | **[136.5](/docs/network/latency/pairs/par-lax-rtt)** | 89.18 ms | 65.3% | 1.58 ms | 0.11% | Good | | 🇸🇬 Singapore (SIN) | **[150.3](/docs/network/latency/pairs/par-sin-rtt)** | 105.19 ms | 70% | 1.32 ms | 0% | Fair | | 🇭🇰 Hong Kong (HKG) | **[160.1](/docs/network/latency/pairs/par-hkg-rtt)** | 94.48 ms | 59% | 1.29 ms | 0.05% | Fair | | 🇹🇼 Taipei (TPE) | **[171.9](/docs/network/latency/pairs/par-tpe-rtt)** | 96.42 ms | 56.1% | 0.85 ms | 0% | Fair | | 🇧🇷 São Paulo (GRU) | **[175.8](/docs/network/latency/pairs/par-gru-rtt)** | 91.83 ms | 52.2% | 1.26 ms | 0% | Fair | | 🇿🇦 Johannesburg (JNB) | **[180.9](/docs/network/latency/pairs/par-jnb-rtt)** | 85.17 ms | 47.1% | 0.94 ms | 0.17% | Fair | | 🇯🇵 Tokyo (TYO) | **[204.7](/docs/network/latency/pairs/par-tyo-rtt)** | 95.34 ms | 46.6% | 2.17 ms | 0% | Fair | | 🇦🇺 Melbourne (MEL) | **[246.4](/docs/network/latency/pairs/par-mel-rtt)** | 164.4 ms | 66.7% | 1.59 ms | 0.17% | Fair | | 🇦🇺 Sydney (SYD) | **[249.7](/docs/network/latency/pairs/par-syd-rtt)** | 166.06 ms | 66.5% | 1.18 ms | 0% | Fair | ## Fastest Routes to Paris (PAR) The 10 fastest measured routes **to Paris**, ranked by average RTT. Min / max / stdev are computed from the same measurement round (50 ICMP echo probes per route, 100 ms interval); routes without a published per-sample round show the measured average RTT only. | Rank | Source | Avg RTT | Min | Max | Stdev | | ---- | -------------------- | ------------------------------------------------------- | --------- | --------- | ------- | | 1 | 🇬🇧 London (LON) | **[6.4 ms](/docs/network/latency/pairs/lon-par-rtt)** | 6.23 ms | 6.9 ms | 0.15 ms | | 2 | 🇳🇱 Amsterdam (AMS) | **[7 ms](/docs/network/latency/pairs/ams-par-rtt)** | 6.61 ms | 8.02 ms | 0.32 ms | | 3 | 🇩🇪 Frankfurt (FRA) | **[7.6 ms](/docs/network/latency/pairs/fra-par-rtt)** | 7.34 ms | 8.24 ms | 0.23 ms | | 4 | 🇫🇷 Marseille (MRS) | **[8.9 ms](/docs/network/latency/pairs/mrs-par-rtt)** | 8.72 ms | 9.67 ms | 0.16 ms | | 5 | 🇩🇪 Berlin (BER) | **[15.6 ms](/docs/network/latency/pairs/ber-par-rtt)** | 14.9 ms | 17.81 ms | 0.63 ms | | 6 | 🇷🇺 Moscow (MOW) | **[44 ms](/docs/network/latency/pairs/mow-par-rtt)** | 42.02 ms | 49.11 ms | 1.64 ms | | 7 | 🇺🇸 New York (NYC) | **[68.9 ms](/docs/network/latency/pairs/nyc-par-rtt)** | 67.48 ms | 73.26 ms | 1.31 ms | | 8 | 🇺🇸 Ashburn (IAD) | **[76.1 ms](/docs/network/latency/pairs/iad-par-rtt)** | 74.36 ms | 85 ms | 1.76 ms | | 9 | 🇺🇸 Miami (MIA) | **[106.2 ms](/docs/network/latency/pairs/mia-par-rtt)** | 104.35 ms | 111.43 ms | 1.69 ms | | 10 | 🇺🇸 Seattle (SEA) | **[129.2 ms](/docs/network/latency/pairs/sea-par-rtt)** | 126.7 ms | 135.02 ms | 2.13 ms | ## Inbound Latency to Paris (PAR) Round-trip time in milliseconds **from all other PoPs to Paris**. The measurement origin matters: inbound and outbound RTT can differ for the same pair. | Source | RTT (ms) | Fiber Floor | Efficiency | Jitter | Loss | Tier | | ----------------------- | ---------------------------------------------------- | ----------- | ---------- | ------- | ----- | --------- | | 🇬🇧 London (LON) | **[6.4](/docs/network/latency/pairs/lon-par-rtt)** | 3.37 ms | 52.7% | 0.12 ms | 0% | Ultra-Low | | 🇳🇱 Amsterdam (AMS) | **[7](/docs/network/latency/pairs/ams-par-rtt)** | 4.22 ms | 60.3% | 0.12 ms | 0.03% | Ultra-Low | | 🇩🇪 Frankfurt (FRA) | **[7.6](/docs/network/latency/pairs/fra-par-rtt)** | 4.7 ms | 61.8% | 0.12 ms | 0% | Ultra-Low | | 🇫🇷 Marseille (MRS) | **[8.9](/docs/network/latency/pairs/mrs-par-rtt)** | 6.47 ms | 72.7% | 0.12 ms | 0% | Ultra-Low | | 🇩🇪 Berlin (BER) | **[15.6](/docs/network/latency/pairs/ber-par-rtt)** | 8.62 ms | 55.3% | 0.16 ms | 0% | Ultra-Low | | 🇷🇺 Moscow (MOW) | **[44](/docs/network/latency/pairs/mow-par-rtt)** | 24.42 ms | 55.5% | 0.29 ms | 0% | Excellent | | 🇺🇸 New York (NYC) | **[68.9](/docs/network/latency/pairs/nyc-par-rtt)** | 57.31 ms | 83.2% | 0.43 ms | 0.15% | Excellent | | 🇺🇸 Ashburn (IAD) | **[76.1](/docs/network/latency/pairs/iad-par-rtt)** | 60.73 ms | 79.8% | 0.7 ms | 0.03% | Excellent | | 🇺🇸 Miami (MIA) | **[106.2](/docs/network/latency/pairs/mia-par-rtt)** | 72.16 ms | 67.9% | 1.07 ms | 0% | Good | | 🇺🇸 Seattle (SEA) | **[129.2](/docs/network/latency/pairs/sea-par-rtt)** | 78.98 ms | 61.1% | 1.22 ms | 0.02% | Good | | 🇺🇸 Los Angeles (LAX) | **[135.9](/docs/network/latency/pairs/lax-par-rtt)** | 89.18 ms | 65.6% | 0.62 ms | 0% | Good | | 🇸🇬 Singapore (SIN) | **[151.4](/docs/network/latency/pairs/sin-par-rtt)** | 105.19 ms | 69.5% | 1.07 ms | 0.12% | Fair | | 🇭🇰 Hong Kong (HKG) | **[159.4](/docs/network/latency/pairs/hkg-par-rtt)** | 94.48 ms | 59.3% | 1.21 ms | 0.07% | Fair | | 🇿🇦 Johannesburg (JNB) | **[164.4](/docs/network/latency/pairs/jnb-par-rtt)** | 85.17 ms | 51.8% | 1.05 ms | 0% | Fair | | 🇹🇼 Taipei (TPE) | **[173.8](/docs/network/latency/pairs/tpe-par-rtt)** | 96.42 ms | 55.5% | 1.13 ms | 0.13% | Fair | | 🇧🇷 São Paulo (GRU) | **[180.2](/docs/network/latency/pairs/gru-par-rtt)** | 91.83 ms | 51% | 1.95 ms | 0.14% | Fair | | 🇯🇵 Tokyo (TYO) | **[204.8](/docs/network/latency/pairs/tyo-par-rtt)** | 95.34 ms | 46.6% | 2.06 ms | 0.03% | Fair | | 🇦🇺 Melbourne (MEL) | **[246.5](/docs/network/latency/pairs/mel-par-rtt)** | 164.4 ms | 66.7% | 1.53 ms | 0.13% | Fair | | 🇦🇺 Sydney (SYD) | **[250.3](/docs/network/latency/pairs/syd-par-rtt)** | 166.06 ms | 66.3% | 2.41 ms | 0% | High | ## Europe Peers Paris is one of 7 Hats Network PoPs in **Europe**. The table lists outbound measurements within the region. | PoP | Code | Country | RTT (ms) | Tier | | --------------------------------------- | ---- | ----------- | -------- | --------- | | [🇳🇱 Amsterdam (AMS)](./ams-amsterdam) | AMS | Netherlands | **7** | Ultra-Low | | [🇩🇪 Berlin (BER)](./ber-berlin) | BER | Germany | **15.5** | Ultra-Low | | [🇩🇪 Frankfurt (FRA)](./fra-frankfurt) | FRA | Germany | **7.6** | Ultra-Low | | [🇬🇧 London (LON)](./lon-london) | LON | UK | **6.4** | Ultra-Low | | [🇫🇷 Marseille (MRS)](./mrs-marseille) | MRS | France | **8.9** | Ultra-Low | | [🇷🇺 Moscow (MOW)](./mow-moscow) | MOW | Russia | **44.7** | Excellent | ## Questions ### What is the fastest route to Paris? The fastest measured route to Paris (PAR) is [London (LON) → Paris (PAR)](/docs/network/latency/pairs/lon-par-rtt), averaging **6.4 ms** RTT (Ultra-Low). ### What is the fastest route from Paris? The fastest measured route from Paris (PAR) is [Paris (PAR) → London (LON)](/docs/network/latency/pairs/par-lon-rtt), averaging **6.4 ms** RTT (Ultra-Low). ### How well connected is Paris to the Hats Network backbone? Paris (PAR) maintains measured routes to all 19 other PoPs across 5 continents. The slowest route, [Paris (PAR) → Sydney (SYD)](/docs/network/latency/pairs/par-syd-rtt), averages **249.7 ms** RTT. *** *Round published August 16, 2026. [All 20 PoPs](/docs/network/latency) · [Dataset and licence](/opendata/latency/).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/par-paris). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Seattle network latency (SEA) Seattle (SEA) is the Pacific Northwest PoP. Its routes provide another West Coast origin for comparison with Los Angeles. Measurement round: `2026-08-16T04:07:28Z`. The published round covers **19 outbound routes** from Seattle. Incoming routes are measured separately; neither direction describes last-mile performance. ## Location * **Location:** Seattle, USA (North America) * **Region:** North America * **Role:** Backbone interconnection point for Hats Network (AS203314) * **Public city reference:** [GeoNames 5809844](https://www.geonames.org/5809844); 47.60621, -122.33207 * **Coverage:** RTT measurements to 19 other PoPs across 5 continents ### Interconnection * **Facility:** SEA1 * **Upstream transit:** HE / Cogent > Interactive content is available on the canonical HTML page. ## Round Summary | Metric | Value | | --------------------- | --------------------------------------------------------------------------------------------------------------- | | Fastest Route | [🇺🇸 Seattle (SEA) → 🇺🇸 Los Angeles (LAX)](/docs/network/latency/pairs/sea-lax-rtt) (**25.9 ms**); Ultra-Low | | Slowest Route | [🇺🇸 Seattle (SEA) → 🇿🇦 Johannesburg (JNB)](/docs/network/latency/pairs/sea-jnb-rtt) (**282.1 ms**) | | Average RTT | **128.5 ms** | | Average Jitter | **1.09 ms** | | Average Packet Loss | **0.03%** | | Best Fiber Efficiency | **88.9%** | | Intra-Region Peers | 4 PoPs in North America | | Total Routes | 19 outbound / 19 inbound (100% coverage) | > Jitter, packet-loss, percentile, and trend fields are estimated comparative indicators. Read the [theoretical fiber latency methodology](/docs/network/latency/theoretical-fiber-latency). ## Regional Routes Seattle's 4 other **North America** PoPs have RTTs from **25.9 ms** (Los Angeles (LAX)) to **81.7 ms** (Miami (MIA)). Seattle is in Washington State in the US Pacific Northwest. ## Outbound Latency from Seattle (SEA) Round-trip time in milliseconds **from Seattle** to all other backbone PoPs, sorted fastest to slowest. | Destination | RTT (ms) | Fiber Floor | Efficiency | Jitter | Loss | Tier | | ----------------------- | ---------------------------------------------------- | ----------- | ---------- | ------- | ----- | --------- | | 🇺🇸 Los Angeles (LAX) | **[25.9](/docs/network/latency/pairs/sea-lax-rtt)** | 15.12 ms | 58.4% | 0.24 ms | 0% | Ultra-Low | | 🇺🇸 New York (NYC) | **[59.5](/docs/network/latency/pairs/sea-nyc-rtt)** | 37.95 ms | 63.8% | 0.66 ms | 0.16% | Excellent | | 🇺🇸 Ashburn (IAD) | **[61.5](/docs/network/latency/pairs/sea-iad-rtt)** | 36.27 ms | 59% | 0.55 ms | 0% | Excellent | | 🇺🇸 Miami (MIA) | **[81.7](/docs/network/latency/pairs/sea-mia-rtt)** | 43.08 ms | 52.7% | 0.84 ms | 0% | Good | | 🇯🇵 Tokyo (TYO) | **[85](/docs/network/latency/pairs/sea-tyo-rtt)** | 75.55 ms | 88.9% | 0.39 ms | 0% | Good | | 🇹🇼 Taipei (TPE) | **[114.7](/docs/network/latency/pairs/sea-tpe-rtt)** | 95.46 ms | 83.2% | 1.36 ms | 0% | Good | | 🇬🇧 London (LON) | **[124.1](/docs/network/latency/pairs/sea-lon-rtt)** | 75.63 ms | 60.9% | 1.38 ms | 0% | Good | | 🇳🇱 Amsterdam (AMS) | **[128.6](/docs/network/latency/pairs/sea-ams-rtt)** | 76.86 ms | 59.8% | 1.46 ms | 0.13% | Good | | 🇫🇷 Paris (PAR) | **[129.2](/docs/network/latency/pairs/sea-par-rtt)** | 78.98 ms | 61.1% | 1.22 ms | 0.02% | Good | | 🇭🇰 Hong Kong (HKG) | **[130.4](/docs/network/latency/pairs/sea-hkg-rtt)** | 102.2 ms | 78.4% | 0.7 ms | 0% | Good | | 🇩🇪 Frankfurt (FRA) | **[133.2](/docs/network/latency/pairs/sea-fra-rtt)** | 80.34 ms | 60.3% | 0.66 ms | 0.11% | Good | | 🇩🇪 Berlin (BER) | **[140.9](/docs/network/latency/pairs/sea-ber-rtt)** | 79.73 ms | 56.6% | 1.36 ms | 0.01% | Good | | 🇫🇷 Marseille (MRS) | **[141.2](/docs/network/latency/pairs/sea-mrs-rtt)** | 85.27 ms | 60.4% | 1.25 ms | 0.04% | Good | | 🇸🇬 Singapore (SIN) | **[151.1](/docs/network/latency/pairs/sea-sin-rtt)** | 127.22 ms | 84.2% | 1.4 ms | 0% | Fair | | 🇧🇷 São Paulo (GRU) | **[155.7](/docs/network/latency/pairs/sea-gru-rtt)** | 106.76 ms | 68.6% | 1.76 ms | 0% | Fair | | 🇦🇺 Sydney (SYD) | **[161.3](/docs/network/latency/pairs/sea-syd-rtt)** | 121.96 ms | 75.6% | 1.49 ms | 0% | Fair | | 🇷🇺 Moscow (MOW) | **[164.6](/docs/network/latency/pairs/sea-mow-rtt)** | 82.23 ms | 50% | 1.42 ms | 0% | Fair | | 🇦🇺 Melbourne (MEL) | **[170.1](/docs/network/latency/pairs/sea-mel-rtt)** | 128.93 ms | 75.8% | 0.98 ms | 0% | Fair | | 🇿🇦 Johannesburg (JNB) | **[282.1](/docs/network/latency/pairs/sea-jnb-rtt)** | 161.61 ms | 57.3% | 1.53 ms | 0.15% | High | ## Fastest Routes to Seattle (SEA) The 10 fastest measured routes **to Seattle**, ranked by average RTT. Min / max / stdev are computed from the same measurement round (50 ICMP echo probes per route, 100 ms interval); routes without a published per-sample round show the measured average RTT only. | Rank | Source | Avg RTT | Min | Max | Stdev | | ---- | ---------------------- | ------------------------------------------------------- | --------- | --------- | ------- | | 1 | 🇺🇸 Los Angeles (LAX) | **[26.9 ms](/docs/network/latency/pairs/lax-sea-rtt)** | 25.5 ms | 29.88 ms | 1.13 ms | | 2 | 🇺🇸 New York (NYC) | **[58.7 ms](/docs/network/latency/pairs/nyc-sea-rtt)** | 56.64 ms | 67.85 ms | 2.09 ms | | 3 | 🇺🇸 Ashburn (IAD) | **[62.3 ms](/docs/network/latency/pairs/iad-sea-rtt)** | 59.37 ms | 75.24 ms | 2.67 ms | | 4 | 🇺🇸 Miami (MIA) | **[81.9 ms](/docs/network/latency/pairs/mia-sea-rtt)** | 78.54 ms | 93.5 ms | 3.48 ms | | 5 | 🇯🇵 Tokyo (TYO) | **[84.8 ms](/docs/network/latency/pairs/tyo-sea-rtt)** | 81.06 ms | 94.31 ms | 2.93 ms | | 6 | 🇹🇼 Taipei (TPE) | **[115.8 ms](/docs/network/latency/pairs/tpe-sea-rtt)** | 111.75 ms | 128.63 ms | 4.07 ms | | 7 | 🇬🇧 London (LON) | **[123.2 ms](/docs/network/latency/pairs/lon-sea-rtt)** | 121.09 ms | 129.68 ms | 1.95 ms | | 8 | 🇫🇷 Paris (PAR) | **[128 ms](/docs/network/latency/pairs/par-sea-rtt)** | 123.76 ms | 137.5 ms | 3.51 ms | | 9 | 🇳🇱 Amsterdam (AMS) | **[130.2 ms](/docs/network/latency/pairs/ams-sea-rtt)** | 123.59 ms | 144.11 ms | 5.22 ms | | 10 | 🇭🇰 Hong Kong (HKG) | **[131.6 ms](/docs/network/latency/pairs/hkg-sea-rtt)** | 125.78 ms | 162.34 ms | 6.2 ms | ## Inbound Latency to Seattle (SEA) Round-trip time in milliseconds **from all other PoPs to Seattle**. The measurement origin matters: inbound and outbound RTT can differ for the same pair. | Source | RTT (ms) | Fiber Floor | Efficiency | Jitter | Loss | Tier | | ----------------------- | ---------------------------------------------------- | ----------- | ---------- | ------- | ----- | --------- | | 🇺🇸 Los Angeles (LAX) | **[26.9](/docs/network/latency/pairs/lax-sea-rtt)** | 15.12 ms | 56.2% | 0.3 ms | 0.13% | Ultra-Low | | 🇺🇸 New York (NYC) | **[58.7](/docs/network/latency/pairs/nyc-sea-rtt)** | 37.95 ms | 64.7% | 0.35 ms | 0.11% | Excellent | | 🇺🇸 Ashburn (IAD) | **[62.3](/docs/network/latency/pairs/iad-sea-rtt)** | 36.27 ms | 58.2% | 0.44 ms | 0.12% | Excellent | | 🇺🇸 Miami (MIA) | **[81.9](/docs/network/latency/pairs/mia-sea-rtt)** | 43.08 ms | 52.6% | 0.47 ms | 0.1% | Good | | 🇯🇵 Tokyo (TYO) | **[84.8](/docs/network/latency/pairs/tyo-sea-rtt)** | 75.55 ms | 89.1% | 0.77 ms | 0.06% | Good | | 🇹🇼 Taipei (TPE) | **[115.8](/docs/network/latency/pairs/tpe-sea-rtt)** | 95.46 ms | 82.4% | 0.97 ms | 0.15% | Good | | 🇬🇧 London (LON) | **[123.2](/docs/network/latency/pairs/lon-sea-rtt)** | 75.63 ms | 61.4% | 0.95 ms | 0% | Good | | 🇫🇷 Paris (PAR) | **[128](/docs/network/latency/pairs/par-sea-rtt)** | 78.98 ms | 61.7% | 0.85 ms | 0.07% | Good | | 🇳🇱 Amsterdam (AMS) | **[130.2](/docs/network/latency/pairs/ams-sea-rtt)** | 76.86 ms | 59% | 1.26 ms | 0.04% | Good | | 🇭🇰 Hong Kong (HKG) | **[131.6](/docs/network/latency/pairs/hkg-sea-rtt)** | 102.2 ms | 77.7% | 0.82 ms | 0.05% | Good | | 🇩🇪 Frankfurt (FRA) | **[135.2](/docs/network/latency/pairs/fra-sea-rtt)** | 80.34 ms | 59.4% | 0.98 ms | 0.16% | Good | | 🇩🇪 Berlin (BER) | **[139.7](/docs/network/latency/pairs/ber-sea-rtt)** | 79.73 ms | 57.1% | 1.63 ms | 0.2% | Good | | 🇫🇷 Marseille (MRS) | **[141.7](/docs/network/latency/pairs/mrs-sea-rtt)** | 85.27 ms | 60.2% | 1.19 ms | 0.12% | Good | | 🇸🇬 Singapore (SIN) | **[151.5](/docs/network/latency/pairs/sin-sea-rtt)** | 127.22 ms | 84% | 1.11 ms | 0% | Fair | | 🇧🇷 São Paulo (GRU) | **[155.9](/docs/network/latency/pairs/gru-sea-rtt)** | 106.76 ms | 68.5% | 1.54 ms | 0.19% | Fair | | 🇦🇺 Sydney (SYD) | **[161.9](/docs/network/latency/pairs/syd-sea-rtt)** | 121.96 ms | 75.3% | 1.74 ms | 0.03% | Fair | | 🇷🇺 Moscow (MOW) | **[164.8](/docs/network/latency/pairs/mow-sea-rtt)** | 82.23 ms | 49.9% | 0.99 ms | 0% | Fair | | 🇦🇺 Melbourne (MEL) | **[172.1](/docs/network/latency/pairs/mel-sea-rtt)** | 128.93 ms | 74.9% | 1.49 ms | 0% | Fair | | 🇿🇦 Johannesburg (JNB) | **[281.2](/docs/network/latency/pairs/jnb-sea-rtt)** | 161.61 ms | 57.5% | 1.91 ms | 0.11% | High | ## North America Peers Seattle is one of 5 Hats Network PoPs in **North America**. The table lists outbound measurements within the region. | PoP | Code | Country | RTT (ms) | Tier | | ------------------------------------------- | ---- | ------- | -------- | --------- | | [🇺🇸 Ashburn (IAD)](./iad-ashburn) | IAD | USA | **61.5** | Excellent | | [🇺🇸 Los Angeles (LAX)](./lax-los-angeles) | LAX | USA | **25.9** | Ultra-Low | | [🇺🇸 Miami (MIA)](./mia-miami) | MIA | USA | **81.7** | Good | | [🇺🇸 New York (NYC)](./nyc-new-york) | NYC | USA | **59.5** | Excellent | ## Questions ### What is the fastest route to Seattle? The fastest measured route to Seattle (SEA) is [Los Angeles (LAX) → Seattle (SEA)](/docs/network/latency/pairs/lax-sea-rtt), averaging **26.9 ms** RTT (Ultra-Low). ### What is the fastest route from Seattle? The fastest measured route from Seattle (SEA) is [Seattle (SEA) → Los Angeles (LAX)](/docs/network/latency/pairs/sea-lax-rtt), averaging **25.9 ms** RTT (Ultra-Low). ### How well connected is Seattle to the Hats Network backbone? Seattle (SEA) maintains measured routes to all 19 other PoPs across 5 continents. The slowest route, [Seattle (SEA) → Johannesburg (JNB)](/docs/network/latency/pairs/sea-jnb-rtt), averages **282.1 ms** RTT. *** *Round published August 16, 2026. [All 20 PoPs](/docs/network/latency) · [Dataset and licence](/opendata/latency/).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/sea-seattle). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Singapore network latency (SIN) Singapore (SIN) is the network's Southeast Asian PoP. Its regional measurements include East Asia and Australia. Measurement round: `2026-08-16T04:07:28Z`. The published round covers **19 outbound routes** from Singapore. Incoming routes are measured separately; neither direction describes last-mile performance. ## Location * **Location:** Singapore, Singapore (Asia Pacific) * **Region:** Asia Pacific * **Role:** Backbone interconnection point for Hats Network (AS203314) * **Public city reference:** [GeoNames 1880252](https://www.geonames.org/1880252); 1.28967, 103.85007 * **Coverage:** RTT measurements to 19 other PoPs across 5 continents ### Interconnection * **Facility:** SIN1 * **Upstream transit:** HE / Cogent / NTT * **Peering / IX:** Equinix IX > Interactive content is available on the canonical HTML page. ## Round Summary | Metric | Value | | --------------------- | --------------------------------------------------------------------------------------------------------------- | | Fastest Route | [🇸🇬 Singapore (SIN) → 🇭🇰 Hong Kong (HKG)](/docs/network/latency/pairs/sin-hkg-rtt) (**30.8 ms**); Excellent | | Slowest Route | [🇸🇬 Singapore (SIN) → 🇿🇦 Johannesburg (JNB)](/docs/network/latency/pairs/sin-jnb-rtt) (**312.6 ms**) | | Average RTT | **156.6 ms** | | Average Jitter | **1.11 ms** | | Average Packet Loss | **0.08%** | | Best Fiber Efficiency | **84.0%** | | Intra-Region Peers | 5 PoPs in Asia Pacific | | Total Routes | 19 outbound / 19 inbound (100% coverage) | > Jitter, packet-loss, percentile, and trend fields are estimated comparative indicators. Read the [theoretical fiber latency methodology](/docs/network/latency/theoretical-fiber-latency). ## Regional Routes Singapore's 5 other **Asia Pacific** PoPs have RTTs from **30.8 ms** (Hong Kong (HKG)) to **94.5 ms** (Sydney (SYD)). Singapore is near the southern end of the Malay Peninsula. ## Outbound Latency from Singapore (SIN) Round-trip time in milliseconds **from Singapore** to all other backbone PoPs, sorted fastest to slowest. | Destination | RTT (ms) | Fiber Floor | Efficiency | Jitter | Loss | Tier | | ----------------------- | ---------------------------------------------------- | ----------- | ---------- | ------- | ----- | --------- | | 🇭🇰 Hong Kong (HKG) | **[30.8](/docs/network/latency/pairs/sin-hkg-rtt)** | 25.24 ms | 81.9% | 0.23 ms | 0.1% | Excellent | | 🇹🇼 Taipei (TPE) | **[45.8](/docs/network/latency/pairs/sin-tpe-rtt)** | 31.77 ms | 69.4% | 0.25 ms | 0% | Excellent | | 🇯🇵 Tokyo (TYO) | **[68.9](/docs/network/latency/pairs/sin-tyo-rtt)** | 52 ms | 75.5% | 0.41 ms | 0% | Excellent | | 🇦🇺 Melbourne (MEL) | **[88.6](/docs/network/latency/pairs/sin-mel-rtt)** | 59.22 ms | 66.8% | 0.7 ms | 0% | Good | | 🇦🇺 Sydney (SYD) | **[94.5](/docs/network/latency/pairs/sin-syd-rtt)** | 61.63 ms | 65.2% | 0.53 ms | 0% | Good | | 🇫🇷 Marseille (MRS) | **[140.6](/docs/network/latency/pairs/sin-mrs-rtt)** | 103.79 ms | 73.8% | 1.48 ms | 0.2% | Good | | 🇷🇺 Moscow (MOW) | **[147.7](/docs/network/latency/pairs/sin-mow-rtt)** | 82.49 ms | 55.8% | 0.99 ms | 0.11% | Good | | 🇫🇷 Paris (PAR) | **[151.4](/docs/network/latency/pairs/sin-par-rtt)** | 105.19 ms | 69.5% | 1.07 ms | 0.12% | Fair | | 🇺🇸 Seattle (SEA) | **[151.5](/docs/network/latency/pairs/sin-sea-rtt)** | 127.22 ms | 84% | 1.11 ms | 0% | Fair | | 🇩🇪 Frankfurt (FRA) | **[151.6](/docs/network/latency/pairs/sin-fra-rtt)** | 100.56 ms | 66.3% | 0.77 ms | 0.04% | Fair | | 🇬🇧 London (LON) | **[155.4](/docs/network/latency/pairs/sin-lon-rtt)** | 106.35 ms | 68.4% | 1.17 ms | 0.06% | Fair | | 🇳🇱 Amsterdam (AMS) | **[155.8](/docs/network/latency/pairs/sin-ams-rtt)** | 102.86 ms | 66% | 1.04 ms | 0.19% | Fair | | 🇩🇪 Berlin (BER) | **[160.3](/docs/network/latency/pairs/sin-ber-rtt)** | 97.19 ms | 60.6% | 0.88 ms | 0.19% | Fair | | 🇺🇸 Los Angeles (LAX) | **[168.2](/docs/network/latency/pairs/sin-lax-rtt)** | 138.43 ms | 82.3% | 1.78 ms | 0.03% | Fair | | 🇺🇸 New York (NYC) | **[209.3](/docs/network/latency/pairs/sin-nyc-rtt)** | 150.29 ms | 71.8% | 1.38 ms | 0% | Fair | | 🇺🇸 Ashburn (IAD) | **[210](/docs/network/latency/pairs/sin-iad-rtt)** | 152.14 ms | 72.4% | 2.32 ms | 0% | Fair | | 🇺🇸 Miami (MIA) | **[221.7](/docs/network/latency/pairs/sin-mia-rtt)** | 166.27 ms | 75% | 1.1 ms | 0.19% | Fair | | 🇧🇷 São Paulo (GRU) | **[311.4](/docs/network/latency/pairs/sin-gru-rtt)** | 156.67 ms | 50.3% | 2.4 ms | 0.1% | High | | 🇿🇦 Johannesburg (JNB) | **[312.6](/docs/network/latency/pairs/sin-jnb-rtt)** | 84.85 ms | 27.1% | 1.51 ms | 0.19% | High | ## Fastest Routes to Singapore (SIN) The 10 fastest measured routes **to Singapore**, ranked by average RTT. Min / max / stdev are computed from the same measurement round (50 ICMP echo probes per route, 100 ms interval); routes without a published per-sample round show the measured average RTT only. | Rank | Source | Avg RTT | Min | Max | Stdev | | ---- | -------------------- | ------------------------------------------------------- | --------- | --------- | ------- | | 1 | 🇭🇰 Hong Kong (HKG) | **[31.5 ms](/docs/network/latency/pairs/hkg-sin-rtt)** | 29.73 ms | 35.06 ms | 1.22 ms | | 2 | 🇹🇼 Taipei (TPE) | **[44.3 ms](/docs/network/latency/pairs/tpe-sin-rtt)** | 42.61 ms | 50.6 ms | 1.64 ms | | 3 | 🇯🇵 Tokyo (TYO) | **[69.4 ms](/docs/network/latency/pairs/tyo-sin-rtt)** | 66.32 ms | 80.7 ms | 3.27 ms | | 4 | 🇦🇺 Melbourne (MEL) | **[88.4 ms](/docs/network/latency/pairs/mel-sin-rtt)** | 83.08 ms | 108.06 ms | 4.37 ms | | 5 | 🇦🇺 Sydney (SYD) | **[94.5 ms](/docs/network/latency/pairs/syd-sin-rtt)** | 89.04 ms | 106.31 ms | 3.56 ms | | 6 | 🇫🇷 Marseille (MRS) | **[139.6 ms](/docs/network/latency/pairs/mrs-sin-rtt)** | 132.36 ms | 163.5 ms | 6.56 ms | | 7 | 🇷🇺 Moscow (MOW) | **[147.3 ms](/docs/network/latency/pairs/mow-sin-rtt)** | 144.41 ms | 156.92 ms | 2.32 ms | | 8 | 🇫🇷 Paris (PAR) | **[150.3 ms](/docs/network/latency/pairs/par-sin-rtt)** | 144.03 ms | 164.23 ms | 4.53 ms | | 9 | 🇺🇸 Seattle (SEA) | **[151.1 ms](/docs/network/latency/pairs/sea-sin-rtt)** | 143.38 ms | 172.46 ms | 6.94 ms | | 10 | 🇩🇪 Frankfurt (FRA) | **[151.4 ms](/docs/network/latency/pairs/fra-sin-rtt)** | 148.19 ms | 159.42 ms | 2.49 ms | ## Inbound Latency to Singapore (SIN) Round-trip time in milliseconds **from all other PoPs to Singapore**. The measurement origin matters: inbound and outbound RTT can differ for the same pair. | Source | RTT (ms) | Fiber Floor | Efficiency | Jitter | Loss | Tier | | ----------------------- | ---------------------------------------------------- | ----------- | ---------- | ------- | ----- | --------- | | 🇭🇰 Hong Kong (HKG) | **[31.5](/docs/network/latency/pairs/hkg-sin-rtt)** | 25.24 ms | 80.1% | 0.33 ms | 0.05% | Excellent | | 🇹🇼 Taipei (TPE) | **[44.3](/docs/network/latency/pairs/tpe-sin-rtt)** | 31.77 ms | 71.7% | 0.22 ms | 0.11% | Excellent | | 🇯🇵 Tokyo (TYO) | **[69.4](/docs/network/latency/pairs/tyo-sin-rtt)** | 52 ms | 74.9% | 0.62 ms | 0% | Excellent | | 🇦🇺 Melbourne (MEL) | **[88.4](/docs/network/latency/pairs/mel-sin-rtt)** | 59.22 ms | 67% | 0.89 ms | 0.02% | Good | | 🇦🇺 Sydney (SYD) | **[94.5](/docs/network/latency/pairs/syd-sin-rtt)** | 61.63 ms | 65.2% | 0.84 ms | 0% | Good | | 🇫🇷 Marseille (MRS) | **[139.6](/docs/network/latency/pairs/mrs-sin-rtt)** | 103.79 ms | 74.3% | 1.19 ms | 0% | Good | | 🇷🇺 Moscow (MOW) | **[147.3](/docs/network/latency/pairs/mow-sin-rtt)** | 82.49 ms | 56% | 1.11 ms | 0.17% | Good | | 🇫🇷 Paris (PAR) | **[150.3](/docs/network/latency/pairs/par-sin-rtt)** | 105.19 ms | 70% | 1.32 ms | 0% | Fair | | 🇺🇸 Seattle (SEA) | **[151.1](/docs/network/latency/pairs/sea-sin-rtt)** | 127.22 ms | 84.2% | 1.4 ms | 0% | Fair | | 🇩🇪 Frankfurt (FRA) | **[151.4](/docs/network/latency/pairs/fra-sin-rtt)** | 100.56 ms | 66.4% | 0.7 ms | 0% | Fair | | 🇬🇧 London (LON) | **[156.5](/docs/network/latency/pairs/lon-sin-rtt)** | 106.35 ms | 68% | 1.57 ms | 0% | Fair | | 🇳🇱 Amsterdam (AMS) | **[157.6](/docs/network/latency/pairs/ams-sin-rtt)** | 102.86 ms | 65.3% | 0.82 ms | 0% | Fair | | 🇩🇪 Berlin (BER) | **[158.3](/docs/network/latency/pairs/ber-sin-rtt)** | 97.19 ms | 61.4% | 0.87 ms | 0.19% | Fair | | 🇺🇸 Los Angeles (LAX) | **[167.7](/docs/network/latency/pairs/lax-sin-rtt)** | 138.43 ms | 82.5% | 1.33 ms | 0% | Fair | | 🇺🇸 New York (NYC) | **[208.7](/docs/network/latency/pairs/nyc-sin-rtt)** | 150.29 ms | 72% | 1.83 ms | 0% | Fair | | 🇺🇸 Ashburn (IAD) | **[211.5](/docs/network/latency/pairs/iad-sin-rtt)** | 152.14 ms | 71.9% | 1.77 ms | 0% | Fair | | 🇺🇸 Miami (MIA) | **[223.1](/docs/network/latency/pairs/mia-sin-rtt)** | 166.27 ms | 74.5% | 1.42 ms | 0.08% | Fair | | 🇧🇷 São Paulo (GRU) | **[297.1](/docs/network/latency/pairs/gru-sin-rtt)** | 156.67 ms | 52.7% | 2.14 ms | 0.08% | High | | 🇿🇦 Johannesburg (JNB) | **[311.6](/docs/network/latency/pairs/jnb-sin-rtt)** | 84.85 ms | 27.2% | 3.35 ms | 0% | High | ## Asia Pacific Peers Singapore is one of 6 Hats Network PoPs in **Asia Pacific**. The table lists outbound measurements within the region. | PoP | Code | Country | RTT (ms) | Tier | | --------------------------------------- | ---- | --------- | -------- | --------- | | [🇭🇰 Hong Kong (HKG)](./hkg-hong-kong) | HKG | Hong Kong | **30.8** | Excellent | | [🇦🇺 Melbourne (MEL)](./mel-melbourne) | MEL | Australia | **88.6** | Good | | [🇦🇺 Sydney (SYD)](./syd-sydney) | SYD | Australia | **94.5** | Good | | [🇹🇼 Taipei (TPE)](./tpe-taipei) | TPE | Taiwan | **45.8** | Excellent | | [🇯🇵 Tokyo (TYO)](./tyo-tokyo) | TYO | Japan | **68.9** | Excellent | ## Questions ### What is the fastest route to Singapore? The fastest measured route to Singapore (SIN) is [Hong Kong (HKG) → Singapore (SIN)](/docs/network/latency/pairs/hkg-sin-rtt), averaging **31.5 ms** RTT (Excellent). ### What is the fastest route from Singapore? The fastest measured route from Singapore (SIN) is [Singapore (SIN) → Hong Kong (HKG)](/docs/network/latency/pairs/sin-hkg-rtt), averaging **30.8 ms** RTT (Excellent). ### How well connected is Singapore to the Hats Network backbone? Singapore (SIN) maintains measured routes to all 19 other PoPs across 5 continents. The slowest route, [Singapore (SIN) → Johannesburg (JNB)](/docs/network/latency/pairs/sin-jnb-rtt), averages **312.6 ms** RTT. *** *Round published August 16, 2026. [All 20 PoPs](/docs/network/latency) · [Dataset and licence](/opendata/latency/).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/sin-singapore). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Sydney network latency (SYD) Sydney (SYD) is an Australian PoP. The round includes Melbourne and the network's overseas destinations. Measurement round: `2026-08-16T04:07:28Z`. The published round covers **19 outbound routes** from Sydney. Incoming routes are measured separately; neither direction describes last-mile performance. ## Location * **Location:** Sydney, Australia (Asia Pacific) * **Region:** Asia Pacific * **Role:** Backbone interconnection point for Hats Network (AS203314) * **Public city reference:** [GeoNames 2147714](https://www.geonames.org/2147714); -33.86785, 151.20732 * **Coverage:** RTT measurements to 19 other PoPs across 5 continents ### Interconnection * **Facility:** SYD1 * **Upstream transit:** HE / Superloop * **Peering / IX:** Equinix SY > Interactive content is available on the canonical HTML page. ## Round Summary | Metric | Value | | --------------------- | ----------------------------------------------------------------------------------------------------------- | | Fastest Route | [🇦🇺 Sydney (SYD) → 🇦🇺 Melbourne (MEL)](/docs/network/latency/pairs/syd-mel-rtt) (**9.8 ms**); Ultra-Low | | Slowest Route | [🇦🇺 Sydney (SYD) → 🇿🇦 Johannesburg (JNB)](/docs/network/latency/pairs/syd-jnb-rtt) (**412.5 ms**) | | Average RTT | **201.2 ms** | | Average Jitter | **1.71 ms** | | Average Packet Loss | **0.06%** | | Best Fiber Efficiency | **86.5%** | | Intra-Region Peers | 5 PoPs in Asia Pacific | | Total Routes | 19 outbound / 19 inbound (100% coverage) | > Jitter, packet-loss, percentile, and trend fields are estimated comparative indicators. Read the [theoretical fiber latency methodology](/docs/network/latency/theoretical-fiber-latency). ## Regional Routes Sydney's 5 other **Asia Pacific** PoPs have RTTs from **9.8 ms** (Melbourne (MEL)) to **135.7 ms** (Hong Kong (HKG)). Sydney is on Australia's east coast, northeast of Melbourne. ## Outbound Latency from Sydney (SYD) Round-trip time in milliseconds **from Sydney** to all other backbone PoPs, sorted fastest to slowest. | Destination | RTT (ms) | Fiber Floor | Efficiency | Jitter | Loss | Tier | | ----------------------- | ---------------------------------------------------- | ----------- | ---------- | ------- | ----- | --------- | | 🇦🇺 Melbourne (MEL) | **[9.8](/docs/network/latency/pairs/syd-mel-rtt)** | 6.99 ms | 71.3% | 0.12 ms | 0.19% | Ultra-Low | | 🇸🇬 Singapore (SIN) | **[94.5](/docs/network/latency/pairs/syd-sin-rtt)** | 61.63 ms | 65.2% | 0.84 ms | 0% | Good | | 🇯🇵 Tokyo (TYO) | **[101.3](/docs/network/latency/pairs/syd-tyo-rtt)** | 76.31 ms | 75.3% | 0.74 ms | 0% | Good | | 🇹🇼 Taipei (TPE) | **[131.7](/docs/network/latency/pairs/syd-tpe-rtt)** | 70.88 ms | 53.8% | 1.51 ms | 0% | Good | | 🇭🇰 Hong Kong (HKG) | **[135.7](/docs/network/latency/pairs/syd-hkg-rtt)** | 71.97 ms | 53% | 0.65 ms | 0.19% | Good | | 🇺🇸 Los Angeles (LAX) | **[136.5](/docs/network/latency/pairs/syd-lax-rtt)** | 118.13 ms | 86.5% | 1.54 ms | 0.12% | Good | | 🇺🇸 Seattle (SEA) | **[161.9](/docs/network/latency/pairs/syd-sea-rtt)** | 121.96 ms | 75.3% | 1.74 ms | 0.03% | Fair | | 🇺🇸 Miami (MIA) | **[191.5](/docs/network/latency/pairs/syd-mia-rtt)** | 147.18 ms | 76.9% | 0.95 ms | 0.12% | Fair | | 🇺🇸 New York (NYC) | **[191.9](/docs/network/latency/pairs/syd-nyc-rtt)** | 156.57 ms | 81.6% | 2.06 ms | 0.07% | Fair | | 🇺🇸 Ashburn (IAD) | **[195.3](/docs/network/latency/pairs/syd-iad-rtt)** | 153.48 ms | 78.6% | 1.26 ms | 0.15% | Fair | | 🇫🇷 Marseille (MRS) | **[240.5](/docs/network/latency/pairs/syd-mrs-rtt)** | 165.41 ms | 68.8% | 2.18 ms | 0% | Fair | | 🇷🇺 Moscow (MOW) | **[242.2](/docs/network/latency/pairs/syd-mow-rtt)** | 141.84 ms | 58.6% | 1.9 ms | 0.11% | Fair | | 🇫🇷 Paris (PAR) | **[250.3](/docs/network/latency/pairs/syd-par-rtt)** | 166.06 ms | 66.3% | 2.41 ms | 0% | High | | 🇩🇪 Frankfurt (FRA) | **[254.4](/docs/network/latency/pairs/syd-fra-rtt)** | 161.37 ms | 63.4% | 2.54 ms | 0% | High | | 🇬🇧 London (LON) | **[256.5](/docs/network/latency/pairs/syd-lon-rtt)** | 166.37 ms | 64.9% | 2.97 ms | 0.09% | High | | 🇳🇱 Amsterdam (AMS) | **[258.9](/docs/network/latency/pairs/syd-ams-rtt)** | 162.93 ms | 62.9% | 1.18 ms | 0.16% | High | | 🇩🇪 Berlin (BER) | **[260.5](/docs/network/latency/pairs/syd-ber-rtt)** | 157.54 ms | 60.5% | 2.14 ms | 0% | High | | 🇧🇷 São Paulo (GRU) | **[296.7](/docs/network/latency/pairs/syd-gru-rtt)** | 131 ms | 44.2% | 1.93 ms | 0% | High | | 🇿🇦 Johannesburg (JNB) | **[412.5](/docs/network/latency/pairs/syd-jnb-rtt)** | 108.32 ms | 26.3% | 3.8 ms | 0% | High | ## Fastest Routes to Sydney (SYD) The 10 fastest measured routes **to Sydney**, ranked by average RTT. Min / max / stdev are computed from the same measurement round (50 ICMP echo probes per route, 100 ms interval); routes without a published per-sample round show the measured average RTT only. | Rank | Source | Avg RTT | Min | Max | Stdev | | ---- | ---------------------- | ------------------------------------------------------- | --------- | --------- | ------- | | 1 | 🇦🇺 Melbourne (MEL) | **[9.8 ms](/docs/network/latency/pairs/mel-syd-rtt)** | 9.36 ms | 11.08 ms | 0.4 ms | | 2 | 🇸🇬 Singapore (SIN) | **[94.5 ms](/docs/network/latency/pairs/sin-syd-rtt)** | 92.37 ms | 99.4 ms | 1.76 ms | | 3 | 🇯🇵 Tokyo (TYO) | **[101.9 ms](/docs/network/latency/pairs/tyo-syd-rtt)** | 97.82 ms | 112.89 ms | 3.32 ms | | 4 | 🇹🇼 Taipei (TPE) | **[133.2 ms](/docs/network/latency/pairs/tpe-syd-rtt)** | 129.47 ms | 143.28 ms | 3.02 ms | | 5 | 🇭🇰 Hong Kong (HKG) | **[137.2 ms](/docs/network/latency/pairs/hkg-syd-rtt)** | 129.36 ms | 162.96 ms | 6.53 ms | | 6 | 🇺🇸 Los Angeles (LAX) | **[137.6 ms](/docs/network/latency/pairs/lax-syd-rtt)** | 134.53 ms | 146.99 ms | 2.45 ms | | 7 | 🇺🇸 Seattle (SEA) | **[161.3 ms](/docs/network/latency/pairs/sea-syd-rtt)** | 157.6 ms | 177.48 ms | 3.32 ms | | 8 | 🇺🇸 Miami (MIA) | **[191.5 ms](/docs/network/latency/pairs/mia-syd-rtt)** | 181.65 ms | 213.36 ms | 7.74 ms | | 9 | 🇺🇸 New York (NYC) | **[192.6 ms](/docs/network/latency/pairs/nyc-syd-rtt)** | 185.89 ms | 212.11 ms | 5.61 ms | | 10 | 🇺🇸 Ashburn (IAD) | **[193.6 ms](/docs/network/latency/pairs/iad-syd-rtt)** | 185.3 ms | 213.55 ms | 6.56 ms | ## Inbound Latency to Sydney (SYD) Round-trip time in milliseconds **from all other PoPs to Sydney**. The measurement origin matters: inbound and outbound RTT can differ for the same pair. | Source | RTT (ms) | Fiber Floor | Efficiency | Jitter | Loss | Tier | | ----------------------- | ---------------------------------------------------- | ----------- | ---------- | ------- | ----- | --------- | | 🇦🇺 Melbourne (MEL) | **[9.8](/docs/network/latency/pairs/mel-syd-rtt)** | 6.99 ms | 71.3% | 0.12 ms | 0% | Ultra-Low | | 🇸🇬 Singapore (SIN) | **[94.5](/docs/network/latency/pairs/sin-syd-rtt)** | 61.63 ms | 65.2% | 0.53 ms | 0% | Good | | 🇯🇵 Tokyo (TYO) | **[101.9](/docs/network/latency/pairs/tyo-syd-rtt)** | 76.31 ms | 74.9% | 0.92 ms | 0.2% | Good | | 🇹🇼 Taipei (TPE) | **[133.2](/docs/network/latency/pairs/tpe-syd-rtt)** | 70.88 ms | 53.2% | 1.48 ms | 0% | Good | | 🇭🇰 Hong Kong (HKG) | **[137.2](/docs/network/latency/pairs/hkg-syd-rtt)** | 71.97 ms | 52.5% | 0.86 ms | 0.17% | Good | | 🇺🇸 Los Angeles (LAX) | **[137.6](/docs/network/latency/pairs/lax-syd-rtt)** | 118.13 ms | 85.9% | 1.65 ms | 0% | Good | | 🇺🇸 Seattle (SEA) | **[161.3](/docs/network/latency/pairs/sea-syd-rtt)** | 121.96 ms | 75.6% | 1.49 ms | 0% | Fair | | 🇺🇸 Miami (MIA) | **[191.5](/docs/network/latency/pairs/mia-syd-rtt)** | 147.18 ms | 76.9% | 0.88 ms | 0% | Fair | | 🇺🇸 New York (NYC) | **[192.6](/docs/network/latency/pairs/nyc-syd-rtt)** | 156.57 ms | 81.3% | 0.91 ms | 0% | Fair | | 🇺🇸 Ashburn (IAD) | **[193.6](/docs/network/latency/pairs/iad-syd-rtt)** | 153.48 ms | 79.3% | 1.68 ms | 0% | Fair | | 🇫🇷 Marseille (MRS) | **[240.6](/docs/network/latency/pairs/mrs-syd-rtt)** | 165.41 ms | 68.7% | 2.14 ms | 0% | Fair | | 🇷🇺 Moscow (MOW) | **[241.8](/docs/network/latency/pairs/mow-syd-rtt)** | 141.84 ms | 58.7% | 2.66 ms | 0% | Fair | | 🇫🇷 Paris (PAR) | **[249.7](/docs/network/latency/pairs/par-syd-rtt)** | 166.06 ms | 66.5% | 1.18 ms | 0% | Fair | | 🇩🇪 Frankfurt (FRA) | **[256](/docs/network/latency/pairs/fra-syd-rtt)** | 161.37 ms | 63% | 1.5 ms | 0.16% | High | | 🇳🇱 Amsterdam (AMS) | **[258](/docs/network/latency/pairs/ams-syd-rtt)** | 162.93 ms | 63.2% | 2.35 ms | 0.11% | High | | 🇬🇧 London (LON) | **[258.1](/docs/network/latency/pairs/lon-syd-rtt)** | 166.37 ms | 64.5% | 1.58 ms | 0.18% | High | | 🇩🇪 Berlin (BER) | **[262.1](/docs/network/latency/pairs/ber-syd-rtt)** | 157.54 ms | 60.1% | 3.1 ms | 0.15% | High | | 🇧🇷 São Paulo (GRU) | **[265.5](/docs/network/latency/pairs/gru-syd-rtt)** | 131 ms | 49.3% | 1.66 ms | 0% | High | | 🇿🇦 Johannesburg (JNB) | **[412.6](/docs/network/latency/pairs/jnb-syd-rtt)** | 108.32 ms | 26.3% | 2.3 ms | 0.07% | High | ## Asia Pacific Peers Sydney is one of 6 Hats Network PoPs in **Asia Pacific**. The table lists outbound measurements within the region. | PoP | Code | Country | RTT (ms) | Tier | | --------------------------------------- | ---- | --------- | --------- | --------- | | [🇭🇰 Hong Kong (HKG)](./hkg-hong-kong) | HKG | Hong Kong | **135.7** | Good | | [🇦🇺 Melbourne (MEL)](./mel-melbourne) | MEL | Australia | **9.8** | Ultra-Low | | [🇸🇬 Singapore (SIN)](./sin-singapore) | SIN | Singapore | **94.5** | Good | | [🇹🇼 Taipei (TPE)](./tpe-taipei) | TPE | Taiwan | **131.7** | Good | | [🇯🇵 Tokyo (TYO)](./tyo-tokyo) | TYO | Japan | **101.3** | Good | ## Questions ### What is the fastest route to Sydney? The fastest measured route to Sydney (SYD) is [Melbourne (MEL) → Sydney (SYD)](/docs/network/latency/pairs/mel-syd-rtt), averaging **9.8 ms** RTT (Ultra-Low). ### What is the fastest route from Sydney? The fastest measured route from Sydney (SYD) is [Sydney (SYD) → Melbourne (MEL)](/docs/network/latency/pairs/syd-mel-rtt), averaging **9.8 ms** RTT (Ultra-Low). ### How well connected is Sydney to the Hats Network backbone? Sydney (SYD) maintains measured routes to all 19 other PoPs across 5 continents. The slowest route, [Sydney (SYD) → Johannesburg (JNB)](/docs/network/latency/pairs/syd-jnb-rtt), averages **412.5 ms** RTT. *** *Round published August 16, 2026. [All 20 PoPs](/docs/network/latency) · [Dataset and licence](/opendata/latency/).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/syd-sydney). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Latency Coordinates & Fiber Model Inputs The latency model starts with public city-centre coordinates, then applies the physical constants and engineering assumptions listed below. The [formula verification](/docs/network/latency/theoretical-latency-verification) works through the equations and a fixed numerical example. > **A city reference, not a facility address** > > Published coordinates identify the GeoNames city centre. They deliberately do not reveal a router, > carrier hotel, cable landing station, or other private infrastructure location. The resulting > value is a city-to-city physical reference, not a reconstruction of the production fiber route. ## 1. Resolving the theoretical physical location Coordinate lookup uses the PoP's public city name and ISO 3166-1 alpha-2 country code. The [GeoNames Search Webservice](https://www.geonames.org/export/geonames-search.html) receives these constraints: | Parameter | Value | Purpose | | -------------- | -------------------- | --------------------------------------------------------- | | `name_equals` | Public PoP city name | Prefer an exact toponym or alternate-name match | | `country` | ISO country code | Prevent collisions such as Paris, France vs. Paris, Texas | | `featureClass` | `P` | Restrict results to populated places | | `style` | `FULL` | Return identity and location fields needed for validation | | `maxRows` | `20` | Permit local disambiguation when names are shared | When the exact query returns no result, the resolver retries with a required-name search. It normalizes diacritics, checks the country and populated-place class, then selects the highest-population exact or alternate-name match. This resolves names such as **São Paulo**, **Frankfurt am Main** and **New York City** without a facility address. The checked snapshot records the GeoNames ID, canonical name, feature code, coordinates, source, date and attribution. Refresh it with `pnpm --filter @hatsnet/docs geonames:refresh` and `GEONAMES_USERNAME`. Normal builds read the snapshot rather than querying GeoNames. > Interactive content is available on the canonical HTML page. The coordinate snapshot uses GeoNames data under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). It is separate from operational topology records. ## 2. Model inputs and engineering evidence Constants, material properties and engineering assumptions have different evidential roles: | Input | Model value | Role and evidence | | -------------------------- | ---------------: | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | | Earth model | WGS-84 ellipsoid | The inverse geodesic uses Charles F. F. Karney's robust [Algorithms for geodesics](https://doi.org/10.1007/s00190-012-0578-z) | | Vacuum light speed | 299,792.458km/s | Exact physical constant used for the absolute propagation floor | | Standard fiber group index | 1.4679 | Conventional single-mode reference from Corning's [fiber latency white paper](https://www.corning.com/media/worldwide/coc/documents/Fiber/white-paper/WP8080_03-11.pdf) | | Low-latency group index | 1.4620 | Optimistic SMF-28 ULL material comparison from the same Corning paper | | Optical path allowance | 1.05× | Transparent engineering reference for slack and small route deviations | | Attenuation reference | 0.20dB/km | Conservative long-haul single-mode planning value | | Optical span | 80km | Lower edge of the 80–120km range in [Cisco's long-haul design guide](https://www.cisco.com/c/en/us/td/docs/optical/ron/4-0/solution/guide/ron-solution-guide-40/deployment-topologies/r-long-haul-topology.html) | | Amplifier transit delay | 0.05µs per pass | Rounded from the roughly 10m equivalent fiber length described by the [GÉANT transport report](https://www.geant3.archive.geant.org/Media%20Library/GN3-13-109_DJ1-1-3_Transport-Network-Technologies-and-Operations.pdf) | | Mapped-fiber reference | 1.33× | Median conduit-to-geodesic comparison reported in [Dissecting Latency in the Internet's Fiber Infrastructure](https://arxiv.org/abs/1811.10737) | [ITU-T G.652](https://www.itu.int/rec/T-REC-G.652/) defines the relevant characteristics of standard single-mode fiber. The broader distinction between the speed-of-light baseline and routed-network path inflation is also discussed in [cISP: A Speed-of-Light Internet Service Provider](https://www.usenix.org/conference/nsdi22/presentation/bhattacherjee). > **Formula derivation and numerical verification** > > See [Theoretical Fiber Latency Formula > Verification](/docs/network/latency/theoretical-latency-verification) for the rendered equations, > unit conversions, boundary checks, and the complete Tokyo (TYO) to Sydney (SYD) worked example. ## 3. Separation of model layers * The **geographic layer** resolves city-centre coordinates and the shortest WGS-84 surface distance. * The **physical layer** compares vacuum, conventional silica, and low-latency silica propagation. * The **engineering layer** adds a 5% path allowance and idealized inline-amplifier transit time. * The **research layer** provides a 1.33× mapped-fiber comparison without treating it as a universal route prediction. * The **comparison layer** relates the published RTT to the standard-fiber physical floor. Attenuation affects optical power, span count and amplifier count, not the speed of propagation. This lower-bound model excludes vendor-specific transponder, FEC, OTN switching, router, serialization and queueing delays. ## 4. Estimated stability indicators City-pair pages report statistics from their published probe round. The separate 24-hour, 7-day and 30-day trend graphic is an estimated comparison, not a history of collected rounds. City-table jitter and loss fields are also estimated indicators. Do not use estimated indicators as an outage record or an SLA measurement. ## 5. Interpretation checklist * Use **vacuum RTT** only as an absolute law-of-physics reference. * Use **fiber RTT** as the shortest possible RTT through standard silica along the WGS-84 geodesic. * Use **engineering floor** as a transparent low-latency optical design reference with small cable and amplifier allowances. * Use **1.33× mapped-fiber RTT** as a research comparison, not as a universal route prediction. * Compare **efficiency** across routes only after considering city-centre coordinates, submarine cable geography, routing policy, and normal measurement variation. * Do not interpret any value on these pages as an SLA or as disclosure of a production facility or cable path. The [formula verification](/docs/network/latency/theoretical-latency-verification) shows the calculation. Published observations are in the [latency matrix](/docs/network/latency) and the [city-pair directory](/docs/network/latency/pairs). --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/theoretical-fiber-latency). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Theoretical Fiber Latency Formula Verification The worked example uses **Tokyo (TYO) to Sydney (SYD)** and a fixed **101.7 ms** RTT recorded on July 18, 2026. Keeping this fixture unchanged lets the regression test check the equations independently of later measurement rounds. Coordinates are public GeoNames city centres. > **Verification scope** > > This is a reproducibility check for the public theoretical model. It verifies the mathematics and > implementation outputs; it does not attempt to reconstruct a deployed cable path. Read [Latency > Coordinates & Fiber Model Inputs](/docs/network/latency/theoretical-fiber-latency) for coordinate > resolution, source selection, and engineering assumptions. ## Symbols and reference inputs | Symbol | Meaning | Verification value | | --------- | ---------------------------------- | ----------------------------------: | | $A$ | Tokyo GeoNames city centre | $35.68950^\circ,\ 139.69171^\circ$ | | $B$ | Sydney GeoNames city centre | $-33.86785^\circ,\ 151.20732^\circ$ | | $d_g$ | WGS-84 inverse-geodesic distance | $7\,792.8\ \mathrm{km}$ | | $c$ | Speed of light in vacuum | $299\,792.458\ \mathrm{km\,s^{-1}}$ | | $n_s$ | Standard-fiber group index | $1.4679$ | | $n_u$ | Low-latency-fiber group index | $1.4620$ | | $\lambda$ | Engineering path multiplier | $1.05$ | | $\alpha$ | Attenuation reference | $0.20\ \mathrm{dB\,km^{-1}}$ | | $S$ | Maximum modeled optical span | $80\ \mathrm{km}$ | | $\tau_a$ | One-pass amplifier transit delay | $0.05\ \mathrm{\mu s}$ | | $RTT_p$ | Reference RTT fixture (2026-07-18) | $101.7\ \mathrm{ms}$ | ## 1. WGS-84 geodesic verification The first step is the WGS-84 inverse geodesic: $$ d_g = \operatorname{InverseGeodesic}_{\mathrm{WGS84}}(A,B) = 7\,792.8\ \mathrm{km} $$ Using the ellipsoid avoids the spherical-Earth approximation. Its surface distance supplies the geographic lower bound for the remaining steps. ## 2. Vacuum propagation floor RTT covers the distance twice. The factor $10^3$ converts seconds to milliseconds: $$ \begin{aligned} RTT_{\mathrm{vac}} &= \frac{2d_g}{c} \times 10^3 \\ &= \frac{2(7\,792.8\ \mathrm{km})} {299\,792.458\ \mathrm{km\,s^{-1}}} \times 10^3\ \mathrm{ms\,s^{-1}} \\ &= 51.99\ \mathrm{ms} \end{aligned} $$ Kilometres cancel in the ratio; the result is in seconds before conversion. ## 3. Silica propagation floors Fiber group velocity is given by: $$ v_g = \frac{c}{n_g} $$ Substituting the standard single-mode group index gives: $$ \begin{aligned} RTT_{\mathrm{fiber}} &= \frac{2d_g n_s}{c} \times 10^3 \\ &= \frac{2(7\,792.8)(1.4679)} {299\,792.458} \times 10^3 \\ &= 76.31\ \mathrm{ms} \end{aligned} $$ For the low-latency material comparison: $$ \begin{aligned} RTT_{\mathrm{ULL}} &= \frac{2d_g n_u}{c} \times 10^3 \\ &= \frac{2(7\,792.8)(1.4620)} {299\,792.458} \times 10^3 \\ &= 76.01\ \mathrm{ms} \end{aligned} $$ The resulting propagation times must follow this ordering: $$ RTT_{\mathrm{vac}} < RTT_{\mathrm{ULL}} < RTT_{\mathrm{fiber}} $$ ## 4. Engineering path, attenuation, and amplifiers The engineering reference applies the stated path allowance: $$ \begin{aligned} d_e &= \lambda d_g \\ &= 1.05(7\,792.8) \\ &= 8\,182.5\ \mathrm{km} \end{aligned} $$ Optical power loss is calculated as: $$ \begin{aligned} L_{\mathrm{dB}} &= \alpha d_e \\ &= (0.20\ \mathrm{dB\,km^{-1}}) (8\,182.5\ \mathrm{km}) \\ &= 1\,636.5\ \mathrm{dB} \end{aligned} $$ Span and inline-amplifier counts are calculated separately: $$ \begin{aligned} N_{\mathrm{span}} &= \left\lceil \frac{d_e}{S} \right\rceil = \left\lceil \frac{8\,182.5}{80} \right\rceil = 103 \\ N_{\mathrm{amp}} &= \max(0,N_{\mathrm{span}}-1) = 102 \end{aligned} $$ Convert each amplifier's microsecond contribution to milliseconds: $$ \begin{aligned} RTT_{\mathrm{amp}} &= 2N_{\mathrm{amp}}\tau_a \\ &= 2(102)(0.05\ \mathrm{\mu s}) \\ &= 10.2\ \mathrm{\mu s} = 0.0102\ \mathrm{ms} \end{aligned} $$ Adding the terms gives the engineering floor: $$ \begin{aligned} RTT_{\mathrm{eng}} &= \frac{2d_e n_s}{c} \times 10^3 + RTT_{\mathrm{amp}} \\ &= \frac{2(8\,182.5)(1.4679)} {299\,792.458} \times 10^3 + 0.0102 \\ &= 80.14\ \mathrm{ms} \end{aligned} $$ Attenuation enters through $N_{\mathrm{span}}$ and $N_{\mathrm{amp}}$. It does not appear in the propagation-speed term, whose units and physical meaning are different. ## 5. Mapped-fiber research reference The separate mapped-fiber research comparison uses 1.33×: $$ \begin{aligned} RTT_{\mathrm{mapped}} &= 1.33RTT_{\mathrm{fiber}} \\ &= 1.33(76.31\ \mathrm{ms}) \\ &= 101.50\ \mathrm{ms} \end{aligned} $$ The result is close to the fixed Tokyo-to-Sydney fixture. That numerical agreement does not establish the route's cable length. ## 6. Efficiency and latency inflation With the fixed $101.7\ \mathrm{ms}$ observation: $$ \begin{aligned} \eta &= 100\frac{RTT_{\mathrm{fiber}}}{RTT_p} \\ &= 100\frac{76.31}{101.7} \\ &= 75.03\% \approx 75.0\% \end{aligned} $$ $$ \begin{aligned} I &= \frac{RTT_p}{RTT_{\mathrm{fiber}}} \\ &= \frac{101.7}{76.31} \\ &= 1.3327 \approx 1.33 \end{aligned} $$ Converting efficiency from percent to a ratio makes it the reciprocal of inflation: $$ \frac{\eta}{100} = \frac{1}{I} $$ ## 7. Implementation reconciliation | Output | Formula result | Expected implementation output | Difference | | ---------------------- | -------------: | -----------------------------: | ---------: | | WGS-84 distance | 7,792.8km | 7,792.8km | 0.0km | | Vacuum RTT | 51.99ms | 51.99ms | 0.00ms | | Standard-fiber RTT | 76.31ms | 76.31ms | 0.00ms | | Low-latency-fiber RTT | 76.01ms | 76.01ms | 0.00ms | | Engineering distance | 8,182.5km | 8,182.5km | 0.0km | | Engineering RTT | 80.14ms | 80.14ms | 0.00ms | | Mapped-fiber reference | 101.50ms | 101.50ms | 0.00ms | | Inline amplifiers | 102 | 102 | 0 | | Amplifier RTT | 0.0102ms | 0.0102ms | 0.0000ms | | Fiber efficiency | 75.0% | 75.0% | 0.0pp | | Latency inflation | 1.33× | 1.33× | 0.00× | ## 8. Boundary checks Automated tests check these invariants: $$ d_g \ge 0 $$ $$ RTT_{\mathrm{vac}} < RTT_{\mathrm{ULL}} < RTT_{\mathrm{fiber}} < RTT_{\mathrm{eng}} $$ $$ N_{\mathrm{span}} \ge 1, \qquad N_{\mathrm{amp}} = N_{\mathrm{span}} - 1 $$ $$ RTT_{\mathrm{amp}} \ge 0, \qquad \eta > 0, \qquad I > 0 $$ A changed constant, coordinate or rounding rule requires a corresponding update to the reconciliation test. The equations and generated results should continue to agree. See the [model inputs](/docs/network/latency/theoretical-fiber-latency), the [published Tokyo to Sydney round](/docs/network/latency/pairs/tyo-syd-rtt), or the [full matrix](/docs/network/latency). --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/theoretical-latency-verification). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Taipei network latency (TPE) Taipei (TPE) is the network's Taiwan PoP. The route table includes Hong Kong, Tokyo and Singapore alongside intercontinental peers. Measurement round: `2026-08-16T04:07:28Z`. The published round covers **19 outbound routes** from Taipei. Incoming routes are measured separately; neither direction describes last-mile performance. ## Location * **Location:** Taipei, Taiwan (Asia Pacific) * **Region:** Asia Pacific * **Role:** Backbone interconnection point for Hats Network (AS203314) * **Public city reference:** [GeoNames 1668341](https://www.geonames.org/1668341); 25.05306, 121.52639 * **Coverage:** RTT measurements to 19 other PoPs across 5 continents ### Interconnection * **Facility:** TPE2 * **Upstream transit:** HE Only * **Peering / IX:** STUIX - Taipei > Interactive content is available on the canonical HTML page. ## Round Summary | Metric | Value | | --------------------- | ------------------------------------------------------------------------------------------------------------ | | Fastest Route | [🇹🇼 Taipei (TPE) → 🇭🇰 Hong Kong (HKG)](/docs/network/latency/pairs/tpe-hkg-rtt) (**15.5 ms**); Ultra-Low | | Slowest Route | [🇹🇼 Taipei (TPE) → 🇿🇦 Johannesburg (JNB)](/docs/network/latency/pairs/tpe-jnb-rtt) (**331 ms**) | | Average RTT | **152.6 ms** | | Average Jitter | **1.24 ms** | | Average Packet Loss | **0.06%** | | Best Fiber Efficiency | **82.4%** | | Intra-Region Peers | 5 PoPs in Asia Pacific | | Total Routes | 19 outbound / 19 inbound (100% coverage) | > Jitter, packet-loss, percentile, and trend fields are estimated comparative indicators. Read the [theoretical fiber latency methodology](/docs/network/latency/theoretical-fiber-latency). ## Regional Routes Taipei's 5 other **Asia Pacific** PoPs have RTTs from **15.5 ms** (Hong Kong (HKG)) to **142 ms** (Melbourne (MEL)). Taipei is in northern Taiwan, between Hong Kong and Tokyo in the regional comparison. ## Outbound Latency from Taipei (TPE) Round-trip time in milliseconds **from Taipei** to all other backbone PoPs, sorted fastest to slowest. | Destination | RTT (ms) | Fiber Floor | Efficiency | Jitter | Loss | Tier | | ----------------------- | ---------------------------------------------------- | ----------- | ---------- | ------- | ----- | --------- | | 🇭🇰 Hong Kong (HKG) | **[15.5](/docs/network/latency/pairs/tpe-hkg-rtt)** | 7.94 ms | 51.2% | 0.16 ms | 0.02% | Ultra-Low | | 🇯🇵 Tokyo (TYO) | **[31.9](/docs/network/latency/pairs/tpe-tyo-rtt)** | 20.58 ms | 64.5% | 0.23 ms | 0.08% | Excellent | | 🇸🇬 Singapore (SIN) | **[44.3](/docs/network/latency/pairs/tpe-sin-rtt)** | 31.77 ms | 71.7% | 0.22 ms | 0.11% | Excellent | | 🇺🇸 Seattle (SEA) | **[115.8](/docs/network/latency/pairs/tpe-sea-rtt)** | 95.46 ms | 82.4% | 0.97 ms | 0.15% | Good | | 🇺🇸 Los Angeles (LAX) | **[132](/docs/network/latency/pairs/tpe-lax-rtt)** | 107 ms | 81.1% | 1.31 ms | 0% | Good | | 🇷🇺 Moscow (MOW) | **[132.7](/docs/network/latency/pairs/tpe-mow-rtt)** | 72.16 ms | 54.4% | 0.68 ms | 0.09% | Good | | 🇦🇺 Sydney (SYD) | **[133.2](/docs/network/latency/pairs/tpe-syd-rtt)** | 70.88 ms | 53.2% | 1.48 ms | 0% | Good | | 🇦🇺 Melbourne (MEL) | **[142](/docs/network/latency/pairs/tpe-mel-rtt)** | 72.25 ms | 50.9% | 1.12 ms | 0.08% | Good | | 🇩🇪 Berlin (BER) | **[158.6](/docs/network/latency/pairs/tpe-ber-rtt)** | 87.84 ms | 55.4% | 1.84 ms | 0.09% | Fair | | 🇩🇪 Frankfurt (FRA) | **[166](/docs/network/latency/pairs/tpe-fra-rtt)** | 91.96 ms | 55.4% | 1.67 ms | 0.13% | Fair | | 🇳🇱 Amsterdam (AMS) | **[166.5](/docs/network/latency/pairs/tpe-ams-rtt)** | 92.72 ms | 55.7% | 0.78 ms | 0% | Fair | | 🇺🇸 Ashburn (IAD) | **[171.6](/docs/network/latency/pairs/tpe-iad-rtt)** | 123.78 ms | 72.1% | 1.7 ms | 0% | Fair | | 🇬🇧 London (LON) | **[173](/docs/network/latency/pairs/tpe-lon-rtt)** | 95.97 ms | 55.5% | 1.51 ms | 0.16% | Fair | | 🇫🇷 Paris (PAR) | **[173.8](/docs/network/latency/pairs/tpe-par-rtt)** | 96.42 ms | 55.5% | 1.13 ms | 0.13% | Fair | | 🇺🇸 New York (NYC) | **[174.2](/docs/network/latency/pairs/tpe-nyc-rtt)** | 122.89 ms | 70.5% | 1.24 ms | 0% | Fair | | 🇫🇷 Marseille (MRS) | **[179.5](/docs/network/latency/pairs/tpe-mrs-rtt)** | 98.2 ms | 54.7% | 1.4 ms | 0% | Fair | | 🇺🇸 Miami (MIA) | **[185.7](/docs/network/latency/pairs/tpe-mia-rtt)** | 136.3 ms | 73.4% | 1.27 ms | 0% | Fair | | 🇧🇷 São Paulo (GRU) | **[273](/docs/network/latency/pairs/tpe-gru-rtt)** | 184.29 ms | 67.5% | 2.95 ms | 0.03% | High | | 🇿🇦 Johannesburg (JNB) | **[331](/docs/network/latency/pairs/tpe-jnb-rtt)** | 112.87 ms | 34.1% | 1.92 ms | 0% | High | ## Fastest Routes to Taipei (TPE) The 10 fastest measured routes **to Taipei**, ranked by average RTT. Min / max / stdev are computed from the same measurement round (50 ICMP echo probes per route, 100 ms interval); routes without a published per-sample round show the measured average RTT only. | Rank | Source | Avg RTT | Min | Max | Stdev | | ---- | ---------------------- | ------------------------------------------------------- | --------- | --------- | ------- | | 1 | 🇭🇰 Hong Kong (HKG) | **[14.7 ms](/docs/network/latency/pairs/hkg-tpe-rtt)** | 13.95 ms | 17.24 ms | 0.72 ms | | 2 | 🇯🇵 Tokyo (TYO) | **[32.3 ms](/docs/network/latency/pairs/tyo-tpe-rtt)** | 31.1 ms | 36.28 ms | 0.96 ms | | 3 | 🇸🇬 Singapore (SIN) | **[45.8 ms](/docs/network/latency/pairs/sin-tpe-rtt)** | 43.12 ms | 50.4 ms | 1.72 ms | | 4 | 🇺🇸 Seattle (SEA) | **[114.7 ms](/docs/network/latency/pairs/sea-tpe-rtt)** | 109.75 ms | 137.83 ms | 5.07 ms | | 5 | 🇦🇺 Sydney (SYD) | **[131.7 ms](/docs/network/latency/pairs/syd-tpe-rtt)** | 126.36 ms | 151.94 ms | 4.46 ms | | 6 | 🇷🇺 Moscow (MOW) | **[132.1 ms](/docs/network/latency/pairs/mow-tpe-rtt)** | 126.75 ms | 145.67 ms | 4.52 ms | | 7 | 🇺🇸 Los Angeles (LAX) | **[132.9 ms](/docs/network/latency/pairs/lax-tpe-rtt)** | 128.75 ms | 148.78 ms | 4.24 ms | | 8 | 🇦🇺 Melbourne (MEL) | **[143.2 ms](/docs/network/latency/pairs/mel-tpe-rtt)** | 135 ms | 166.26 ms | 7.11 ms | | 9 | 🇩🇪 Berlin (BER) | **[156.8 ms](/docs/network/latency/pairs/ber-tpe-rtt)** | 150.23 ms | 174.22 ms | 5.07 ms | | 10 | 🇩🇪 Frankfurt (FRA) | **[163.8 ms](/docs/network/latency/pairs/fra-tpe-rtt)** | 159.37 ms | 176 ms | 3.63 ms | ## Inbound Latency to Taipei (TPE) Round-trip time in milliseconds **from all other PoPs to Taipei**. The measurement origin matters: inbound and outbound RTT can differ for the same pair. | Source | RTT (ms) | Fiber Floor | Efficiency | Jitter | Loss | Tier | | ----------------------- | ---------------------------------------------------- | ----------- | ---------- | ------- | ----- | --------- | | 🇭🇰 Hong Kong (HKG) | **[14.7](/docs/network/latency/pairs/hkg-tpe-rtt)** | 7.94 ms | 54% | 0.14 ms | 0.08% | Ultra-Low | | 🇯🇵 Tokyo (TYO) | **[32.3](/docs/network/latency/pairs/tyo-tpe-rtt)** | 20.58 ms | 63.7% | 0.34 ms | 0% | Excellent | | 🇸🇬 Singapore (SIN) | **[45.8](/docs/network/latency/pairs/sin-tpe-rtt)** | 31.77 ms | 69.4% | 0.25 ms | 0% | Excellent | | 🇺🇸 Seattle (SEA) | **[114.7](/docs/network/latency/pairs/sea-tpe-rtt)** | 95.46 ms | 83.2% | 1.36 ms | 0% | Good | | 🇦🇺 Sydney (SYD) | **[131.7](/docs/network/latency/pairs/syd-tpe-rtt)** | 70.88 ms | 53.8% | 1.51 ms | 0% | Good | | 🇷🇺 Moscow (MOW) | **[132.1](/docs/network/latency/pairs/mow-tpe-rtt)** | 72.16 ms | 54.6% | 1.49 ms | 0% | Good | | 🇺🇸 Los Angeles (LAX) | **[132.9](/docs/network/latency/pairs/lax-tpe-rtt)** | 107 ms | 80.5% | 0.87 ms | 0.08% | Good | | 🇦🇺 Melbourne (MEL) | **[143.2](/docs/network/latency/pairs/mel-tpe-rtt)** | 72.25 ms | 50.5% | 0.65 ms | 0% | Good | | 🇩🇪 Berlin (BER) | **[156.8](/docs/network/latency/pairs/ber-tpe-rtt)** | 87.84 ms | 56% | 1.45 ms | 0.17% | Fair | | 🇩🇪 Frankfurt (FRA) | **[163.8](/docs/network/latency/pairs/fra-tpe-rtt)** | 91.96 ms | 56.1% | 1.28 ms | 0.04% | Fair | | 🇳🇱 Amsterdam (AMS) | **[168.7](/docs/network/latency/pairs/ams-tpe-rtt)** | 92.72 ms | 55% | 1.29 ms | 0.11% | Fair | | 🇫🇷 Paris (PAR) | **[171.9](/docs/network/latency/pairs/par-tpe-rtt)** | 96.42 ms | 56.1% | 0.85 ms | 0% | Fair | | 🇺🇸 Ashburn (IAD) | **[172.3](/docs/network/latency/pairs/iad-tpe-rtt)** | 123.78 ms | 71.8% | 1.59 ms | 0.04% | Fair | | 🇺🇸 New York (NYC) | **[173.4](/docs/network/latency/pairs/nyc-tpe-rtt)** | 122.89 ms | 70.9% | 1.21 ms | 0% | Fair | | 🇬🇧 London (LON) | **[173.8](/docs/network/latency/pairs/lon-tpe-rtt)** | 95.97 ms | 55.2% | 1.58 ms | 0.09% | Fair | | 🇫🇷 Marseille (MRS) | **[180.9](/docs/network/latency/pairs/mrs-tpe-rtt)** | 98.2 ms | 54.3% | 1.86 ms | 0% | Fair | | 🇺🇸 Miami (MIA) | **[185.1](/docs/network/latency/pairs/mia-tpe-rtt)** | 136.3 ms | 73.6% | 1.17 ms | 0.09% | Fair | | 🇧🇷 São Paulo (GRU) | **[259.1](/docs/network/latency/pairs/gru-tpe-rtt)** | 184.29 ms | 71.1% | 3.01 ms | 0% | High | | 🇿🇦 Johannesburg (JNB) | **[329.8](/docs/network/latency/pairs/jnb-tpe-rtt)** | 112.87 ms | 34.2% | 3.33 ms | 0% | High | ## Asia Pacific Peers Taipei is one of 6 Hats Network PoPs in **Asia Pacific**. The table lists outbound measurements within the region. | PoP | Code | Country | RTT (ms) | Tier | | --------------------------------------- | ---- | --------- | --------- | --------- | | [🇭🇰 Hong Kong (HKG)](./hkg-hong-kong) | HKG | Hong Kong | **15.5** | Ultra-Low | | [🇦🇺 Melbourne (MEL)](./mel-melbourne) | MEL | Australia | **142** | Good | | [🇸🇬 Singapore (SIN)](./sin-singapore) | SIN | Singapore | **44.3** | Excellent | | [🇦🇺 Sydney (SYD)](./syd-sydney) | SYD | Australia | **133.2** | Good | | [🇯🇵 Tokyo (TYO)](./tyo-tokyo) | TYO | Japan | **31.9** | Excellent | ## Questions ### What is the fastest route to Taipei? The fastest measured route to Taipei (TPE) is [Hong Kong (HKG) → Taipei (TPE)](/docs/network/latency/pairs/hkg-tpe-rtt), averaging **14.7 ms** RTT (Ultra-Low). ### What is the fastest route from Taipei? The fastest measured route from Taipei (TPE) is [Taipei (TPE) → Hong Kong (HKG)](/docs/network/latency/pairs/tpe-hkg-rtt), averaging **15.5 ms** RTT (Ultra-Low). ### How well connected is Taipei to the Hats Network backbone? Taipei (TPE) maintains measured routes to all 19 other PoPs across 5 continents. The slowest route, [Taipei (TPE) → Johannesburg (JNB)](/docs/network/latency/pairs/tpe-jnb-rtt), averages **331 ms** RTT. *** *Round published August 16, 2026. [All 20 PoPs](/docs/network/latency) · [Dataset and licence](/opendata/latency/).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/tpe-taipei). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Tokyo network latency (TYO) Tokyo (TYO) is the Japanese PoP in this dataset. Regional and trans-Pacific measurements are available in both directions. Measurement round: `2026-08-16T04:07:28Z`. The published round covers **19 outbound routes** from Tokyo. Incoming routes are measured separately; neither direction describes last-mile performance. ## Location * **Location:** Tokyo, Japan (Asia Pacific) * **Region:** Asia Pacific * **Role:** Backbone interconnection point for Hats Network (AS203314) * **Public city reference:** [GeoNames 1850147](https://www.geonames.org/1850147); 35.68950, 139.69171 * **Coverage:** RTT measurements to 19 other PoPs across 5 continents ### Interconnection * **Facility:** TYO2 > Interactive content is available on the canonical HTML page. ## Round Summary | Metric | Value | | --------------------- | -------------------------------------------------------------------------------------------------------- | | Fastest Route | [🇯🇵 Tokyo (TYO) → 🇹🇼 Taipei (TPE)](/docs/network/latency/pairs/tyo-tpe-rtt) (**32.3 ms**); Excellent | | Slowest Route | [🇯🇵 Tokyo (TYO) → 🇿🇦 Johannesburg (JNB)](/docs/network/latency/pairs/tyo-jnb-rtt) (**359.2 ms**) | | Average RTT | **154.4 ms** | | Average Jitter | **1.25 ms** | | Average Packet Loss | **0.05%** | | Best Fiber Efficiency | **89.1%** | | Intra-Region Peers | 5 PoPs in Asia Pacific | | Total Routes | 19 outbound / 19 inbound (100% coverage) | > Jitter, packet-loss, percentile, and trend fields are estimated comparative indicators. Read the [theoretical fiber latency methodology](/docs/network/latency/theoretical-fiber-latency). ## Regional Routes Tokyo's 5 other **Asia Pacific** PoPs have RTTs from **32.3 ms** (Taipei (TPE)) to **113.9 ms** (Melbourne (MEL)). Tokyo is on Japan's Pacific coast, northeast of Taipei. ## Outbound Latency from Tokyo (TYO) Round-trip time in milliseconds **from Tokyo** to all other backbone PoPs, sorted fastest to slowest. | Destination | RTT (ms) | Fiber Floor | Efficiency | Jitter | Loss | Tier | | ----------------------- | ---------------------------------------------------- | ----------- | ---------- | ------- | ----- | --------- | | 🇹🇼 Taipei (TPE) | **[32.3](/docs/network/latency/pairs/tyo-tpe-rtt)** | 20.58 ms | 63.7% | 0.34 ms | 0% | Excellent | | 🇭🇰 Hong Kong (HKG) | **[44.8](/docs/network/latency/pairs/tyo-hkg-rtt)** | 28.24 ms | 63% | 0.45 ms | 0.07% | Excellent | | 🇸🇬 Singapore (SIN) | **[69.4](/docs/network/latency/pairs/tyo-sin-rtt)** | 52 ms | 74.9% | 0.62 ms | 0% | Excellent | | 🇺🇸 Seattle (SEA) | **[84.8](/docs/network/latency/pairs/tyo-sea-rtt)** | 75.55 ms | 89.1% | 0.77 ms | 0.06% | Good | | 🇺🇸 Los Angeles (LAX) | **[101.2](/docs/network/latency/pairs/tyo-lax-rtt)** | 86.51 ms | 85.5% | 0.97 ms | 0.02% | Good | | 🇦🇺 Sydney (SYD) | **[101.9](/docs/network/latency/pairs/tyo-syd-rtt)** | 76.31 ms | 74.9% | 0.92 ms | 0.2% | Good | | 🇦🇺 Melbourne (MEL) | **[113.9](/docs/network/latency/pairs/tyo-mel-rtt)** | 79.87 ms | 70.1% | 1.25 ms | 0.05% | Good | | 🇺🇸 Ashburn (IAD) | **[142.9](/docs/network/latency/pairs/tyo-iad-rtt)** | 106.69 ms | 74.7% | 0.69 ms | 0% | Good | | 🇺🇸 New York (NYC) | **[144.6](/docs/network/latency/pairs/tyo-nyc-rtt)** | 106.47 ms | 73.6% | 1.33 ms | 0.17% | Good | | 🇺🇸 Miami (MIA) | **[156.5](/docs/network/latency/pairs/tyo-mia-rtt)** | 117.71 ms | 75.2% | 1.36 ms | 0.14% | Fair | | 🇷🇺 Moscow (MOW) | **[161.7](/docs/network/latency/pairs/tyo-mow-rtt)** | 73.42 ms | 45.4% | 1.62 ms | 0.09% | Fair | | 🇩🇪 Berlin (BER) | **[186.1](/docs/network/latency/pairs/tyo-ber-rtt)** | 87.53 ms | 47% | 1.64 ms | 0% | Fair | | 🇩🇪 Frankfurt (FRA) | **[193.2](/docs/network/latency/pairs/tyo-fra-rtt)** | 91.61 ms | 47.4% | 2.16 ms | 0% | Fair | | 🇳🇱 Amsterdam (AMS) | **[197.4](/docs/network/latency/pairs/tyo-ams-rtt)** | 91.18 ms | 46.2% | 1.09 ms | 0% | Fair | | 🇬🇧 London (LON) | **[202.7](/docs/network/latency/pairs/tyo-lon-rtt)** | 93.84 ms | 46.3% | 0.93 ms | 0% | Fair | | 🇫🇷 Paris (PAR) | **[204.8](/docs/network/latency/pairs/tyo-par-rtt)** | 95.34 ms | 46.6% | 2.06 ms | 0.03% | Fair | | 🇫🇷 Marseille (MRS) | **[205.6](/docs/network/latency/pairs/tyo-mrs-rtt)** | 99.03 ms | 48.2% | 1.86 ms | 0.14% | Fair | | 🇧🇷 São Paulo (GRU) | **[230.5](/docs/network/latency/pairs/tyo-gru-rtt)** | 181.46 ms | 78.7% | 1.53 ms | 0% | Fair | | 🇿🇦 Johannesburg (JNB) | **[359.2](/docs/network/latency/pairs/tyo-jnb-rtt)** | 132.57 ms | 36.9% | 2.19 ms | 0.05% | High | ## Fastest Routes to Tokyo (TYO) The 10 fastest measured routes **to Tokyo**, ranked by average RTT. Min / max / stdev are computed from the same measurement round (50 ICMP echo probes per route, 100 ms interval); routes without a published per-sample round show the measured average RTT only. | Rank | Source | Avg RTT | Min | Max | Stdev | | ---- | ---------------------- | ------------------------------------------------------- | --------- | --------- | ------- | | 1 | 🇹🇼 Taipei (TPE) | **[31.9 ms](/docs/network/latency/pairs/tpe-tyo-rtt)** | 30.57 ms | 35.67 ms | 1.12 ms | | 2 | 🇭🇰 Hong Kong (HKG) | **[44.9 ms](/docs/network/latency/pairs/hkg-tyo-rtt)** | 42.32 ms | 50.91 ms | 1.82 ms | | 3 | 🇸🇬 Singapore (SIN) | **[68.9 ms](/docs/network/latency/pairs/sin-tyo-rtt)** | 65.47 ms | 78.16 ms | 2.72 ms | | 4 | 🇺🇸 Seattle (SEA) | **[85 ms](/docs/network/latency/pairs/sea-tyo-rtt)** | 81.03 ms | 97.43 ms | 3.15 ms | | 5 | 🇺🇸 Los Angeles (LAX) | **[101.2 ms](/docs/network/latency/pairs/lax-tyo-rtt)** | 97.1 ms | 112.57 ms | 3.72 ms | | 6 | 🇦🇺 Sydney (SYD) | **[101.3 ms](/docs/network/latency/pairs/syd-tyo-rtt)** | 99.25 ms | 108.62 ms | 2.05 ms | | 7 | 🇦🇺 Melbourne (MEL) | **[112 ms](/docs/network/latency/pairs/mel-tyo-rtt)** | 105.34 ms | 140.34 ms | 5.58 ms | | 8 | 🇺🇸 Ashburn (IAD) | **[141.9 ms](/docs/network/latency/pairs/iad-tyo-rtt)** | 136.1 ms | 154.55 ms | 4.32 ms | | 9 | 🇺🇸 New York (NYC) | **[143.1 ms](/docs/network/latency/pairs/nyc-tyo-rtt)** | 136.69 ms | 171.34 ms | 6.58 ms | | 10 | 🇺🇸 Miami (MIA) | **[156.3 ms](/docs/network/latency/pairs/mia-tyo-rtt)** | 149.17 ms | 178.72 ms | 6.19 ms | ## Inbound Latency to Tokyo (TYO) Round-trip time in milliseconds **from all other PoPs to Tokyo**. The measurement origin matters: inbound and outbound RTT can differ for the same pair. | Source | RTT (ms) | Fiber Floor | Efficiency | Jitter | Loss | Tier | | ----------------------- | ---------------------------------------------------- | ----------- | ---------- | ------- | ----- | --------- | | 🇹🇼 Taipei (TPE) | **[31.9](/docs/network/latency/pairs/tpe-tyo-rtt)** | 20.58 ms | 64.5% | 0.23 ms | 0.08% | Excellent | | 🇭🇰 Hong Kong (HKG) | **[44.9](/docs/network/latency/pairs/hkg-tyo-rtt)** | 28.24 ms | 62.9% | 0.23 ms | 0.08% | Excellent | | 🇸🇬 Singapore (SIN) | **[68.9](/docs/network/latency/pairs/sin-tyo-rtt)** | 52 ms | 75.5% | 0.41 ms | 0% | Excellent | | 🇺🇸 Seattle (SEA) | **[85](/docs/network/latency/pairs/sea-tyo-rtt)** | 75.55 ms | 88.9% | 0.39 ms | 0% | Good | | 🇺🇸 Los Angeles (LAX) | **[101.2](/docs/network/latency/pairs/lax-tyo-rtt)** | 86.51 ms | 85.5% | 0.82 ms | 0% | Good | | 🇦🇺 Sydney (SYD) | **[101.3](/docs/network/latency/pairs/syd-tyo-rtt)** | 76.31 ms | 75.3% | 0.74 ms | 0% | Good | | 🇦🇺 Melbourne (MEL) | **[112](/docs/network/latency/pairs/mel-tyo-rtt)** | 79.87 ms | 71.3% | 0.86 ms | 0% | Good | | 🇺🇸 Ashburn (IAD) | **[141.9](/docs/network/latency/pairs/iad-tyo-rtt)** | 106.69 ms | 75.2% | 1.23 ms | 0.02% | Good | | 🇺🇸 New York (NYC) | **[143.1](/docs/network/latency/pairs/nyc-tyo-rtt)** | 106.47 ms | 74.4% | 0.8 ms | 0% | Good | | 🇺🇸 Miami (MIA) | **[156.3](/docs/network/latency/pairs/mia-tyo-rtt)** | 117.71 ms | 75.3% | 0.95 ms | 0.03% | Fair | | 🇷🇺 Moscow (MOW) | **[162.2](/docs/network/latency/pairs/mow-tyo-rtt)** | 73.42 ms | 45.3% | 1.92 ms | 0.08% | Fair | | 🇩🇪 Berlin (BER) | **[187.5](/docs/network/latency/pairs/ber-tyo-rtt)** | 87.53 ms | 46.7% | 1.21 ms | 0% | Fair | | 🇩🇪 Frankfurt (FRA) | **[194.5](/docs/network/latency/pairs/fra-tyo-rtt)** | 91.61 ms | 47.1% | 1.29 ms | 0.12% | Fair | | 🇳🇱 Amsterdam (AMS) | **[198.2](/docs/network/latency/pairs/ams-tyo-rtt)** | 91.18 ms | 46% | 1.27 ms | 0.14% | Fair | | 🇬🇧 London (LON) | **[202.9](/docs/network/latency/pairs/lon-tyo-rtt)** | 93.84 ms | 46.2% | 1.61 ms | 0% | Fair | | 🇫🇷 Marseille (MRS) | **[203.8](/docs/network/latency/pairs/mrs-tyo-rtt)** | 99.03 ms | 48.6% | 2.03 ms | 0.18% | Fair | | 🇫🇷 Paris (PAR) | **[204.7](/docs/network/latency/pairs/par-tyo-rtt)** | 95.34 ms | 46.6% | 2.17 ms | 0% | Fair | | 🇧🇷 São Paulo (GRU) | **[230.3](/docs/network/latency/pairs/gru-tyo-rtt)** | 181.46 ms | 78.8% | 2.42 ms | 0% | Fair | | 🇿🇦 Johannesburg (JNB) | **[360.5](/docs/network/latency/pairs/jnb-tyo-rtt)** | 132.57 ms | 36.8% | 2.79 ms | 0.06% | High | ## Asia Pacific Peers Tokyo is one of 6 Hats Network PoPs in **Asia Pacific**. The table lists outbound measurements within the region. | PoP | Code | Country | RTT (ms) | Tier | | --------------------------------------- | ---- | --------- | --------- | --------- | | [🇭🇰 Hong Kong (HKG)](./hkg-hong-kong) | HKG | Hong Kong | **44.8** | Excellent | | [🇦🇺 Melbourne (MEL)](./mel-melbourne) | MEL | Australia | **113.9** | Good | | [🇸🇬 Singapore (SIN)](./sin-singapore) | SIN | Singapore | **69.4** | Excellent | | [🇦🇺 Sydney (SYD)](./syd-sydney) | SYD | Australia | **101.9** | Good | | [🇹🇼 Taipei (TPE)](./tpe-taipei) | TPE | Taiwan | **32.3** | Excellent | ## Questions ### What is the fastest route to Tokyo? The fastest measured route to Tokyo (TYO) is [Taipei (TPE) → Tokyo (TYO)](/docs/network/latency/pairs/tpe-tyo-rtt), averaging **31.9 ms** RTT (Excellent). ### What is the fastest route from Tokyo? The fastest measured route from Tokyo (TYO) is [Tokyo (TYO) → Taipei (TPE)](/docs/network/latency/pairs/tyo-tpe-rtt), averaging **32.3 ms** RTT (Excellent). ### How well connected is Tokyo to the Hats Network backbone? Tokyo (TYO) maintains measured routes to all 19 other PoPs across 5 continents. The slowest route, [Tokyo (TYO) → Johannesburg (JNB)](/docs/network/latency/pairs/tyo-jnb-rtt), averages **359.2 ms** RTT. *** *Round published August 16, 2026. [All 20 PoPs](/docs/network/latency) · [Dataset and licence](/opendata/latency/).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/tyo-tokyo). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Africa Ping & RTT Latency | AS203314 Johannesburg is the dataset's only African PoP. Its comparisons therefore use destinations in other regions. Measurement round: `2026-08-16T04:07:28Z`. ## Africa to Other Regions The five lowest recorded RTTs to each destination region. ### To Europe | Route | RTT (ms) | Tier | | ------------------------------------------------------------------------------------------------------------------------------- | --------- | ---- | | 🇿🇦 [Johannesburg (JNB)](/docs/network/latency/jnb-johannesburg) → 🇬🇧 [London (LON)](/docs/network/latency/lon-london) | **158** | Fair | | 🇿🇦 [Johannesburg (JNB)](/docs/network/latency/jnb-johannesburg) → 🇳🇱 [Amsterdam (AMS)](/docs/network/latency/ams-amsterdam) | **163.2** | Fair | | 🇿🇦 [Johannesburg (JNB)](/docs/network/latency/jnb-johannesburg) → 🇫🇷 [Paris (PAR)](/docs/network/latency/par-paris) | **164.4** | Fair | | 🇿🇦 [Johannesburg (JNB)](/docs/network/latency/jnb-johannesburg) → 🇩🇪 [Frankfurt (FRA)](/docs/network/latency/fra-frankfurt) | **166** | Fair | | 🇿🇦 [Johannesburg (JNB)](/docs/network/latency/jnb-johannesburg) → 🇫🇷 [Marseille (MRS)](/docs/network/latency/mrs-marseille) | **172** | Fair | ### To North America | Route | RTT (ms) | Tier | | ----------------------------------------------------------------------------------------------------------------------------------- | --------- | ---- | | 🇿🇦 [Johannesburg (JNB)](/docs/network/latency/jnb-johannesburg) → 🇺🇸 [New York (NYC)](/docs/network/latency/nyc-new-york) | **221.8** | Fair | | 🇿🇦 [Johannesburg (JNB)](/docs/network/latency/jnb-johannesburg) → 🇺🇸 [Ashburn (IAD)](/docs/network/latency/iad-ashburn) | **229** | Fair | | 🇿🇦 [Johannesburg (JNB)](/docs/network/latency/jnb-johannesburg) → 🇺🇸 [Miami (MIA)](/docs/network/latency/mia-miami) | **259.9** | High | | 🇿🇦 [Johannesburg (JNB)](/docs/network/latency/jnb-johannesburg) → 🇺🇸 [Seattle (SEA)](/docs/network/latency/sea-seattle) | **281.2** | High | | 🇿🇦 [Johannesburg (JNB)](/docs/network/latency/jnb-johannesburg) → 🇺🇸 [Los Angeles (LAX)](/docs/network/latency/lax-los-angeles) | **285.5** | High | ### To South America | Route | RTT (ms) | Tier | | ------------------------------------------------------------------------------------------------------------------------------- | --------- | ---- | | 🇿🇦 [Johannesburg (JNB)](/docs/network/latency/jnb-johannesburg) → 🇧🇷 [São Paulo (GRU)](/docs/network/latency/gru-sao-paulo) | **328.7** | High | ### To Asia Pacific | Route | RTT (ms) | Tier | | ------------------------------------------------------------------------------------------------------------------------------- | --------- | ---- | | 🇿🇦 [Johannesburg (JNB)](/docs/network/latency/jnb-johannesburg) → 🇸🇬 [Singapore (SIN)](/docs/network/latency/sin-singapore) | **311.6** | High | | 🇿🇦 [Johannesburg (JNB)](/docs/network/latency/jnb-johannesburg) → 🇭🇰 [Hong Kong (HKG)](/docs/network/latency/hkg-hong-kong) | **318** | High | | 🇿🇦 [Johannesburg (JNB)](/docs/network/latency/jnb-johannesburg) → 🇹🇼 [Taipei (TPE)](/docs/network/latency/tpe-taipei) | **329.8** | High | | 🇿🇦 [Johannesburg (JNB)](/docs/network/latency/jnb-johannesburg) → 🇯🇵 [Tokyo (TYO)](/docs/network/latency/tyo-tokyo) | **360.5** | High | | 🇿🇦 [Johannesburg (JNB)](/docs/network/latency/jnb-johannesburg) → 🇦🇺 [Melbourne (MEL)](/docs/network/latency/mel-melbourne) | **406.2** | High | ## PoPs in this Region | PoP | City | Country | | ------------------------------------------------- | ----------------- | ------------ | | **[JNB](/docs/network/latency/jnb-johannesburg)** | 🇿🇦 Johannesburg | South Africa | *** *Round published August 16, 2026. See the [full global latency matrix](/docs/network/latency) or browse by [city pair](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/regions/africa). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Asia Pacific to South America Ping & RTT | AS203314 Asia-Pacific origins are compared with São Paulo. Long distances contribute to propagation delay, but the measurements do not reveal the transit path. Measurement round: `2026-08-16T04:07:28Z`. ## Quick Stats | Metric | Value | | ------------------- | ------------------------------------------------------------------------------------------ | | Fastest Route | **[Tokyo (TYO) → São Paulo (GRU)](/docs/network/latency/pairs/tyo-gru-rtt)**: 230.5 ms | | Slowest Route | **[Singapore (SIN) → São Paulo (GRU)](/docs/network/latency/pairs/sin-gru-rtt)**: 311.4 ms | | Average RTT | **284.1 ms** | | City Pairs Measured | 6 | ## Asia Pacific to South America City-Pair RTT | Route | RTT (ms) | Tier | | ------------------------------------------------------------------------------------------------------------------------- | ---------------------------------------------------- | ---- | | 🇯🇵 [Tokyo (TYO)](/docs/network/latency/tyo-tokyo) → 🇧🇷 [São Paulo (GRU)](/docs/network/latency/gru-sao-paulo) | **[230.5](/docs/network/latency/pairs/tyo-gru-rtt)** | Fair | | 🇹🇼 [Taipei (TPE)](/docs/network/latency/tpe-taipei) → 🇧🇷 [São Paulo (GRU)](/docs/network/latency/gru-sao-paulo) | **[273](/docs/network/latency/pairs/tpe-gru-rtt)** | High | | 🇭🇰 [Hong Kong (HKG)](/docs/network/latency/hkg-hong-kong) → 🇧🇷 [São Paulo (GRU)](/docs/network/latency/gru-sao-paulo) | **[287.6](/docs/network/latency/pairs/hkg-gru-rtt)** | High | | 🇦🇺 [Sydney (SYD)](/docs/network/latency/syd-sydney) → 🇧🇷 [São Paulo (GRU)](/docs/network/latency/gru-sao-paulo) | **[296.7](/docs/network/latency/pairs/syd-gru-rtt)** | High | | 🇦🇺 [Melbourne (MEL)](/docs/network/latency/mel-melbourne) → 🇧🇷 [São Paulo (GRU)](/docs/network/latency/gru-sao-paulo) | **[305.3](/docs/network/latency/pairs/mel-gru-rtt)** | High | | 🇸🇬 [Singapore (SIN)](/docs/network/latency/sin-singapore) → 🇧🇷 [São Paulo (GRU)](/docs/network/latency/gru-sao-paulo) | **[311.4](/docs/network/latency/pairs/sin-gru-rtt)** | High | ## Related Regions * [Asia Pacific backbone latency](./asia-pacific) * [South America backbone latency](./south-america) *** *Round published August 16, 2026. See the [full global latency matrix](/docs/network/latency).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/regions/asia-pacific-to-south-america). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Asia Pacific Ping & RTT Latency | AS203314 Asia-Pacific measurements cover Hong Kong, Singapore, Taipei, Tokyo, Sydney and Melbourne. Distances within this group range from regional links to crossings between Asia and Australia. Measurement round: `2026-08-16T04:07:28Z`. ## Intra-Asia Pacific RTT Matrix Round-trip times in milliseconds between PoPs within the region. | Route | HKG | MEL | SIN | SYD | TPE | TYO | | ----------------------------------------------------------- | ----- | ----- | ---- | ----- | ----- | ----- | | 🇭🇰 [Hong Kong (HKG)](/docs/network/latency/hkg-hong-kong) | N/A | 138.4 | 31.5 | 137.2 | 14.7 | 44.9 | | 🇦🇺 [Melbourne (MEL)](/docs/network/latency/mel-melbourne) | 138.5 | N/A | 88.4 | 9.8 | 143.2 | 112 | | 🇸🇬 [Singapore (SIN)](/docs/network/latency/sin-singapore) | 30.8 | 88.6 | N/A | 94.5 | 45.8 | 68.9 | | 🇦🇺 [Sydney (SYD)](/docs/network/latency/syd-sydney) | 135.7 | 9.8 | 94.5 | N/A | 131.7 | 101.3 | | 🇹🇼 [Taipei (TPE)](/docs/network/latency/tpe-taipei) | 15.5 | 142 | 44.3 | 133.2 | N/A | 31.9 | | 🇯🇵 [Tokyo (TYO)](/docs/network/latency/tyo-tokyo) | 44.8 | 113.9 | 69.4 | 101.9 | 32.3 | N/A | ## Asia Pacific to Other Regions The five lowest recorded RTTs to each destination region. ### To Europe | Route | RTT (ms) | Tier | | ------------------------------------------------------------------------------------------------------------------------- | --------- | ---- | | 🇭🇰 [Hong Kong (HKG)](/docs/network/latency/hkg-hong-kong) → 🇷🇺 [Moscow (MOW)](/docs/network/latency/mow-moscow) | **118.2** | Good | | 🇹🇼 [Taipei (TPE)](/docs/network/latency/tpe-taipei) → 🇷🇺 [Moscow (MOW)](/docs/network/latency/mow-moscow) | **132.7** | Good | | 🇸🇬 [Singapore (SIN)](/docs/network/latency/sin-singapore) → 🇫🇷 [Marseille (MRS)](/docs/network/latency/mrs-marseille) | **140.6** | Good | | 🇭🇰 [Hong Kong (HKG)](/docs/network/latency/hkg-hong-kong) → 🇩🇪 [Berlin (BER)](/docs/network/latency/ber-berlin) | **145.1** | Good | | 🇸🇬 [Singapore (SIN)](/docs/network/latency/sin-singapore) → 🇷🇺 [Moscow (MOW)](/docs/network/latency/mow-moscow) | **147.7** | Good | ### To North America | Route | RTT (ms) | Tier | | ----------------------------------------------------------------------------------------------------------------------- | --------- | ---- | | 🇯🇵 [Tokyo (TYO)](/docs/network/latency/tyo-tokyo) → 🇺🇸 [Seattle (SEA)](/docs/network/latency/sea-seattle) | **84.8** | Good | | 🇯🇵 [Tokyo (TYO)](/docs/network/latency/tyo-tokyo) → 🇺🇸 [Los Angeles (LAX)](/docs/network/latency/lax-los-angeles) | **101.2** | Good | | 🇹🇼 [Taipei (TPE)](/docs/network/latency/tpe-taipei) → 🇺🇸 [Seattle (SEA)](/docs/network/latency/sea-seattle) | **115.8** | Good | | 🇭🇰 [Hong Kong (HKG)](/docs/network/latency/hkg-hong-kong) → 🇺🇸 [Seattle (SEA)](/docs/network/latency/sea-seattle) | **131.6** | Good | | 🇹🇼 [Taipei (TPE)](/docs/network/latency/tpe-taipei) → 🇺🇸 [Los Angeles (LAX)](/docs/network/latency/lax-los-angeles) | **132** | Good | ### To South America | Route | RTT (ms) | Tier | | ------------------------------------------------------------------------------------------------------------------------- | --------- | ---- | | 🇯🇵 [Tokyo (TYO)](/docs/network/latency/tyo-tokyo) → 🇧🇷 [São Paulo (GRU)](/docs/network/latency/gru-sao-paulo) | **230.5** | Fair | | 🇹🇼 [Taipei (TPE)](/docs/network/latency/tpe-taipei) → 🇧🇷 [São Paulo (GRU)](/docs/network/latency/gru-sao-paulo) | **273** | High | | 🇭🇰 [Hong Kong (HKG)](/docs/network/latency/hkg-hong-kong) → 🇧🇷 [São Paulo (GRU)](/docs/network/latency/gru-sao-paulo) | **287.6** | High | | 🇦🇺 [Sydney (SYD)](/docs/network/latency/syd-sydney) → 🇧🇷 [São Paulo (GRU)](/docs/network/latency/gru-sao-paulo) | **296.7** | High | | 🇦🇺 [Melbourne (MEL)](/docs/network/latency/mel-melbourne) → 🇧🇷 [São Paulo (GRU)](/docs/network/latency/gru-sao-paulo) | **305.3** | High | ### To Africa | Route | RTT (ms) | Tier | | ------------------------------------------------------------------------------------------------------------------------------- | --------- | ---- | | 🇸🇬 [Singapore (SIN)](/docs/network/latency/sin-singapore) → 🇿🇦 [Johannesburg (JNB)](/docs/network/latency/jnb-johannesburg) | **312.6** | High | | 🇭🇰 [Hong Kong (HKG)](/docs/network/latency/hkg-hong-kong) → 🇿🇦 [Johannesburg (JNB)](/docs/network/latency/jnb-johannesburg) | **316.7** | High | | 🇹🇼 [Taipei (TPE)](/docs/network/latency/tpe-taipei) → 🇿🇦 [Johannesburg (JNB)](/docs/network/latency/jnb-johannesburg) | **331** | High | | 🇯🇵 [Tokyo (TYO)](/docs/network/latency/tyo-tokyo) → 🇿🇦 [Johannesburg (JNB)](/docs/network/latency/jnb-johannesburg) | **359.2** | High | | 🇦🇺 [Melbourne (MEL)](/docs/network/latency/mel-melbourne) → 🇿🇦 [Johannesburg (JNB)](/docs/network/latency/jnb-johannesburg) | **405.5** | High | ## PoPs in this Region | PoP | City | Country | | ---------------------------------------------- | -------------- | --------- | | **[HKG](/docs/network/latency/hkg-hong-kong)** | 🇭🇰 Hong Kong | Hong Kong | | **[MEL](/docs/network/latency/mel-melbourne)** | 🇦🇺 Melbourne | Australia | | **[SIN](/docs/network/latency/sin-singapore)** | 🇸🇬 Singapore | Singapore | | **[SYD](/docs/network/latency/syd-sydney)** | 🇦🇺 Sydney | Australia | | **[TPE](/docs/network/latency/tpe-taipei)** | 🇹🇼 Taipei | Taiwan | | **[TYO](/docs/network/latency/tyo-tokyo)** | 🇯🇵 Tokyo | Japan | *** *Round published August 16, 2026. See the [full global latency matrix](/docs/network/latency) or browse by [city pair](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/regions/asia-pacific). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Europe to Asia Pacific Ping & RTT | AS203314 European origins are compared with destinations in East Asia, Southeast Asia and Australia. The table records endpoint RTT, not the path across Eurasia or the oceans. Measurement round: `2026-08-16T04:07:28Z`. ## Quick Stats | Metric | Value | | ------------------- | --------------------------------------------------------------------------------------- | | Fastest Route | **[Moscow (MOW) → Hong Kong (HKG)](/docs/network/latency/pairs/mow-hkg-rtt)**: 118.6 ms | | Slowest Route | **[Berlin (BER) → Sydney (SYD)](/docs/network/latency/pairs/ber-syd-rtt)**: 262.1 ms | | Average RTT | **193.2 ms** | | City Pairs Measured | 42 | ## Europe to Asia Pacific City-Pair RTT | Route | RTT (ms) | Tier | | ------------------------------------------------------------------------------------------------------------------------- | ---------------------------------------------------- | ---- | | 🇷🇺 [Moscow (MOW)](/docs/network/latency/mow-moscow) → 🇭🇰 [Hong Kong (HKG)](/docs/network/latency/hkg-hong-kong) | **[118.6](/docs/network/latency/pairs/mow-hkg-rtt)** | Good | | 🇷🇺 [Moscow (MOW)](/docs/network/latency/mow-moscow) → 🇹🇼 [Taipei (TPE)](/docs/network/latency/tpe-taipei) | **[132.1](/docs/network/latency/pairs/mow-tpe-rtt)** | Good | | 🇫🇷 [Marseille (MRS)](/docs/network/latency/mrs-marseille) → 🇸🇬 [Singapore (SIN)](/docs/network/latency/sin-singapore) | **[139.6](/docs/network/latency/pairs/mrs-sin-rtt)** | Good | | 🇩🇪 [Berlin (BER)](/docs/network/latency/ber-berlin) → 🇭🇰 [Hong Kong (HKG)](/docs/network/latency/hkg-hong-kong) | **[144.9](/docs/network/latency/pairs/ber-hkg-rtt)** | Good | | 🇷🇺 [Moscow (MOW)](/docs/network/latency/mow-moscow) → 🇸🇬 [Singapore (SIN)](/docs/network/latency/sin-singapore) | **[147.3](/docs/network/latency/pairs/mow-sin-rtt)** | Good | | 🇫🇷 [Paris (PAR)](/docs/network/latency/par-paris) → 🇸🇬 [Singapore (SIN)](/docs/network/latency/sin-singapore) | **[150.3](/docs/network/latency/pairs/par-sin-rtt)** | Fair | | 🇩🇪 [Frankfurt (FRA)](/docs/network/latency/fra-frankfurt) → 🇸🇬 [Singapore (SIN)](/docs/network/latency/sin-singapore) | **[151.4](/docs/network/latency/pairs/fra-sin-rtt)** | Fair | | 🇩🇪 [Frankfurt (FRA)](/docs/network/latency/fra-frankfurt) → 🇭🇰 [Hong Kong (HKG)](/docs/network/latency/hkg-hong-kong) | **[152](/docs/network/latency/pairs/fra-hkg-rtt)** | Fair | | 🇳🇱 [Amsterdam (AMS)](/docs/network/latency/ams-amsterdam) → 🇭🇰 [Hong Kong (HKG)](/docs/network/latency/hkg-hong-kong) | **[155.9](/docs/network/latency/pairs/ams-hkg-rtt)** | Fair | | 🇬🇧 [London (LON)](/docs/network/latency/lon-london) → 🇸🇬 [Singapore (SIN)](/docs/network/latency/sin-singapore) | **[156.5](/docs/network/latency/pairs/lon-sin-rtt)** | Fair | | 🇩🇪 [Berlin (BER)](/docs/network/latency/ber-berlin) → 🇹🇼 [Taipei (TPE)](/docs/network/latency/tpe-taipei) | **[156.8](/docs/network/latency/pairs/ber-tpe-rtt)** | Fair | | 🇳🇱 [Amsterdam (AMS)](/docs/network/latency/ams-amsterdam) → 🇸🇬 [Singapore (SIN)](/docs/network/latency/sin-singapore) | **[157.6](/docs/network/latency/pairs/ams-sin-rtt)** | Fair | | 🇩🇪 [Berlin (BER)](/docs/network/latency/ber-berlin) → 🇸🇬 [Singapore (SIN)](/docs/network/latency/sin-singapore) | **[158.3](/docs/network/latency/pairs/ber-sin-rtt)** | Fair | | 🇫🇷 [Paris (PAR)](/docs/network/latency/par-paris) → 🇭🇰 [Hong Kong (HKG)](/docs/network/latency/hkg-hong-kong) | **[160.1](/docs/network/latency/pairs/par-hkg-rtt)** | Fair | | 🇬🇧 [London (LON)](/docs/network/latency/lon-london) → 🇭🇰 [Hong Kong (HKG)](/docs/network/latency/hkg-hong-kong) | **[160.6](/docs/network/latency/pairs/lon-hkg-rtt)** | Fair | | 🇷🇺 [Moscow (MOW)](/docs/network/latency/mow-moscow) → 🇯🇵 [Tokyo (TYO)](/docs/network/latency/tyo-tokyo) | **[162.2](/docs/network/latency/pairs/mow-tyo-rtt)** | Fair | | 🇩🇪 [Frankfurt (FRA)](/docs/network/latency/fra-frankfurt) → 🇹🇼 [Taipei (TPE)](/docs/network/latency/tpe-taipei) | **[163.8](/docs/network/latency/pairs/fra-tpe-rtt)** | Fair | | 🇫🇷 [Marseille (MRS)](/docs/network/latency/mrs-marseille) → 🇭🇰 [Hong Kong (HKG)](/docs/network/latency/hkg-hong-kong) | **[166.4](/docs/network/latency/pairs/mrs-hkg-rtt)** | Fair | | 🇳🇱 [Amsterdam (AMS)](/docs/network/latency/ams-amsterdam) → 🇹🇼 [Taipei (TPE)](/docs/network/latency/tpe-taipei) | **[168.7](/docs/network/latency/pairs/ams-tpe-rtt)** | Fair | | 🇫🇷 [Paris (PAR)](/docs/network/latency/par-paris) → 🇹🇼 [Taipei (TPE)](/docs/network/latency/tpe-taipei) | **[171.9](/docs/network/latency/pairs/par-tpe-rtt)** | Fair | | 🇬🇧 [London (LON)](/docs/network/latency/lon-london) → 🇹🇼 [Taipei (TPE)](/docs/network/latency/tpe-taipei) | **[173.8](/docs/network/latency/pairs/lon-tpe-rtt)** | Fair | | 🇫🇷 [Marseille (MRS)](/docs/network/latency/mrs-marseille) → 🇹🇼 [Taipei (TPE)](/docs/network/latency/tpe-taipei) | **[180.9](/docs/network/latency/pairs/mrs-tpe-rtt)** | Fair | | 🇩🇪 [Berlin (BER)](/docs/network/latency/ber-berlin) → 🇯🇵 [Tokyo (TYO)](/docs/network/latency/tyo-tokyo) | **[187.5](/docs/network/latency/pairs/ber-tyo-rtt)** | Fair | | 🇩🇪 [Frankfurt (FRA)](/docs/network/latency/fra-frankfurt) → 🇯🇵 [Tokyo (TYO)](/docs/network/latency/tyo-tokyo) | **[194.5](/docs/network/latency/pairs/fra-tyo-rtt)** | Fair | | 🇳🇱 [Amsterdam (AMS)](/docs/network/latency/ams-amsterdam) → 🇯🇵 [Tokyo (TYO)](/docs/network/latency/tyo-tokyo) | **[198.2](/docs/network/latency/pairs/ams-tyo-rtt)** | Fair | | 🇬🇧 [London (LON)](/docs/network/latency/lon-london) → 🇯🇵 [Tokyo (TYO)](/docs/network/latency/tyo-tokyo) | **[202.9](/docs/network/latency/pairs/lon-tyo-rtt)** | Fair | | 🇫🇷 [Marseille (MRS)](/docs/network/latency/mrs-marseille) → 🇯🇵 [Tokyo (TYO)](/docs/network/latency/tyo-tokyo) | **[203.8](/docs/network/latency/pairs/mrs-tyo-rtt)** | Fair | | 🇫🇷 [Paris (PAR)](/docs/network/latency/par-paris) → 🇯🇵 [Tokyo (TYO)](/docs/network/latency/tyo-tokyo) | **[204.7](/docs/network/latency/pairs/par-tyo-rtt)** | Fair | | 🇫🇷 [Marseille (MRS)](/docs/network/latency/mrs-marseille) → 🇦🇺 [Melbourne (MEL)](/docs/network/latency/mel-melbourne) | **[234.2](/docs/network/latency/pairs/mrs-mel-rtt)** | Fair | | 🇷🇺 [Moscow (MOW)](/docs/network/latency/mow-moscow) → 🇦🇺 [Melbourne (MEL)](/docs/network/latency/mel-melbourne) | **[235.9](/docs/network/latency/pairs/mow-mel-rtt)** | Fair | | 🇫🇷 [Marseille (MRS)](/docs/network/latency/mrs-marseille) → 🇦🇺 [Sydney (SYD)](/docs/network/latency/syd-sydney) | **[240.6](/docs/network/latency/pairs/mrs-syd-rtt)** | Fair | | 🇷🇺 [Moscow (MOW)](/docs/network/latency/mow-moscow) → 🇦🇺 [Sydney (SYD)](/docs/network/latency/syd-sydney) | **[241.8](/docs/network/latency/pairs/mow-syd-rtt)** | Fair | | 🇫🇷 [Paris (PAR)](/docs/network/latency/par-paris) → 🇦🇺 [Melbourne (MEL)](/docs/network/latency/mel-melbourne) | **[246.4](/docs/network/latency/pairs/par-mel-rtt)** | Fair | | 🇬🇧 [London (LON)](/docs/network/latency/lon-london) → 🇦🇺 [Melbourne (MEL)](/docs/network/latency/mel-melbourne) | **[249.6](/docs/network/latency/pairs/lon-mel-rtt)** | Fair | | 🇫🇷 [Paris (PAR)](/docs/network/latency/par-paris) → 🇦🇺 [Sydney (SYD)](/docs/network/latency/syd-sydney) | **[249.7](/docs/network/latency/pairs/par-syd-rtt)** | Fair | | 🇩🇪 [Frankfurt (FRA)](/docs/network/latency/fra-frankfurt) → 🇦🇺 [Melbourne (MEL)](/docs/network/latency/mel-melbourne) | **[250.1](/docs/network/latency/pairs/fra-mel-rtt)** | High | | 🇳🇱 [Amsterdam (AMS)](/docs/network/latency/ams-amsterdam) → 🇦🇺 [Melbourne (MEL)](/docs/network/latency/mel-melbourne) | **[252.5](/docs/network/latency/pairs/ams-mel-rtt)** | High | | 🇩🇪 [Frankfurt (FRA)](/docs/network/latency/fra-frankfurt) → 🇦🇺 [Sydney (SYD)](/docs/network/latency/syd-sydney) | **[256](/docs/network/latency/pairs/fra-syd-rtt)** | High | | 🇩🇪 [Berlin (BER)](/docs/network/latency/ber-berlin) → 🇦🇺 [Melbourne (MEL)](/docs/network/latency/mel-melbourne) | **[256.2](/docs/network/latency/pairs/ber-mel-rtt)** | High | | 🇳🇱 [Amsterdam (AMS)](/docs/network/latency/ams-amsterdam) → 🇦🇺 [Sydney (SYD)](/docs/network/latency/syd-sydney) | **[258](/docs/network/latency/pairs/ams-syd-rtt)** | High | | 🇬🇧 [London (LON)](/docs/network/latency/lon-london) → 🇦🇺 [Sydney (SYD)](/docs/network/latency/syd-sydney) | **[258.1](/docs/network/latency/pairs/lon-syd-rtt)** | High | | 🇩🇪 [Berlin (BER)](/docs/network/latency/ber-berlin) → 🇦🇺 [Sydney (SYD)](/docs/network/latency/syd-sydney) | **[262.1](/docs/network/latency/pairs/ber-syd-rtt)** | High | ## Related Regions * [Europe backbone latency](./europe) * [Asia Pacific backbone latency](./asia-pacific) *** *Round published August 16, 2026. See the [full global latency matrix](/docs/network/latency).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/regions/europe-to-asia-pacific). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Europe to North America Ping & RTT | AS203314 These measurements run from European PoPs to US destinations. RTT alone does not identify the transatlantic cable or the intermediate networks. Measurement round: `2026-08-16T04:07:28Z`. ## Quick Stats | Metric | Value | | ------------------- | ----------------------------------------------------------------------------------------- | | Fastest Route | **[London (LON) → New York (NYC)](/docs/network/latency/pairs/lon-nyc-rtt)**: 63.8 ms | | Slowest Route | **[Moscow (MOW) → Los Angeles (LAX)](/docs/network/latency/pairs/mow-lax-rtt)**: 177.7 ms | | Average RTT | **111.4 ms** | | City Pairs Measured | 35 | ## Europe to North America City-Pair RTT | Route | RTT (ms) | Tier | | ----------------------------------------------------------------------------------------------------------------------------- | ---------------------------------------------------- | --------- | | 🇬🇧 [London (LON)](/docs/network/latency/lon-london) → 🇺🇸 [New York (NYC)](/docs/network/latency/nyc-new-york) | **[63.8](/docs/network/latency/pairs/lon-nyc-rtt)** | Excellent | | 🇳🇱 [Amsterdam (AMS)](/docs/network/latency/ams-amsterdam) → 🇺🇸 [New York (NYC)](/docs/network/latency/nyc-new-york) | **[68.6](/docs/network/latency/pairs/ams-nyc-rtt)** | Excellent | | 🇫🇷 [Paris (PAR)](/docs/network/latency/par-paris) → 🇺🇸 [New York (NYC)](/docs/network/latency/nyc-new-york) | **[68.9](/docs/network/latency/pairs/par-nyc-rtt)** | Excellent | | 🇬🇧 [London (LON)](/docs/network/latency/lon-london) → 🇺🇸 [Ashburn (IAD)](/docs/network/latency/iad-ashburn) | **[71](/docs/network/latency/pairs/lon-iad-rtt)** | Excellent | | 🇩🇪 [Frankfurt (FRA)](/docs/network/latency/fra-frankfurt) → 🇺🇸 [New York (NYC)](/docs/network/latency/nyc-new-york) | **[74.4](/docs/network/latency/pairs/fra-nyc-rtt)** | Excellent | | 🇫🇷 [Paris (PAR)](/docs/network/latency/par-paris) → 🇺🇸 [Ashburn (IAD)](/docs/network/latency/iad-ashburn) | **[75.5](/docs/network/latency/pairs/par-iad-rtt)** | Excellent | | 🇳🇱 [Amsterdam (AMS)](/docs/network/latency/ams-amsterdam) → 🇺🇸 [Ashburn (IAD)](/docs/network/latency/iad-ashburn) | **[75.7](/docs/network/latency/pairs/ams-iad-rtt)** | Excellent | | 🇩🇪 [Berlin (BER)](/docs/network/latency/ber-berlin) → 🇺🇸 [New York (NYC)](/docs/network/latency/nyc-new-york) | **[79.6](/docs/network/latency/pairs/ber-nyc-rtt)** | Excellent | | 🇫🇷 [Marseille (MRS)](/docs/network/latency/mrs-marseille) → 🇺🇸 [New York (NYC)](/docs/network/latency/nyc-new-york) | **[79.8](/docs/network/latency/pairs/mrs-nyc-rtt)** | Excellent | | 🇩🇪 [Frankfurt (FRA)](/docs/network/latency/fra-frankfurt) → 🇺🇸 [Ashburn (IAD)](/docs/network/latency/iad-ashburn) | **[80.7](/docs/network/latency/pairs/fra-iad-rtt)** | Good | | 🇩🇪 [Berlin (BER)](/docs/network/latency/ber-berlin) → 🇺🇸 [Ashburn (IAD)](/docs/network/latency/iad-ashburn) | **[85.7](/docs/network/latency/pairs/ber-iad-rtt)** | Good | | 🇫🇷 [Marseille (MRS)](/docs/network/latency/mrs-marseille) → 🇺🇸 [Ashburn (IAD)](/docs/network/latency/iad-ashburn) | **[85.7](/docs/network/latency/pairs/mrs-iad-rtt)** | Good | | 🇬🇧 [London (LON)](/docs/network/latency/lon-london) → 🇺🇸 [Miami (MIA)](/docs/network/latency/mia-miami) | **[101.9](/docs/network/latency/pairs/lon-mia-rtt)** | Good | | 🇳🇱 [Amsterdam (AMS)](/docs/network/latency/ams-amsterdam) → 🇺🇸 [Miami (MIA)](/docs/network/latency/mia-miami) | **[106](/docs/network/latency/pairs/ams-mia-rtt)** | Good | | 🇫🇷 [Paris (PAR)](/docs/network/latency/par-paris) → 🇺🇸 [Miami (MIA)](/docs/network/latency/mia-miami) | **[106.8](/docs/network/latency/pairs/par-mia-rtt)** | Good | | 🇷🇺 [Moscow (MOW)](/docs/network/latency/mow-moscow) → 🇺🇸 [New York (NYC)](/docs/network/latency/nyc-new-york) | **[107.5](/docs/network/latency/pairs/mow-nyc-rtt)** | Good | | 🇩🇪 [Frankfurt (FRA)](/docs/network/latency/fra-frankfurt) → 🇺🇸 [Miami (MIA)](/docs/network/latency/mia-miami) | **[112](/docs/network/latency/pairs/fra-mia-rtt)** | Good | | 🇫🇷 [Marseille (MRS)](/docs/network/latency/mrs-marseille) → 🇺🇸 [Miami (MIA)](/docs/network/latency/mia-miami) | **[112.4](/docs/network/latency/pairs/mrs-mia-rtt)** | Good | | 🇷🇺 [Moscow (MOW)](/docs/network/latency/mow-moscow) → 🇺🇸 [Ashburn (IAD)](/docs/network/latency/iad-ashburn) | **[113.9](/docs/network/latency/pairs/mow-iad-rtt)** | Good | | 🇩🇪 [Berlin (BER)](/docs/network/latency/ber-berlin) → 🇺🇸 [Miami (MIA)](/docs/network/latency/mia-miami) | **[118.3](/docs/network/latency/pairs/ber-mia-rtt)** | Good | | 🇬🇧 [London (LON)](/docs/network/latency/lon-london) → 🇺🇸 [Seattle (SEA)](/docs/network/latency/sea-seattle) | **[123.2](/docs/network/latency/pairs/lon-sea-rtt)** | Good | | 🇬🇧 [London (LON)](/docs/network/latency/lon-london) → 🇺🇸 [Los Angeles (LAX)](/docs/network/latency/lax-los-angeles) | **[127.5](/docs/network/latency/pairs/lon-lax-rtt)** | Good | | 🇫🇷 [Paris (PAR)](/docs/network/latency/par-paris) → 🇺🇸 [Seattle (SEA)](/docs/network/latency/sea-seattle) | **[128](/docs/network/latency/pairs/par-sea-rtt)** | Good | | 🇳🇱 [Amsterdam (AMS)](/docs/network/latency/ams-amsterdam) → 🇺🇸 [Seattle (SEA)](/docs/network/latency/sea-seattle) | **[130.2](/docs/network/latency/pairs/ams-sea-rtt)** | Good | | 🇳🇱 [Amsterdam (AMS)](/docs/network/latency/ams-amsterdam) → 🇺🇸 [Los Angeles (LAX)](/docs/network/latency/lax-los-angeles) | **[133.7](/docs/network/latency/pairs/ams-lax-rtt)** | Good | | 🇩🇪 [Frankfurt (FRA)](/docs/network/latency/fra-frankfurt) → 🇺🇸 [Seattle (SEA)](/docs/network/latency/sea-seattle) | **[135.2](/docs/network/latency/pairs/fra-sea-rtt)** | Good | | 🇫🇷 [Paris (PAR)](/docs/network/latency/par-paris) → 🇺🇸 [Los Angeles (LAX)](/docs/network/latency/lax-los-angeles) | **[136.5](/docs/network/latency/pairs/par-lax-rtt)** | Good | | 🇩🇪 [Berlin (BER)](/docs/network/latency/ber-berlin) → 🇺🇸 [Seattle (SEA)](/docs/network/latency/sea-seattle) | **[139.7](/docs/network/latency/pairs/ber-sea-rtt)** | Good | | 🇩🇪 [Berlin (BER)](/docs/network/latency/ber-berlin) → 🇺🇸 [Los Angeles (LAX)](/docs/network/latency/lax-los-angeles) | **[140.9](/docs/network/latency/pairs/ber-lax-rtt)** | Good | | 🇫🇷 [Marseille (MRS)](/docs/network/latency/mrs-marseille) → 🇺🇸 [Seattle (SEA)](/docs/network/latency/sea-seattle) | **[141.7](/docs/network/latency/pairs/mrs-sea-rtt)** | Good | | 🇩🇪 [Frankfurt (FRA)](/docs/network/latency/fra-frankfurt) → 🇺🇸 [Los Angeles (LAX)](/docs/network/latency/lax-los-angeles) | **[141.8](/docs/network/latency/pairs/fra-lax-rtt)** | Good | | 🇫🇷 [Marseille (MRS)](/docs/network/latency/mrs-marseille) → 🇺🇸 [Los Angeles (LAX)](/docs/network/latency/lax-los-angeles) | **[144.5](/docs/network/latency/pairs/mrs-lax-rtt)** | Good | | 🇷🇺 [Moscow (MOW)](/docs/network/latency/mow-moscow) → 🇺🇸 [Miami (MIA)](/docs/network/latency/mia-miami) | **[144.5](/docs/network/latency/pairs/mow-mia-rtt)** | Good | | 🇷🇺 [Moscow (MOW)](/docs/network/latency/mow-moscow) → 🇺🇸 [Seattle (SEA)](/docs/network/latency/sea-seattle) | **[164.8](/docs/network/latency/pairs/mow-sea-rtt)** | Fair | | 🇷🇺 [Moscow (MOW)](/docs/network/latency/mow-moscow) → 🇺🇸 [Los Angeles (LAX)](/docs/network/latency/lax-los-angeles) | **[177.7](/docs/network/latency/pairs/mow-lax-rtt)** | Fair | ## Related Regions * [Europe backbone latency](./europe) * [North America backbone latency](./north-america) *** *Round published August 16, 2026. See the [full global latency matrix](/docs/network/latency).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/regions/europe-to-north-america). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Europe to South America Ping & RTT | AS203314 These routes start in Europe and end at São Paulo. The measurements do not establish whether traffic crosses the South Atlantic directly or passes through another region. Measurement round: `2026-08-16T04:07:28Z`. ## Quick Stats | Metric | Value | | ------------------- | --------------------------------------------------------------------------------------- | | Fastest Route | **[London (LON) → São Paulo (GRU)](/docs/network/latency/pairs/lon-gru-rtt)**: 170.7 ms | | Slowest Route | **[Moscow (MOW) → São Paulo (GRU)](/docs/network/latency/pairs/mow-gru-rtt)**: 214.4 ms | | Average RTT | **184.4 ms** | | City Pairs Measured | 7 | ## Europe to South America City-Pair RTT | Route | RTT (ms) | Tier | | ------------------------------------------------------------------------------------------------------------------------- | ---------------------------------------------------- | ---- | | 🇬🇧 [London (LON)](/docs/network/latency/lon-london) → 🇧🇷 [São Paulo (GRU)](/docs/network/latency/gru-sao-paulo) | **[170.7](/docs/network/latency/pairs/lon-gru-rtt)** | Fair | | 🇳🇱 [Amsterdam (AMS)](/docs/network/latency/ams-amsterdam) → 🇧🇷 [São Paulo (GRU)](/docs/network/latency/gru-sao-paulo) | **[175.5](/docs/network/latency/pairs/ams-gru-rtt)** | Fair | | 🇫🇷 [Paris (PAR)](/docs/network/latency/par-paris) → 🇧🇷 [São Paulo (GRU)](/docs/network/latency/gru-sao-paulo) | **[175.8](/docs/network/latency/pairs/par-gru-rtt)** | Fair | | 🇩🇪 [Frankfurt (FRA)](/docs/network/latency/fra-frankfurt) → 🇧🇷 [São Paulo (GRU)](/docs/network/latency/gru-sao-paulo) | **[181.3](/docs/network/latency/pairs/fra-gru-rtt)** | Fair | | 🇩🇪 [Berlin (BER)](/docs/network/latency/ber-berlin) → 🇧🇷 [São Paulo (GRU)](/docs/network/latency/gru-sao-paulo) | **[186.5](/docs/network/latency/pairs/ber-gru-rtt)** | Fair | | 🇫🇷 [Marseille (MRS)](/docs/network/latency/mrs-marseille) → 🇧🇷 [São Paulo (GRU)](/docs/network/latency/gru-sao-paulo) | **[186.7](/docs/network/latency/pairs/mrs-gru-rtt)** | Fair | | 🇷🇺 [Moscow (MOW)](/docs/network/latency/mow-moscow) → 🇧🇷 [São Paulo (GRU)](/docs/network/latency/gru-sao-paulo) | **[214.4](/docs/network/latency/pairs/mow-gru-rtt)** | Fair | ## Related Regions * [Europe backbone latency](./europe) * [South America backbone latency](./south-america) *** *Round published August 16, 2026. See the [full global latency matrix](/docs/network/latency).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/regions/europe-to-south-america). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Europe Ping & RTT Latency | AS203314 European measurements cover Amsterdam, Berlin, Frankfurt, London, Marseille, Moscow and Paris. The regional matrix keeps each direction separate. Measurement round: `2026-08-16T04:07:28Z`. ## Intra-Europe RTT Matrix Round-trip times in milliseconds between PoPs within the region. | Route | AMS | BER | FRA | LON | MRS | MOW | PAR | | ----------------------------------------------------------- | ---- | ---- | ---- | ---- | ---- | ---- | ---- | | 🇳🇱 [Amsterdam (AMS)](/docs/network/latency/ams-amsterdam) | N/A | 7.2 | 5.9 | 5.2 | 19.6 | 38.1 | 7 | | 🇩🇪 [Berlin (BER)](/docs/network/latency/ber-berlin) | 7.2 | N/A | 6.1 | 15.2 | 20.2 | 28.6 | 15.6 | | 🇩🇪 [Frankfurt (FRA)](/docs/network/latency/fra-frankfurt) | 5.9 | 6.1 | N/A | 12.9 | 16.4 | 34.7 | 7.6 | | 🇬🇧 [London (LON)](/docs/network/latency/lon-london) | 5.2 | 16.2 | 13.6 | N/A | 17.9 | 43.5 | 6.4 | | 🇫🇷 [Marseille (MRS)](/docs/network/latency/mrs-marseille) | 20.3 | 20.6 | 16 | 18.9 | N/A | 49.6 | 8.9 | | 🇷🇺 [Moscow (MOW)](/docs/network/latency/mow-moscow) | 37.6 | 27.5 | 35.9 | 42.4 | 49.9 | N/A | 44 | | 🇫🇷 [Paris (PAR)](/docs/network/latency/par-paris) | 7 | 15.5 | 7.6 | 6.4 | 8.9 | 44.7 | N/A | ## Europe to Other Regions The five lowest recorded RTTs to each destination region. ### To North America | Route | RTT (ms) | Tier | | ----------------------------------------------------------------------------------------------------------------------- | -------- | --------- | | 🇬🇧 [London (LON)](/docs/network/latency/lon-london) → 🇺🇸 [New York (NYC)](/docs/network/latency/nyc-new-york) | **63.8** | Excellent | | 🇳🇱 [Amsterdam (AMS)](/docs/network/latency/ams-amsterdam) → 🇺🇸 [New York (NYC)](/docs/network/latency/nyc-new-york) | **68.6** | Excellent | | 🇫🇷 [Paris (PAR)](/docs/network/latency/par-paris) → 🇺🇸 [New York (NYC)](/docs/network/latency/nyc-new-york) | **68.9** | Excellent | | 🇬🇧 [London (LON)](/docs/network/latency/lon-london) → 🇺🇸 [Ashburn (IAD)](/docs/network/latency/iad-ashburn) | **71** | Excellent | | 🇩🇪 [Frankfurt (FRA)](/docs/network/latency/fra-frankfurt) → 🇺🇸 [New York (NYC)](/docs/network/latency/nyc-new-york) | **74.4** | Excellent | ### To South America | Route | RTT (ms) | Tier | | ------------------------------------------------------------------------------------------------------------------------- | --------- | ---- | | 🇬🇧 [London (LON)](/docs/network/latency/lon-london) → 🇧🇷 [São Paulo (GRU)](/docs/network/latency/gru-sao-paulo) | **170.7** | Fair | | 🇳🇱 [Amsterdam (AMS)](/docs/network/latency/ams-amsterdam) → 🇧🇷 [São Paulo (GRU)](/docs/network/latency/gru-sao-paulo) | **175.5** | Fair | | 🇫🇷 [Paris (PAR)](/docs/network/latency/par-paris) → 🇧🇷 [São Paulo (GRU)](/docs/network/latency/gru-sao-paulo) | **175.8** | Fair | | 🇩🇪 [Frankfurt (FRA)](/docs/network/latency/fra-frankfurt) → 🇧🇷 [São Paulo (GRU)](/docs/network/latency/gru-sao-paulo) | **181.3** | Fair | | 🇩🇪 [Berlin (BER)](/docs/network/latency/ber-berlin) → 🇧🇷 [São Paulo (GRU)](/docs/network/latency/gru-sao-paulo) | **186.5** | Fair | ### To Asia Pacific | Route | RTT (ms) | Tier | | ------------------------------------------------------------------------------------------------------------------------- | --------- | ---- | | 🇷🇺 [Moscow (MOW)](/docs/network/latency/mow-moscow) → 🇭🇰 [Hong Kong (HKG)](/docs/network/latency/hkg-hong-kong) | **118.6** | Good | | 🇷🇺 [Moscow (MOW)](/docs/network/latency/mow-moscow) → 🇹🇼 [Taipei (TPE)](/docs/network/latency/tpe-taipei) | **132.1** | Good | | 🇫🇷 [Marseille (MRS)](/docs/network/latency/mrs-marseille) → 🇸🇬 [Singapore (SIN)](/docs/network/latency/sin-singapore) | **139.6** | Good | | 🇩🇪 [Berlin (BER)](/docs/network/latency/ber-berlin) → 🇭🇰 [Hong Kong (HKG)](/docs/network/latency/hkg-hong-kong) | **144.9** | Good | | 🇷🇺 [Moscow (MOW)](/docs/network/latency/mow-moscow) → 🇸🇬 [Singapore (SIN)](/docs/network/latency/sin-singapore) | **147.3** | Good | ### To Africa | Route | RTT (ms) | Tier | | ------------------------------------------------------------------------------------------------------------------------------- | --------- | ---- | | 🇳🇱 [Amsterdam (AMS)](/docs/network/latency/ams-amsterdam) → 🇿🇦 [Johannesburg (JNB)](/docs/network/latency/jnb-johannesburg) | **171.9** | Fair | | 🇩🇪 [Berlin (BER)](/docs/network/latency/ber-berlin) → 🇿🇦 [Johannesburg (JNB)](/docs/network/latency/jnb-johannesburg) | **172.1** | Fair | | 🇬🇧 [London (LON)](/docs/network/latency/lon-london) → 🇿🇦 [Johannesburg (JNB)](/docs/network/latency/jnb-johannesburg) | **179.6** | Fair | | 🇫🇷 [Paris (PAR)](/docs/network/latency/par-paris) → 🇿🇦 [Johannesburg (JNB)](/docs/network/latency/jnb-johannesburg) | **180.9** | Fair | | 🇩🇪 [Frankfurt (FRA)](/docs/network/latency/fra-frankfurt) → 🇿🇦 [Johannesburg (JNB)](/docs/network/latency/jnb-johannesburg) | **188.4** | Fair | ## PoPs in this Region | PoP | City | Country | | ---------------------------------------------- | -------------- | ----------- | | **[AMS](/docs/network/latency/ams-amsterdam)** | 🇳🇱 Amsterdam | Netherlands | | **[BER](/docs/network/latency/ber-berlin)** | 🇩🇪 Berlin | Germany | | **[FRA](/docs/network/latency/fra-frankfurt)** | 🇩🇪 Frankfurt | Germany | | **[LON](/docs/network/latency/lon-london)** | 🇬🇧 London | UK | | **[MRS](/docs/network/latency/mrs-marseille)** | 🇫🇷 Marseille | France | | **[MOW](/docs/network/latency/mow-moscow)** | 🇷🇺 Moscow | Russia | | **[PAR](/docs/network/latency/par-paris)** | 🇫🇷 Paris | France | *** *Round published August 16, 2026. See the [full global latency matrix](/docs/network/latency) or browse by [city pair](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/regions/europe). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Regional Ping & RTT Corridors | AS203314 Regional pages group PoPs by location. Corridor pages compare one origin region with another destination region. Each table uses the published round below; a corridor name does not describe a physical cable route. Measurement round: `2026-08-16T04:07:28Z`. ## Regions | Region | PoPs | Description | | -------------------------------- | ---- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | | [Europe](./europe) | 7 | European measurements cover Amsterdam, Berlin, Frankfurt, London, Marseille, Moscow and Paris. The regional matrix keeps each direction separate. | | [North America](./north-america) | 5 | The North American dataset includes five US PoPs: Ashburn, Los Angeles, Miami, New York and Seattle. | | [South America](./south-america) | 1 | São Paulo is the only South American PoP in this dataset, so all of its listed routes cross a regional boundary. | | [Asia Pacific](./asia-pacific) | 6 | Asia-Pacific measurements cover Hong Kong, Singapore, Taipei, Tokyo, Sydney and Melbourne. Distances within this group range from regional links to crossings between Asia and Australia. | | [Africa](./africa) | 1 | Johannesburg is the dataset's only African PoP. Its comparisons therefore use destinations in other regions. | ## Corridors | Corridor | Description | | ----------------------------------------------------------------- | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | | [Europe → North America](./europe-to-north-america) | These measurements run from European PoPs to US destinations. RTT alone does not identify the transatlantic cable or the intermediate networks. | | [Europe → Asia Pacific](./europe-to-asia-pacific) | European origins are compared with destinations in East Asia, Southeast Asia and Australia. The table records endpoint RTT, not the path across Eurasia or the oceans. | | [North America → Asia Pacific](./north-america-to-asia-pacific) | US origins are compared with the six Asia-Pacific PoPs. West Coast and East Coast results are kept separate. | | [Europe → South America](./europe-to-south-america) | These routes start in Europe and end at São Paulo. The measurements do not establish whether traffic crosses the South Atlantic directly or passes through another region. | | [North America → South America](./north-america-to-south-america) | These routes start at the five US PoPs and end at São Paulo. Compare the origins using the same published measurement round. | | [Asia Pacific → South America](./asia-pacific-to-south-america) | Asia-Pacific origins are compared with São Paulo. Long distances contribute to propagation delay, but the measurements do not reveal the transit path. | *** *Round published August 16, 2026. See the [full global latency matrix](/docs/network/latency) or browse individual [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/regions). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # North America to Asia Pacific Ping & RTT | AS203314 US origins are compared with the six Asia-Pacific PoPs. West Coast and East Coast results are kept separate. Measurement round: `2026-08-16T04:07:28Z`. ## Quick Stats | Metric | Value | | ------------------- | -------------------------------------------------------------------------------------- | | Fastest Route | **[Seattle (SEA) → Tokyo (TYO)](/docs/network/latency/pairs/sea-tyo-rtt)**: 85 ms | | Slowest Route | **[Miami (MIA) → Singapore (SIN)](/docs/network/latency/pairs/mia-sin-rtt)**: 223.1 ms | | Average RTT | **167.4 ms** | | City Pairs Measured | 30 | ## North America to Asia Pacific City-Pair RTT | Route | RTT (ms) | Tier | | ----------------------------------------------------------------------------------------------------------------------------- | ---------------------------------------------------- | ---- | | 🇺🇸 [Seattle (SEA)](/docs/network/latency/sea-seattle) → 🇯🇵 [Tokyo (TYO)](/docs/network/latency/tyo-tokyo) | **[85](/docs/network/latency/pairs/sea-tyo-rtt)** | Good | | 🇺🇸 [Los Angeles (LAX)](/docs/network/latency/lax-los-angeles) → 🇯🇵 [Tokyo (TYO)](/docs/network/latency/tyo-tokyo) | **[101.2](/docs/network/latency/pairs/lax-tyo-rtt)** | Good | | 🇺🇸 [Seattle (SEA)](/docs/network/latency/sea-seattle) → 🇹🇼 [Taipei (TPE)](/docs/network/latency/tpe-taipei) | **[114.7](/docs/network/latency/pairs/sea-tpe-rtt)** | Good | | 🇺🇸 [Seattle (SEA)](/docs/network/latency/sea-seattle) → 🇭🇰 [Hong Kong (HKG)](/docs/network/latency/hkg-hong-kong) | **[130.4](/docs/network/latency/pairs/sea-hkg-rtt)** | Good | | 🇺🇸 [Los Angeles (LAX)](/docs/network/latency/lax-los-angeles) → 🇹🇼 [Taipei (TPE)](/docs/network/latency/tpe-taipei) | **[132.9](/docs/network/latency/pairs/lax-tpe-rtt)** | Good | | 🇺🇸 [Los Angeles (LAX)](/docs/network/latency/lax-los-angeles) → 🇦🇺 [Sydney (SYD)](/docs/network/latency/syd-sydney) | **[137.6](/docs/network/latency/pairs/lax-syd-rtt)** | Good | | 🇺🇸 [Ashburn (IAD)](/docs/network/latency/iad-ashburn) → 🇯🇵 [Tokyo (TYO)](/docs/network/latency/tyo-tokyo) | **[141.9](/docs/network/latency/pairs/iad-tyo-rtt)** | Good | | 🇺🇸 [New York (NYC)](/docs/network/latency/nyc-new-york) → 🇯🇵 [Tokyo (TYO)](/docs/network/latency/tyo-tokyo) | **[143.1](/docs/network/latency/pairs/nyc-tyo-rtt)** | Good | | 🇺🇸 [Los Angeles (LAX)](/docs/network/latency/lax-los-angeles) → 🇭🇰 [Hong Kong (HKG)](/docs/network/latency/hkg-hong-kong) | **[146.5](/docs/network/latency/pairs/lax-hkg-rtt)** | Good | | 🇺🇸 [Los Angeles (LAX)](/docs/network/latency/lax-los-angeles) → 🇦🇺 [Melbourne (MEL)](/docs/network/latency/mel-melbourne) | **[148.2](/docs/network/latency/pairs/lax-mel-rtt)** | Good | | 🇺🇸 [Seattle (SEA)](/docs/network/latency/sea-seattle) → 🇸🇬 [Singapore (SIN)](/docs/network/latency/sin-singapore) | **[151.1](/docs/network/latency/pairs/sea-sin-rtt)** | Fair | | 🇺🇸 [Miami (MIA)](/docs/network/latency/mia-miami) → 🇯🇵 [Tokyo (TYO)](/docs/network/latency/tyo-tokyo) | **[156.3](/docs/network/latency/pairs/mia-tyo-rtt)** | Fair | | 🇺🇸 [Seattle (SEA)](/docs/network/latency/sea-seattle) → 🇦🇺 [Sydney (SYD)](/docs/network/latency/syd-sydney) | **[161.3](/docs/network/latency/pairs/sea-syd-rtt)** | Fair | | 🇺🇸 [Los Angeles (LAX)](/docs/network/latency/lax-los-angeles) → 🇸🇬 [Singapore (SIN)](/docs/network/latency/sin-singapore) | **[167.7](/docs/network/latency/pairs/lax-sin-rtt)** | Fair | | 🇺🇸 [Seattle (SEA)](/docs/network/latency/sea-seattle) → 🇦🇺 [Melbourne (MEL)](/docs/network/latency/mel-melbourne) | **[170.1](/docs/network/latency/pairs/sea-mel-rtt)** | Fair | | 🇺🇸 [Ashburn (IAD)](/docs/network/latency/iad-ashburn) → 🇹🇼 [Taipei (TPE)](/docs/network/latency/tpe-taipei) | **[172.3](/docs/network/latency/pairs/iad-tpe-rtt)** | Fair | | 🇺🇸 [New York (NYC)](/docs/network/latency/nyc-new-york) → 🇹🇼 [Taipei (TPE)](/docs/network/latency/tpe-taipei) | **[173.4](/docs/network/latency/pairs/nyc-tpe-rtt)** | Fair | | 🇺🇸 [Miami (MIA)](/docs/network/latency/mia-miami) → 🇹🇼 [Taipei (TPE)](/docs/network/latency/tpe-taipei) | **[185.1](/docs/network/latency/pairs/mia-tpe-rtt)** | Fair | | 🇺🇸 [New York (NYC)](/docs/network/latency/nyc-new-york) → 🇭🇰 [Hong Kong (HKG)](/docs/network/latency/hkg-hong-kong) | **[187.1](/docs/network/latency/pairs/nyc-hkg-rtt)** | Fair | | 🇺🇸 [Ashburn (IAD)](/docs/network/latency/iad-ashburn) → 🇭🇰 [Hong Kong (HKG)](/docs/network/latency/hkg-hong-kong) | **[187.2](/docs/network/latency/pairs/iad-hkg-rtt)** | Fair | | 🇺🇸 [Miami (MIA)](/docs/network/latency/mia-miami) → 🇦🇺 [Sydney (SYD)](/docs/network/latency/syd-sydney) | **[191.5](/docs/network/latency/pairs/mia-syd-rtt)** | Fair | | 🇺🇸 [New York (NYC)](/docs/network/latency/nyc-new-york) → 🇦🇺 [Sydney (SYD)](/docs/network/latency/syd-sydney) | **[192.6](/docs/network/latency/pairs/nyc-syd-rtt)** | Fair | | 🇺🇸 [Ashburn (IAD)](/docs/network/latency/iad-ashburn) → 🇦🇺 [Sydney (SYD)](/docs/network/latency/syd-sydney) | **[193.6](/docs/network/latency/pairs/iad-syd-rtt)** | Fair | | 🇺🇸 [Miami (MIA)](/docs/network/latency/mia-miami) → 🇭🇰 [Hong Kong (HKG)](/docs/network/latency/hkg-hong-kong) | **[199.5](/docs/network/latency/pairs/mia-hkg-rtt)** | Fair | | 🇺🇸 [Miami (MIA)](/docs/network/latency/mia-miami) → 🇦🇺 [Melbourne (MEL)](/docs/network/latency/mel-melbourne) | **[200.7](/docs/network/latency/pairs/mia-mel-rtt)** | Fair | | 🇺🇸 [Ashburn (IAD)](/docs/network/latency/iad-ashburn) → 🇦🇺 [Melbourne (MEL)](/docs/network/latency/mel-melbourne) | **[203.9](/docs/network/latency/pairs/iad-mel-rtt)** | Fair | | 🇺🇸 [New York (NYC)](/docs/network/latency/nyc-new-york) → 🇦🇺 [Melbourne (MEL)](/docs/network/latency/mel-melbourne) | **[205.1](/docs/network/latency/pairs/nyc-mel-rtt)** | Fair | | 🇺🇸 [New York (NYC)](/docs/network/latency/nyc-new-york) → 🇸🇬 [Singapore (SIN)](/docs/network/latency/sin-singapore) | **[208.7](/docs/network/latency/pairs/nyc-sin-rtt)** | Fair | | 🇺🇸 [Ashburn (IAD)](/docs/network/latency/iad-ashburn) → 🇸🇬 [Singapore (SIN)](/docs/network/latency/sin-singapore) | **[211.5](/docs/network/latency/pairs/iad-sin-rtt)** | Fair | | 🇺🇸 [Miami (MIA)](/docs/network/latency/mia-miami) → 🇸🇬 [Singapore (SIN)](/docs/network/latency/sin-singapore) | **[223.1](/docs/network/latency/pairs/mia-sin-rtt)** | Fair | ## Related Regions * [North America backbone latency](./north-america) * [Asia Pacific backbone latency](./asia-pacific) *** *Round published August 16, 2026. See the [full global latency matrix](/docs/network/latency).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/regions/north-america-to-asia-pacific). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # North America to South America Ping & RTT | AS203314 These routes start at the five US PoPs and end at São Paulo. Compare the origins using the same published measurement round. Measurement round: `2026-08-16T04:07:28Z`. ## Quick Stats | Metric | Value | | ------------------- | ---------------------------------------------------------------------------------------- | | Fastest Route | **[Ashburn (IAD) → São Paulo (GRU)](/docs/network/latency/pairs/iad-gru-rtt)**: 101.4 ms | | Slowest Route | **[Seattle (SEA) → São Paulo (GRU)](/docs/network/latency/pairs/sea-gru-rtt)**: 155.7 ms | | Average RTT | **124.8 ms** | | City Pairs Measured | 5 | ## North America to South America City-Pair RTT | Route | RTT (ms) | Tier | | ----------------------------------------------------------------------------------------------------------------------------- | ---------------------------------------------------- | ---- | | 🇺🇸 [Ashburn (IAD)](/docs/network/latency/iad-ashburn) → 🇧🇷 [São Paulo (GRU)](/docs/network/latency/gru-sao-paulo) | **[101.4](/docs/network/latency/pairs/iad-gru-rtt)** | Good | | 🇺🇸 [New York (NYC)](/docs/network/latency/nyc-new-york) → 🇧🇷 [São Paulo (GRU)](/docs/network/latency/gru-sao-paulo) | **[106.9](/docs/network/latency/pairs/nyc-gru-rtt)** | Good | | 🇺🇸 [Miami (MIA)](/docs/network/latency/mia-miami) → 🇧🇷 [São Paulo (GRU)](/docs/network/latency/gru-sao-paulo) | **[128.7](/docs/network/latency/pairs/mia-gru-rtt)** | Good | | 🇺🇸 [Los Angeles (LAX)](/docs/network/latency/lax-los-angeles) → 🇧🇷 [São Paulo (GRU)](/docs/network/latency/gru-sao-paulo) | **[131.3](/docs/network/latency/pairs/lax-gru-rtt)** | Good | | 🇺🇸 [Seattle (SEA)](/docs/network/latency/sea-seattle) → 🇧🇷 [São Paulo (GRU)](/docs/network/latency/gru-sao-paulo) | **[155.7](/docs/network/latency/pairs/sea-gru-rtt)** | Fair | ## Related Regions * [North America backbone latency](./north-america) * [South America backbone latency](./south-america) *** *Round published August 16, 2026. See the [full global latency matrix](/docs/network/latency).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/regions/north-america-to-south-america). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # North America Ping & RTT Latency | AS203314 The North American dataset includes five US PoPs: Ashburn, Los Angeles, Miami, New York and Seattle. Measurement round: `2026-08-16T04:07:28Z`. ## Intra-North America RTT Matrix Round-trip times in milliseconds between PoPs within the region. | Route | IAD | LAX | MIA | NYC | SEA | | --------------------------------------------------------------- | ---- | ---- | ---- | ---- | ---- | | 🇺🇸 [Ashburn (IAD)](/docs/network/latency/iad-ashburn) | N/A | 60.7 | 27.4 | 6 | 62.3 | | 🇺🇸 [Los Angeles (LAX)](/docs/network/latency/lax-los-angeles) | 60.4 | N/A | 57.3 | 57.9 | 26.9 | | 🇺🇸 [Miami (MIA)](/docs/network/latency/mia-miami) | 27.3 | 56.8 | N/A | 33.3 | 81.9 | | 🇺🇸 [New York (NYC)](/docs/network/latency/nyc-new-york) | 6 | 58.9 | 32.9 | N/A | 58.7 | | 🇺🇸 [Seattle (SEA)](/docs/network/latency/sea-seattle) | 61.5 | 25.9 | 81.7 | 59.5 | N/A | ## North America to Other Regions The five lowest recorded RTTs to each destination region. ### To Europe | Route | RTT (ms) | Tier | | ----------------------------------------------------------------------------------------------------------------------- | -------- | --------- | | 🇺🇸 [New York (NYC)](/docs/network/latency/nyc-new-york) → 🇬🇧 [London (LON)](/docs/network/latency/lon-london) | **63.5** | Excellent | | 🇺🇸 [New York (NYC)](/docs/network/latency/nyc-new-york) → 🇳🇱 [Amsterdam (AMS)](/docs/network/latency/ams-amsterdam) | **68.7** | Excellent | | 🇺🇸 [New York (NYC)](/docs/network/latency/nyc-new-york) → 🇫🇷 [Paris (PAR)](/docs/network/latency/par-paris) | **68.9** | Excellent | | 🇺🇸 [Ashburn (IAD)](/docs/network/latency/iad-ashburn) → 🇬🇧 [London (LON)](/docs/network/latency/lon-london) | **70.3** | Excellent | | 🇺🇸 [New York (NYC)](/docs/network/latency/nyc-new-york) → 🇩🇪 [Frankfurt (FRA)](/docs/network/latency/fra-frankfurt) | **73.9** | Excellent | ### To South America | Route | RTT (ms) | Tier | | ----------------------------------------------------------------------------------------------------------------------------- | --------- | ---- | | 🇺🇸 [Ashburn (IAD)](/docs/network/latency/iad-ashburn) → 🇧🇷 [São Paulo (GRU)](/docs/network/latency/gru-sao-paulo) | **101.4** | Good | | 🇺🇸 [New York (NYC)](/docs/network/latency/nyc-new-york) → 🇧🇷 [São Paulo (GRU)](/docs/network/latency/gru-sao-paulo) | **106.9** | Good | | 🇺🇸 [Miami (MIA)](/docs/network/latency/mia-miami) → 🇧🇷 [São Paulo (GRU)](/docs/network/latency/gru-sao-paulo) | **128.7** | Good | | 🇺🇸 [Los Angeles (LAX)](/docs/network/latency/lax-los-angeles) → 🇧🇷 [São Paulo (GRU)](/docs/network/latency/gru-sao-paulo) | **131.3** | Good | | 🇺🇸 [Seattle (SEA)](/docs/network/latency/sea-seattle) → 🇧🇷 [São Paulo (GRU)](/docs/network/latency/gru-sao-paulo) | **155.7** | Fair | ### To Asia Pacific | Route | RTT (ms) | Tier | | ----------------------------------------------------------------------------------------------------------------------- | --------- | ---- | | 🇺🇸 [Seattle (SEA)](/docs/network/latency/sea-seattle) → 🇯🇵 [Tokyo (TYO)](/docs/network/latency/tyo-tokyo) | **85** | Good | | 🇺🇸 [Los Angeles (LAX)](/docs/network/latency/lax-los-angeles) → 🇯🇵 [Tokyo (TYO)](/docs/network/latency/tyo-tokyo) | **101.2** | Good | | 🇺🇸 [Seattle (SEA)](/docs/network/latency/sea-seattle) → 🇹🇼 [Taipei (TPE)](/docs/network/latency/tpe-taipei) | **114.7** | Good | | 🇺🇸 [Seattle (SEA)](/docs/network/latency/sea-seattle) → 🇭🇰 [Hong Kong (HKG)](/docs/network/latency/hkg-hong-kong) | **130.4** | Good | | 🇺🇸 [Los Angeles (LAX)](/docs/network/latency/lax-los-angeles) → 🇹🇼 [Taipei (TPE)](/docs/network/latency/tpe-taipei) | **132.9** | Good | ### To Africa | Route | RTT (ms) | Tier | | ----------------------------------------------------------------------------------------------------------------------------------- | --------- | ---- | | 🇺🇸 [New York (NYC)](/docs/network/latency/nyc-new-york) → 🇿🇦 [Johannesburg (JNB)](/docs/network/latency/jnb-johannesburg) | **221.5** | Fair | | 🇺🇸 [Ashburn (IAD)](/docs/network/latency/iad-ashburn) → 🇿🇦 [Johannesburg (JNB)](/docs/network/latency/jnb-johannesburg) | **228.3** | Fair | | 🇺🇸 [Miami (MIA)](/docs/network/latency/mia-miami) → 🇿🇦 [Johannesburg (JNB)](/docs/network/latency/jnb-johannesburg) | **259.8** | High | | 🇺🇸 [Seattle (SEA)](/docs/network/latency/sea-seattle) → 🇿🇦 [Johannesburg (JNB)](/docs/network/latency/jnb-johannesburg) | **282.1** | High | | 🇺🇸 [Los Angeles (LAX)](/docs/network/latency/lax-los-angeles) → 🇿🇦 [Johannesburg (JNB)](/docs/network/latency/jnb-johannesburg) | **287.2** | High | ## PoPs in this Region | PoP | City | Country | | ------------------------------------------------ | ---------------- | ------- | | **[IAD](/docs/network/latency/iad-ashburn)** | 🇺🇸 Ashburn | USA | | **[LAX](/docs/network/latency/lax-los-angeles)** | 🇺🇸 Los Angeles | USA | | **[MIA](/docs/network/latency/mia-miami)** | 🇺🇸 Miami | USA | | **[NYC](/docs/network/latency/nyc-new-york)** | 🇺🇸 New York | USA | | **[SEA](/docs/network/latency/sea-seattle)** | 🇺🇸 Seattle | USA | *** *Round published August 16, 2026. See the [full global latency matrix](/docs/network/latency) or browse by [city pair](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/regions/north-america). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # South America Ping & RTT Latency | AS203314 São Paulo is the only South American PoP in this dataset, so all of its listed routes cross a regional boundary. Measurement round: `2026-08-16T04:07:28Z`. ## South America to Other Regions The five lowest recorded RTTs to each destination region. ### To Europe | Route | RTT (ms) | Tier | | ------------------------------------------------------------------------------------------------------------------------- | --------- | ---- | | 🇧🇷 [São Paulo (GRU)](/docs/network/latency/gru-sao-paulo) → 🇬🇧 [London (LON)](/docs/network/latency/lon-london) | **175.8** | Fair | | 🇧🇷 [São Paulo (GRU)](/docs/network/latency/gru-sao-paulo) → 🇳🇱 [Amsterdam (AMS)](/docs/network/latency/ams-amsterdam) | **180** | Fair | | 🇧🇷 [São Paulo (GRU)](/docs/network/latency/gru-sao-paulo) → 🇫🇷 [Paris (PAR)](/docs/network/latency/par-paris) | **180.2** | Fair | | 🇧🇷 [São Paulo (GRU)](/docs/network/latency/gru-sao-paulo) → 🇩🇪 [Frankfurt (FRA)](/docs/network/latency/fra-frankfurt) | **186.2** | Fair | | 🇧🇷 [São Paulo (GRU)](/docs/network/latency/gru-sao-paulo) → 🇫🇷 [Marseille (MRS)](/docs/network/latency/mrs-marseille) | **187.7** | Fair | ### To North America | Route | RTT (ms) | Tier | | ----------------------------------------------------------------------------------------------------------------------------- | --------- | --------- | | 🇧🇷 [São Paulo (GRU)](/docs/network/latency/gru-sao-paulo) → 🇺🇸 [Miami (MIA)](/docs/network/latency/mia-miami) | **74** | Excellent | | 🇧🇷 [São Paulo (GRU)](/docs/network/latency/gru-sao-paulo) → 🇺🇸 [New York (NYC)](/docs/network/latency/nyc-new-york) | **122** | Good | | 🇧🇷 [São Paulo (GRU)](/docs/network/latency/gru-sao-paulo) → 🇺🇸 [Ashburn (IAD)](/docs/network/latency/iad-ashburn) | **128** | Good | | 🇧🇷 [São Paulo (GRU)](/docs/network/latency/gru-sao-paulo) → 🇺🇸 [Los Angeles (LAX)](/docs/network/latency/lax-los-angeles) | **130.8** | Good | | 🇧🇷 [São Paulo (GRU)](/docs/network/latency/gru-sao-paulo) → 🇺🇸 [Seattle (SEA)](/docs/network/latency/sea-seattle) | **155.9** | Fair | ### To Asia Pacific | Route | RTT (ms) | Tier | | ------------------------------------------------------------------------------------------------------------------------- | --------- | ---- | | 🇧🇷 [São Paulo (GRU)](/docs/network/latency/gru-sao-paulo) → 🇯🇵 [Tokyo (TYO)](/docs/network/latency/tyo-tokyo) | **230.3** | Fair | | 🇧🇷 [São Paulo (GRU)](/docs/network/latency/gru-sao-paulo) → 🇹🇼 [Taipei (TPE)](/docs/network/latency/tpe-taipei) | **259.1** | High | | 🇧🇷 [São Paulo (GRU)](/docs/network/latency/gru-sao-paulo) → 🇦🇺 [Sydney (SYD)](/docs/network/latency/syd-sydney) | **265.5** | High | | 🇧🇷 [São Paulo (GRU)](/docs/network/latency/gru-sao-paulo) → 🇭🇰 [Hong Kong (HKG)](/docs/network/latency/hkg-hong-kong) | **273.5** | High | | 🇧🇷 [São Paulo (GRU)](/docs/network/latency/gru-sao-paulo) → 🇦🇺 [Melbourne (MEL)](/docs/network/latency/mel-melbourne) | **274.7** | High | ### To Africa | Route | RTT (ms) | Tier | | ------------------------------------------------------------------------------------------------------------------------------- | --------- | ---- | | 🇧🇷 [São Paulo (GRU)](/docs/network/latency/gru-sao-paulo) → 🇿🇦 [Johannesburg (JNB)](/docs/network/latency/jnb-johannesburg) | **333.8** | High | ## PoPs in this Region | PoP | City | Country | | ---------------------------------------------- | -------------- | ------- | | **[GRU](/docs/network/latency/gru-sao-paulo)** | 🇧🇷 São Paulo | Brazil | *** *Round published August 16, 2026. See the [full global latency matrix](/docs/network/latency) or browse by [city pair](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/regions/south-america). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Amsterdam, Netherlands → Berlin, Germany RTT 🇳🇱 **Amsterdam, Netherlands (AMS)** → 🇩🇪 **Berlin, Germany (BER)** Measurement round: `2026-08-16T04:07:28Z`. Amsterdam to Berlin averaged 7.2 ms across 50 replies. The observed range was 6.88 to 8.21 ms; these figures describe one probe round, not a long-term service target. This is number 4 of 19 outgoing routes by average RTT. That ranking compares different destinations and is not adjusted for distance. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------- | | RTT | **7.2 ms** | | Jitter | **0.23 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **5.67 ms** | | Fiber Efficiency | **78.8%** | | Latency Tier | Ultra-Low | | Source Region | [Europe](/docs/network/latency/regions/europe) | | Destination Region | [Europe](/docs/network/latency/regions/europe) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Amsterdam](https://www.geonames.org/2759794) (52.37403, 4.88969) and [Berlin](https://www.geonames.org/2950159) (52.52437, 13.41053). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | ------------ | | WGS-84 geodesic distance | **579.2 km** | | Vacuum RTT floor | **3.86 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **5.67 ms** | | Low-latency fiber material floor | **5.65 ms** | | Engineering floor (5% path allowance) | **5.96 ms** | | Research 1.33× mapped-fiber reference | **7.54 ms** | | Estimated unamplified path loss | **121.6 dB** | | Transparent optical spans / inline amplifiers | **8 / 7** | | Published RTT inflation over fiber floor | **1.27×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ----------- | | Minimum RTT | **6.88 ms** | | Average RTT | **7.2 ms** | | Maximum RTT | **8.21 ms** | | Standard deviation | **0.31 ms** | | Stdev / average | **4.3%** | ## Route Context * Departure PoP: [Amsterdam (AMS)](/docs/network/latency/ams-amsterdam) * Destination PoP: [Berlin (BER)](/docs/network/latency/ber-berlin) * Region overview: [Europe](/docs/network/latency/regions/europe) → [Europe](/docs/network/latency/regions/europe) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Berlin to Amsterdam latency and RTT](/docs/network/latency/pairs/ber-ams-rtt): 7.2 ms **Fastest routes departing Amsterdam (AMS)** * [Amsterdam to London latency and RTT](/docs/network/latency/pairs/ams-lon-rtt): 5.2 ms * [Amsterdam to Frankfurt latency and RTT](/docs/network/latency/pairs/ams-fra-rtt): 5.9 ms * [Amsterdam to Paris latency and RTT](/docs/network/latency/pairs/ams-par-rtt): 7 ms * [Amsterdam to Marseille latency and RTT](/docs/network/latency/pairs/ams-mrs-rtt): 19.6 ms * [Amsterdam to Moscow latency and RTT](/docs/network/latency/pairs/ams-mow-rtt): 38.1 ms **Fastest routes arriving at Berlin (BER)** * [Frankfurt to Berlin latency and RTT](/docs/network/latency/pairs/fra-ber-rtt): 6.1 ms * [Paris to Berlin latency and RTT](/docs/network/latency/pairs/par-ber-rtt): 15.5 ms * [London to Berlin latency and RTT](/docs/network/latency/pairs/lon-ber-rtt): 16.2 ms * [Marseille to Berlin latency and RTT](/docs/network/latency/pairs/mrs-ber-rtt): 20.6 ms * [Moscow to Berlin latency and RTT](/docs/network/latency/pairs/mow-ber-rtt): 27.5 ms **Same corridor (Europe → Europe)** * [London to Amsterdam latency and RTT](/docs/network/latency/pairs/lon-ams-rtt): 5.2 ms * [Frankfurt to Amsterdam latency and RTT](/docs/network/latency/pairs/fra-ams-rtt): 5.9 ms * [Berlin to Frankfurt latency and RTT](/docs/network/latency/pairs/ber-fra-rtt): 6.1 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/ams-ber.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Amsterdam → Berlin round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/ams-ber-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Amsterdam, Netherlands → Frankfurt, Germany RTT 🇳🇱 **Amsterdam, Netherlands (AMS)** → 🇩🇪 **Frankfurt, Germany (FRA)** Measurement round: `2026-08-16T04:07:28Z`. Amsterdam to Frankfurt averaged 5.9 ms across 50 replies. The observed range was 5.67 to 7.15 ms; these figures describe one probe round, not a long-term service target. This is number 2 of 19 outgoing routes by average RTT. That ranking compares different destinations and is not adjusted for distance. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------- | | RTT | **5.9 ms** | | Jitter | **0.17 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **3.57 ms** | | Fiber Efficiency | **60.6%** | | Latency Tier | Ultra-Low | | Source Region | [Europe](/docs/network/latency/regions/europe) | | Destination Region | [Europe](/docs/network/latency/regions/europe) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Amsterdam](https://www.geonames.org/2759794) (52.37403, 4.88969) and [Frankfurt am Main](https://www.geonames.org/2925533) (50.11552, 8.68417). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | ------------ | | WGS-84 geodesic distance | **365.1 km** | | Vacuum RTT floor | **2.44 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **3.57 ms** | | Low-latency fiber material floor | **3.56 ms** | | Engineering floor (5% path allowance) | **3.75 ms** | | Research 1.33× mapped-fiber reference | **4.75 ms** | | Estimated unamplified path loss | **76.7 dB** | | Transparent optical spans / inline amplifiers | **5 / 4** | | Published RTT inflation over fiber floor | **1.65×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ----------- | | Minimum RTT | **5.67 ms** | | Average RTT | **5.9 ms** | | Maximum RTT | **7.15 ms** | | Standard deviation | **0.23 ms** | | Stdev / average | **3.9%** | ## Route Context * Departure PoP: [Amsterdam (AMS)](/docs/network/latency/ams-amsterdam) * Destination PoP: [Frankfurt (FRA)](/docs/network/latency/fra-frankfurt) * Region overview: [Europe](/docs/network/latency/regions/europe) → [Europe](/docs/network/latency/regions/europe) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Frankfurt to Amsterdam latency and RTT](/docs/network/latency/pairs/fra-ams-rtt): 5.9 ms **Fastest routes departing Amsterdam (AMS)** * [Amsterdam to London latency and RTT](/docs/network/latency/pairs/ams-lon-rtt): 5.2 ms * [Amsterdam to Paris latency and RTT](/docs/network/latency/pairs/ams-par-rtt): 7 ms * [Amsterdam to Berlin latency and RTT](/docs/network/latency/pairs/ams-ber-rtt): 7.2 ms * [Amsterdam to Marseille latency and RTT](/docs/network/latency/pairs/ams-mrs-rtt): 19.6 ms * [Amsterdam to Moscow latency and RTT](/docs/network/latency/pairs/ams-mow-rtt): 38.1 ms **Fastest routes arriving at Frankfurt (FRA)** * [Berlin to Frankfurt latency and RTT](/docs/network/latency/pairs/ber-fra-rtt): 6.1 ms * [Paris to Frankfurt latency and RTT](/docs/network/latency/pairs/par-fra-rtt): 7.6 ms * [London to Frankfurt latency and RTT](/docs/network/latency/pairs/lon-fra-rtt): 13.6 ms * [Marseille to Frankfurt latency and RTT](/docs/network/latency/pairs/mrs-fra-rtt): 16 ms * [Moscow to Frankfurt latency and RTT](/docs/network/latency/pairs/mow-fra-rtt): 35.9 ms **Same corridor (Europe → Europe)** * [London to Amsterdam latency and RTT](/docs/network/latency/pairs/lon-ams-rtt): 5.2 ms * [Frankfurt to Berlin latency and RTT](/docs/network/latency/pairs/fra-ber-rtt): 6.1 ms * [London to Paris latency and RTT](/docs/network/latency/pairs/lon-par-rtt): 6.4 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/ams-fra.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Amsterdam → Frankfurt round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/ams-fra-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Amsterdam, Netherlands → São Paulo, Brazil RTT 🇳🇱 **Amsterdam, Netherlands (AMS)** → 🇧🇷 **São Paulo, Brazil (GRU)** Measurement round: `2026-08-16T04:07:28Z`. Amsterdam to São Paulo averaged 175.5 ms across 50 replies. The observed range was 166.82 to 195.11 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 6.86 ms, or 3.9% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **175.5 ms** | | Jitter | **6.92 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **95.79 ms** | | Fiber Efficiency | **54.6%** | | Latency Tier | Fair | | Source Region | [Europe](/docs/network/latency/regions/europe) | | Destination Region | [South America](/docs/network/latency/regions/south-america) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Amsterdam](https://www.geonames.org/2759794) (52.37403, 4.88969) and [São Paulo](https://www.geonames.org/3448439) (-23.54750, -46.63611). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **9,781.6 km** | | Vacuum RTT floor | **65.26 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **95.79 ms** | | Low-latency fiber material floor | **95.4 ms** | | Engineering floor (5% path allowance) | **100.59 ms** | | Research 1.33× mapped-fiber reference | **127.4 ms** | | Estimated unamplified path loss | **2054.1 dB** | | Transparent optical spans / inline amplifiers | **129 / 128** | | Published RTT inflation over fiber floor | **1.83×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **166.82 ms** | | Average RTT | **175.5 ms** | | Maximum RTT | **195.11 ms** | | Standard deviation | **6.86 ms** | | Stdev / average | **3.9%** | ## Route Context * Departure PoP: [Amsterdam (AMS)](/docs/network/latency/ams-amsterdam) * Destination PoP: [São Paulo (GRU)](/docs/network/latency/gru-sao-paulo) * Region overview: [Europe](/docs/network/latency/regions/europe) → [South America](/docs/network/latency/regions/south-america) * Corridor overview: [Europe to South America](/docs/network/latency/regions/europe-to-south-america) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [São Paulo to Amsterdam latency and RTT](/docs/network/latency/pairs/gru-ams-rtt): 180 ms **Fastest routes departing Amsterdam (AMS)** * [Amsterdam to London latency and RTT](/docs/network/latency/pairs/ams-lon-rtt): 5.2 ms * [Amsterdam to Frankfurt latency and RTT](/docs/network/latency/pairs/ams-fra-rtt): 5.9 ms * [Amsterdam to Paris latency and RTT](/docs/network/latency/pairs/ams-par-rtt): 7 ms * [Amsterdam to Berlin latency and RTT](/docs/network/latency/pairs/ams-ber-rtt): 7.2 ms * [Amsterdam to Marseille latency and RTT](/docs/network/latency/pairs/ams-mrs-rtt): 19.6 ms **Fastest routes arriving at São Paulo (GRU)** * [Ashburn to São Paulo latency and RTT](/docs/network/latency/pairs/iad-gru-rtt): 101.4 ms * [New York to São Paulo latency and RTT](/docs/network/latency/pairs/nyc-gru-rtt): 106.9 ms * [Miami to São Paulo latency and RTT](/docs/network/latency/pairs/mia-gru-rtt): 128.7 ms * [Los Angeles to São Paulo latency and RTT](/docs/network/latency/pairs/lax-gru-rtt): 131.3 ms * [Seattle to São Paulo latency and RTT](/docs/network/latency/pairs/sea-gru-rtt): 155.7 ms **Same corridor (Europe → South America)** * [London to São Paulo latency and RTT](/docs/network/latency/pairs/lon-gru-rtt): 170.7 ms * [Paris to São Paulo latency and RTT](/docs/network/latency/pairs/par-gru-rtt): 175.8 ms * [Frankfurt to São Paulo latency and RTT](/docs/network/latency/pairs/fra-gru-rtt): 181.3 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/ams-gru.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Amsterdam → São Paulo round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/ams-gru-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Amsterdam, Netherlands → Hong Kong RTT 🇳🇱 **Amsterdam, Netherlands (AMS)** → 🇭🇰 **Hong Kong (HKG)** Measurement round: `2026-08-16T04:07:28Z`. Amsterdam to Hong Kong averaged 155.9 ms across 50 replies. The observed range was 149.39 to 177.73 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 5.82 ms, or 3.7% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------------------- | | RTT | **155.9 ms** | | Jitter | **4.48 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **91.05 ms** | | Fiber Efficiency | **58.4%** | | Latency Tier | Fair | | Source Region | [Europe](/docs/network/latency/regions/europe) | | Destination Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Amsterdam](https://www.geonames.org/2759794) (52.37403, 4.88969) and [Hong Kong](https://www.geonames.org/1819729) (22.27832, 114.17469). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **9,297.6 km** | | Vacuum RTT floor | **62.03 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **91.05 ms** | | Low-latency fiber material floor | **90.68 ms** | | Engineering floor (5% path allowance) | **95.61 ms** | | Research 1.33× mapped-fiber reference | **121.1 ms** | | Estimated unamplified path loss | **1952.5 dB** | | Transparent optical spans / inline amplifiers | **123 / 122** | | Published RTT inflation over fiber floor | **1.71×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **149.39 ms** | | Average RTT | **155.9 ms** | | Maximum RTT | **177.73 ms** | | Standard deviation | **5.82 ms** | | Stdev / average | **3.7%** | ## Route Context * Departure PoP: [Amsterdam (AMS)](/docs/network/latency/ams-amsterdam) * Destination PoP: [Hong Kong (HKG)](/docs/network/latency/hkg-hong-kong) * Region overview: [Europe](/docs/network/latency/regions/europe) → [Asia Pacific](/docs/network/latency/regions/asia-pacific) * Corridor overview: [Europe to Asia Pacific](/docs/network/latency/regions/europe-to-asia-pacific) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Hong Kong to Amsterdam latency and RTT](/docs/network/latency/pairs/hkg-ams-rtt): 154.4 ms **Fastest routes departing Amsterdam (AMS)** * [Amsterdam to London latency and RTT](/docs/network/latency/pairs/ams-lon-rtt): 5.2 ms * [Amsterdam to Frankfurt latency and RTT](/docs/network/latency/pairs/ams-fra-rtt): 5.9 ms * [Amsterdam to Paris latency and RTT](/docs/network/latency/pairs/ams-par-rtt): 7 ms * [Amsterdam to Berlin latency and RTT](/docs/network/latency/pairs/ams-ber-rtt): 7.2 ms * [Amsterdam to Marseille latency and RTT](/docs/network/latency/pairs/ams-mrs-rtt): 19.6 ms **Fastest routes arriving at Hong Kong (HKG)** * [Taipei to Hong Kong latency and RTT](/docs/network/latency/pairs/tpe-hkg-rtt): 15.5 ms * [Singapore to Hong Kong latency and RTT](/docs/network/latency/pairs/sin-hkg-rtt): 30.8 ms * [Tokyo to Hong Kong latency and RTT](/docs/network/latency/pairs/tyo-hkg-rtt): 44.8 ms * [Moscow to Hong Kong latency and RTT](/docs/network/latency/pairs/mow-hkg-rtt): 118.6 ms * [Seattle to Hong Kong latency and RTT](/docs/network/latency/pairs/sea-hkg-rtt): 130.4 ms **Same corridor (Europe → Asia Pacific)** * [Moscow to Taipei latency and RTT](/docs/network/latency/pairs/mow-tpe-rtt): 132.1 ms * [Marseille to Singapore latency and RTT](/docs/network/latency/pairs/mrs-sin-rtt): 139.6 ms * [Berlin to Hong Kong latency and RTT](/docs/network/latency/pairs/ber-hkg-rtt): 144.9 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/ams-hkg.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Amsterdam → Hong Kong round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/ams-hkg-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Amsterdam, Netherlands → Ashburn, USA RTT 🇳🇱 **Amsterdam, Netherlands (AMS)** → 🇺🇸 **Ashburn, USA (IAD)** Measurement round: `2026-08-16T04:07:28Z`. Amsterdam to Ashburn averaged 75.7 ms across 50 replies. The observed range was 72.76 to 82.67 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 2.28 ms, or 3.0% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **75.7 ms** | | Jitter | **2.14 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **60.94 ms** | | Fiber Efficiency | **80.5%** | | Latency Tier | Excellent | | Source Region | [Europe](/docs/network/latency/regions/europe) | | Destination Region | [North America](/docs/network/latency/regions/north-america) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Amsterdam](https://www.geonames.org/2759794) (52.37403, 4.88969) and [Ashburn](https://www.geonames.org/4744870) (39.04372, -77.48749). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **6,223.2 km** | | Vacuum RTT floor | **41.52 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **60.94 ms** | | Low-latency fiber material floor | **60.7 ms** | | Engineering floor (5% path allowance) | **64 ms** | | Research 1.33× mapped-fiber reference | **81.05 ms** | | Estimated unamplified path loss | **1306.9 dB** | | Transparent optical spans / inline amplifiers | **82 / 81** | | Published RTT inflation over fiber floor | **1.24×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------ | | Minimum RTT | **72.76 ms** | | Average RTT | **75.7 ms** | | Maximum RTT | **82.67 ms** | | Standard deviation | **2.28 ms** | | Stdev / average | **3.0%** | ## Route Context * Departure PoP: [Amsterdam (AMS)](/docs/network/latency/ams-amsterdam) * Destination PoP: [Ashburn (IAD)](/docs/network/latency/iad-ashburn) * Region overview: [Europe](/docs/network/latency/regions/europe) → [North America](/docs/network/latency/regions/north-america) * Corridor overview: [Europe to North America](/docs/network/latency/regions/europe-to-north-america) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Ashburn to Amsterdam latency and RTT](/docs/network/latency/pairs/iad-ams-rtt): 75.2 ms **Fastest routes departing Amsterdam (AMS)** * [Amsterdam to London latency and RTT](/docs/network/latency/pairs/ams-lon-rtt): 5.2 ms * [Amsterdam to Frankfurt latency and RTT](/docs/network/latency/pairs/ams-fra-rtt): 5.9 ms * [Amsterdam to Paris latency and RTT](/docs/network/latency/pairs/ams-par-rtt): 7 ms * [Amsterdam to Berlin latency and RTT](/docs/network/latency/pairs/ams-ber-rtt): 7.2 ms * [Amsterdam to Marseille latency and RTT](/docs/network/latency/pairs/ams-mrs-rtt): 19.6 ms **Fastest routes arriving at Ashburn (IAD)** * [New York to Ashburn latency and RTT](/docs/network/latency/pairs/nyc-iad-rtt): 6 ms * [Miami to Ashburn latency and RTT](/docs/network/latency/pairs/mia-iad-rtt): 27.3 ms * [Los Angeles to Ashburn latency and RTT](/docs/network/latency/pairs/lax-iad-rtt): 60.4 ms * [Seattle to Ashburn latency and RTT](/docs/network/latency/pairs/sea-iad-rtt): 61.5 ms * [London to Ashburn latency and RTT](/docs/network/latency/pairs/lon-iad-rtt): 71 ms **Same corridor (Europe → North America)** * [London to New York latency and RTT](/docs/network/latency/pairs/lon-nyc-rtt): 63.8 ms * [Amsterdam to New York latency and RTT](/docs/network/latency/pairs/ams-nyc-rtt): 68.6 ms * [Paris to New York latency and RTT](/docs/network/latency/pairs/par-nyc-rtt): 68.9 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/ams-iad.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Amsterdam → Ashburn round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/ams-iad-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Amsterdam, Netherlands → Johannesburg, South Africa RTT 🇳🇱 **Amsterdam, Netherlands (AMS)** → 🇿🇦 **Johannesburg, South Africa (JNB)** Measurement round: `2026-08-16T04:07:28Z`. Amsterdam to Johannesburg averaged 171.9 ms across 50 replies. The observed range was 163.56 to 197.47 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 6.53 ms, or 3.8% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------- | | RTT | **171.9 ms** | | Jitter | **5.46 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **88.05 ms** | | Fiber Efficiency | **51.2%** | | Latency Tier | Fair | | Source Region | [Europe](/docs/network/latency/regions/europe) | | Destination Region | [Africa](/docs/network/latency/regions/africa) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Amsterdam](https://www.geonames.org/2759794) (52.37403, 4.88969) and [Johannesburg](https://www.geonames.org/993800) (-26.20227, 28.04363). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **8,991.8 km** | | Vacuum RTT floor | **59.99 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **88.05 ms** | | Low-latency fiber material floor | **87.7 ms** | | Engineering floor (5% path allowance) | **92.47 ms** | | Research 1.33× mapped-fiber reference | **117.11 ms** | | Estimated unamplified path loss | **1888.3 dB** | | Transparent optical spans / inline amplifiers | **119 / 118** | | Published RTT inflation over fiber floor | **1.95×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **163.56 ms** | | Average RTT | **171.9 ms** | | Maximum RTT | **197.47 ms** | | Standard deviation | **6.53 ms** | | Stdev / average | **3.8%** | ## Route Context * Departure PoP: [Amsterdam (AMS)](/docs/network/latency/ams-amsterdam) * Destination PoP: [Johannesburg (JNB)](/docs/network/latency/jnb-johannesburg) * Region overview: [Europe](/docs/network/latency/regions/europe) → [Africa](/docs/network/latency/regions/africa) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Johannesburg to Amsterdam latency and RTT](/docs/network/latency/pairs/jnb-ams-rtt): 163.2 ms **Fastest routes departing Amsterdam (AMS)** * [Amsterdam to London latency and RTT](/docs/network/latency/pairs/ams-lon-rtt): 5.2 ms * [Amsterdam to Frankfurt latency and RTT](/docs/network/latency/pairs/ams-fra-rtt): 5.9 ms * [Amsterdam to Paris latency and RTT](/docs/network/latency/pairs/ams-par-rtt): 7 ms * [Amsterdam to Berlin latency and RTT](/docs/network/latency/pairs/ams-ber-rtt): 7.2 ms * [Amsterdam to Marseille latency and RTT](/docs/network/latency/pairs/ams-mrs-rtt): 19.6 ms **Fastest routes arriving at Johannesburg (JNB)** * [Berlin to Johannesburg latency and RTT](/docs/network/latency/pairs/ber-jnb-rtt): 172.1 ms * [London to Johannesburg latency and RTT](/docs/network/latency/pairs/lon-jnb-rtt): 179.6 ms * [Paris to Johannesburg latency and RTT](/docs/network/latency/pairs/par-jnb-rtt): 180.9 ms * [Frankfurt to Johannesburg latency and RTT](/docs/network/latency/pairs/fra-jnb-rtt): 188.4 ms * [Marseille to Johannesburg latency and RTT](/docs/network/latency/pairs/mrs-jnb-rtt): 198.5 ms **Same corridor (Europe → Africa)** * [Moscow to Johannesburg latency and RTT](/docs/network/latency/pairs/mow-jnb-rtt): 201.9 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/ams-jnb.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Amsterdam → Johannesburg round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/ams-jnb-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Amsterdam, Netherlands → Los Angeles, USA RTT 🇳🇱 **Amsterdam, Netherlands (AMS)** → 🇺🇸 **Los Angeles, USA (LAX)** Measurement round: `2026-08-16T04:07:28Z`. Amsterdam to Los Angeles averaged 133.7 ms across 50 replies. The observed range was 130.58 to 141.4 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 2.49 ms, or 1.9% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **133.7 ms** | | Jitter | **2.11 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **87.74 ms** | | Fiber Efficiency | **65.6%** | | Latency Tier | Good | | Source Region | [Europe](/docs/network/latency/regions/europe) | | Destination Region | [North America](/docs/network/latency/regions/north-america) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Amsterdam](https://www.geonames.org/2759794) (52.37403, 4.88969) and [Los Angeles](https://www.geonames.org/5368361) (34.05223, -118.24368). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | ------------- | | WGS-84 geodesic distance | **8,960 km** | | Vacuum RTT floor | **59.77 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **87.74 ms** | | Low-latency fiber material floor | **87.39 ms** | | Engineering floor (5% path allowance) | **92.14 ms** | | Research 1.33× mapped-fiber reference | **116.7 ms** | | Estimated unamplified path loss | **1881.6 dB** | | Transparent optical spans / inline amplifiers | **118 / 117** | | Published RTT inflation over fiber floor | **1.52×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **130.58 ms** | | Average RTT | **133.7 ms** | | Maximum RTT | **141.4 ms** | | Standard deviation | **2.49 ms** | | Stdev / average | **1.9%** | ## Route Context * Departure PoP: [Amsterdam (AMS)](/docs/network/latency/ams-amsterdam) * Destination PoP: [Los Angeles (LAX)](/docs/network/latency/lax-los-angeles) * Region overview: [Europe](/docs/network/latency/regions/europe) → [North America](/docs/network/latency/regions/north-america) * Corridor overview: [Europe to North America](/docs/network/latency/regions/europe-to-north-america) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Los Angeles to Amsterdam latency and RTT](/docs/network/latency/pairs/lax-ams-rtt): 132.1 ms **Fastest routes departing Amsterdam (AMS)** * [Amsterdam to London latency and RTT](/docs/network/latency/pairs/ams-lon-rtt): 5.2 ms * [Amsterdam to Frankfurt latency and RTT](/docs/network/latency/pairs/ams-fra-rtt): 5.9 ms * [Amsterdam to Paris latency and RTT](/docs/network/latency/pairs/ams-par-rtt): 7 ms * [Amsterdam to Berlin latency and RTT](/docs/network/latency/pairs/ams-ber-rtt): 7.2 ms * [Amsterdam to Marseille latency and RTT](/docs/network/latency/pairs/ams-mrs-rtt): 19.6 ms **Fastest routes arriving at Los Angeles (LAX)** * [Seattle to Los Angeles latency and RTT](/docs/network/latency/pairs/sea-lax-rtt): 25.9 ms * [Miami to Los Angeles latency and RTT](/docs/network/latency/pairs/mia-lax-rtt): 56.8 ms * [New York to Los Angeles latency and RTT](/docs/network/latency/pairs/nyc-lax-rtt): 58.9 ms * [Ashburn to Los Angeles latency and RTT](/docs/network/latency/pairs/iad-lax-rtt): 60.7 ms * [Tokyo to Los Angeles latency and RTT](/docs/network/latency/pairs/tyo-lax-rtt): 101.2 ms **Same corridor (Europe → North America)** * [London to New York latency and RTT](/docs/network/latency/pairs/lon-nyc-rtt): 63.8 ms * [Amsterdam to New York latency and RTT](/docs/network/latency/pairs/ams-nyc-rtt): 68.6 ms * [Paris to New York latency and RTT](/docs/network/latency/pairs/par-nyc-rtt): 68.9 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/ams-lax.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Amsterdam → Los Angeles round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/ams-lax-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Amsterdam, Netherlands → London, UK RTT 🇳🇱 **Amsterdam, Netherlands (AMS)** → 🇬🇧 **London, UK (LON)** Measurement round: `2026-08-16T04:07:28Z`. Amsterdam to London averaged 5.2 ms across 50 replies. The observed range was 4.89 to 5.85 ms; these figures describe one probe round, not a long-term service target. This is number 1 of 19 outgoing routes by average RTT. That ranking compares different destinations and is not adjusted for distance. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------- | | RTT | **5.2 ms** | | Jitter | **0.22 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **3.51 ms** | | Fiber Efficiency | **67.4%** | | Latency Tier | Ultra-Low | | Source Region | [Europe](/docs/network/latency/regions/europe) | | Destination Region | [Europe](/docs/network/latency/regions/europe) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Amsterdam](https://www.geonames.org/2759794) (52.37403, 4.88969) and [London](https://www.geonames.org/2643743) (51.50853, -0.12574). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | ----------- | | WGS-84 geodesic distance | **358 km** | | Vacuum RTT floor | **2.39 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **3.51 ms** | | Low-latency fiber material floor | **3.49 ms** | | Engineering floor (5% path allowance) | **3.68 ms** | | Research 1.33× mapped-fiber reference | **4.66 ms** | | Estimated unamplified path loss | **75.2 dB** | | Transparent optical spans / inline amplifiers | **5 / 4** | | Published RTT inflation over fiber floor | **1.48×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ----------- | | Minimum RTT | **4.89 ms** | | Average RTT | **5.2 ms** | | Maximum RTT | **5.85 ms** | | Standard deviation | **0.25 ms** | | Stdev / average | **4.8%** | ## Route Context * Departure PoP: [Amsterdam (AMS)](/docs/network/latency/ams-amsterdam) * Destination PoP: [London (LON)](/docs/network/latency/lon-london) * Region overview: [Europe](/docs/network/latency/regions/europe) → [Europe](/docs/network/latency/regions/europe) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [London to Amsterdam latency and RTT](/docs/network/latency/pairs/lon-ams-rtt): 5.2 ms **Fastest routes departing Amsterdam (AMS)** * [Amsterdam to Frankfurt latency and RTT](/docs/network/latency/pairs/ams-fra-rtt): 5.9 ms * [Amsterdam to Paris latency and RTT](/docs/network/latency/pairs/ams-par-rtt): 7 ms * [Amsterdam to Berlin latency and RTT](/docs/network/latency/pairs/ams-ber-rtt): 7.2 ms * [Amsterdam to Marseille latency and RTT](/docs/network/latency/pairs/ams-mrs-rtt): 19.6 ms * [Amsterdam to Moscow latency and RTT](/docs/network/latency/pairs/ams-mow-rtt): 38.1 ms **Fastest routes arriving at London (LON)** * [Paris to London latency and RTT](/docs/network/latency/pairs/par-lon-rtt): 6.4 ms * [Frankfurt to London latency and RTT](/docs/network/latency/pairs/fra-lon-rtt): 12.9 ms * [Berlin to London latency and RTT](/docs/network/latency/pairs/ber-lon-rtt): 15.2 ms * [Marseille to London latency and RTT](/docs/network/latency/pairs/mrs-lon-rtt): 18.9 ms * [Moscow to London latency and RTT](/docs/network/latency/pairs/mow-lon-rtt): 42.4 ms **Same corridor (Europe → Europe)** * [Frankfurt to Amsterdam latency and RTT](/docs/network/latency/pairs/fra-ams-rtt): 5.9 ms * [Berlin to Frankfurt latency and RTT](/docs/network/latency/pairs/ber-fra-rtt): 6.1 ms * [Frankfurt to Berlin latency and RTT](/docs/network/latency/pairs/fra-ber-rtt): 6.1 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/ams-lon.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Amsterdam → London round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/ams-lon-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Amsterdam, Netherlands → Melbourne, Australia RTT 🇳🇱 **Amsterdam, Netherlands (AMS)** → 🇦🇺 **Melbourne, Australia (MEL)** Measurement round: `2026-08-16T04:07:28Z`. Amsterdam to Melbourne averaged 252.5 ms across 50 replies. The observed range was 235.69 to 297.33 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 12.24 ms, or 4.8% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------------------- | | RTT | **252.5 ms** | | Jitter | **10.79 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **161.98 ms** | | Fiber Efficiency | **64.1%** | | Latency Tier | High | | Source Region | [Europe](/docs/network/latency/regions/europe) | | Destination Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Amsterdam](https://www.geonames.org/2759794) (52.37403, 4.88969) and [Melbourne](https://www.geonames.org/2158177) (-37.81400, 144.96332). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | --------------- | | WGS-84 geodesic distance | **16,540.5 km** | | Vacuum RTT floor | **110.35 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **161.98 ms** | | Low-latency fiber material floor | **161.33 ms** | | Engineering floor (5% path allowance) | **170.1 ms** | | Research 1.33× mapped-fiber reference | **215.43 ms** | | Estimated unamplified path loss | **3473.5 dB** | | Transparent optical spans / inline amplifiers | **218 / 217** | | Published RTT inflation over fiber floor | **1.56×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **235.69 ms** | | Average RTT | **252.5 ms** | | Maximum RTT | **297.33 ms** | | Standard deviation | **12.24 ms** | | Stdev / average | **4.8%** | ## Route Context * Departure PoP: [Amsterdam (AMS)](/docs/network/latency/ams-amsterdam) * Destination PoP: [Melbourne (MEL)](/docs/network/latency/mel-melbourne) * Region overview: [Europe](/docs/network/latency/regions/europe) → [Asia Pacific](/docs/network/latency/regions/asia-pacific) * Corridor overview: [Europe to Asia Pacific](/docs/network/latency/regions/europe-to-asia-pacific) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Melbourne to Amsterdam latency and RTT](/docs/network/latency/pairs/mel-ams-rtt): 252.8 ms **Fastest routes departing Amsterdam (AMS)** * [Amsterdam to London latency and RTT](/docs/network/latency/pairs/ams-lon-rtt): 5.2 ms * [Amsterdam to Frankfurt latency and RTT](/docs/network/latency/pairs/ams-fra-rtt): 5.9 ms * [Amsterdam to Paris latency and RTT](/docs/network/latency/pairs/ams-par-rtt): 7 ms * [Amsterdam to Berlin latency and RTT](/docs/network/latency/pairs/ams-ber-rtt): 7.2 ms * [Amsterdam to Marseille latency and RTT](/docs/network/latency/pairs/ams-mrs-rtt): 19.6 ms **Fastest routes arriving at Melbourne (MEL)** * [Sydney to Melbourne latency and RTT](/docs/network/latency/pairs/syd-mel-rtt): 9.8 ms * [Singapore to Melbourne latency and RTT](/docs/network/latency/pairs/sin-mel-rtt): 88.6 ms * [Tokyo to Melbourne latency and RTT](/docs/network/latency/pairs/tyo-mel-rtt): 113.9 ms * [Hong Kong to Melbourne latency and RTT](/docs/network/latency/pairs/hkg-mel-rtt): 138.4 ms * [Taipei to Melbourne latency and RTT](/docs/network/latency/pairs/tpe-mel-rtt): 142 ms **Same corridor (Europe → Asia Pacific)** * [Moscow to Hong Kong latency and RTT](/docs/network/latency/pairs/mow-hkg-rtt): 118.6 ms * [Moscow to Taipei latency and RTT](/docs/network/latency/pairs/mow-tpe-rtt): 132.1 ms * [Marseille to Singapore latency and RTT](/docs/network/latency/pairs/mrs-sin-rtt): 139.6 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/ams-mel.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Amsterdam → Melbourne round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/ams-mel-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Amsterdam, Netherlands → Miami, USA RTT 🇳🇱 **Amsterdam, Netherlands (AMS)** → 🇺🇸 **Miami, USA (MIA)** Measurement round: `2026-08-16T04:07:28Z`. Amsterdam to Miami averaged 106 ms across 50 replies. The observed range was 99.94 to 119.16 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 4.46 ms, or 4.2% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **106 ms** | | Jitter | **4.12 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **73.03 ms** | | Fiber Efficiency | **68.9%** | | Latency Tier | Good | | Source Region | [Europe](/docs/network/latency/regions/europe) | | Destination Region | [North America](/docs/network/latency/regions/north-america) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Amsterdam](https://www.geonames.org/2759794) (52.37403, 4.88969) and [Miami](https://www.geonames.org/4164138) (25.77427, -80.19366). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **7,457.8 km** | | Vacuum RTT floor | **49.75 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **73.03 ms** | | Low-latency fiber material floor | **72.74 ms** | | Engineering floor (5% path allowance) | **76.69 ms** | | Research 1.33× mapped-fiber reference | **97.13 ms** | | Estimated unamplified path loss | **1566.1 dB** | | Transparent optical spans / inline amplifiers | **98 / 97** | | Published RTT inflation over fiber floor | **1.45×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **99.94 ms** | | Average RTT | **106 ms** | | Maximum RTT | **119.16 ms** | | Standard deviation | **4.46 ms** | | Stdev / average | **4.2%** | ## Route Context * Departure PoP: [Amsterdam (AMS)](/docs/network/latency/ams-amsterdam) * Destination PoP: [Miami (MIA)](/docs/network/latency/mia-miami) * Region overview: [Europe](/docs/network/latency/regions/europe) → [North America](/docs/network/latency/regions/north-america) * Corridor overview: [Europe to North America](/docs/network/latency/regions/europe-to-north-america) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Miami to Amsterdam latency and RTT](/docs/network/latency/pairs/mia-ams-rtt): 106 ms **Fastest routes departing Amsterdam (AMS)** * [Amsterdam to London latency and RTT](/docs/network/latency/pairs/ams-lon-rtt): 5.2 ms * [Amsterdam to Frankfurt latency and RTT](/docs/network/latency/pairs/ams-fra-rtt): 5.9 ms * [Amsterdam to Paris latency and RTT](/docs/network/latency/pairs/ams-par-rtt): 7 ms * [Amsterdam to Berlin latency and RTT](/docs/network/latency/pairs/ams-ber-rtt): 7.2 ms * [Amsterdam to Marseille latency and RTT](/docs/network/latency/pairs/ams-mrs-rtt): 19.6 ms **Fastest routes arriving at Miami (MIA)** * [Ashburn to Miami latency and RTT](/docs/network/latency/pairs/iad-mia-rtt): 27.4 ms * [New York to Miami latency and RTT](/docs/network/latency/pairs/nyc-mia-rtt): 32.9 ms * [Los Angeles to Miami latency and RTT](/docs/network/latency/pairs/lax-mia-rtt): 57.3 ms * [São Paulo to Miami latency and RTT](/docs/network/latency/pairs/gru-mia-rtt): 74 ms * [Seattle to Miami latency and RTT](/docs/network/latency/pairs/sea-mia-rtt): 81.7 ms **Same corridor (Europe → North America)** * [London to New York latency and RTT](/docs/network/latency/pairs/lon-nyc-rtt): 63.8 ms * [Amsterdam to New York latency and RTT](/docs/network/latency/pairs/ams-nyc-rtt): 68.6 ms * [Paris to New York latency and RTT](/docs/network/latency/pairs/par-nyc-rtt): 68.9 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/ams-mia.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Amsterdam → Miami round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/ams-mia-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Amsterdam, Netherlands → Moscow, Russia RTT 🇳🇱 **Amsterdam, Netherlands (AMS)** → 🇷🇺 **Moscow, Russia (MOW)** Measurement round: `2026-08-16T04:07:28Z`. Amsterdam to Moscow averaged 38.1 ms across 50 replies. The observed range was 35.69 to 44.57 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 1.9 ms, or 5.0% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------- | | RTT | **38.1 ms** | | Jitter | **1.52 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **21.1 ms** | | Fiber Efficiency | **55.4%** | | Latency Tier | Excellent | | Source Region | [Europe](/docs/network/latency/regions/europe) | | Destination Region | [Europe](/docs/network/latency/regions/europe) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Amsterdam](https://www.geonames.org/2759794) (52.37403, 4.88969) and [Moscow](https://www.geonames.org/524901) (55.75204, 37.61781). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **2,154.6 km** | | Vacuum RTT floor | **14.37 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **21.1 ms** | | Low-latency fiber material floor | **21.02 ms** | | Engineering floor (5% path allowance) | **22.16 ms** | | Research 1.33× mapped-fiber reference | **28.06 ms** | | Estimated unamplified path loss | **452.5 dB** | | Transparent optical spans / inline amplifiers | **29 / 28** | | Published RTT inflation over fiber floor | **1.81×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------ | | Minimum RTT | **35.69 ms** | | Average RTT | **38.1 ms** | | Maximum RTT | **44.57 ms** | | Standard deviation | **1.9 ms** | | Stdev / average | **5.0%** | ## Route Context * Departure PoP: [Amsterdam (AMS)](/docs/network/latency/ams-amsterdam) * Destination PoP: [Moscow (MOW)](/docs/network/latency/mow-moscow) * Region overview: [Europe](/docs/network/latency/regions/europe) → [Europe](/docs/network/latency/regions/europe) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Moscow to Amsterdam latency and RTT](/docs/network/latency/pairs/mow-ams-rtt): 37.6 ms **Fastest routes departing Amsterdam (AMS)** * [Amsterdam to London latency and RTT](/docs/network/latency/pairs/ams-lon-rtt): 5.2 ms * [Amsterdam to Frankfurt latency and RTT](/docs/network/latency/pairs/ams-fra-rtt): 5.9 ms * [Amsterdam to Paris latency and RTT](/docs/network/latency/pairs/ams-par-rtt): 7 ms * [Amsterdam to Berlin latency and RTT](/docs/network/latency/pairs/ams-ber-rtt): 7.2 ms * [Amsterdam to Marseille latency and RTT](/docs/network/latency/pairs/ams-mrs-rtt): 19.6 ms **Fastest routes arriving at Moscow (MOW)** * [Berlin to Moscow latency and RTT](/docs/network/latency/pairs/ber-mow-rtt): 28.6 ms * [Frankfurt to Moscow latency and RTT](/docs/network/latency/pairs/fra-mow-rtt): 34.7 ms * [London to Moscow latency and RTT](/docs/network/latency/pairs/lon-mow-rtt): 43.5 ms * [Paris to Moscow latency and RTT](/docs/network/latency/pairs/par-mow-rtt): 44.7 ms * [Marseille to Moscow latency and RTT](/docs/network/latency/pairs/mrs-mow-rtt): 49.6 ms **Same corridor (Europe → Europe)** * [London to Amsterdam latency and RTT](/docs/network/latency/pairs/lon-ams-rtt): 5.2 ms * [Frankfurt to Amsterdam latency and RTT](/docs/network/latency/pairs/fra-ams-rtt): 5.9 ms * [Berlin to Frankfurt latency and RTT](/docs/network/latency/pairs/ber-fra-rtt): 6.1 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/ams-mow.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Amsterdam → Moscow round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/ams-mow-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Amsterdam, Netherlands → Marseille, France RTT 🇳🇱 **Amsterdam, Netherlands (AMS)** → 🇫🇷 **Marseille, France (MRS)** Measurement round: `2026-08-16T04:07:28Z`. Amsterdam to Marseille averaged 19.6 ms across 50 replies. The observed range was 19.03 to 20.95 ms; these figures describe one probe round, not a long-term service target. This is number 5 of 19 outgoing routes by average RTT. That ranking compares different destinations and is not adjusted for distance. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------- | | RTT | **19.6 ms** | | Jitter | **0.41 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **9.89 ms** | | Fiber Efficiency | **50.5%** | | Latency Tier | Ultra-Low | | Source Region | [Europe](/docs/network/latency/regions/europe) | | Destination Region | [Europe](/docs/network/latency/regions/europe) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Amsterdam](https://www.geonames.org/2759794) (52.37403, 4.88969) and [Marseille](https://www.geonames.org/2995469) (43.29695, 5.38107). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **1,009.9 km** | | Vacuum RTT floor | **6.74 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **9.89 ms** | | Low-latency fiber material floor | **9.85 ms** | | Engineering floor (5% path allowance) | **10.39 ms** | | Research 1.33× mapped-fiber reference | **13.15 ms** | | Estimated unamplified path loss | **212.1 dB** | | Transparent optical spans / inline amplifiers | **14 / 13** | | Published RTT inflation over fiber floor | **1.98×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------ | | Minimum RTT | **19.03 ms** | | Average RTT | **19.6 ms** | | Maximum RTT | **20.95 ms** | | Standard deviation | **0.5 ms** | | Stdev / average | **2.6%** | ## Route Context * Departure PoP: [Amsterdam (AMS)](/docs/network/latency/ams-amsterdam) * Destination PoP: [Marseille (MRS)](/docs/network/latency/mrs-marseille) * Region overview: [Europe](/docs/network/latency/regions/europe) → [Europe](/docs/network/latency/regions/europe) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Marseille to Amsterdam latency and RTT](/docs/network/latency/pairs/mrs-ams-rtt): 20.3 ms **Fastest routes departing Amsterdam (AMS)** * [Amsterdam to London latency and RTT](/docs/network/latency/pairs/ams-lon-rtt): 5.2 ms * [Amsterdam to Frankfurt latency and RTT](/docs/network/latency/pairs/ams-fra-rtt): 5.9 ms * [Amsterdam to Paris latency and RTT](/docs/network/latency/pairs/ams-par-rtt): 7 ms * [Amsterdam to Berlin latency and RTT](/docs/network/latency/pairs/ams-ber-rtt): 7.2 ms * [Amsterdam to Moscow latency and RTT](/docs/network/latency/pairs/ams-mow-rtt): 38.1 ms **Fastest routes arriving at Marseille (MRS)** * [Paris to Marseille latency and RTT](/docs/network/latency/pairs/par-mrs-rtt): 8.9 ms * [Frankfurt to Marseille latency and RTT](/docs/network/latency/pairs/fra-mrs-rtt): 16.4 ms * [London to Marseille latency and RTT](/docs/network/latency/pairs/lon-mrs-rtt): 17.9 ms * [Berlin to Marseille latency and RTT](/docs/network/latency/pairs/ber-mrs-rtt): 20.2 ms * [Moscow to Marseille latency and RTT](/docs/network/latency/pairs/mow-mrs-rtt): 49.9 ms **Same corridor (Europe → Europe)** * [London to Amsterdam latency and RTT](/docs/network/latency/pairs/lon-ams-rtt): 5.2 ms * [Frankfurt to Amsterdam latency and RTT](/docs/network/latency/pairs/fra-ams-rtt): 5.9 ms * [Berlin to Frankfurt latency and RTT](/docs/network/latency/pairs/ber-fra-rtt): 6.1 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/ams-mrs.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Amsterdam → Marseille round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/ams-mrs-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Amsterdam, Netherlands → New York, USA RTT 🇳🇱 **Amsterdam, Netherlands (AMS)** → 🇺🇸 **New York, USA (NYC)** Measurement round: `2026-08-16T04:07:28Z`. Amsterdam to New York averaged 68.6 ms across 50 replies. The observed range was 67.28 to 73 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 1.16 ms, or 1.7% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **68.6 ms** | | Jitter | **1.05 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **57.56 ms** | | Fiber Efficiency | **83.9%** | | Latency Tier | Excellent | | Source Region | [Europe](/docs/network/latency/regions/europe) | | Destination Region | [North America](/docs/network/latency/regions/north-america) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Amsterdam](https://www.geonames.org/2759794) (52.37403, 4.88969) and [New York City](https://www.geonames.org/5128581) (40.71427, -74.00597). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **5,877.8 km** | | Vacuum RTT floor | **39.21 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **57.56 ms** | | Low-latency fiber material floor | **57.33 ms** | | Engineering floor (5% path allowance) | **60.45 ms** | | Research 1.33× mapped-fiber reference | **76.56 ms** | | Estimated unamplified path loss | **1234.3 dB** | | Transparent optical spans / inline amplifiers | **78 / 77** | | Published RTT inflation over fiber floor | **1.19×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------ | | Minimum RTT | **67.28 ms** | | Average RTT | **68.6 ms** | | Maximum RTT | **73 ms** | | Standard deviation | **1.16 ms** | | Stdev / average | **1.7%** | ## Route Context * Departure PoP: [Amsterdam (AMS)](/docs/network/latency/ams-amsterdam) * Destination PoP: [New York (NYC)](/docs/network/latency/nyc-new-york) * Region overview: [Europe](/docs/network/latency/regions/europe) → [North America](/docs/network/latency/regions/north-america) * Corridor overview: [Europe to North America](/docs/network/latency/regions/europe-to-north-america) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [New York to Amsterdam latency and RTT](/docs/network/latency/pairs/nyc-ams-rtt): 68.7 ms **Fastest routes departing Amsterdam (AMS)** * [Amsterdam to London latency and RTT](/docs/network/latency/pairs/ams-lon-rtt): 5.2 ms * [Amsterdam to Frankfurt latency and RTT](/docs/network/latency/pairs/ams-fra-rtt): 5.9 ms * [Amsterdam to Paris latency and RTT](/docs/network/latency/pairs/ams-par-rtt): 7 ms * [Amsterdam to Berlin latency and RTT](/docs/network/latency/pairs/ams-ber-rtt): 7.2 ms * [Amsterdam to Marseille latency and RTT](/docs/network/latency/pairs/ams-mrs-rtt): 19.6 ms **Fastest routes arriving at New York (NYC)** * [Ashburn to New York latency and RTT](/docs/network/latency/pairs/iad-nyc-rtt): 6 ms * [Miami to New York latency and RTT](/docs/network/latency/pairs/mia-nyc-rtt): 33.3 ms * [Los Angeles to New York latency and RTT](/docs/network/latency/pairs/lax-nyc-rtt): 57.9 ms * [Seattle to New York latency and RTT](/docs/network/latency/pairs/sea-nyc-rtt): 59.5 ms * [London to New York latency and RTT](/docs/network/latency/pairs/lon-nyc-rtt): 63.8 ms **Same corridor (Europe → North America)** * [Paris to New York latency and RTT](/docs/network/latency/pairs/par-nyc-rtt): 68.9 ms * [London to Ashburn latency and RTT](/docs/network/latency/pairs/lon-iad-rtt): 71 ms * [Frankfurt to New York latency and RTT](/docs/network/latency/pairs/fra-nyc-rtt): 74.4 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/ams-nyc.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Amsterdam → New York round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/ams-nyc-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Amsterdam, Netherlands → Paris, France RTT 🇳🇱 **Amsterdam, Netherlands (AMS)** → 🇫🇷 **Paris, France (PAR)** Measurement round: `2026-08-16T04:07:28Z`. Amsterdam to Paris averaged 7 ms across 50 replies. The observed range was 6.61 to 8.02 ms; these figures describe one probe round, not a long-term service target. This is number 3 of 19 outgoing routes by average RTT. That ranking compares different destinations and is not adjusted for distance. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------- | | RTT | **7 ms** | | Jitter | **0.27 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **4.22 ms** | | Fiber Efficiency | **60.3%** | | Latency Tier | Ultra-Low | | Source Region | [Europe](/docs/network/latency/regions/europe) | | Destination Region | [Europe](/docs/network/latency/regions/europe) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Amsterdam](https://www.geonames.org/2759794) (52.37403, 4.88969) and [Paris](https://www.geonames.org/2988507) (48.85341, 2.34880). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | ------------ | | WGS-84 geodesic distance | **430.9 km** | | Vacuum RTT floor | **2.87 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **4.22 ms** | | Low-latency fiber material floor | **4.2 ms** | | Engineering floor (5% path allowance) | **4.43 ms** | | Research 1.33× mapped-fiber reference | **5.61 ms** | | Estimated unamplified path loss | **90.5 dB** | | Transparent optical spans / inline amplifiers | **6 / 5** | | Published RTT inflation over fiber floor | **1.66×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ----------- | | Minimum RTT | **6.61 ms** | | Average RTT | **7 ms** | | Maximum RTT | **8.02 ms** | | Standard deviation | **0.32 ms** | | Stdev / average | **4.6%** | ## Route Context * Departure PoP: [Amsterdam (AMS)](/docs/network/latency/ams-amsterdam) * Destination PoP: [Paris (PAR)](/docs/network/latency/par-paris) * Region overview: [Europe](/docs/network/latency/regions/europe) → [Europe](/docs/network/latency/regions/europe) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Paris to Amsterdam latency and RTT](/docs/network/latency/pairs/par-ams-rtt): 7 ms **Fastest routes departing Amsterdam (AMS)** * [Amsterdam to London latency and RTT](/docs/network/latency/pairs/ams-lon-rtt): 5.2 ms * [Amsterdam to Frankfurt latency and RTT](/docs/network/latency/pairs/ams-fra-rtt): 5.9 ms * [Amsterdam to Berlin latency and RTT](/docs/network/latency/pairs/ams-ber-rtt): 7.2 ms * [Amsterdam to Marseille latency and RTT](/docs/network/latency/pairs/ams-mrs-rtt): 19.6 ms * [Amsterdam to Moscow latency and RTT](/docs/network/latency/pairs/ams-mow-rtt): 38.1 ms **Fastest routes arriving at Paris (PAR)** * [London to Paris latency and RTT](/docs/network/latency/pairs/lon-par-rtt): 6.4 ms * [Frankfurt to Paris latency and RTT](/docs/network/latency/pairs/fra-par-rtt): 7.6 ms * [Marseille to Paris latency and RTT](/docs/network/latency/pairs/mrs-par-rtt): 8.9 ms * [Berlin to Paris latency and RTT](/docs/network/latency/pairs/ber-par-rtt): 15.6 ms * [Moscow to Paris latency and RTT](/docs/network/latency/pairs/mow-par-rtt): 44 ms **Same corridor (Europe → Europe)** * [London to Amsterdam latency and RTT](/docs/network/latency/pairs/lon-ams-rtt): 5.2 ms * [Frankfurt to Amsterdam latency and RTT](/docs/network/latency/pairs/fra-ams-rtt): 5.9 ms * [Berlin to Frankfurt latency and RTT](/docs/network/latency/pairs/ber-fra-rtt): 6.1 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/ams-par.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Amsterdam → Paris round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/ams-par-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Amsterdam, Netherlands → Seattle, USA RTT 🇳🇱 **Amsterdam, Netherlands (AMS)** → 🇺🇸 **Seattle, USA (SEA)** Measurement round: `2026-08-16T04:07:28Z`. Amsterdam to Seattle averaged 130.2 ms across 50 replies. The observed range was 123.59 to 144.11 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 5.22 ms, or 4.0% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **130.2 ms** | | Jitter | **4.46 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **76.86 ms** | | Fiber Efficiency | **59%** | | Latency Tier | Good | | Source Region | [Europe](/docs/network/latency/regions/europe) | | Destination Region | [North America](/docs/network/latency/regions/north-america) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Amsterdam](https://www.geonames.org/2759794) (52.37403, 4.88969) and [Seattle](https://www.geonames.org/5809844) (47.60621, -122.33207). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **7,848.5 km** | | Vacuum RTT floor | **52.36 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **76.86 ms** | | Low-latency fiber material floor | **76.55 ms** | | Engineering floor (5% path allowance) | **80.71 ms** | | Research 1.33× mapped-fiber reference | **102.22 ms** | | Estimated unamplified path loss | **1648.2 dB** | | Transparent optical spans / inline amplifiers | **104 / 103** | | Published RTT inflation over fiber floor | **1.69×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **123.59 ms** | | Average RTT | **130.2 ms** | | Maximum RTT | **144.11 ms** | | Standard deviation | **5.22 ms** | | Stdev / average | **4.0%** | ## Route Context * Departure PoP: [Amsterdam (AMS)](/docs/network/latency/ams-amsterdam) * Destination PoP: [Seattle (SEA)](/docs/network/latency/sea-seattle) * Region overview: [Europe](/docs/network/latency/regions/europe) → [North America](/docs/network/latency/regions/north-america) * Corridor overview: [Europe to North America](/docs/network/latency/regions/europe-to-north-america) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Seattle to Amsterdam latency and RTT](/docs/network/latency/pairs/sea-ams-rtt): 128.6 ms **Fastest routes departing Amsterdam (AMS)** * [Amsterdam to London latency and RTT](/docs/network/latency/pairs/ams-lon-rtt): 5.2 ms * [Amsterdam to Frankfurt latency and RTT](/docs/network/latency/pairs/ams-fra-rtt): 5.9 ms * [Amsterdam to Paris latency and RTT](/docs/network/latency/pairs/ams-par-rtt): 7 ms * [Amsterdam to Berlin latency and RTT](/docs/network/latency/pairs/ams-ber-rtt): 7.2 ms * [Amsterdam to Marseille latency and RTT](/docs/network/latency/pairs/ams-mrs-rtt): 19.6 ms **Fastest routes arriving at Seattle (SEA)** * [Los Angeles to Seattle latency and RTT](/docs/network/latency/pairs/lax-sea-rtt): 26.9 ms * [New York to Seattle latency and RTT](/docs/network/latency/pairs/nyc-sea-rtt): 58.7 ms * [Ashburn to Seattle latency and RTT](/docs/network/latency/pairs/iad-sea-rtt): 62.3 ms * [Miami to Seattle latency and RTT](/docs/network/latency/pairs/mia-sea-rtt): 81.9 ms * [Tokyo to Seattle latency and RTT](/docs/network/latency/pairs/tyo-sea-rtt): 84.8 ms **Same corridor (Europe → North America)** * [London to New York latency and RTT](/docs/network/latency/pairs/lon-nyc-rtt): 63.8 ms * [Amsterdam to New York latency and RTT](/docs/network/latency/pairs/ams-nyc-rtt): 68.6 ms * [Paris to New York latency and RTT](/docs/network/latency/pairs/par-nyc-rtt): 68.9 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/ams-sea.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Amsterdam → Seattle round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/ams-sea-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Amsterdam, Netherlands → Singapore RTT 🇳🇱 **Amsterdam, Netherlands (AMS)** → 🇸🇬 **Singapore (SIN)** Measurement round: `2026-08-16T04:07:28Z`. Amsterdam to Singapore averaged 157.6 ms across 50 replies. The observed range was 150.92 to 175.59 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 6.2 ms, or 3.9% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------------------- | | RTT | **157.6 ms** | | Jitter | **4.44 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **102.86 ms** | | Fiber Efficiency | **65.3%** | | Latency Tier | Fair | | Source Region | [Europe](/docs/network/latency/regions/europe) | | Destination Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Amsterdam](https://www.geonames.org/2759794) (52.37403, 4.88969) and [Singapore](https://www.geonames.org/1880252) (1.28967, 103.85007). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | --------------- | | WGS-84 geodesic distance | **10,503.4 km** | | Vacuum RTT floor | **70.07 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **102.86 ms** | | Low-latency fiber material floor | **102.44 ms** | | Engineering floor (5% path allowance) | **108.01 ms** | | Research 1.33× mapped-fiber reference | **136.8 ms** | | Estimated unamplified path loss | **2205.7 dB** | | Transparent optical spans / inline amplifiers | **138 / 137** | | Published RTT inflation over fiber floor | **1.53×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **150.92 ms** | | Average RTT | **157.6 ms** | | Maximum RTT | **175.59 ms** | | Standard deviation | **6.2 ms** | | Stdev / average | **3.9%** | ## Route Context * Departure PoP: [Amsterdam (AMS)](/docs/network/latency/ams-amsterdam) * Destination PoP: [Singapore (SIN)](/docs/network/latency/sin-singapore) * Region overview: [Europe](/docs/network/latency/regions/europe) → [Asia Pacific](/docs/network/latency/regions/asia-pacific) * Corridor overview: [Europe to Asia Pacific](/docs/network/latency/regions/europe-to-asia-pacific) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Singapore to Amsterdam latency and RTT](/docs/network/latency/pairs/sin-ams-rtt): 155.8 ms **Fastest routes departing Amsterdam (AMS)** * [Amsterdam to London latency and RTT](/docs/network/latency/pairs/ams-lon-rtt): 5.2 ms * [Amsterdam to Frankfurt latency and RTT](/docs/network/latency/pairs/ams-fra-rtt): 5.9 ms * [Amsterdam to Paris latency and RTT](/docs/network/latency/pairs/ams-par-rtt): 7 ms * [Amsterdam to Berlin latency and RTT](/docs/network/latency/pairs/ams-ber-rtt): 7.2 ms * [Amsterdam to Marseille latency and RTT](/docs/network/latency/pairs/ams-mrs-rtt): 19.6 ms **Fastest routes arriving at Singapore (SIN)** * [Hong Kong to Singapore latency and RTT](/docs/network/latency/pairs/hkg-sin-rtt): 31.5 ms * [Taipei to Singapore latency and RTT](/docs/network/latency/pairs/tpe-sin-rtt): 44.3 ms * [Tokyo to Singapore latency and RTT](/docs/network/latency/pairs/tyo-sin-rtt): 69.4 ms * [Melbourne to Singapore latency and RTT](/docs/network/latency/pairs/mel-sin-rtt): 88.4 ms * [Sydney to Singapore latency and RTT](/docs/network/latency/pairs/syd-sin-rtt): 94.5 ms **Same corridor (Europe → Asia Pacific)** * [Moscow to Hong Kong latency and RTT](/docs/network/latency/pairs/mow-hkg-rtt): 118.6 ms * [Moscow to Taipei latency and RTT](/docs/network/latency/pairs/mow-tpe-rtt): 132.1 ms * [Marseille to Singapore latency and RTT](/docs/network/latency/pairs/mrs-sin-rtt): 139.6 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/ams-sin.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Amsterdam → Singapore round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/ams-sin-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Amsterdam, Netherlands → Sydney, Australia RTT 🇳🇱 **Amsterdam, Netherlands (AMS)** → 🇦🇺 **Sydney, Australia (SYD)** Measurement round: `2026-08-16T04:07:28Z`. Amsterdam to Sydney averaged 258 ms across 50 replies. The observed range was 246.95 to 290.31 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 9.9 ms, or 3.8% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------------------- | | RTT | **258 ms** | | Jitter | **7.93 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **162.93 ms** | | Fiber Efficiency | **63.2%** | | Latency Tier | High | | Source Region | [Europe](/docs/network/latency/regions/europe) | | Destination Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Amsterdam](https://www.geonames.org/2759794) (52.37403, 4.88969) and [Sydney](https://www.geonames.org/2147714) (-33.86785, 151.20732). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | --------------- | | WGS-84 geodesic distance | **16,638.1 km** | | Vacuum RTT floor | **111 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **162.93 ms** | | Low-latency fiber material floor | **162.28 ms** | | Engineering floor (5% path allowance) | **171.1 ms** | | Research 1.33× mapped-fiber reference | **216.7 ms** | | Estimated unamplified path loss | **3494 dB** | | Transparent optical spans / inline amplifiers | **219 / 218** | | Published RTT inflation over fiber floor | **1.58×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **246.95 ms** | | Average RTT | **258 ms** | | Maximum RTT | **290.31 ms** | | Standard deviation | **9.9 ms** | | Stdev / average | **3.8%** | ## Route Context * Departure PoP: [Amsterdam (AMS)](/docs/network/latency/ams-amsterdam) * Destination PoP: [Sydney (SYD)](/docs/network/latency/syd-sydney) * Region overview: [Europe](/docs/network/latency/regions/europe) → [Asia Pacific](/docs/network/latency/regions/asia-pacific) * Corridor overview: [Europe to Asia Pacific](/docs/network/latency/regions/europe-to-asia-pacific) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Sydney to Amsterdam latency and RTT](/docs/network/latency/pairs/syd-ams-rtt): 258.9 ms **Fastest routes departing Amsterdam (AMS)** * [Amsterdam to London latency and RTT](/docs/network/latency/pairs/ams-lon-rtt): 5.2 ms * [Amsterdam to Frankfurt latency and RTT](/docs/network/latency/pairs/ams-fra-rtt): 5.9 ms * [Amsterdam to Paris latency and RTT](/docs/network/latency/pairs/ams-par-rtt): 7 ms * [Amsterdam to Berlin latency and RTT](/docs/network/latency/pairs/ams-ber-rtt): 7.2 ms * [Amsterdam to Marseille latency and RTT](/docs/network/latency/pairs/ams-mrs-rtt): 19.6 ms **Fastest routes arriving at Sydney (SYD)** * [Melbourne to Sydney latency and RTT](/docs/network/latency/pairs/mel-syd-rtt): 9.8 ms * [Singapore to Sydney latency and RTT](/docs/network/latency/pairs/sin-syd-rtt): 94.5 ms * [Tokyo to Sydney latency and RTT](/docs/network/latency/pairs/tyo-syd-rtt): 101.9 ms * [Taipei to Sydney latency and RTT](/docs/network/latency/pairs/tpe-syd-rtt): 133.2 ms * [Hong Kong to Sydney latency and RTT](/docs/network/latency/pairs/hkg-syd-rtt): 137.2 ms **Same corridor (Europe → Asia Pacific)** * [Moscow to Hong Kong latency and RTT](/docs/network/latency/pairs/mow-hkg-rtt): 118.6 ms * [Moscow to Taipei latency and RTT](/docs/network/latency/pairs/mow-tpe-rtt): 132.1 ms * [Marseille to Singapore latency and RTT](/docs/network/latency/pairs/mrs-sin-rtt): 139.6 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/ams-syd.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Amsterdam → Sydney round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/ams-syd-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Amsterdam, Netherlands → Taipei, Taiwan RTT 🇳🇱 **Amsterdam, Netherlands (AMS)** → 🇹🇼 **Taipei, Taiwan (TPE)** Measurement round: `2026-08-16T04:07:28Z`. Amsterdam to Taipei averaged 168.7 ms across 50 replies. The observed range was 163.09 to 184.23 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 4.36 ms, or 2.6% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------------------- | | RTT | **168.7 ms** | | Jitter | **3.84 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **92.72 ms** | | Fiber Efficiency | **55%** | | Latency Tier | Fair | | Source Region | [Europe](/docs/network/latency/regions/europe) | | Destination Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Amsterdam](https://www.geonames.org/2759794) (52.37403, 4.88969) and [Taipei](https://www.geonames.org/1668341) (25.05306, 121.52639). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **9,467.8 km** | | Vacuum RTT floor | **63.16 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **92.72 ms** | | Low-latency fiber material floor | **92.34 ms** | | Engineering floor (5% path allowance) | **97.36 ms** | | Research 1.33× mapped-fiber reference | **123.31 ms** | | Estimated unamplified path loss | **1988.2 dB** | | Transparent optical spans / inline amplifiers | **125 / 124** | | Published RTT inflation over fiber floor | **1.82×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **163.09 ms** | | Average RTT | **168.7 ms** | | Maximum RTT | **184.23 ms** | | Standard deviation | **4.36 ms** | | Stdev / average | **2.6%** | ## Route Context * Departure PoP: [Amsterdam (AMS)](/docs/network/latency/ams-amsterdam) * Destination PoP: [Taipei (TPE)](/docs/network/latency/tpe-taipei) * Region overview: [Europe](/docs/network/latency/regions/europe) → [Asia Pacific](/docs/network/latency/regions/asia-pacific) * Corridor overview: [Europe to Asia Pacific](/docs/network/latency/regions/europe-to-asia-pacific) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Taipei to Amsterdam latency and RTT](/docs/network/latency/pairs/tpe-ams-rtt): 166.5 ms **Fastest routes departing Amsterdam (AMS)** * [Amsterdam to London latency and RTT](/docs/network/latency/pairs/ams-lon-rtt): 5.2 ms * [Amsterdam to Frankfurt latency and RTT](/docs/network/latency/pairs/ams-fra-rtt): 5.9 ms * [Amsterdam to Paris latency and RTT](/docs/network/latency/pairs/ams-par-rtt): 7 ms * [Amsterdam to Berlin latency and RTT](/docs/network/latency/pairs/ams-ber-rtt): 7.2 ms * [Amsterdam to Marseille latency and RTT](/docs/network/latency/pairs/ams-mrs-rtt): 19.6 ms **Fastest routes arriving at Taipei (TPE)** * [Hong Kong to Taipei latency and RTT](/docs/network/latency/pairs/hkg-tpe-rtt): 14.7 ms * [Tokyo to Taipei latency and RTT](/docs/network/latency/pairs/tyo-tpe-rtt): 32.3 ms * [Singapore to Taipei latency and RTT](/docs/network/latency/pairs/sin-tpe-rtt): 45.8 ms * [Seattle to Taipei latency and RTT](/docs/network/latency/pairs/sea-tpe-rtt): 114.7 ms * [Sydney to Taipei latency and RTT](/docs/network/latency/pairs/syd-tpe-rtt): 131.7 ms **Same corridor (Europe → Asia Pacific)** * [Moscow to Hong Kong latency and RTT](/docs/network/latency/pairs/mow-hkg-rtt): 118.6 ms * [Moscow to Taipei latency and RTT](/docs/network/latency/pairs/mow-tpe-rtt): 132.1 ms * [Marseille to Singapore latency and RTT](/docs/network/latency/pairs/mrs-sin-rtt): 139.6 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/ams-tpe.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Amsterdam → Taipei round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/ams-tpe-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Amsterdam, Netherlands → Tokyo, Japan RTT 🇳🇱 **Amsterdam, Netherlands (AMS)** → 🇯🇵 **Tokyo, Japan (TYO)** Measurement round: `2026-08-16T04:07:28Z`. Amsterdam to Tokyo averaged 198.2 ms, compared with a 91.18 ms city-centre fiber reference. The difference cannot be assigned to a particular cable or carrier from RTT alone. Standard deviation was 3.99 ms, or 2.0% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------------------- | | RTT | **198.2 ms** | | Jitter | **2.6 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **91.18 ms** | | Fiber Efficiency | **46%** | | Latency Tier | Fair | | Source Region | [Europe](/docs/network/latency/regions/europe) | | Destination Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Amsterdam](https://www.geonames.org/2759794) (52.37403, 4.88969) and [Tokyo](https://www.geonames.org/1850147) (35.68950, 139.69171). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **9,311.4 km** | | Vacuum RTT floor | **62.12 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **91.18 ms** | | Low-latency fiber material floor | **90.82 ms** | | Engineering floor (5% path allowance) | **95.76 ms** | | Research 1.33× mapped-fiber reference | **121.28 ms** | | Estimated unamplified path loss | **1955.4 dB** | | Transparent optical spans / inline amplifiers | **123 / 122** | | Published RTT inflation over fiber floor | **2.17×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **194.91 ms** | | Average RTT | **198.2 ms** | | Maximum RTT | **216.11 ms** | | Standard deviation | **3.99 ms** | | Stdev / average | **2.0%** | ## Route Context * Departure PoP: [Amsterdam (AMS)](/docs/network/latency/ams-amsterdam) * Destination PoP: [Tokyo (TYO)](/docs/network/latency/tyo-tokyo) * Region overview: [Europe](/docs/network/latency/regions/europe) → [Asia Pacific](/docs/network/latency/regions/asia-pacific) * Corridor overview: [Europe to Asia Pacific](/docs/network/latency/regions/europe-to-asia-pacific) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Tokyo to Amsterdam latency and RTT](/docs/network/latency/pairs/tyo-ams-rtt): 197.4 ms **Fastest routes departing Amsterdam (AMS)** * [Amsterdam to London latency and RTT](/docs/network/latency/pairs/ams-lon-rtt): 5.2 ms * [Amsterdam to Frankfurt latency and RTT](/docs/network/latency/pairs/ams-fra-rtt): 5.9 ms * [Amsterdam to Paris latency and RTT](/docs/network/latency/pairs/ams-par-rtt): 7 ms * [Amsterdam to Berlin latency and RTT](/docs/network/latency/pairs/ams-ber-rtt): 7.2 ms * [Amsterdam to Marseille latency and RTT](/docs/network/latency/pairs/ams-mrs-rtt): 19.6 ms **Fastest routes arriving at Tokyo (TYO)** * [Taipei to Tokyo latency and RTT](/docs/network/latency/pairs/tpe-tyo-rtt): 31.9 ms * [Hong Kong to Tokyo latency and RTT](/docs/network/latency/pairs/hkg-tyo-rtt): 44.9 ms * [Singapore to Tokyo latency and RTT](/docs/network/latency/pairs/sin-tyo-rtt): 68.9 ms * [Seattle to Tokyo latency and RTT](/docs/network/latency/pairs/sea-tyo-rtt): 85 ms * [Los Angeles to Tokyo latency and RTT](/docs/network/latency/pairs/lax-tyo-rtt): 101.2 ms **Same corridor (Europe → Asia Pacific)** * [Moscow to Hong Kong latency and RTT](/docs/network/latency/pairs/mow-hkg-rtt): 118.6 ms * [Moscow to Taipei latency and RTT](/docs/network/latency/pairs/mow-tpe-rtt): 132.1 ms * [Marseille to Singapore latency and RTT](/docs/network/latency/pairs/mrs-sin-rtt): 139.6 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/ams-tyo.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Amsterdam → Tokyo round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/ams-tyo-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Berlin, Germany → Amsterdam, Netherlands RTT 🇩🇪 **Berlin, Germany (BER)** → 🇳🇱 **Amsterdam, Netherlands (AMS)** Measurement round: `2026-08-16T04:07:28Z`. Berlin to Amsterdam averaged 7.2 ms across 50 replies. The observed range was 6.91 to 7.91 ms; these figures describe one probe round, not a long-term service target. This is number 2 of 19 outgoing routes by average RTT. That ranking compares different destinations and is not adjusted for distance. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------- | | RTT | **7.2 ms** | | Jitter | **0.22 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **5.67 ms** | | Fiber Efficiency | **78.8%** | | Latency Tier | Ultra-Low | | Source Region | [Europe](/docs/network/latency/regions/europe) | | Destination Region | [Europe](/docs/network/latency/regions/europe) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Berlin](https://www.geonames.org/2950159) (52.52437, 13.41053) and [Amsterdam](https://www.geonames.org/2759794) (52.37403, 4.88969). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | ------------ | | WGS-84 geodesic distance | **579.2 km** | | Vacuum RTT floor | **3.86 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **5.67 ms** | | Low-latency fiber material floor | **5.65 ms** | | Engineering floor (5% path allowance) | **5.96 ms** | | Research 1.33× mapped-fiber reference | **7.54 ms** | | Estimated unamplified path loss | **121.6 dB** | | Transparent optical spans / inline amplifiers | **8 / 7** | | Published RTT inflation over fiber floor | **1.27×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ----------- | | Minimum RTT | **6.91 ms** | | Average RTT | **7.2 ms** | | Maximum RTT | **7.91 ms** | | Standard deviation | **0.23 ms** | | Stdev / average | **3.2%** | ## Route Context * Departure PoP: [Berlin (BER)](/docs/network/latency/ber-berlin) * Destination PoP: [Amsterdam (AMS)](/docs/network/latency/ams-amsterdam) * Region overview: [Europe](/docs/network/latency/regions/europe) → [Europe](/docs/network/latency/regions/europe) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Amsterdam to Berlin latency and RTT](/docs/network/latency/pairs/ams-ber-rtt): 7.2 ms **Fastest routes departing Berlin (BER)** * [Berlin to Frankfurt latency and RTT](/docs/network/latency/pairs/ber-fra-rtt): 6.1 ms * [Berlin to London latency and RTT](/docs/network/latency/pairs/ber-lon-rtt): 15.2 ms * [Berlin to Paris latency and RTT](/docs/network/latency/pairs/ber-par-rtt): 15.6 ms * [Berlin to Marseille latency and RTT](/docs/network/latency/pairs/ber-mrs-rtt): 20.2 ms * [Berlin to Moscow latency and RTT](/docs/network/latency/pairs/ber-mow-rtt): 28.6 ms **Fastest routes arriving at Amsterdam (AMS)** * [London to Amsterdam latency and RTT](/docs/network/latency/pairs/lon-ams-rtt): 5.2 ms * [Frankfurt to Amsterdam latency and RTT](/docs/network/latency/pairs/fra-ams-rtt): 5.9 ms * [Paris to Amsterdam latency and RTT](/docs/network/latency/pairs/par-ams-rtt): 7 ms * [Marseille to Amsterdam latency and RTT](/docs/network/latency/pairs/mrs-ams-rtt): 20.3 ms * [Moscow to Amsterdam latency and RTT](/docs/network/latency/pairs/mow-ams-rtt): 37.6 ms **Same corridor (Europe → Europe)** * [Amsterdam to London latency and RTT](/docs/network/latency/pairs/ams-lon-rtt): 5.2 ms * [Amsterdam to Frankfurt latency and RTT](/docs/network/latency/pairs/ams-fra-rtt): 5.9 ms * [Frankfurt to Berlin latency and RTT](/docs/network/latency/pairs/fra-ber-rtt): 6.1 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/ber-ams.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Berlin → Amsterdam round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/ber-ams-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Berlin, Germany → Frankfurt, Germany RTT 🇩🇪 **Berlin, Germany (BER)** → 🇩🇪 **Frankfurt, Germany (FRA)** Measurement round: `2026-08-16T04:07:28Z`. Berlin to Frankfurt averaged 6.1 ms across 50 replies. The observed range was 5.9 to 6.73 ms; these figures describe one probe round, not a long-term service target. This is number 1 of 19 outgoing routes by average RTT. That ranking compares different destinations and is not adjusted for distance. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------- | | RTT | **6.1 ms** | | Jitter | **0.15 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **4.16 ms** | | Fiber Efficiency | **68.2%** | | Latency Tier | Ultra-Low | | Source Region | [Europe](/docs/network/latency/regions/europe) | | Destination Region | [Europe](/docs/network/latency/regions/europe) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Berlin](https://www.geonames.org/2950159) (52.52437, 13.41053) and [Frankfurt am Main](https://www.geonames.org/2925533) (50.11552, 8.68417). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | ------------ | | WGS-84 geodesic distance | **424.6 km** | | Vacuum RTT floor | **2.83 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **4.16 ms** | | Low-latency fiber material floor | **4.14 ms** | | Engineering floor (5% path allowance) | **4.37 ms** | | Research 1.33× mapped-fiber reference | **5.53 ms** | | Estimated unamplified path loss | **89.2 dB** | | Transparent optical spans / inline amplifiers | **6 / 5** | | Published RTT inflation over fiber floor | **1.47×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ----------- | | Minimum RTT | **5.9 ms** | | Average RTT | **6.1 ms** | | Maximum RTT | **6.73 ms** | | Standard deviation | **0.18 ms** | | Stdev / average | **3.0%** | ## Route Context * Departure PoP: [Berlin (BER)](/docs/network/latency/ber-berlin) * Destination PoP: [Frankfurt (FRA)](/docs/network/latency/fra-frankfurt) * Region overview: [Europe](/docs/network/latency/regions/europe) → [Europe](/docs/network/latency/regions/europe) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Frankfurt to Berlin latency and RTT](/docs/network/latency/pairs/fra-ber-rtt): 6.1 ms **Fastest routes departing Berlin (BER)** * [Berlin to Amsterdam latency and RTT](/docs/network/latency/pairs/ber-ams-rtt): 7.2 ms * [Berlin to London latency and RTT](/docs/network/latency/pairs/ber-lon-rtt): 15.2 ms * [Berlin to Paris latency and RTT](/docs/network/latency/pairs/ber-par-rtt): 15.6 ms * [Berlin to Marseille latency and RTT](/docs/network/latency/pairs/ber-mrs-rtt): 20.2 ms * [Berlin to Moscow latency and RTT](/docs/network/latency/pairs/ber-mow-rtt): 28.6 ms **Fastest routes arriving at Frankfurt (FRA)** * [Amsterdam to Frankfurt latency and RTT](/docs/network/latency/pairs/ams-fra-rtt): 5.9 ms * [Paris to Frankfurt latency and RTT](/docs/network/latency/pairs/par-fra-rtt): 7.6 ms * [London to Frankfurt latency and RTT](/docs/network/latency/pairs/lon-fra-rtt): 13.6 ms * [Marseille to Frankfurt latency and RTT](/docs/network/latency/pairs/mrs-fra-rtt): 16 ms * [Moscow to Frankfurt latency and RTT](/docs/network/latency/pairs/mow-fra-rtt): 35.9 ms **Same corridor (Europe → Europe)** * [Amsterdam to London latency and RTT](/docs/network/latency/pairs/ams-lon-rtt): 5.2 ms * [London to Amsterdam latency and RTT](/docs/network/latency/pairs/lon-ams-rtt): 5.2 ms * [Frankfurt to Amsterdam latency and RTT](/docs/network/latency/pairs/fra-ams-rtt): 5.9 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/ber-fra.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Berlin → Frankfurt round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/ber-fra-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Berlin, Germany → São Paulo, Brazil RTT 🇩🇪 **Berlin, Germany (BER)** → 🇧🇷 **São Paulo, Brazil (GRU)** Measurement round: `2026-08-16T04:07:28Z`. Berlin to São Paulo averaged 186.5 ms across 50 replies. The observed range was 179.2 to 206.69 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 6.78 ms, or 3.6% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **186.5 ms** | | Jitter | **5.27 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **100.2 ms** | | Fiber Efficiency | **53.7%** | | Latency Tier | Fair | | Source Region | [Europe](/docs/network/latency/regions/europe) | | Destination Region | [South America](/docs/network/latency/regions/south-america) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Berlin](https://www.geonames.org/2950159) (52.52437, 13.41053) and [São Paulo](https://www.geonames.org/3448439) (-23.54750, -46.63611). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | --------------- | | WGS-84 geodesic distance | **10,232.3 km** | | Vacuum RTT floor | **68.26 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **100.2 ms** | | Low-latency fiber material floor | **99.8 ms** | | Engineering floor (5% path allowance) | **105.23 ms** | | Research 1.33× mapped-fiber reference | **133.27 ms** | | Estimated unamplified path loss | **2148.8 dB** | | Transparent optical spans / inline amplifiers | **135 / 134** | | Published RTT inflation over fiber floor | **1.86×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **179.2 ms** | | Average RTT | **186.5 ms** | | Maximum RTT | **206.69 ms** | | Standard deviation | **6.78 ms** | | Stdev / average | **3.6%** | ## Route Context * Departure PoP: [Berlin (BER)](/docs/network/latency/ber-berlin) * Destination PoP: [São Paulo (GRU)](/docs/network/latency/gru-sao-paulo) * Region overview: [Europe](/docs/network/latency/regions/europe) → [South America](/docs/network/latency/regions/south-america) * Corridor overview: [Europe to South America](/docs/network/latency/regions/europe-to-south-america) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [São Paulo to Berlin latency and RTT](/docs/network/latency/pairs/gru-ber-rtt): 191.3 ms **Fastest routes departing Berlin (BER)** * [Berlin to Frankfurt latency and RTT](/docs/network/latency/pairs/ber-fra-rtt): 6.1 ms * [Berlin to Amsterdam latency and RTT](/docs/network/latency/pairs/ber-ams-rtt): 7.2 ms * [Berlin to London latency and RTT](/docs/network/latency/pairs/ber-lon-rtt): 15.2 ms * [Berlin to Paris latency and RTT](/docs/network/latency/pairs/ber-par-rtt): 15.6 ms * [Berlin to Marseille latency and RTT](/docs/network/latency/pairs/ber-mrs-rtt): 20.2 ms **Fastest routes arriving at São Paulo (GRU)** * [Ashburn to São Paulo latency and RTT](/docs/network/latency/pairs/iad-gru-rtt): 101.4 ms * [New York to São Paulo latency and RTT](/docs/network/latency/pairs/nyc-gru-rtt): 106.9 ms * [Miami to São Paulo latency and RTT](/docs/network/latency/pairs/mia-gru-rtt): 128.7 ms * [Los Angeles to São Paulo latency and RTT](/docs/network/latency/pairs/lax-gru-rtt): 131.3 ms * [Seattle to São Paulo latency and RTT](/docs/network/latency/pairs/sea-gru-rtt): 155.7 ms **Same corridor (Europe → South America)** * [London to São Paulo latency and RTT](/docs/network/latency/pairs/lon-gru-rtt): 170.7 ms * [Amsterdam to São Paulo latency and RTT](/docs/network/latency/pairs/ams-gru-rtt): 175.5 ms * [Paris to São Paulo latency and RTT](/docs/network/latency/pairs/par-gru-rtt): 175.8 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/ber-gru.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Berlin → São Paulo round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/ber-gru-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Berlin, Germany → Hong Kong RTT 🇩🇪 **Berlin, Germany (BER)** → 🇭🇰 **Hong Kong (HKG)** Measurement round: `2026-08-16T04:07:28Z`. Berlin to Hong Kong averaged 144.9 ms across 50 replies. The observed range was 140.42 to 158.73 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 3.7 ms, or 2.6% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------------------- | | RTT | **144.9 ms** | | Jitter | **2.94 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **85.85 ms** | | Fiber Efficiency | **59.3%** | | Latency Tier | Good | | Source Region | [Europe](/docs/network/latency/regions/europe) | | Destination Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Berlin](https://www.geonames.org/2950159) (52.52437, 13.41053) and [Hong Kong](https://www.geonames.org/1819729) (22.27832, 114.17469). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **8,767.1 km** | | Vacuum RTT floor | **58.49 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **85.85 ms** | | Low-latency fiber material floor | **85.51 ms** | | Engineering floor (5% path allowance) | **90.16 ms** | | Research 1.33× mapped-fiber reference | **114.19 ms** | | Estimated unamplified path loss | **1841.1 dB** | | Transparent optical spans / inline amplifiers | **116 / 115** | | Published RTT inflation over fiber floor | **1.69×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **140.42 ms** | | Average RTT | **144.9 ms** | | Maximum RTT | **158.73 ms** | | Standard deviation | **3.7 ms** | | Stdev / average | **2.6%** | ## Route Context * Departure PoP: [Berlin (BER)](/docs/network/latency/ber-berlin) * Destination PoP: [Hong Kong (HKG)](/docs/network/latency/hkg-hong-kong) * Region overview: [Europe](/docs/network/latency/regions/europe) → [Asia Pacific](/docs/network/latency/regions/asia-pacific) * Corridor overview: [Europe to Asia Pacific](/docs/network/latency/regions/europe-to-asia-pacific) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Hong Kong to Berlin latency and RTT](/docs/network/latency/pairs/hkg-ber-rtt): 145.1 ms **Fastest routes departing Berlin (BER)** * [Berlin to Frankfurt latency and RTT](/docs/network/latency/pairs/ber-fra-rtt): 6.1 ms * [Berlin to Amsterdam latency and RTT](/docs/network/latency/pairs/ber-ams-rtt): 7.2 ms * [Berlin to London latency and RTT](/docs/network/latency/pairs/ber-lon-rtt): 15.2 ms * [Berlin to Paris latency and RTT](/docs/network/latency/pairs/ber-par-rtt): 15.6 ms * [Berlin to Marseille latency and RTT](/docs/network/latency/pairs/ber-mrs-rtt): 20.2 ms **Fastest routes arriving at Hong Kong (HKG)** * [Taipei to Hong Kong latency and RTT](/docs/network/latency/pairs/tpe-hkg-rtt): 15.5 ms * [Singapore to Hong Kong latency and RTT](/docs/network/latency/pairs/sin-hkg-rtt): 30.8 ms * [Tokyo to Hong Kong latency and RTT](/docs/network/latency/pairs/tyo-hkg-rtt): 44.8 ms * [Moscow to Hong Kong latency and RTT](/docs/network/latency/pairs/mow-hkg-rtt): 118.6 ms * [Seattle to Hong Kong latency and RTT](/docs/network/latency/pairs/sea-hkg-rtt): 130.4 ms **Same corridor (Europe → Asia Pacific)** * [Moscow to Taipei latency and RTT](/docs/network/latency/pairs/mow-tpe-rtt): 132.1 ms * [Marseille to Singapore latency and RTT](/docs/network/latency/pairs/mrs-sin-rtt): 139.6 ms * [Moscow to Singapore latency and RTT](/docs/network/latency/pairs/mow-sin-rtt): 147.3 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/ber-hkg.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Berlin → Hong Kong round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/ber-hkg-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Berlin, Germany → Ashburn, USA RTT 🇩🇪 **Berlin, Germany (BER)** → 🇺🇸 **Ashburn, USA (IAD)** Measurement round: `2026-08-16T04:07:28Z`. Berlin to Ashburn averaged 85.7 ms across 50 replies. The observed range was 83.7 to 92.94 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 2.12 ms, or 2.5% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **85.7 ms** | | Jitter | **1.62 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **66.05 ms** | | Fiber Efficiency | **77.1%** | | Latency Tier | Good | | Source Region | [Europe](/docs/network/latency/regions/europe) | | Destination Region | [North America](/docs/network/latency/regions/north-america) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Berlin](https://www.geonames.org/2950159) (52.52437, 13.41053) and [Ashburn](https://www.geonames.org/4744870) (39.04372, -77.48749). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **6,744.5 km** | | Vacuum RTT floor | **44.99 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **66.05 ms** | | Low-latency fiber material floor | **65.78 ms** | | Engineering floor (5% path allowance) | **69.36 ms** | | Research 1.33× mapped-fiber reference | **87.84 ms** | | Estimated unamplified path loss | **1416.4 dB** | | Transparent optical spans / inline amplifiers | **89 / 88** | | Published RTT inflation over fiber floor | **1.3×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------ | | Minimum RTT | **83.7 ms** | | Average RTT | **85.7 ms** | | Maximum RTT | **92.94 ms** | | Standard deviation | **2.12 ms** | | Stdev / average | **2.5%** | ## Route Context * Departure PoP: [Berlin (BER)](/docs/network/latency/ber-berlin) * Destination PoP: [Ashburn (IAD)](/docs/network/latency/iad-ashburn) * Region overview: [Europe](/docs/network/latency/regions/europe) → [North America](/docs/network/latency/regions/north-america) * Corridor overview: [Europe to North America](/docs/network/latency/regions/europe-to-north-america) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Ashburn to Berlin latency and RTT](/docs/network/latency/pairs/iad-ber-rtt): 86.1 ms **Fastest routes departing Berlin (BER)** * [Berlin to Frankfurt latency and RTT](/docs/network/latency/pairs/ber-fra-rtt): 6.1 ms * [Berlin to Amsterdam latency and RTT](/docs/network/latency/pairs/ber-ams-rtt): 7.2 ms * [Berlin to London latency and RTT](/docs/network/latency/pairs/ber-lon-rtt): 15.2 ms * [Berlin to Paris latency and RTT](/docs/network/latency/pairs/ber-par-rtt): 15.6 ms * [Berlin to Marseille latency and RTT](/docs/network/latency/pairs/ber-mrs-rtt): 20.2 ms **Fastest routes arriving at Ashburn (IAD)** * [New York to Ashburn latency and RTT](/docs/network/latency/pairs/nyc-iad-rtt): 6 ms * [Miami to Ashburn latency and RTT](/docs/network/latency/pairs/mia-iad-rtt): 27.3 ms * [Los Angeles to Ashburn latency and RTT](/docs/network/latency/pairs/lax-iad-rtt): 60.4 ms * [Seattle to Ashburn latency and RTT](/docs/network/latency/pairs/sea-iad-rtt): 61.5 ms * [London to Ashburn latency and RTT](/docs/network/latency/pairs/lon-iad-rtt): 71 ms **Same corridor (Europe → North America)** * [London to New York latency and RTT](/docs/network/latency/pairs/lon-nyc-rtt): 63.8 ms * [Amsterdam to New York latency and RTT](/docs/network/latency/pairs/ams-nyc-rtt): 68.6 ms * [Paris to New York latency and RTT](/docs/network/latency/pairs/par-nyc-rtt): 68.9 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/ber-iad.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Berlin → Ashburn round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/ber-iad-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Berlin, Germany → Johannesburg, South Africa RTT 🇩🇪 **Berlin, Germany (BER)** → 🇿🇦 **Johannesburg, South Africa (JNB)** Measurement round: `2026-08-16T04:07:28Z`. Berlin to Johannesburg averaged 172.1 ms across 50 replies. The observed range was 163.72 to 199.65 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 7.34 ms, or 4.3% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------- | | RTT | **172.1 ms** | | Jitter | **6.16 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **86.53 ms** | | Fiber Efficiency | **50.3%** | | Latency Tier | Fair | | Source Region | [Europe](/docs/network/latency/regions/europe) | | Destination Region | [Africa](/docs/network/latency/regions/africa) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Berlin](https://www.geonames.org/2950159) (52.52437, 13.41053) and [Johannesburg](https://www.geonames.org/993800) (-26.20227, 28.04363). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **8,836.6 km** | | Vacuum RTT floor | **58.95 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **86.53 ms** | | Low-latency fiber material floor | **86.19 ms** | | Engineering floor (5% path allowance) | **90.87 ms** | | Research 1.33× mapped-fiber reference | **115.09 ms** | | Estimated unamplified path loss | **1855.7 dB** | | Transparent optical spans / inline amplifiers | **116 / 115** | | Published RTT inflation over fiber floor | **1.99×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **163.72 ms** | | Average RTT | **172.1 ms** | | Maximum RTT | **199.65 ms** | | Standard deviation | **7.34 ms** | | Stdev / average | **4.3%** | ## Route Context * Departure PoP: [Berlin (BER)](/docs/network/latency/ber-berlin) * Destination PoP: [Johannesburg (JNB)](/docs/network/latency/jnb-johannesburg) * Region overview: [Europe](/docs/network/latency/regions/europe) → [Africa](/docs/network/latency/regions/africa) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Johannesburg to Berlin latency and RTT](/docs/network/latency/pairs/jnb-ber-rtt): 172.1 ms **Fastest routes departing Berlin (BER)** * [Berlin to Frankfurt latency and RTT](/docs/network/latency/pairs/ber-fra-rtt): 6.1 ms * [Berlin to Amsterdam latency and RTT](/docs/network/latency/pairs/ber-ams-rtt): 7.2 ms * [Berlin to London latency and RTT](/docs/network/latency/pairs/ber-lon-rtt): 15.2 ms * [Berlin to Paris latency and RTT](/docs/network/latency/pairs/ber-par-rtt): 15.6 ms * [Berlin to Marseille latency and RTT](/docs/network/latency/pairs/ber-mrs-rtt): 20.2 ms **Fastest routes arriving at Johannesburg (JNB)** * [Amsterdam to Johannesburg latency and RTT](/docs/network/latency/pairs/ams-jnb-rtt): 171.9 ms * [London to Johannesburg latency and RTT](/docs/network/latency/pairs/lon-jnb-rtt): 179.6 ms * [Paris to Johannesburg latency and RTT](/docs/network/latency/pairs/par-jnb-rtt): 180.9 ms * [Frankfurt to Johannesburg latency and RTT](/docs/network/latency/pairs/fra-jnb-rtt): 188.4 ms * [Marseille to Johannesburg latency and RTT](/docs/network/latency/pairs/mrs-jnb-rtt): 198.5 ms **Same corridor (Europe → Africa)** * [Moscow to Johannesburg latency and RTT](/docs/network/latency/pairs/mow-jnb-rtt): 201.9 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/ber-jnb.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Berlin → Johannesburg round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/ber-jnb-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Berlin, Germany → Los Angeles, USA RTT 🇩🇪 **Berlin, Germany (BER)** → 🇺🇸 **Los Angeles, USA (LAX)** Measurement round: `2026-08-16T04:07:28Z`. Berlin to Los Angeles averaged 140.9 ms across 50 replies. The observed range was 133.87 to 162.61 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 5.85 ms, or 4.2% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **140.9 ms** | | Jitter | **4.81 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **91.39 ms** | | Fiber Efficiency | **64.9%** | | Latency Tier | Good | | Source Region | [Europe](/docs/network/latency/regions/europe) | | Destination Region | [North America](/docs/network/latency/regions/north-america) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Berlin](https://www.geonames.org/2950159) (52.52437, 13.41053) and [Los Angeles](https://www.geonames.org/5368361) (34.05223, -118.24368). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **9,331.9 km** | | Vacuum RTT floor | **62.26 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **91.39 ms** | | Low-latency fiber material floor | **91.02 ms** | | Engineering floor (5% path allowance) | **95.97 ms** | | Research 1.33× mapped-fiber reference | **121.54 ms** | | Estimated unamplified path loss | **1959.7 dB** | | Transparent optical spans / inline amplifiers | **123 / 122** | | Published RTT inflation over fiber floor | **1.54×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **133.87 ms** | | Average RTT | **140.9 ms** | | Maximum RTT | **162.61 ms** | | Standard deviation | **5.85 ms** | | Stdev / average | **4.2%** | ## Route Context * Departure PoP: [Berlin (BER)](/docs/network/latency/ber-berlin) * Destination PoP: [Los Angeles (LAX)](/docs/network/latency/lax-los-angeles) * Region overview: [Europe](/docs/network/latency/regions/europe) → [North America](/docs/network/latency/regions/north-america) * Corridor overview: [Europe to North America](/docs/network/latency/regions/europe-to-north-america) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Los Angeles to Berlin latency and RTT](/docs/network/latency/pairs/lax-ber-rtt): 139.3 ms **Fastest routes departing Berlin (BER)** * [Berlin to Frankfurt latency and RTT](/docs/network/latency/pairs/ber-fra-rtt): 6.1 ms * [Berlin to Amsterdam latency and RTT](/docs/network/latency/pairs/ber-ams-rtt): 7.2 ms * [Berlin to London latency and RTT](/docs/network/latency/pairs/ber-lon-rtt): 15.2 ms * [Berlin to Paris latency and RTT](/docs/network/latency/pairs/ber-par-rtt): 15.6 ms * [Berlin to Marseille latency and RTT](/docs/network/latency/pairs/ber-mrs-rtt): 20.2 ms **Fastest routes arriving at Los Angeles (LAX)** * [Seattle to Los Angeles latency and RTT](/docs/network/latency/pairs/sea-lax-rtt): 25.9 ms * [Miami to Los Angeles latency and RTT](/docs/network/latency/pairs/mia-lax-rtt): 56.8 ms * [New York to Los Angeles latency and RTT](/docs/network/latency/pairs/nyc-lax-rtt): 58.9 ms * [Ashburn to Los Angeles latency and RTT](/docs/network/latency/pairs/iad-lax-rtt): 60.7 ms * [Tokyo to Los Angeles latency and RTT](/docs/network/latency/pairs/tyo-lax-rtt): 101.2 ms **Same corridor (Europe → North America)** * [London to New York latency and RTT](/docs/network/latency/pairs/lon-nyc-rtt): 63.8 ms * [Amsterdam to New York latency and RTT](/docs/network/latency/pairs/ams-nyc-rtt): 68.6 ms * [Paris to New York latency and RTT](/docs/network/latency/pairs/par-nyc-rtt): 68.9 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/ber-lax.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Berlin → Los Angeles round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/ber-lax-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Berlin, Germany → London, UK RTT 🇩🇪 **Berlin, Germany (BER)** → 🇬🇧 **London, UK (LON)** Measurement round: `2026-08-16T04:07:28Z`. Berlin to London averaged 15.2 ms across 50 replies. The observed range was 14.57 to 16.96 ms; these figures describe one probe round, not a long-term service target. This is number 3 of 19 outgoing routes by average RTT. That ranking compares different destinations and is not adjusted for distance. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------- | | RTT | **15.2 ms** | | Jitter | **0.4 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **9.15 ms** | | Fiber Efficiency | **60.2%** | | Latency Tier | Ultra-Low | | Source Region | [Europe](/docs/network/latency/regions/europe) | | Destination Region | [Europe](/docs/network/latency/regions/europe) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Berlin](https://www.geonames.org/2950159) (52.52437, 13.41053) and [London](https://www.geonames.org/2643743) (51.50853, -0.12574). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | ------------ | | WGS-84 geodesic distance | **934.7 km** | | Vacuum RTT floor | **6.24 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **9.15 ms** | | Low-latency fiber material floor | **9.12 ms** | | Engineering floor (5% path allowance) | **9.61 ms** | | Research 1.33× mapped-fiber reference | **12.17 ms** | | Estimated unamplified path loss | **196.3 dB** | | Transparent optical spans / inline amplifiers | **13 / 12** | | Published RTT inflation over fiber floor | **1.66×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------ | | Minimum RTT | **14.57 ms** | | Average RTT | **15.2 ms** | | Maximum RTT | **16.96 ms** | | Standard deviation | **0.49 ms** | | Stdev / average | **3.2%** | ## Route Context * Departure PoP: [Berlin (BER)](/docs/network/latency/ber-berlin) * Destination PoP: [London (LON)](/docs/network/latency/lon-london) * Region overview: [Europe](/docs/network/latency/regions/europe) → [Europe](/docs/network/latency/regions/europe) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [London to Berlin latency and RTT](/docs/network/latency/pairs/lon-ber-rtt): 16.2 ms **Fastest routes departing Berlin (BER)** * [Berlin to Frankfurt latency and RTT](/docs/network/latency/pairs/ber-fra-rtt): 6.1 ms * [Berlin to Amsterdam latency and RTT](/docs/network/latency/pairs/ber-ams-rtt): 7.2 ms * [Berlin to Paris latency and RTT](/docs/network/latency/pairs/ber-par-rtt): 15.6 ms * [Berlin to Marseille latency and RTT](/docs/network/latency/pairs/ber-mrs-rtt): 20.2 ms * [Berlin to Moscow latency and RTT](/docs/network/latency/pairs/ber-mow-rtt): 28.6 ms **Fastest routes arriving at London (LON)** * [Amsterdam to London latency and RTT](/docs/network/latency/pairs/ams-lon-rtt): 5.2 ms * [Paris to London latency and RTT](/docs/network/latency/pairs/par-lon-rtt): 6.4 ms * [Frankfurt to London latency and RTT](/docs/network/latency/pairs/fra-lon-rtt): 12.9 ms * [Marseille to London latency and RTT](/docs/network/latency/pairs/mrs-lon-rtt): 18.9 ms * [Moscow to London latency and RTT](/docs/network/latency/pairs/mow-lon-rtt): 42.4 ms **Same corridor (Europe → Europe)** * [London to Amsterdam latency and RTT](/docs/network/latency/pairs/lon-ams-rtt): 5.2 ms * [Amsterdam to Frankfurt latency and RTT](/docs/network/latency/pairs/ams-fra-rtt): 5.9 ms * [Frankfurt to Amsterdam latency and RTT](/docs/network/latency/pairs/fra-ams-rtt): 5.9 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/ber-lon.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Berlin → London round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/ber-lon-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Berlin, Germany → Melbourne, Australia RTT 🇩🇪 **Berlin, Germany (BER)** → 🇦🇺 **Melbourne, Australia (MEL)** Measurement round: `2026-08-16T04:07:28Z`. Berlin to Melbourne averaged 256.2 ms across 50 replies. The observed range was 246.67 to 285.1 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 7.82 ms, or 3.1% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------------------- | | RTT | **256.2 ms** | | Jitter | **5.78 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **156.34 ms** | | Fiber Efficiency | **61%** | | Latency Tier | High | | Source Region | [Europe](/docs/network/latency/regions/europe) | | Destination Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Berlin](https://www.geonames.org/2950159) (52.52437, 13.41053) and [Melbourne](https://www.geonames.org/2158177) (-37.81400, 144.96332). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | --------------- | | WGS-84 geodesic distance | **15,965.1 km** | | Vacuum RTT floor | **106.51 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **156.34 ms** | | Low-latency fiber material floor | **155.71 ms** | | Engineering floor (5% path allowance) | **164.18 ms** | | Research 1.33× mapped-fiber reference | **207.94 ms** | | Estimated unamplified path loss | **3352.7 dB** | | Transparent optical spans / inline amplifiers | **210 / 209** | | Published RTT inflation over fiber floor | **1.64×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **246.67 ms** | | Average RTT | **256.2 ms** | | Maximum RTT | **285.1 ms** | | Standard deviation | **7.82 ms** | | Stdev / average | **3.1%** | ## Route Context * Departure PoP: [Berlin (BER)](/docs/network/latency/ber-berlin) * Destination PoP: [Melbourne (MEL)](/docs/network/latency/mel-melbourne) * Region overview: [Europe](/docs/network/latency/regions/europe) → [Asia Pacific](/docs/network/latency/regions/asia-pacific) * Corridor overview: [Europe to Asia Pacific](/docs/network/latency/regions/europe-to-asia-pacific) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Melbourne to Berlin latency and RTT](/docs/network/latency/pairs/mel-ber-rtt): 254.7 ms **Fastest routes departing Berlin (BER)** * [Berlin to Frankfurt latency and RTT](/docs/network/latency/pairs/ber-fra-rtt): 6.1 ms * [Berlin to Amsterdam latency and RTT](/docs/network/latency/pairs/ber-ams-rtt): 7.2 ms * [Berlin to London latency and RTT](/docs/network/latency/pairs/ber-lon-rtt): 15.2 ms * [Berlin to Paris latency and RTT](/docs/network/latency/pairs/ber-par-rtt): 15.6 ms * [Berlin to Marseille latency and RTT](/docs/network/latency/pairs/ber-mrs-rtt): 20.2 ms **Fastest routes arriving at Melbourne (MEL)** * [Sydney to Melbourne latency and RTT](/docs/network/latency/pairs/syd-mel-rtt): 9.8 ms * [Singapore to Melbourne latency and RTT](/docs/network/latency/pairs/sin-mel-rtt): 88.6 ms * [Tokyo to Melbourne latency and RTT](/docs/network/latency/pairs/tyo-mel-rtt): 113.9 ms * [Hong Kong to Melbourne latency and RTT](/docs/network/latency/pairs/hkg-mel-rtt): 138.4 ms * [Taipei to Melbourne latency and RTT](/docs/network/latency/pairs/tpe-mel-rtt): 142 ms **Same corridor (Europe → Asia Pacific)** * [Moscow to Hong Kong latency and RTT](/docs/network/latency/pairs/mow-hkg-rtt): 118.6 ms * [Moscow to Taipei latency and RTT](/docs/network/latency/pairs/mow-tpe-rtt): 132.1 ms * [Marseille to Singapore latency and RTT](/docs/network/latency/pairs/mrs-sin-rtt): 139.6 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/ber-mel.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Berlin → Melbourne round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/ber-mel-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Berlin, Germany → Miami, USA RTT 🇩🇪 **Berlin, Germany (BER)** → 🇺🇸 **Miami, USA (MIA)** Measurement round: `2026-08-16T04:07:28Z`. Berlin to Miami averaged 118.3 ms across 50 replies. The observed range was 115.44 to 127.36 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 2.6 ms, or 2.2% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **118.3 ms** | | Jitter | **2.21 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **78.44 ms** | | Fiber Efficiency | **66.3%** | | Latency Tier | Good | | Source Region | [Europe](/docs/network/latency/regions/europe) | | Destination Region | [North America](/docs/network/latency/regions/north-america) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Berlin](https://www.geonames.org/2950159) (52.52437, 13.41053) and [Miami](https://www.geonames.org/4164138) (25.77427, -80.19366). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **8,009.8 km** | | Vacuum RTT floor | **53.44 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **78.44 ms** | | Low-latency fiber material floor | **78.12 ms** | | Engineering floor (5% path allowance) | **82.37 ms** | | Research 1.33× mapped-fiber reference | **104.32 ms** | | Estimated unamplified path loss | **1682.1 dB** | | Transparent optical spans / inline amplifiers | **106 / 105** | | Published RTT inflation over fiber floor | **1.51×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **115.44 ms** | | Average RTT | **118.3 ms** | | Maximum RTT | **127.36 ms** | | Standard deviation | **2.6 ms** | | Stdev / average | **2.2%** | ## Route Context * Departure PoP: [Berlin (BER)](/docs/network/latency/ber-berlin) * Destination PoP: [Miami (MIA)](/docs/network/latency/mia-miami) * Region overview: [Europe](/docs/network/latency/regions/europe) → [North America](/docs/network/latency/regions/north-america) * Corridor overview: [Europe to North America](/docs/network/latency/regions/europe-to-north-america) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Miami to Berlin latency and RTT](/docs/network/latency/pairs/mia-ber-rtt): 117.3 ms **Fastest routes departing Berlin (BER)** * [Berlin to Frankfurt latency and RTT](/docs/network/latency/pairs/ber-fra-rtt): 6.1 ms * [Berlin to Amsterdam latency and RTT](/docs/network/latency/pairs/ber-ams-rtt): 7.2 ms * [Berlin to London latency and RTT](/docs/network/latency/pairs/ber-lon-rtt): 15.2 ms * [Berlin to Paris latency and RTT](/docs/network/latency/pairs/ber-par-rtt): 15.6 ms * [Berlin to Marseille latency and RTT](/docs/network/latency/pairs/ber-mrs-rtt): 20.2 ms **Fastest routes arriving at Miami (MIA)** * [Ashburn to Miami latency and RTT](/docs/network/latency/pairs/iad-mia-rtt): 27.4 ms * [New York to Miami latency and RTT](/docs/network/latency/pairs/nyc-mia-rtt): 32.9 ms * [Los Angeles to Miami latency and RTT](/docs/network/latency/pairs/lax-mia-rtt): 57.3 ms * [São Paulo to Miami latency and RTT](/docs/network/latency/pairs/gru-mia-rtt): 74 ms * [Seattle to Miami latency and RTT](/docs/network/latency/pairs/sea-mia-rtt): 81.7 ms **Same corridor (Europe → North America)** * [London to New York latency and RTT](/docs/network/latency/pairs/lon-nyc-rtt): 63.8 ms * [Amsterdam to New York latency and RTT](/docs/network/latency/pairs/ams-nyc-rtt): 68.6 ms * [Paris to New York latency and RTT](/docs/network/latency/pairs/par-nyc-rtt): 68.9 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/ber-mia.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Berlin → Miami round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/ber-mia-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Berlin, Germany → Moscow, Russia RTT 🇩🇪 **Berlin, Germany (BER)** → 🇷🇺 **Moscow, Russia (MOW)** Measurement round: `2026-08-16T04:07:28Z`. Berlin to Moscow averaged 28.6 ms across 50 replies. The observed range was 27.35 to 31.58 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 1.1 ms, or 3.8% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------- | | RTT | **28.6 ms** | | Jitter | **0.83 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **15.8 ms** | | Fiber Efficiency | **55.2%** | | Latency Tier | Ultra-Low | | Source Region | [Europe](/docs/network/latency/regions/europe) | | Destination Region | [Europe](/docs/network/latency/regions/europe) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Berlin](https://www.geonames.org/2950159) (52.52437, 13.41053) and [Moscow](https://www.geonames.org/524901) (55.75204, 37.61781). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **1,613.5 km** | | Vacuum RTT floor | **10.76 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **15.8 ms** | | Low-latency fiber material floor | **15.74 ms** | | Engineering floor (5% path allowance) | **16.59 ms** | | Research 1.33× mapped-fiber reference | **21.01 ms** | | Estimated unamplified path loss | **338.8 dB** | | Transparent optical spans / inline amplifiers | **22 / 21** | | Published RTT inflation over fiber floor | **1.81×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------ | | Minimum RTT | **27.35 ms** | | Average RTT | **28.6 ms** | | Maximum RTT | **31.58 ms** | | Standard deviation | **1.1 ms** | | Stdev / average | **3.8%** | ## Route Context * Departure PoP: [Berlin (BER)](/docs/network/latency/ber-berlin) * Destination PoP: [Moscow (MOW)](/docs/network/latency/mow-moscow) * Region overview: [Europe](/docs/network/latency/regions/europe) → [Europe](/docs/network/latency/regions/europe) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Moscow to Berlin latency and RTT](/docs/network/latency/pairs/mow-ber-rtt): 27.5 ms **Fastest routes departing Berlin (BER)** * [Berlin to Frankfurt latency and RTT](/docs/network/latency/pairs/ber-fra-rtt): 6.1 ms * [Berlin to Amsterdam latency and RTT](/docs/network/latency/pairs/ber-ams-rtt): 7.2 ms * [Berlin to London latency and RTT](/docs/network/latency/pairs/ber-lon-rtt): 15.2 ms * [Berlin to Paris latency and RTT](/docs/network/latency/pairs/ber-par-rtt): 15.6 ms * [Berlin to Marseille latency and RTT](/docs/network/latency/pairs/ber-mrs-rtt): 20.2 ms **Fastest routes arriving at Moscow (MOW)** * [Frankfurt to Moscow latency and RTT](/docs/network/latency/pairs/fra-mow-rtt): 34.7 ms * [Amsterdam to Moscow latency and RTT](/docs/network/latency/pairs/ams-mow-rtt): 38.1 ms * [London to Moscow latency and RTT](/docs/network/latency/pairs/lon-mow-rtt): 43.5 ms * [Paris to Moscow latency and RTT](/docs/network/latency/pairs/par-mow-rtt): 44.7 ms * [Marseille to Moscow latency and RTT](/docs/network/latency/pairs/mrs-mow-rtt): 49.6 ms **Same corridor (Europe → Europe)** * [Amsterdam to London latency and RTT](/docs/network/latency/pairs/ams-lon-rtt): 5.2 ms * [London to Amsterdam latency and RTT](/docs/network/latency/pairs/lon-ams-rtt): 5.2 ms * [Amsterdam to Frankfurt latency and RTT](/docs/network/latency/pairs/ams-fra-rtt): 5.9 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/ber-mow.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Berlin → Moscow round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/ber-mow-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Berlin, Germany → Marseille, France RTT 🇩🇪 **Berlin, Germany (BER)** → 🇫🇷 **Marseille, France (MRS)** Measurement round: `2026-08-16T04:07:28Z`. Berlin to Marseille averaged 20.2 ms across 50 replies. The observed range was 19.92 to 21.65 ms; these figures describe one probe round, not a long-term service target. This is number 5 of 19 outgoing routes by average RTT. That ranking compares different destinations and is not adjusted for distance. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------- | | RTT | **20.2 ms** | | Jitter | **0.22 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **11.62 ms** | | Fiber Efficiency | **57.5%** | | Latency Tier | Ultra-Low | | Source Region | [Europe](/docs/network/latency/regions/europe) | | Destination Region | [Europe](/docs/network/latency/regions/europe) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Berlin](https://www.geonames.org/2950159) (52.52437, 13.41053) and [Marseille](https://www.geonames.org/2995469) (43.29695, 5.38107). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | ------------ | | WGS-84 geodesic distance | **1,187 km** | | Vacuum RTT floor | **7.92 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **11.62 ms** | | Low-latency fiber material floor | **11.58 ms** | | Engineering floor (5% path allowance) | **12.21 ms** | | Research 1.33× mapped-fiber reference | **15.46 ms** | | Estimated unamplified path loss | **249.3 dB** | | Transparent optical spans / inline amplifiers | **16 / 15** | | Published RTT inflation over fiber floor | **1.74×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------ | | Minimum RTT | **19.92 ms** | | Average RTT | **20.2 ms** | | Maximum RTT | **21.65 ms** | | Standard deviation | **0.31 ms** | | Stdev / average | **1.5%** | ## Route Context * Departure PoP: [Berlin (BER)](/docs/network/latency/ber-berlin) * Destination PoP: [Marseille (MRS)](/docs/network/latency/mrs-marseille) * Region overview: [Europe](/docs/network/latency/regions/europe) → [Europe](/docs/network/latency/regions/europe) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Marseille to Berlin latency and RTT](/docs/network/latency/pairs/mrs-ber-rtt): 20.6 ms **Fastest routes departing Berlin (BER)** * [Berlin to Frankfurt latency and RTT](/docs/network/latency/pairs/ber-fra-rtt): 6.1 ms * [Berlin to Amsterdam latency and RTT](/docs/network/latency/pairs/ber-ams-rtt): 7.2 ms * [Berlin to London latency and RTT](/docs/network/latency/pairs/ber-lon-rtt): 15.2 ms * [Berlin to Paris latency and RTT](/docs/network/latency/pairs/ber-par-rtt): 15.6 ms * [Berlin to Moscow latency and RTT](/docs/network/latency/pairs/ber-mow-rtt): 28.6 ms **Fastest routes arriving at Marseille (MRS)** * [Paris to Marseille latency and RTT](/docs/network/latency/pairs/par-mrs-rtt): 8.9 ms * [Frankfurt to Marseille latency and RTT](/docs/network/latency/pairs/fra-mrs-rtt): 16.4 ms * [London to Marseille latency and RTT](/docs/network/latency/pairs/lon-mrs-rtt): 17.9 ms * [Amsterdam to Marseille latency and RTT](/docs/network/latency/pairs/ams-mrs-rtt): 19.6 ms * [Moscow to Marseille latency and RTT](/docs/network/latency/pairs/mow-mrs-rtt): 49.9 ms **Same corridor (Europe → Europe)** * [Amsterdam to London latency and RTT](/docs/network/latency/pairs/ams-lon-rtt): 5.2 ms * [London to Amsterdam latency and RTT](/docs/network/latency/pairs/lon-ams-rtt): 5.2 ms * [Amsterdam to Frankfurt latency and RTT](/docs/network/latency/pairs/ams-fra-rtt): 5.9 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/ber-mrs.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Berlin → Marseille round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/ber-mrs-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Berlin, Germany → New York, USA RTT 🇩🇪 **Berlin, Germany (BER)** → 🇺🇸 **New York, USA (NYC)** Measurement round: `2026-08-16T04:07:28Z`. Berlin to New York averaged 79.6 ms across 50 replies. The observed range was 78.03 to 83.78 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 1.28 ms, or 1.6% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **79.6 ms** | | Jitter | **1.07 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **62.7 ms** | | Fiber Efficiency | **78.8%** | | Latency Tier | Excellent | | Source Region | [Europe](/docs/network/latency/regions/europe) | | Destination Region | [North America](/docs/network/latency/regions/north-america) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Berlin](https://www.geonames.org/2950159) (52.52437, 13.41053) and [New York City](https://www.geonames.org/5128581) (40.71427, -74.00597). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **6,402.4 km** | | Vacuum RTT floor | **42.71 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **62.7 ms** | | Low-latency fiber material floor | **62.45 ms** | | Engineering floor (5% path allowance) | **65.84 ms** | | Research 1.33× mapped-fiber reference | **83.39 ms** | | Estimated unamplified path loss | **1344.5 dB** | | Transparent optical spans / inline amplifiers | **85 / 84** | | Published RTT inflation over fiber floor | **1.27×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------ | | Minimum RTT | **78.03 ms** | | Average RTT | **79.6 ms** | | Maximum RTT | **83.78 ms** | | Standard deviation | **1.28 ms** | | Stdev / average | **1.6%** | ## Route Context * Departure PoP: [Berlin (BER)](/docs/network/latency/ber-berlin) * Destination PoP: [New York (NYC)](/docs/network/latency/nyc-new-york) * Region overview: [Europe](/docs/network/latency/regions/europe) → [North America](/docs/network/latency/regions/north-america) * Corridor overview: [Europe to North America](/docs/network/latency/regions/europe-to-north-america) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [New York to Berlin latency and RTT](/docs/network/latency/pairs/nyc-ber-rtt): 78.8 ms **Fastest routes departing Berlin (BER)** * [Berlin to Frankfurt latency and RTT](/docs/network/latency/pairs/ber-fra-rtt): 6.1 ms * [Berlin to Amsterdam latency and RTT](/docs/network/latency/pairs/ber-ams-rtt): 7.2 ms * [Berlin to London latency and RTT](/docs/network/latency/pairs/ber-lon-rtt): 15.2 ms * [Berlin to Paris latency and RTT](/docs/network/latency/pairs/ber-par-rtt): 15.6 ms * [Berlin to Marseille latency and RTT](/docs/network/latency/pairs/ber-mrs-rtt): 20.2 ms **Fastest routes arriving at New York (NYC)** * [Ashburn to New York latency and RTT](/docs/network/latency/pairs/iad-nyc-rtt): 6 ms * [Miami to New York latency and RTT](/docs/network/latency/pairs/mia-nyc-rtt): 33.3 ms * [Los Angeles to New York latency and RTT](/docs/network/latency/pairs/lax-nyc-rtt): 57.9 ms * [Seattle to New York latency and RTT](/docs/network/latency/pairs/sea-nyc-rtt): 59.5 ms * [London to New York latency and RTT](/docs/network/latency/pairs/lon-nyc-rtt): 63.8 ms **Same corridor (Europe → North America)** * [Amsterdam to New York latency and RTT](/docs/network/latency/pairs/ams-nyc-rtt): 68.6 ms * [Paris to New York latency and RTT](/docs/network/latency/pairs/par-nyc-rtt): 68.9 ms * [London to Ashburn latency and RTT](/docs/network/latency/pairs/lon-iad-rtt): 71 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/ber-nyc.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Berlin → New York round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/ber-nyc-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Berlin, Germany → Paris, France RTT 🇩🇪 **Berlin, Germany (BER)** → 🇫🇷 **Paris, France (PAR)** Measurement round: `2026-08-16T04:07:28Z`. Berlin to Paris averaged 15.6 ms across 50 replies. The observed range was 14.9 to 17.81 ms; these figures describe one probe round, not a long-term service target. This is number 4 of 19 outgoing routes by average RTT. That ranking compares different destinations and is not adjusted for distance. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------- | | RTT | **15.6 ms** | | Jitter | **0.52 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **8.62 ms** | | Fiber Efficiency | **55.3%** | | Latency Tier | Ultra-Low | | Source Region | [Europe](/docs/network/latency/regions/europe) | | Destination Region | [Europe](/docs/network/latency/regions/europe) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Berlin](https://www.geonames.org/2950159) (52.52437, 13.41053) and [Paris](https://www.geonames.org/2988507) (48.85341, 2.34880). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | ------------ | | WGS-84 geodesic distance | **880.6 km** | | Vacuum RTT floor | **5.87 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **8.62 ms** | | Low-latency fiber material floor | **8.59 ms** | | Engineering floor (5% path allowance) | **9.06 ms** | | Research 1.33× mapped-fiber reference | **11.47 ms** | | Estimated unamplified path loss | **184.9 dB** | | Transparent optical spans / inline amplifiers | **12 / 11** | | Published RTT inflation over fiber floor | **1.81×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------ | | Minimum RTT | **14.9 ms** | | Average RTT | **15.6 ms** | | Maximum RTT | **17.81 ms** | | Standard deviation | **0.63 ms** | | Stdev / average | **4.0%** | ## Route Context * Departure PoP: [Berlin (BER)](/docs/network/latency/ber-berlin) * Destination PoP: [Paris (PAR)](/docs/network/latency/par-paris) * Region overview: [Europe](/docs/network/latency/regions/europe) → [Europe](/docs/network/latency/regions/europe) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Paris to Berlin latency and RTT](/docs/network/latency/pairs/par-ber-rtt): 15.5 ms **Fastest routes departing Berlin (BER)** * [Berlin to Frankfurt latency and RTT](/docs/network/latency/pairs/ber-fra-rtt): 6.1 ms * [Berlin to Amsterdam latency and RTT](/docs/network/latency/pairs/ber-ams-rtt): 7.2 ms * [Berlin to London latency and RTT](/docs/network/latency/pairs/ber-lon-rtt): 15.2 ms * [Berlin to Marseille latency and RTT](/docs/network/latency/pairs/ber-mrs-rtt): 20.2 ms * [Berlin to Moscow latency and RTT](/docs/network/latency/pairs/ber-mow-rtt): 28.6 ms **Fastest routes arriving at Paris (PAR)** * [London to Paris latency and RTT](/docs/network/latency/pairs/lon-par-rtt): 6.4 ms * [Amsterdam to Paris latency and RTT](/docs/network/latency/pairs/ams-par-rtt): 7 ms * [Frankfurt to Paris latency and RTT](/docs/network/latency/pairs/fra-par-rtt): 7.6 ms * [Marseille to Paris latency and RTT](/docs/network/latency/pairs/mrs-par-rtt): 8.9 ms * [Moscow to Paris latency and RTT](/docs/network/latency/pairs/mow-par-rtt): 44 ms **Same corridor (Europe → Europe)** * [Amsterdam to London latency and RTT](/docs/network/latency/pairs/ams-lon-rtt): 5.2 ms * [London to Amsterdam latency and RTT](/docs/network/latency/pairs/lon-ams-rtt): 5.2 ms * [Amsterdam to Frankfurt latency and RTT](/docs/network/latency/pairs/ams-fra-rtt): 5.9 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/ber-par.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Berlin → Paris round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/ber-par-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Berlin, Germany → Seattle, USA RTT 🇩🇪 **Berlin, Germany (BER)** → 🇺🇸 **Seattle, USA (SEA)** Measurement round: `2026-08-16T04:07:28Z`. Berlin to Seattle averaged 139.7 ms across 50 replies. The observed range was 135.21 to 156.43 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 4.65 ms, or 3.3% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **139.7 ms** | | Jitter | **3.54 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **79.73 ms** | | Fiber Efficiency | **57.1%** | | Latency Tier | Good | | Source Region | [Europe](/docs/network/latency/regions/europe) | | Destination Region | [North America](/docs/network/latency/regions/north-america) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Berlin](https://www.geonames.org/2950159) (52.52437, 13.41053) and [Seattle](https://www.geonames.org/5809844) (47.60621, -122.33207). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **8,142.2 km** | | Vacuum RTT floor | **54.32 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **79.73 ms** | | Low-latency fiber material floor | **79.41 ms** | | Engineering floor (5% path allowance) | **83.73 ms** | | Research 1.33× mapped-fiber reference | **106.05 ms** | | Estimated unamplified path loss | **1709.9 dB** | | Transparent optical spans / inline amplifiers | **107 / 106** | | Published RTT inflation over fiber floor | **1.75×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **135.21 ms** | | Average RTT | **139.7 ms** | | Maximum RTT | **156.43 ms** | | Standard deviation | **4.65 ms** | | Stdev / average | **3.3%** | ## Route Context * Departure PoP: [Berlin (BER)](/docs/network/latency/ber-berlin) * Destination PoP: [Seattle (SEA)](/docs/network/latency/sea-seattle) * Region overview: [Europe](/docs/network/latency/regions/europe) → [North America](/docs/network/latency/regions/north-america) * Corridor overview: [Europe to North America](/docs/network/latency/regions/europe-to-north-america) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Seattle to Berlin latency and RTT](/docs/network/latency/pairs/sea-ber-rtt): 140.9 ms **Fastest routes departing Berlin (BER)** * [Berlin to Frankfurt latency and RTT](/docs/network/latency/pairs/ber-fra-rtt): 6.1 ms * [Berlin to Amsterdam latency and RTT](/docs/network/latency/pairs/ber-ams-rtt): 7.2 ms * [Berlin to London latency and RTT](/docs/network/latency/pairs/ber-lon-rtt): 15.2 ms * [Berlin to Paris latency and RTT](/docs/network/latency/pairs/ber-par-rtt): 15.6 ms * [Berlin to Marseille latency and RTT](/docs/network/latency/pairs/ber-mrs-rtt): 20.2 ms **Fastest routes arriving at Seattle (SEA)** * [Los Angeles to Seattle latency and RTT](/docs/network/latency/pairs/lax-sea-rtt): 26.9 ms * [New York to Seattle latency and RTT](/docs/network/latency/pairs/nyc-sea-rtt): 58.7 ms * [Ashburn to Seattle latency and RTT](/docs/network/latency/pairs/iad-sea-rtt): 62.3 ms * [Miami to Seattle latency and RTT](/docs/network/latency/pairs/mia-sea-rtt): 81.9 ms * [Tokyo to Seattle latency and RTT](/docs/network/latency/pairs/tyo-sea-rtt): 84.8 ms **Same corridor (Europe → North America)** * [London to New York latency and RTT](/docs/network/latency/pairs/lon-nyc-rtt): 63.8 ms * [Amsterdam to New York latency and RTT](/docs/network/latency/pairs/ams-nyc-rtt): 68.6 ms * [Paris to New York latency and RTT](/docs/network/latency/pairs/par-nyc-rtt): 68.9 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/ber-sea.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Berlin → Seattle round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/ber-sea-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Berlin, Germany → Singapore RTT 🇩🇪 **Berlin, Germany (BER)** → 🇸🇬 **Singapore (SIN)** Measurement round: `2026-08-16T04:07:28Z`. Berlin to Singapore averaged 158.3 ms across 50 replies. The observed range was 154.42 to 169.59 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 3.09 ms, or 2.0% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------------------- | | RTT | **158.3 ms** | | Jitter | **2.74 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **97.19 ms** | | Fiber Efficiency | **61.4%** | | Latency Tier | Fair | | Source Region | [Europe](/docs/network/latency/regions/europe) | | Destination Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Berlin](https://www.geonames.org/2950159) (52.52437, 13.41053) and [Singapore](https://www.geonames.org/1880252) (1.28967, 103.85007). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | ------------- | | WGS-84 geodesic distance | **9,925 km** | | Vacuum RTT floor | **66.21 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **97.19 ms** | | Low-latency fiber material floor | **96.8 ms** | | Engineering floor (5% path allowance) | **102.07 ms** | | Research 1.33× mapped-fiber reference | **129.27 ms** | | Estimated unamplified path loss | **2084.3 dB** | | Transparent optical spans / inline amplifiers | **131 / 130** | | Published RTT inflation over fiber floor | **1.63×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **154.42 ms** | | Average RTT | **158.3 ms** | | Maximum RTT | **169.59 ms** | | Standard deviation | **3.09 ms** | | Stdev / average | **2.0%** | ## Route Context * Departure PoP: [Berlin (BER)](/docs/network/latency/ber-berlin) * Destination PoP: [Singapore (SIN)](/docs/network/latency/sin-singapore) * Region overview: [Europe](/docs/network/latency/regions/europe) → [Asia Pacific](/docs/network/latency/regions/asia-pacific) * Corridor overview: [Europe to Asia Pacific](/docs/network/latency/regions/europe-to-asia-pacific) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Singapore to Berlin latency and RTT](/docs/network/latency/pairs/sin-ber-rtt): 160.3 ms **Fastest routes departing Berlin (BER)** * [Berlin to Frankfurt latency and RTT](/docs/network/latency/pairs/ber-fra-rtt): 6.1 ms * [Berlin to Amsterdam latency and RTT](/docs/network/latency/pairs/ber-ams-rtt): 7.2 ms * [Berlin to London latency and RTT](/docs/network/latency/pairs/ber-lon-rtt): 15.2 ms * [Berlin to Paris latency and RTT](/docs/network/latency/pairs/ber-par-rtt): 15.6 ms * [Berlin to Marseille latency and RTT](/docs/network/latency/pairs/ber-mrs-rtt): 20.2 ms **Fastest routes arriving at Singapore (SIN)** * [Hong Kong to Singapore latency and RTT](/docs/network/latency/pairs/hkg-sin-rtt): 31.5 ms * [Taipei to Singapore latency and RTT](/docs/network/latency/pairs/tpe-sin-rtt): 44.3 ms * [Tokyo to Singapore latency and RTT](/docs/network/latency/pairs/tyo-sin-rtt): 69.4 ms * [Melbourne to Singapore latency and RTT](/docs/network/latency/pairs/mel-sin-rtt): 88.4 ms * [Sydney to Singapore latency and RTT](/docs/network/latency/pairs/syd-sin-rtt): 94.5 ms **Same corridor (Europe → Asia Pacific)** * [Moscow to Hong Kong latency and RTT](/docs/network/latency/pairs/mow-hkg-rtt): 118.6 ms * [Moscow to Taipei latency and RTT](/docs/network/latency/pairs/mow-tpe-rtt): 132.1 ms * [Marseille to Singapore latency and RTT](/docs/network/latency/pairs/mrs-sin-rtt): 139.6 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/ber-sin.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Berlin → Singapore round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/ber-sin-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Berlin, Germany → Sydney, Australia RTT 🇩🇪 **Berlin, Germany (BER)** → 🇦🇺 **Sydney, Australia (SYD)** Measurement round: `2026-08-16T04:07:28Z`. Berlin to Sydney averaged 262.1 ms across 50 replies. The observed range was 255.96 to 280.93 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 5.08 ms, or 1.9% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------------------- | | RTT | **262.1 ms** | | Jitter | **4.49 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **157.54 ms** | | Fiber Efficiency | **60.1%** | | Latency Tier | High | | Source Region | [Europe](/docs/network/latency/regions/europe) | | Destination Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Berlin](https://www.geonames.org/2950159) (52.52437, 13.41053) and [Sydney](https://www.geonames.org/2147714) (-33.86785, 151.20732). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | --------------- | | WGS-84 geodesic distance | **16,087.7 km** | | Vacuum RTT floor | **107.33 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **157.54 ms** | | Low-latency fiber material floor | **156.91 ms** | | Engineering floor (5% path allowance) | **165.44 ms** | | Research 1.33× mapped-fiber reference | **209.53 ms** | | Estimated unamplified path loss | **3378.4 dB** | | Transparent optical spans / inline amplifiers | **212 / 211** | | Published RTT inflation over fiber floor | **1.66×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **255.96 ms** | | Average RTT | **262.1 ms** | | Maximum RTT | **280.93 ms** | | Standard deviation | **5.08 ms** | | Stdev / average | **1.9%** | ## Route Context * Departure PoP: [Berlin (BER)](/docs/network/latency/ber-berlin) * Destination PoP: [Sydney (SYD)](/docs/network/latency/syd-sydney) * Region overview: [Europe](/docs/network/latency/regions/europe) → [Asia Pacific](/docs/network/latency/regions/asia-pacific) * Corridor overview: [Europe to Asia Pacific](/docs/network/latency/regions/europe-to-asia-pacific) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Sydney to Berlin latency and RTT](/docs/network/latency/pairs/syd-ber-rtt): 260.5 ms **Fastest routes departing Berlin (BER)** * [Berlin to Frankfurt latency and RTT](/docs/network/latency/pairs/ber-fra-rtt): 6.1 ms * [Berlin to Amsterdam latency and RTT](/docs/network/latency/pairs/ber-ams-rtt): 7.2 ms * [Berlin to London latency and RTT](/docs/network/latency/pairs/ber-lon-rtt): 15.2 ms * [Berlin to Paris latency and RTT](/docs/network/latency/pairs/ber-par-rtt): 15.6 ms * [Berlin to Marseille latency and RTT](/docs/network/latency/pairs/ber-mrs-rtt): 20.2 ms **Fastest routes arriving at Sydney (SYD)** * [Melbourne to Sydney latency and RTT](/docs/network/latency/pairs/mel-syd-rtt): 9.8 ms * [Singapore to Sydney latency and RTT](/docs/network/latency/pairs/sin-syd-rtt): 94.5 ms * [Tokyo to Sydney latency and RTT](/docs/network/latency/pairs/tyo-syd-rtt): 101.9 ms * [Taipei to Sydney latency and RTT](/docs/network/latency/pairs/tpe-syd-rtt): 133.2 ms * [Hong Kong to Sydney latency and RTT](/docs/network/latency/pairs/hkg-syd-rtt): 137.2 ms **Same corridor (Europe → Asia Pacific)** * [Moscow to Hong Kong latency and RTT](/docs/network/latency/pairs/mow-hkg-rtt): 118.6 ms * [Moscow to Taipei latency and RTT](/docs/network/latency/pairs/mow-tpe-rtt): 132.1 ms * [Marseille to Singapore latency and RTT](/docs/network/latency/pairs/mrs-sin-rtt): 139.6 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/ber-syd.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Berlin → Sydney round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/ber-syd-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Berlin, Germany → Taipei, Taiwan RTT 🇩🇪 **Berlin, Germany (BER)** → 🇹🇼 **Taipei, Taiwan (TPE)** Measurement round: `2026-08-16T04:07:28Z`. Berlin to Taipei averaged 156.8 ms across 50 replies. The observed range was 150.23 to 174.22 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 5.07 ms, or 3.2% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------------------- | | RTT | **156.8 ms** | | Jitter | **4.24 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **87.84 ms** | | Fiber Efficiency | **56%** | | Latency Tier | Fair | | Source Region | [Europe](/docs/network/latency/regions/europe) | | Destination Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Berlin](https://www.geonames.org/2950159) (52.52437, 13.41053) and [Taipei](https://www.geonames.org/1668341) (25.05306, 121.52639). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **8,970.4 km** | | Vacuum RTT floor | **59.84 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **87.84 ms** | | Low-latency fiber material floor | **87.49 ms** | | Engineering floor (5% path allowance) | **92.25 ms** | | Research 1.33× mapped-fiber reference | **116.83 ms** | | Estimated unamplified path loss | **1883.8 dB** | | Transparent optical spans / inline amplifiers | **118 / 117** | | Published RTT inflation over fiber floor | **1.78×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **150.23 ms** | | Average RTT | **156.8 ms** | | Maximum RTT | **174.22 ms** | | Standard deviation | **5.07 ms** | | Stdev / average | **3.2%** | ## Route Context * Departure PoP: [Berlin (BER)](/docs/network/latency/ber-berlin) * Destination PoP: [Taipei (TPE)](/docs/network/latency/tpe-taipei) * Region overview: [Europe](/docs/network/latency/regions/europe) → [Asia Pacific](/docs/network/latency/regions/asia-pacific) * Corridor overview: [Europe to Asia Pacific](/docs/network/latency/regions/europe-to-asia-pacific) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Taipei to Berlin latency and RTT](/docs/network/latency/pairs/tpe-ber-rtt): 158.6 ms **Fastest routes departing Berlin (BER)** * [Berlin to Frankfurt latency and RTT](/docs/network/latency/pairs/ber-fra-rtt): 6.1 ms * [Berlin to Amsterdam latency and RTT](/docs/network/latency/pairs/ber-ams-rtt): 7.2 ms * [Berlin to London latency and RTT](/docs/network/latency/pairs/ber-lon-rtt): 15.2 ms * [Berlin to Paris latency and RTT](/docs/network/latency/pairs/ber-par-rtt): 15.6 ms * [Berlin to Marseille latency and RTT](/docs/network/latency/pairs/ber-mrs-rtt): 20.2 ms **Fastest routes arriving at Taipei (TPE)** * [Hong Kong to Taipei latency and RTT](/docs/network/latency/pairs/hkg-tpe-rtt): 14.7 ms * [Tokyo to Taipei latency and RTT](/docs/network/latency/pairs/tyo-tpe-rtt): 32.3 ms * [Singapore to Taipei latency and RTT](/docs/network/latency/pairs/sin-tpe-rtt): 45.8 ms * [Seattle to Taipei latency and RTT](/docs/network/latency/pairs/sea-tpe-rtt): 114.7 ms * [Sydney to Taipei latency and RTT](/docs/network/latency/pairs/syd-tpe-rtt): 131.7 ms **Same corridor (Europe → Asia Pacific)** * [Moscow to Hong Kong latency and RTT](/docs/network/latency/pairs/mow-hkg-rtt): 118.6 ms * [Moscow to Taipei latency and RTT](/docs/network/latency/pairs/mow-tpe-rtt): 132.1 ms * [Marseille to Singapore latency and RTT](/docs/network/latency/pairs/mrs-sin-rtt): 139.6 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/ber-tpe.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Berlin → Taipei round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/ber-tpe-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Berlin, Germany → Tokyo, Japan RTT 🇩🇪 **Berlin, Germany (BER)** → 🇯🇵 **Tokyo, Japan (TYO)** Measurement round: `2026-08-16T04:07:28Z`. Berlin to Tokyo averaged 187.5 ms, compared with a 87.53 ms city-centre fiber reference. The difference cannot be assigned to a particular cable or carrier from RTT alone. Standard deviation was 4.64 ms, or 2.5% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------------------- | | RTT | **187.5 ms** | | Jitter | **3.12 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **87.53 ms** | | Fiber Efficiency | **46.7%** | | Latency Tier | Fair | | Source Region | [Europe](/docs/network/latency/regions/europe) | | Destination Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Berlin](https://www.geonames.org/2950159) (52.52437, 13.41053) and [Tokyo](https://www.geonames.org/1850147) (35.68950, 139.69171). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **8,937.7 km** | | Vacuum RTT floor | **59.63 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **87.53 ms** | | Low-latency fiber material floor | **87.17 ms** | | Engineering floor (5% path allowance) | **91.91 ms** | | Research 1.33× mapped-fiber reference | **116.41 ms** | | Estimated unamplified path loss | **1876.9 dB** | | Transparent optical spans / inline amplifiers | **118 / 117** | | Published RTT inflation over fiber floor | **2.14×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **181.19 ms** | | Average RTT | **187.5 ms** | | Maximum RTT | **200.11 ms** | | Standard deviation | **4.64 ms** | | Stdev / average | **2.5%** | ## Route Context * Departure PoP: [Berlin (BER)](/docs/network/latency/ber-berlin) * Destination PoP: [Tokyo (TYO)](/docs/network/latency/tyo-tokyo) * Region overview: [Europe](/docs/network/latency/regions/europe) → [Asia Pacific](/docs/network/latency/regions/asia-pacific) * Corridor overview: [Europe to Asia Pacific](/docs/network/latency/regions/europe-to-asia-pacific) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Tokyo to Berlin latency and RTT](/docs/network/latency/pairs/tyo-ber-rtt): 186.1 ms **Fastest routes departing Berlin (BER)** * [Berlin to Frankfurt latency and RTT](/docs/network/latency/pairs/ber-fra-rtt): 6.1 ms * [Berlin to Amsterdam latency and RTT](/docs/network/latency/pairs/ber-ams-rtt): 7.2 ms * [Berlin to London latency and RTT](/docs/network/latency/pairs/ber-lon-rtt): 15.2 ms * [Berlin to Paris latency and RTT](/docs/network/latency/pairs/ber-par-rtt): 15.6 ms * [Berlin to Marseille latency and RTT](/docs/network/latency/pairs/ber-mrs-rtt): 20.2 ms **Fastest routes arriving at Tokyo (TYO)** * [Taipei to Tokyo latency and RTT](/docs/network/latency/pairs/tpe-tyo-rtt): 31.9 ms * [Hong Kong to Tokyo latency and RTT](/docs/network/latency/pairs/hkg-tyo-rtt): 44.9 ms * [Singapore to Tokyo latency and RTT](/docs/network/latency/pairs/sin-tyo-rtt): 68.9 ms * [Seattle to Tokyo latency and RTT](/docs/network/latency/pairs/sea-tyo-rtt): 85 ms * [Los Angeles to Tokyo latency and RTT](/docs/network/latency/pairs/lax-tyo-rtt): 101.2 ms **Same corridor (Europe → Asia Pacific)** * [Moscow to Hong Kong latency and RTT](/docs/network/latency/pairs/mow-hkg-rtt): 118.6 ms * [Moscow to Taipei latency and RTT](/docs/network/latency/pairs/mow-tpe-rtt): 132.1 ms * [Marseille to Singapore latency and RTT](/docs/network/latency/pairs/mrs-sin-rtt): 139.6 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/ber-tyo.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Berlin → Tokyo round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/ber-tyo-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Frankfurt, Germany → Amsterdam, Netherlands RTT 🇩🇪 **Frankfurt, Germany (FRA)** → 🇳🇱 **Amsterdam, Netherlands (AMS)** Measurement round: `2026-08-16T04:07:28Z`. Frankfurt to Amsterdam averaged 5.9 ms across 50 replies. The observed range was 5.62 to 6.84 ms; these figures describe one probe round, not a long-term service target. This is number 1 of 19 outgoing routes by average RTT. That ranking compares different destinations and is not adjusted for distance. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------- | | RTT | **5.9 ms** | | Jitter | **0.17 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **3.57 ms** | | Fiber Efficiency | **60.6%** | | Latency Tier | Ultra-Low | | Source Region | [Europe](/docs/network/latency/regions/europe) | | Destination Region | [Europe](/docs/network/latency/regions/europe) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Frankfurt am Main](https://www.geonames.org/2925533) (50.11552, 8.68417) and [Amsterdam](https://www.geonames.org/2759794) (52.37403, 4.88969). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | ------------ | | WGS-84 geodesic distance | **365.1 km** | | Vacuum RTT floor | **2.44 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **3.57 ms** | | Low-latency fiber material floor | **3.56 ms** | | Engineering floor (5% path allowance) | **3.75 ms** | | Research 1.33× mapped-fiber reference | **4.75 ms** | | Estimated unamplified path loss | **76.7 dB** | | Transparent optical spans / inline amplifiers | **5 / 4** | | Published RTT inflation over fiber floor | **1.65×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ----------- | | Minimum RTT | **5.62 ms** | | Average RTT | **5.9 ms** | | Maximum RTT | **6.84 ms** | | Standard deviation | **0.25 ms** | | Stdev / average | **4.2%** | ## Route Context * Departure PoP: [Frankfurt (FRA)](/docs/network/latency/fra-frankfurt) * Destination PoP: [Amsterdam (AMS)](/docs/network/latency/ams-amsterdam) * Region overview: [Europe](/docs/network/latency/regions/europe) → [Europe](/docs/network/latency/regions/europe) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Amsterdam to Frankfurt latency and RTT](/docs/network/latency/pairs/ams-fra-rtt): 5.9 ms **Fastest routes departing Frankfurt (FRA)** * [Frankfurt to Berlin latency and RTT](/docs/network/latency/pairs/fra-ber-rtt): 6.1 ms * [Frankfurt to Paris latency and RTT](/docs/network/latency/pairs/fra-par-rtt): 7.6 ms * [Frankfurt to London latency and RTT](/docs/network/latency/pairs/fra-lon-rtt): 12.9 ms * [Frankfurt to Marseille latency and RTT](/docs/network/latency/pairs/fra-mrs-rtt): 16.4 ms * [Frankfurt to Moscow latency and RTT](/docs/network/latency/pairs/fra-mow-rtt): 34.7 ms **Fastest routes arriving at Amsterdam (AMS)** * [London to Amsterdam latency and RTT](/docs/network/latency/pairs/lon-ams-rtt): 5.2 ms * [Paris to Amsterdam latency and RTT](/docs/network/latency/pairs/par-ams-rtt): 7 ms * [Berlin to Amsterdam latency and RTT](/docs/network/latency/pairs/ber-ams-rtt): 7.2 ms * [Marseille to Amsterdam latency and RTT](/docs/network/latency/pairs/mrs-ams-rtt): 20.3 ms * [Moscow to Amsterdam latency and RTT](/docs/network/latency/pairs/mow-ams-rtt): 37.6 ms **Same corridor (Europe → Europe)** * [Amsterdam to London latency and RTT](/docs/network/latency/pairs/ams-lon-rtt): 5.2 ms * [Berlin to Frankfurt latency and RTT](/docs/network/latency/pairs/ber-fra-rtt): 6.1 ms * [London to Paris latency and RTT](/docs/network/latency/pairs/lon-par-rtt): 6.4 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/fra-ams.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Frankfurt → Amsterdam round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/fra-ams-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Frankfurt, Germany → Berlin, Germany RTT 🇩🇪 **Frankfurt, Germany (FRA)** → 🇩🇪 **Berlin, Germany (BER)** Measurement round: `2026-08-16T04:07:28Z`. Frankfurt to Berlin averaged 6.1 ms across 50 replies. The observed range was 5.91 to 6.79 ms; these figures describe one probe round, not a long-term service target. This is number 2 of 19 outgoing routes by average RTT. That ranking compares different destinations and is not adjusted for distance. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------- | | RTT | **6.1 ms** | | Jitter | **0.16 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **4.16 ms** | | Fiber Efficiency | **68.2%** | | Latency Tier | Ultra-Low | | Source Region | [Europe](/docs/network/latency/regions/europe) | | Destination Region | [Europe](/docs/network/latency/regions/europe) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Frankfurt am Main](https://www.geonames.org/2925533) (50.11552, 8.68417) and [Berlin](https://www.geonames.org/2950159) (52.52437, 13.41053). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | ------------ | | WGS-84 geodesic distance | **424.6 km** | | Vacuum RTT floor | **2.83 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **4.16 ms** | | Low-latency fiber material floor | **4.14 ms** | | Engineering floor (5% path allowance) | **4.37 ms** | | Research 1.33× mapped-fiber reference | **5.53 ms** | | Estimated unamplified path loss | **89.2 dB** | | Transparent optical spans / inline amplifiers | **6 / 5** | | Published RTT inflation over fiber floor | **1.47×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ----------- | | Minimum RTT | **5.91 ms** | | Average RTT | **6.1 ms** | | Maximum RTT | **6.79 ms** | | Standard deviation | **0.18 ms** | | Stdev / average | **3.0%** | ## Route Context * Departure PoP: [Frankfurt (FRA)](/docs/network/latency/fra-frankfurt) * Destination PoP: [Berlin (BER)](/docs/network/latency/ber-berlin) * Region overview: [Europe](/docs/network/latency/regions/europe) → [Europe](/docs/network/latency/regions/europe) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Berlin to Frankfurt latency and RTT](/docs/network/latency/pairs/ber-fra-rtt): 6.1 ms **Fastest routes departing Frankfurt (FRA)** * [Frankfurt to Amsterdam latency and RTT](/docs/network/latency/pairs/fra-ams-rtt): 5.9 ms * [Frankfurt to Paris latency and RTT](/docs/network/latency/pairs/fra-par-rtt): 7.6 ms * [Frankfurt to London latency and RTT](/docs/network/latency/pairs/fra-lon-rtt): 12.9 ms * [Frankfurt to Marseille latency and RTT](/docs/network/latency/pairs/fra-mrs-rtt): 16.4 ms * [Frankfurt to Moscow latency and RTT](/docs/network/latency/pairs/fra-mow-rtt): 34.7 ms **Fastest routes arriving at Berlin (BER)** * [Amsterdam to Berlin latency and RTT](/docs/network/latency/pairs/ams-ber-rtt): 7.2 ms * [Paris to Berlin latency and RTT](/docs/network/latency/pairs/par-ber-rtt): 15.5 ms * [London to Berlin latency and RTT](/docs/network/latency/pairs/lon-ber-rtt): 16.2 ms * [Marseille to Berlin latency and RTT](/docs/network/latency/pairs/mrs-ber-rtt): 20.6 ms * [Moscow to Berlin latency and RTT](/docs/network/latency/pairs/mow-ber-rtt): 27.5 ms **Same corridor (Europe → Europe)** * [Amsterdam to London latency and RTT](/docs/network/latency/pairs/ams-lon-rtt): 5.2 ms * [London to Amsterdam latency and RTT](/docs/network/latency/pairs/lon-ams-rtt): 5.2 ms * [Amsterdam to Frankfurt latency and RTT](/docs/network/latency/pairs/ams-fra-rtt): 5.9 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/fra-ber.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Frankfurt → Berlin round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/fra-ber-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Frankfurt, Germany → São Paulo, Brazil RTT 🇩🇪 **Frankfurt, Germany (FRA)** → 🇧🇷 **São Paulo, Brazil (GRU)** Measurement round: `2026-08-16T04:07:28Z`. Frankfurt to São Paulo averaged 181.3 ms across 50 replies. The observed range was 176.28 to 197.75 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 3.73 ms, or 2.1% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **181.3 ms** | | Jitter | **3.34 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **96.04 ms** | | Fiber Efficiency | **53%** | | Latency Tier | Fair | | Source Region | [Europe](/docs/network/latency/regions/europe) | | Destination Region | [South America](/docs/network/latency/regions/south-america) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Frankfurt am Main](https://www.geonames.org/2925533) (50.11552, 8.68417) and [São Paulo](https://www.geonames.org/3448439) (-23.54750, -46.63611). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **9,807.7 km** | | Vacuum RTT floor | **65.43 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **96.04 ms** | | Low-latency fiber material floor | **95.66 ms** | | Engineering floor (5% path allowance) | **100.86 ms** | | Research 1.33× mapped-fiber reference | **127.74 ms** | | Estimated unamplified path loss | **2059.6 dB** | | Transparent optical spans / inline amplifiers | **129 / 128** | | Published RTT inflation over fiber floor | **1.89×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **176.28 ms** | | Average RTT | **181.3 ms** | | Maximum RTT | **197.75 ms** | | Standard deviation | **3.73 ms** | | Stdev / average | **2.1%** | ## Route Context * Departure PoP: [Frankfurt (FRA)](/docs/network/latency/fra-frankfurt) * Destination PoP: [São Paulo (GRU)](/docs/network/latency/gru-sao-paulo) * Region overview: [Europe](/docs/network/latency/regions/europe) → [South America](/docs/network/latency/regions/south-america) * Corridor overview: [Europe to South America](/docs/network/latency/regions/europe-to-south-america) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [São Paulo to Frankfurt latency and RTT](/docs/network/latency/pairs/gru-fra-rtt): 186.2 ms **Fastest routes departing Frankfurt (FRA)** * [Frankfurt to Amsterdam latency and RTT](/docs/network/latency/pairs/fra-ams-rtt): 5.9 ms * [Frankfurt to Berlin latency and RTT](/docs/network/latency/pairs/fra-ber-rtt): 6.1 ms * [Frankfurt to Paris latency and RTT](/docs/network/latency/pairs/fra-par-rtt): 7.6 ms * [Frankfurt to London latency and RTT](/docs/network/latency/pairs/fra-lon-rtt): 12.9 ms * [Frankfurt to Marseille latency and RTT](/docs/network/latency/pairs/fra-mrs-rtt): 16.4 ms **Fastest routes arriving at São Paulo (GRU)** * [Ashburn to São Paulo latency and RTT](/docs/network/latency/pairs/iad-gru-rtt): 101.4 ms * [New York to São Paulo latency and RTT](/docs/network/latency/pairs/nyc-gru-rtt): 106.9 ms * [Miami to São Paulo latency and RTT](/docs/network/latency/pairs/mia-gru-rtt): 128.7 ms * [Los Angeles to São Paulo latency and RTT](/docs/network/latency/pairs/lax-gru-rtt): 131.3 ms * [Seattle to São Paulo latency and RTT](/docs/network/latency/pairs/sea-gru-rtt): 155.7 ms **Same corridor (Europe → South America)** * [London to São Paulo latency and RTT](/docs/network/latency/pairs/lon-gru-rtt): 170.7 ms * [Amsterdam to São Paulo latency and RTT](/docs/network/latency/pairs/ams-gru-rtt): 175.5 ms * [Paris to São Paulo latency and RTT](/docs/network/latency/pairs/par-gru-rtt): 175.8 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/fra-gru.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Frankfurt → São Paulo round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/fra-gru-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Frankfurt, Germany → Hong Kong RTT 🇩🇪 **Frankfurt, Germany (FRA)** → 🇭🇰 **Hong Kong (HKG)** Measurement round: `2026-08-16T04:07:28Z`. Frankfurt to Hong Kong averaged 152 ms across 50 replies. The observed range was 144.15 to 171.03 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 6.84 ms, or 4.5% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------------------- | | RTT | **152 ms** | | Jitter | **4.6 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **89.87 ms** | | Fiber Efficiency | **59.1%** | | Latency Tier | Fair | | Source Region | [Europe](/docs/network/latency/regions/europe) | | Destination Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Frankfurt am Main](https://www.geonames.org/2925533) (50.11552, 8.68417) and [Hong Kong](https://www.geonames.org/1819729) (22.27832, 114.17469). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **9,177.1 km** | | Vacuum RTT floor | **61.22 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **89.87 ms** | | Low-latency fiber material floor | **89.51 ms** | | Engineering floor (5% path allowance) | **94.37 ms** | | Research 1.33× mapped-fiber reference | **119.53 ms** | | Estimated unamplified path loss | **1927.2 dB** | | Transparent optical spans / inline amplifiers | **121 / 120** | | Published RTT inflation over fiber floor | **1.69×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **144.15 ms** | | Average RTT | **152 ms** | | Maximum RTT | **171.03 ms** | | Standard deviation | **6.84 ms** | | Stdev / average | **4.5%** | ## Route Context * Departure PoP: [Frankfurt (FRA)](/docs/network/latency/fra-frankfurt) * Destination PoP: [Hong Kong (HKG)](/docs/network/latency/hkg-hong-kong) * Region overview: [Europe](/docs/network/latency/regions/europe) → [Asia Pacific](/docs/network/latency/regions/asia-pacific) * Corridor overview: [Europe to Asia Pacific](/docs/network/latency/regions/europe-to-asia-pacific) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Hong Kong to Frankfurt latency and RTT](/docs/network/latency/pairs/hkg-fra-rtt): 150.7 ms **Fastest routes departing Frankfurt (FRA)** * [Frankfurt to Amsterdam latency and RTT](/docs/network/latency/pairs/fra-ams-rtt): 5.9 ms * [Frankfurt to Berlin latency and RTT](/docs/network/latency/pairs/fra-ber-rtt): 6.1 ms * [Frankfurt to Paris latency and RTT](/docs/network/latency/pairs/fra-par-rtt): 7.6 ms * [Frankfurt to London latency and RTT](/docs/network/latency/pairs/fra-lon-rtt): 12.9 ms * [Frankfurt to Marseille latency and RTT](/docs/network/latency/pairs/fra-mrs-rtt): 16.4 ms **Fastest routes arriving at Hong Kong (HKG)** * [Taipei to Hong Kong latency and RTT](/docs/network/latency/pairs/tpe-hkg-rtt): 15.5 ms * [Singapore to Hong Kong latency and RTT](/docs/network/latency/pairs/sin-hkg-rtt): 30.8 ms * [Tokyo to Hong Kong latency and RTT](/docs/network/latency/pairs/tyo-hkg-rtt): 44.8 ms * [Moscow to Hong Kong latency and RTT](/docs/network/latency/pairs/mow-hkg-rtt): 118.6 ms * [Seattle to Hong Kong latency and RTT](/docs/network/latency/pairs/sea-hkg-rtt): 130.4 ms **Same corridor (Europe → Asia Pacific)** * [Moscow to Taipei latency and RTT](/docs/network/latency/pairs/mow-tpe-rtt): 132.1 ms * [Marseille to Singapore latency and RTT](/docs/network/latency/pairs/mrs-sin-rtt): 139.6 ms * [Berlin to Hong Kong latency and RTT](/docs/network/latency/pairs/ber-hkg-rtt): 144.9 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/fra-hkg.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Frankfurt → Hong Kong round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/fra-hkg-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Frankfurt, Germany → Ashburn, USA RTT 🇩🇪 **Frankfurt, Germany (FRA)** → 🇺🇸 **Ashburn, USA (IAD)** Measurement round: `2026-08-16T04:07:28Z`. Frankfurt to Ashburn averaged 80.7 ms across 50 replies. The observed range was 79.19 to 87.49 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 1.42 ms, or 1.8% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **80.7 ms** | | Jitter | **1.11 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **64.3 ms** | | Fiber Efficiency | **79.7%** | | Latency Tier | Good | | Source Region | [Europe](/docs/network/latency/regions/europe) | | Destination Region | [North America](/docs/network/latency/regions/north-america) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Frankfurt am Main](https://www.geonames.org/2925533) (50.11552, 8.68417) and [Ashburn](https://www.geonames.org/4744870) (39.04372, -77.48749). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | ------------- | | WGS-84 geodesic distance | **6,566 km** | | Vacuum RTT floor | **43.8 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **64.3 ms** | | Low-latency fiber material floor | **64.04 ms** | | Engineering floor (5% path allowance) | **67.52 ms** | | Research 1.33× mapped-fiber reference | **85.52 ms** | | Estimated unamplified path loss | **1378.9 dB** | | Transparent optical spans / inline amplifiers | **87 / 86** | | Published RTT inflation over fiber floor | **1.26×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------ | | Minimum RTT | **79.19 ms** | | Average RTT | **80.7 ms** | | Maximum RTT | **87.49 ms** | | Standard deviation | **1.42 ms** | | Stdev / average | **1.8%** | ## Route Context * Departure PoP: [Frankfurt (FRA)](/docs/network/latency/fra-frankfurt) * Destination PoP: [Ashburn (IAD)](/docs/network/latency/iad-ashburn) * Region overview: [Europe](/docs/network/latency/regions/europe) → [North America](/docs/network/latency/regions/north-america) * Corridor overview: [Europe to North America](/docs/network/latency/regions/europe-to-north-america) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Ashburn to Frankfurt latency and RTT](/docs/network/latency/pairs/iad-fra-rtt): 81.2 ms **Fastest routes departing Frankfurt (FRA)** * [Frankfurt to Amsterdam latency and RTT](/docs/network/latency/pairs/fra-ams-rtt): 5.9 ms * [Frankfurt to Berlin latency and RTT](/docs/network/latency/pairs/fra-ber-rtt): 6.1 ms * [Frankfurt to Paris latency and RTT](/docs/network/latency/pairs/fra-par-rtt): 7.6 ms * [Frankfurt to London latency and RTT](/docs/network/latency/pairs/fra-lon-rtt): 12.9 ms * [Frankfurt to Marseille latency and RTT](/docs/network/latency/pairs/fra-mrs-rtt): 16.4 ms **Fastest routes arriving at Ashburn (IAD)** * [New York to Ashburn latency and RTT](/docs/network/latency/pairs/nyc-iad-rtt): 6 ms * [Miami to Ashburn latency and RTT](/docs/network/latency/pairs/mia-iad-rtt): 27.3 ms * [Los Angeles to Ashburn latency and RTT](/docs/network/latency/pairs/lax-iad-rtt): 60.4 ms * [Seattle to Ashburn latency and RTT](/docs/network/latency/pairs/sea-iad-rtt): 61.5 ms * [London to Ashburn latency and RTT](/docs/network/latency/pairs/lon-iad-rtt): 71 ms **Same corridor (Europe → North America)** * [London to New York latency and RTT](/docs/network/latency/pairs/lon-nyc-rtt): 63.8 ms * [Amsterdam to New York latency and RTT](/docs/network/latency/pairs/ams-nyc-rtt): 68.6 ms * [Paris to New York latency and RTT](/docs/network/latency/pairs/par-nyc-rtt): 68.9 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/fra-iad.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Frankfurt → Ashburn round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/fra-iad-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Frankfurt, Germany → Johannesburg, South Africa RTT 🇩🇪 **Frankfurt, Germany (FRA)** → 🇿🇦 **Johannesburg, South Africa (JNB)** Measurement round: `2026-08-16T04:07:28Z`. Frankfurt to Johannesburg averaged 188.4 ms, compared with a 84.87 ms city-centre fiber reference. The difference cannot be assigned to a particular cable or carrier from RTT alone. Standard deviation was 6.75 ms, or 3.6% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------- | | RTT | **188.4 ms** | | Jitter | **4.87 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **84.87 ms** | | Fiber Efficiency | **45%** | | Latency Tier | Fair | | Source Region | [Europe](/docs/network/latency/regions/europe) | | Destination Region | [Africa](/docs/network/latency/regions/africa) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Frankfurt am Main](https://www.geonames.org/2925533) (50.11552, 8.68417) and [Johannesburg](https://www.geonames.org/993800) (-26.20227, 28.04363). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **8,666.9 km** | | Vacuum RTT floor | **57.82 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **84.87 ms** | | Low-latency fiber material floor | **84.53 ms** | | Engineering floor (5% path allowance) | **89.13 ms** | | Research 1.33× mapped-fiber reference | **112.88 ms** | | Estimated unamplified path loss | **1820.1 dB** | | Transparent optical spans / inline amplifiers | **114 / 113** | | Published RTT inflation over fiber floor | **2.22×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **182.33 ms** | | Average RTT | **188.4 ms** | | Maximum RTT | **218.93 ms** | | Standard deviation | **6.75 ms** | | Stdev / average | **3.6%** | ## Route Context * Departure PoP: [Frankfurt (FRA)](/docs/network/latency/fra-frankfurt) * Destination PoP: [Johannesburg (JNB)](/docs/network/latency/jnb-johannesburg) * Region overview: [Europe](/docs/network/latency/regions/europe) → [Africa](/docs/network/latency/regions/africa) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Johannesburg to Frankfurt latency and RTT](/docs/network/latency/pairs/jnb-fra-rtt): 166 ms **Fastest routes departing Frankfurt (FRA)** * [Frankfurt to Amsterdam latency and RTT](/docs/network/latency/pairs/fra-ams-rtt): 5.9 ms * [Frankfurt to Berlin latency and RTT](/docs/network/latency/pairs/fra-ber-rtt): 6.1 ms * [Frankfurt to Paris latency and RTT](/docs/network/latency/pairs/fra-par-rtt): 7.6 ms * [Frankfurt to London latency and RTT](/docs/network/latency/pairs/fra-lon-rtt): 12.9 ms * [Frankfurt to Marseille latency and RTT](/docs/network/latency/pairs/fra-mrs-rtt): 16.4 ms **Fastest routes arriving at Johannesburg (JNB)** * [Amsterdam to Johannesburg latency and RTT](/docs/network/latency/pairs/ams-jnb-rtt): 171.9 ms * [Berlin to Johannesburg latency and RTT](/docs/network/latency/pairs/ber-jnb-rtt): 172.1 ms * [London to Johannesburg latency and RTT](/docs/network/latency/pairs/lon-jnb-rtt): 179.6 ms * [Paris to Johannesburg latency and RTT](/docs/network/latency/pairs/par-jnb-rtt): 180.9 ms * [Marseille to Johannesburg latency and RTT](/docs/network/latency/pairs/mrs-jnb-rtt): 198.5 ms **Same corridor (Europe → Africa)** * [Moscow to Johannesburg latency and RTT](/docs/network/latency/pairs/mow-jnb-rtt): 201.9 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/fra-jnb.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Frankfurt → Johannesburg round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/fra-jnb-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Frankfurt, Germany → Los Angeles, USA RTT 🇩🇪 **Frankfurt, Germany (FRA)** → 🇺🇸 **Los Angeles, USA (LAX)** Measurement round: `2026-08-16T04:07:28Z`. Frankfurt to Los Angeles averaged 141.8 ms across 50 replies. The observed range was 135.2 to 158.63 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 5.99 ms, or 4.2% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **141.8 ms** | | Jitter | **5.12 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **91.31 ms** | | Fiber Efficiency | **64.4%** | | Latency Tier | Good | | Source Region | [Europe](/docs/network/latency/regions/europe) | | Destination Region | [North America](/docs/network/latency/regions/north-america) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Frankfurt am Main](https://www.geonames.org/2925533) (50.11552, 8.68417) and [Los Angeles](https://www.geonames.org/5368361) (34.05223, -118.24368). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **9,324.5 km** | | Vacuum RTT floor | **62.21 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **91.31 ms** | | Low-latency fiber material floor | **90.95 ms** | | Engineering floor (5% path allowance) | **95.89 ms** | | Research 1.33× mapped-fiber reference | **121.45 ms** | | Estimated unamplified path loss | **1958.1 dB** | | Transparent optical spans / inline amplifiers | **123 / 122** | | Published RTT inflation over fiber floor | **1.55×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **135.2 ms** | | Average RTT | **141.8 ms** | | Maximum RTT | **158.63 ms** | | Standard deviation | **5.99 ms** | | Stdev / average | **4.2%** | ## Route Context * Departure PoP: [Frankfurt (FRA)](/docs/network/latency/fra-frankfurt) * Destination PoP: [Los Angeles (LAX)](/docs/network/latency/lax-los-angeles) * Region overview: [Europe](/docs/network/latency/regions/europe) → [North America](/docs/network/latency/regions/north-america) * Corridor overview: [Europe to North America](/docs/network/latency/regions/europe-to-north-america) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Los Angeles to Frankfurt latency and RTT](/docs/network/latency/pairs/lax-fra-rtt): 142.3 ms **Fastest routes departing Frankfurt (FRA)** * [Frankfurt to Amsterdam latency and RTT](/docs/network/latency/pairs/fra-ams-rtt): 5.9 ms * [Frankfurt to Berlin latency and RTT](/docs/network/latency/pairs/fra-ber-rtt): 6.1 ms * [Frankfurt to Paris latency and RTT](/docs/network/latency/pairs/fra-par-rtt): 7.6 ms * [Frankfurt to London latency and RTT](/docs/network/latency/pairs/fra-lon-rtt): 12.9 ms * [Frankfurt to Marseille latency and RTT](/docs/network/latency/pairs/fra-mrs-rtt): 16.4 ms **Fastest routes arriving at Los Angeles (LAX)** * [Seattle to Los Angeles latency and RTT](/docs/network/latency/pairs/sea-lax-rtt): 25.9 ms * [Miami to Los Angeles latency and RTT](/docs/network/latency/pairs/mia-lax-rtt): 56.8 ms * [New York to Los Angeles latency and RTT](/docs/network/latency/pairs/nyc-lax-rtt): 58.9 ms * [Ashburn to Los Angeles latency and RTT](/docs/network/latency/pairs/iad-lax-rtt): 60.7 ms * [Tokyo to Los Angeles latency and RTT](/docs/network/latency/pairs/tyo-lax-rtt): 101.2 ms **Same corridor (Europe → North America)** * [London to New York latency and RTT](/docs/network/latency/pairs/lon-nyc-rtt): 63.8 ms * [Amsterdam to New York latency and RTT](/docs/network/latency/pairs/ams-nyc-rtt): 68.6 ms * [Paris to New York latency and RTT](/docs/network/latency/pairs/par-nyc-rtt): 68.9 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/fra-lax.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Frankfurt → Los Angeles round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/fra-lax-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Frankfurt, Germany → London, UK RTT 🇩🇪 **Frankfurt, Germany (FRA)** → 🇬🇧 **London, UK (LON)** Measurement round: `2026-08-16T04:07:28Z`. Frankfurt to London averaged 12.9 ms, compared with a 6.26 ms city-centre fiber reference. The difference cannot be assigned to a particular cable or carrier from RTT alone. This is number 4 of 19 outgoing routes by average RTT. That ranking compares different destinations and is not adjusted for distance. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------- | | RTT | **12.9 ms** | | Jitter | **0.33 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **6.26 ms** | | Fiber Efficiency | **48.5%** | | Latency Tier | Ultra-Low | | Source Region | [Europe](/docs/network/latency/regions/europe) | | Destination Region | [Europe](/docs/network/latency/regions/europe) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Frankfurt am Main](https://www.geonames.org/2925533) (50.11552, 8.68417) and [London](https://www.geonames.org/2643743) (51.50853, -0.12574). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | ------------ | | WGS-84 geodesic distance | **639.5 km** | | Vacuum RTT floor | **4.27 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **6.26 ms** | | Low-latency fiber material floor | **6.24 ms** | | Engineering floor (5% path allowance) | **6.58 ms** | | Research 1.33× mapped-fiber reference | **8.33 ms** | | Estimated unamplified path loss | **134.3 dB** | | Transparent optical spans / inline amplifiers | **9 / 8** | | Published RTT inflation over fiber floor | **2.06×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------ | | Minimum RTT | **12.38 ms** | | Average RTT | **12.9 ms** | | Maximum RTT | **14.46 ms** | | Standard deviation | **0.43 ms** | | Stdev / average | **3.3%** | ## Route Context * Departure PoP: [Frankfurt (FRA)](/docs/network/latency/fra-frankfurt) * Destination PoP: [London (LON)](/docs/network/latency/lon-london) * Region overview: [Europe](/docs/network/latency/regions/europe) → [Europe](/docs/network/latency/regions/europe) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [London to Frankfurt latency and RTT](/docs/network/latency/pairs/lon-fra-rtt): 13.6 ms **Fastest routes departing Frankfurt (FRA)** * [Frankfurt to Amsterdam latency and RTT](/docs/network/latency/pairs/fra-ams-rtt): 5.9 ms * [Frankfurt to Berlin latency and RTT](/docs/network/latency/pairs/fra-ber-rtt): 6.1 ms * [Frankfurt to Paris latency and RTT](/docs/network/latency/pairs/fra-par-rtt): 7.6 ms * [Frankfurt to Marseille latency and RTT](/docs/network/latency/pairs/fra-mrs-rtt): 16.4 ms * [Frankfurt to Moscow latency and RTT](/docs/network/latency/pairs/fra-mow-rtt): 34.7 ms **Fastest routes arriving at London (LON)** * [Amsterdam to London latency and RTT](/docs/network/latency/pairs/ams-lon-rtt): 5.2 ms * [Paris to London latency and RTT](/docs/network/latency/pairs/par-lon-rtt): 6.4 ms * [Berlin to London latency and RTT](/docs/network/latency/pairs/ber-lon-rtt): 15.2 ms * [Marseille to London latency and RTT](/docs/network/latency/pairs/mrs-lon-rtt): 18.9 ms * [Moscow to London latency and RTT](/docs/network/latency/pairs/mow-lon-rtt): 42.4 ms **Same corridor (Europe → Europe)** * [London to Amsterdam latency and RTT](/docs/network/latency/pairs/lon-ams-rtt): 5.2 ms * [Amsterdam to Frankfurt latency and RTT](/docs/network/latency/pairs/ams-fra-rtt): 5.9 ms * [Berlin to Frankfurt latency and RTT](/docs/network/latency/pairs/ber-fra-rtt): 6.1 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/fra-lon.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Frankfurt → London round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/fra-lon-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Frankfurt, Germany → Melbourne, Australia RTT 🇩🇪 **Frankfurt, Germany (FRA)** → 🇦🇺 **Melbourne, Australia (MEL)** Measurement round: `2026-08-16T04:07:28Z`. Frankfurt to Melbourne averaged 250.1 ms across 50 replies. The observed range was 236.31 to 287.22 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 11.19 ms, or 4.5% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------------------- | | RTT | **250.1 ms** | | Jitter | **9.69 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **159.77 ms** | | Fiber Efficiency | **63.9%** | | Latency Tier | High | | Source Region | [Europe](/docs/network/latency/regions/europe) | | Destination Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Frankfurt am Main](https://www.geonames.org/2925533) (50.11552, 8.68417) and [Melbourne](https://www.geonames.org/2158177) (-37.81400, 144.96332). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | --------------- | | WGS-84 geodesic distance | **16,315.5 km** | | Vacuum RTT floor | **108.85 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **159.77 ms** | | Low-latency fiber material floor | **159.13 ms** | | Engineering floor (5% path allowance) | **167.78 ms** | | Research 1.33× mapped-fiber reference | **212.5 ms** | | Estimated unamplified path loss | **3426.2 dB** | | Transparent optical spans / inline amplifiers | **215 / 214** | | Published RTT inflation over fiber floor | **1.57×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **236.31 ms** | | Average RTT | **250.1 ms** | | Maximum RTT | **287.22 ms** | | Standard deviation | **11.19 ms** | | Stdev / average | **4.5%** | ## Route Context * Departure PoP: [Frankfurt (FRA)](/docs/network/latency/fra-frankfurt) * Destination PoP: [Melbourne (MEL)](/docs/network/latency/mel-melbourne) * Region overview: [Europe](/docs/network/latency/regions/europe) → [Asia Pacific](/docs/network/latency/regions/asia-pacific) * Corridor overview: [Europe to Asia Pacific](/docs/network/latency/regions/europe-to-asia-pacific) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Melbourne to Frankfurt latency and RTT](/docs/network/latency/pairs/mel-fra-rtt): 248.6 ms **Fastest routes departing Frankfurt (FRA)** * [Frankfurt to Amsterdam latency and RTT](/docs/network/latency/pairs/fra-ams-rtt): 5.9 ms * [Frankfurt to Berlin latency and RTT](/docs/network/latency/pairs/fra-ber-rtt): 6.1 ms * [Frankfurt to Paris latency and RTT](/docs/network/latency/pairs/fra-par-rtt): 7.6 ms * [Frankfurt to London latency and RTT](/docs/network/latency/pairs/fra-lon-rtt): 12.9 ms * [Frankfurt to Marseille latency and RTT](/docs/network/latency/pairs/fra-mrs-rtt): 16.4 ms **Fastest routes arriving at Melbourne (MEL)** * [Sydney to Melbourne latency and RTT](/docs/network/latency/pairs/syd-mel-rtt): 9.8 ms * [Singapore to Melbourne latency and RTT](/docs/network/latency/pairs/sin-mel-rtt): 88.6 ms * [Tokyo to Melbourne latency and RTT](/docs/network/latency/pairs/tyo-mel-rtt): 113.9 ms * [Hong Kong to Melbourne latency and RTT](/docs/network/latency/pairs/hkg-mel-rtt): 138.4 ms * [Taipei to Melbourne latency and RTT](/docs/network/latency/pairs/tpe-mel-rtt): 142 ms **Same corridor (Europe → Asia Pacific)** * [Moscow to Hong Kong latency and RTT](/docs/network/latency/pairs/mow-hkg-rtt): 118.6 ms * [Moscow to Taipei latency and RTT](/docs/network/latency/pairs/mow-tpe-rtt): 132.1 ms * [Marseille to Singapore latency and RTT](/docs/network/latency/pairs/mrs-sin-rtt): 139.6 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/fra-mel.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Frankfurt → Melbourne round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/fra-mel-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Frankfurt, Germany → Miami, USA RTT 🇩🇪 **Frankfurt, Germany (FRA)** → 🇺🇸 **Miami, USA (MIA)** Measurement round: `2026-08-16T04:07:28Z`. Frankfurt to Miami averaged 112 ms across 50 replies. The observed range was 110.01 to 117.35 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 1.8 ms, or 1.6% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **112 ms** | | Jitter | **1.53 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **76.17 ms** | | Fiber Efficiency | **68%** | | Latency Tier | Good | | Source Region | [Europe](/docs/network/latency/regions/europe) | | Destination Region | [North America](/docs/network/latency/regions/north-america) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Frankfurt am Main](https://www.geonames.org/2925533) (50.11552, 8.68417) and [Miami](https://www.geonames.org/4164138) (25.77427, -80.19366). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **7,777.9 km** | | Vacuum RTT floor | **51.89 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **76.17 ms** | | Low-latency fiber material floor | **75.86 ms** | | Engineering floor (5% path allowance) | **79.99 ms** | | Research 1.33× mapped-fiber reference | **101.3 ms** | | Estimated unamplified path loss | **1633.4 dB** | | Transparent optical spans / inline amplifiers | **103 / 102** | | Published RTT inflation over fiber floor | **1.47×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **110.01 ms** | | Average RTT | **112 ms** | | Maximum RTT | **117.35 ms** | | Standard deviation | **1.8 ms** | | Stdev / average | **1.6%** | ## Route Context * Departure PoP: [Frankfurt (FRA)](/docs/network/latency/fra-frankfurt) * Destination PoP: [Miami (MIA)](/docs/network/latency/mia-miami) * Region overview: [Europe](/docs/network/latency/regions/europe) → [North America](/docs/network/latency/regions/north-america) * Corridor overview: [Europe to North America](/docs/network/latency/regions/europe-to-north-america) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Miami to Frankfurt latency and RTT](/docs/network/latency/pairs/mia-fra-rtt): 112.2 ms **Fastest routes departing Frankfurt (FRA)** * [Frankfurt to Amsterdam latency and RTT](/docs/network/latency/pairs/fra-ams-rtt): 5.9 ms * [Frankfurt to Berlin latency and RTT](/docs/network/latency/pairs/fra-ber-rtt): 6.1 ms * [Frankfurt to Paris latency and RTT](/docs/network/latency/pairs/fra-par-rtt): 7.6 ms * [Frankfurt to London latency and RTT](/docs/network/latency/pairs/fra-lon-rtt): 12.9 ms * [Frankfurt to Marseille latency and RTT](/docs/network/latency/pairs/fra-mrs-rtt): 16.4 ms **Fastest routes arriving at Miami (MIA)** * [Ashburn to Miami latency and RTT](/docs/network/latency/pairs/iad-mia-rtt): 27.4 ms * [New York to Miami latency and RTT](/docs/network/latency/pairs/nyc-mia-rtt): 32.9 ms * [Los Angeles to Miami latency and RTT](/docs/network/latency/pairs/lax-mia-rtt): 57.3 ms * [São Paulo to Miami latency and RTT](/docs/network/latency/pairs/gru-mia-rtt): 74 ms * [Seattle to Miami latency and RTT](/docs/network/latency/pairs/sea-mia-rtt): 81.7 ms **Same corridor (Europe → North America)** * [London to New York latency and RTT](/docs/network/latency/pairs/lon-nyc-rtt): 63.8 ms * [Amsterdam to New York latency and RTT](/docs/network/latency/pairs/ams-nyc-rtt): 68.6 ms * [Paris to New York latency and RTT](/docs/network/latency/pairs/par-nyc-rtt): 68.9 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/fra-mia.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Frankfurt → Miami round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/fra-mia-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Frankfurt, Germany → Moscow, Russia RTT 🇩🇪 **Frankfurt, Germany (FRA)** → 🇷🇺 **Moscow, Russia (MOW)** Measurement round: `2026-08-16T04:07:28Z`. Frankfurt to Moscow averaged 34.7 ms across 50 replies. The observed range was 32.66 to 40.02 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 1.68 ms, or 4.8% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------- | | RTT | **34.7 ms** | | Jitter | **1.68 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **19.85 ms** | | Fiber Efficiency | **57.2%** | | Latency Tier | Excellent | | Source Region | [Europe](/docs/network/latency/regions/europe) | | Destination Region | [Europe](/docs/network/latency/regions/europe) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Frankfurt am Main](https://www.geonames.org/2925533) (50.11552, 8.68417) and [Moscow](https://www.geonames.org/524901) (55.75204, 37.61781). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **2,026.8 km** | | Vacuum RTT floor | **13.52 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **19.85 ms** | | Low-latency fiber material floor | **19.77 ms** | | Engineering floor (5% path allowance) | **20.84 ms** | | Research 1.33× mapped-fiber reference | **26.4 ms** | | Estimated unamplified path loss | **425.6 dB** | | Transparent optical spans / inline amplifiers | **27 / 26** | | Published RTT inflation over fiber floor | **1.75×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------ | | Minimum RTT | **32.66 ms** | | Average RTT | **34.7 ms** | | Maximum RTT | **40.02 ms** | | Standard deviation | **1.68 ms** | | Stdev / average | **4.8%** | ## Route Context * Departure PoP: [Frankfurt (FRA)](/docs/network/latency/fra-frankfurt) * Destination PoP: [Moscow (MOW)](/docs/network/latency/mow-moscow) * Region overview: [Europe](/docs/network/latency/regions/europe) → [Europe](/docs/network/latency/regions/europe) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Moscow to Frankfurt latency and RTT](/docs/network/latency/pairs/mow-fra-rtt): 35.9 ms **Fastest routes departing Frankfurt (FRA)** * [Frankfurt to Amsterdam latency and RTT](/docs/network/latency/pairs/fra-ams-rtt): 5.9 ms * [Frankfurt to Berlin latency and RTT](/docs/network/latency/pairs/fra-ber-rtt): 6.1 ms * [Frankfurt to Paris latency and RTT](/docs/network/latency/pairs/fra-par-rtt): 7.6 ms * [Frankfurt to London latency and RTT](/docs/network/latency/pairs/fra-lon-rtt): 12.9 ms * [Frankfurt to Marseille latency and RTT](/docs/network/latency/pairs/fra-mrs-rtt): 16.4 ms **Fastest routes arriving at Moscow (MOW)** * [Berlin to Moscow latency and RTT](/docs/network/latency/pairs/ber-mow-rtt): 28.6 ms * [Amsterdam to Moscow latency and RTT](/docs/network/latency/pairs/ams-mow-rtt): 38.1 ms * [London to Moscow latency and RTT](/docs/network/latency/pairs/lon-mow-rtt): 43.5 ms * [Paris to Moscow latency and RTT](/docs/network/latency/pairs/par-mow-rtt): 44.7 ms * [Marseille to Moscow latency and RTT](/docs/network/latency/pairs/mrs-mow-rtt): 49.6 ms **Same corridor (Europe → Europe)** * [Amsterdam to London latency and RTT](/docs/network/latency/pairs/ams-lon-rtt): 5.2 ms * [London to Amsterdam latency and RTT](/docs/network/latency/pairs/lon-ams-rtt): 5.2 ms * [Amsterdam to Frankfurt latency and RTT](/docs/network/latency/pairs/ams-fra-rtt): 5.9 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/fra-mow.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Frankfurt → Moscow round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/fra-mow-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Frankfurt, Germany → Marseille, France RTT 🇩🇪 **Frankfurt, Germany (FRA)** → 🇫🇷 **Marseille, France (MRS)** Measurement round: `2026-08-16T04:07:28Z`. Frankfurt to Marseille averaged 16.4 ms, compared with a 7.82 ms city-centre fiber reference. The difference cannot be assigned to a particular cable or carrier from RTT alone. This is number 5 of 19 outgoing routes by average RTT. That ranking compares different destinations and is not adjusted for distance. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------- | | RTT | **16.4 ms** | | Jitter | **0.31 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **7.82 ms** | | Fiber Efficiency | **47.7%** | | Latency Tier | Ultra-Low | | Source Region | [Europe](/docs/network/latency/regions/europe) | | Destination Region | [Europe](/docs/network/latency/regions/europe) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Frankfurt am Main](https://www.geonames.org/2925533) (50.11552, 8.68417) and [Marseille](https://www.geonames.org/2995469) (43.29695, 5.38107). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | ------------ | | WGS-84 geodesic distance | **798.8 km** | | Vacuum RTT floor | **5.33 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **7.82 ms** | | Low-latency fiber material floor | **7.79 ms** | | Engineering floor (5% path allowance) | **8.21 ms** | | Research 1.33× mapped-fiber reference | **10.4 ms** | | Estimated unamplified path loss | **167.7 dB** | | Transparent optical spans / inline amplifiers | **11 / 10** | | Published RTT inflation over fiber floor | **2.1×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------ | | Minimum RTT | **15.96 ms** | | Average RTT | **16.4 ms** | | Maximum RTT | **17.74 ms** | | Standard deviation | **0.37 ms** | | Stdev / average | **2.3%** | ## Route Context * Departure PoP: [Frankfurt (FRA)](/docs/network/latency/fra-frankfurt) * Destination PoP: [Marseille (MRS)](/docs/network/latency/mrs-marseille) * Region overview: [Europe](/docs/network/latency/regions/europe) → [Europe](/docs/network/latency/regions/europe) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Marseille to Frankfurt latency and RTT](/docs/network/latency/pairs/mrs-fra-rtt): 16 ms **Fastest routes departing Frankfurt (FRA)** * [Frankfurt to Amsterdam latency and RTT](/docs/network/latency/pairs/fra-ams-rtt): 5.9 ms * [Frankfurt to Berlin latency and RTT](/docs/network/latency/pairs/fra-ber-rtt): 6.1 ms * [Frankfurt to Paris latency and RTT](/docs/network/latency/pairs/fra-par-rtt): 7.6 ms * [Frankfurt to London latency and RTT](/docs/network/latency/pairs/fra-lon-rtt): 12.9 ms * [Frankfurt to Moscow latency and RTT](/docs/network/latency/pairs/fra-mow-rtt): 34.7 ms **Fastest routes arriving at Marseille (MRS)** * [Paris to Marseille latency and RTT](/docs/network/latency/pairs/par-mrs-rtt): 8.9 ms * [London to Marseille latency and RTT](/docs/network/latency/pairs/lon-mrs-rtt): 17.9 ms * [Amsterdam to Marseille latency and RTT](/docs/network/latency/pairs/ams-mrs-rtt): 19.6 ms * [Berlin to Marseille latency and RTT](/docs/network/latency/pairs/ber-mrs-rtt): 20.2 ms * [Moscow to Marseille latency and RTT](/docs/network/latency/pairs/mow-mrs-rtt): 49.9 ms **Same corridor (Europe → Europe)** * [Amsterdam to London latency and RTT](/docs/network/latency/pairs/ams-lon-rtt): 5.2 ms * [London to Amsterdam latency and RTT](/docs/network/latency/pairs/lon-ams-rtt): 5.2 ms * [Amsterdam to Frankfurt latency and RTT](/docs/network/latency/pairs/ams-fra-rtt): 5.9 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/fra-mrs.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Frankfurt → Marseille round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/fra-mrs-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Frankfurt, Germany → New York, USA RTT 🇩🇪 **Frankfurt, Germany (FRA)** → 🇺🇸 **New York, USA (NYC)** Measurement round: `2026-08-16T04:07:28Z`. Frankfurt to New York averaged 74.4 ms across 50 replies. The observed range was 72.04 to 80.52 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 2.09 ms, or 2.8% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **74.4 ms** | | Jitter | **1.69 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **60.9 ms** | | Fiber Efficiency | **81.9%** | | Latency Tier | Excellent | | Source Region | [Europe](/docs/network/latency/regions/europe) | | Destination Region | [North America](/docs/network/latency/regions/north-america) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Frankfurt am Main](https://www.geonames.org/2925533) (50.11552, 8.68417) and [New York City](https://www.geonames.org/5128581) (40.71427, -74.00597). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **6,219.3 km** | | Vacuum RTT floor | **41.49 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **60.9 ms** | | Low-latency fiber material floor | **60.66 ms** | | Engineering floor (5% path allowance) | **63.96 ms** | | Research 1.33× mapped-fiber reference | **81 ms** | | Estimated unamplified path loss | **1306 dB** | | Transparent optical spans / inline amplifiers | **82 / 81** | | Published RTT inflation over fiber floor | **1.22×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------ | | Minimum RTT | **72.04 ms** | | Average RTT | **74.4 ms** | | Maximum RTT | **80.52 ms** | | Standard deviation | **2.09 ms** | | Stdev / average | **2.8%** | ## Route Context * Departure PoP: [Frankfurt (FRA)](/docs/network/latency/fra-frankfurt) * Destination PoP: [New York (NYC)](/docs/network/latency/nyc-new-york) * Region overview: [Europe](/docs/network/latency/regions/europe) → [North America](/docs/network/latency/regions/north-america) * Corridor overview: [Europe to North America](/docs/network/latency/regions/europe-to-north-america) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [New York to Frankfurt latency and RTT](/docs/network/latency/pairs/nyc-fra-rtt): 73.9 ms **Fastest routes departing Frankfurt (FRA)** * [Frankfurt to Amsterdam latency and RTT](/docs/network/latency/pairs/fra-ams-rtt): 5.9 ms * [Frankfurt to Berlin latency and RTT](/docs/network/latency/pairs/fra-ber-rtt): 6.1 ms * [Frankfurt to Paris latency and RTT](/docs/network/latency/pairs/fra-par-rtt): 7.6 ms * [Frankfurt to London latency and RTT](/docs/network/latency/pairs/fra-lon-rtt): 12.9 ms * [Frankfurt to Marseille latency and RTT](/docs/network/latency/pairs/fra-mrs-rtt): 16.4 ms **Fastest routes arriving at New York (NYC)** * [Ashburn to New York latency and RTT](/docs/network/latency/pairs/iad-nyc-rtt): 6 ms * [Miami to New York latency and RTT](/docs/network/latency/pairs/mia-nyc-rtt): 33.3 ms * [Los Angeles to New York latency and RTT](/docs/network/latency/pairs/lax-nyc-rtt): 57.9 ms * [Seattle to New York latency and RTT](/docs/network/latency/pairs/sea-nyc-rtt): 59.5 ms * [London to New York latency and RTT](/docs/network/latency/pairs/lon-nyc-rtt): 63.8 ms **Same corridor (Europe → North America)** * [Amsterdam to New York latency and RTT](/docs/network/latency/pairs/ams-nyc-rtt): 68.6 ms * [Paris to New York latency and RTT](/docs/network/latency/pairs/par-nyc-rtt): 68.9 ms * [London to Ashburn latency and RTT](/docs/network/latency/pairs/lon-iad-rtt): 71 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/fra-nyc.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Frankfurt → New York round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/fra-nyc-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Frankfurt, Germany → Paris, France RTT 🇩🇪 **Frankfurt, Germany (FRA)** → 🇫🇷 **Paris, France (PAR)** Measurement round: `2026-08-16T04:07:28Z`. Frankfurt to Paris averaged 7.6 ms across 50 replies. The observed range was 7.34 to 8.24 ms; these figures describe one probe round, not a long-term service target. This is number 3 of 19 outgoing routes by average RTT. That ranking compares different destinations and is not adjusted for distance. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------- | | RTT | **7.6 ms** | | Jitter | **0.15 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **4.7 ms** | | Fiber Efficiency | **61.8%** | | Latency Tier | Ultra-Low | | Source Region | [Europe](/docs/network/latency/regions/europe) | | Destination Region | [Europe](/docs/network/latency/regions/europe) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Frankfurt am Main](https://www.geonames.org/2925533) (50.11552, 8.68417) and [Paris](https://www.geonames.org/2988507) (48.85341, 2.34880). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | ------------ | | WGS-84 geodesic distance | **479.9 km** | | Vacuum RTT floor | **3.2 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **4.7 ms** | | Low-latency fiber material floor | **4.68 ms** | | Engineering floor (5% path allowance) | **4.93 ms** | | Research 1.33× mapped-fiber reference | **6.25 ms** | | Estimated unamplified path loss | **100.8 dB** | | Transparent optical spans / inline amplifiers | **7 / 6** | | Published RTT inflation over fiber floor | **1.62×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ----------- | | Minimum RTT | **7.34 ms** | | Average RTT | **7.6 ms** | | Maximum RTT | **8.24 ms** | | Standard deviation | **0.23 ms** | | Stdev / average | **3.0%** | ## Route Context * Departure PoP: [Frankfurt (FRA)](/docs/network/latency/fra-frankfurt) * Destination PoP: [Paris (PAR)](/docs/network/latency/par-paris) * Region overview: [Europe](/docs/network/latency/regions/europe) → [Europe](/docs/network/latency/regions/europe) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Paris to Frankfurt latency and RTT](/docs/network/latency/pairs/par-fra-rtt): 7.6 ms **Fastest routes departing Frankfurt (FRA)** * [Frankfurt to Amsterdam latency and RTT](/docs/network/latency/pairs/fra-ams-rtt): 5.9 ms * [Frankfurt to Berlin latency and RTT](/docs/network/latency/pairs/fra-ber-rtt): 6.1 ms * [Frankfurt to London latency and RTT](/docs/network/latency/pairs/fra-lon-rtt): 12.9 ms * [Frankfurt to Marseille latency and RTT](/docs/network/latency/pairs/fra-mrs-rtt): 16.4 ms * [Frankfurt to Moscow latency and RTT](/docs/network/latency/pairs/fra-mow-rtt): 34.7 ms **Fastest routes arriving at Paris (PAR)** * [London to Paris latency and RTT](/docs/network/latency/pairs/lon-par-rtt): 6.4 ms * [Amsterdam to Paris latency and RTT](/docs/network/latency/pairs/ams-par-rtt): 7 ms * [Marseille to Paris latency and RTT](/docs/network/latency/pairs/mrs-par-rtt): 8.9 ms * [Berlin to Paris latency and RTT](/docs/network/latency/pairs/ber-par-rtt): 15.6 ms * [Moscow to Paris latency and RTT](/docs/network/latency/pairs/mow-par-rtt): 44 ms **Same corridor (Europe → Europe)** * [Amsterdam to London latency and RTT](/docs/network/latency/pairs/ams-lon-rtt): 5.2 ms * [London to Amsterdam latency and RTT](/docs/network/latency/pairs/lon-ams-rtt): 5.2 ms * [Amsterdam to Frankfurt latency and RTT](/docs/network/latency/pairs/ams-fra-rtt): 5.9 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/fra-par.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Frankfurt → Paris round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/fra-par-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Frankfurt, Germany → Seattle, USA RTT 🇩🇪 **Frankfurt, Germany (FRA)** → 🇺🇸 **Seattle, USA (SEA)** Measurement round: `2026-08-16T04:07:28Z`. Frankfurt to Seattle averaged 135.2 ms across 50 replies. The observed range was 129.14 to 151.39 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 5.97 ms, or 4.4% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **135.2 ms** | | Jitter | **4.05 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **80.34 ms** | | Fiber Efficiency | **59.4%** | | Latency Tier | Good | | Source Region | [Europe](/docs/network/latency/regions/europe) | | Destination Region | [North America](/docs/network/latency/regions/north-america) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Frankfurt am Main](https://www.geonames.org/2925533) (50.11552, 8.68417) and [Seattle](https://www.geonames.org/5809844) (47.60621, -122.33207). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | ------------- | | WGS-84 geodesic distance | **8,204 km** | | Vacuum RTT floor | **54.73 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **80.34 ms** | | Low-latency fiber material floor | **80.02 ms** | | Engineering floor (5% path allowance) | **84.37 ms** | | Research 1.33× mapped-fiber reference | **106.85 ms** | | Estimated unamplified path loss | **1722.8 dB** | | Transparent optical spans / inline amplifiers | **108 / 107** | | Published RTT inflation over fiber floor | **1.68×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **129.14 ms** | | Average RTT | **135.2 ms** | | Maximum RTT | **151.39 ms** | | Standard deviation | **5.97 ms** | | Stdev / average | **4.4%** | ## Route Context * Departure PoP: [Frankfurt (FRA)](/docs/network/latency/fra-frankfurt) * Destination PoP: [Seattle (SEA)](/docs/network/latency/sea-seattle) * Region overview: [Europe](/docs/network/latency/regions/europe) → [North America](/docs/network/latency/regions/north-america) * Corridor overview: [Europe to North America](/docs/network/latency/regions/europe-to-north-america) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Seattle to Frankfurt latency and RTT](/docs/network/latency/pairs/sea-fra-rtt): 133.2 ms **Fastest routes departing Frankfurt (FRA)** * [Frankfurt to Amsterdam latency and RTT](/docs/network/latency/pairs/fra-ams-rtt): 5.9 ms * [Frankfurt to Berlin latency and RTT](/docs/network/latency/pairs/fra-ber-rtt): 6.1 ms * [Frankfurt to Paris latency and RTT](/docs/network/latency/pairs/fra-par-rtt): 7.6 ms * [Frankfurt to London latency and RTT](/docs/network/latency/pairs/fra-lon-rtt): 12.9 ms * [Frankfurt to Marseille latency and RTT](/docs/network/latency/pairs/fra-mrs-rtt): 16.4 ms **Fastest routes arriving at Seattle (SEA)** * [Los Angeles to Seattle latency and RTT](/docs/network/latency/pairs/lax-sea-rtt): 26.9 ms * [New York to Seattle latency and RTT](/docs/network/latency/pairs/nyc-sea-rtt): 58.7 ms * [Ashburn to Seattle latency and RTT](/docs/network/latency/pairs/iad-sea-rtt): 62.3 ms * [Miami to Seattle latency and RTT](/docs/network/latency/pairs/mia-sea-rtt): 81.9 ms * [Tokyo to Seattle latency and RTT](/docs/network/latency/pairs/tyo-sea-rtt): 84.8 ms **Same corridor (Europe → North America)** * [London to New York latency and RTT](/docs/network/latency/pairs/lon-nyc-rtt): 63.8 ms * [Amsterdam to New York latency and RTT](/docs/network/latency/pairs/ams-nyc-rtt): 68.6 ms * [Paris to New York latency and RTT](/docs/network/latency/pairs/par-nyc-rtt): 68.9 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/fra-sea.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Frankfurt → Seattle round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/fra-sea-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Frankfurt, Germany → Singapore RTT 🇩🇪 **Frankfurt, Germany (FRA)** → 🇸🇬 **Singapore (SIN)** Measurement round: `2026-08-16T04:07:28Z`. Frankfurt to Singapore averaged 151.4 ms across 50 replies. The observed range was 148.19 to 159.42 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 2.49 ms, or 1.6% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------------------- | | RTT | **151.4 ms** | | Jitter | **2.17 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **100.56 ms** | | Fiber Efficiency | **66.4%** | | Latency Tier | Fair | | Source Region | [Europe](/docs/network/latency/regions/europe) | | Destination Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Frankfurt am Main](https://www.geonames.org/2925533) (50.11552, 8.68417) and [Singapore](https://www.geonames.org/1880252) (1.28967, 103.85007). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | --------------- | | WGS-84 geodesic distance | **10,268.5 km** | | Vacuum RTT floor | **68.5 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **100.56 ms** | | Low-latency fiber material floor | **100.15 ms** | | Engineering floor (5% path allowance) | **105.6 ms** | | Research 1.33× mapped-fiber reference | **133.74 ms** | | Estimated unamplified path loss | **2156.4 dB** | | Transparent optical spans / inline amplifiers | **135 / 134** | | Published RTT inflation over fiber floor | **1.51×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **148.19 ms** | | Average RTT | **151.4 ms** | | Maximum RTT | **159.42 ms** | | Standard deviation | **2.49 ms** | | Stdev / average | **1.6%** | ## Route Context * Departure PoP: [Frankfurt (FRA)](/docs/network/latency/fra-frankfurt) * Destination PoP: [Singapore (SIN)](/docs/network/latency/sin-singapore) * Region overview: [Europe](/docs/network/latency/regions/europe) → [Asia Pacific](/docs/network/latency/regions/asia-pacific) * Corridor overview: [Europe to Asia Pacific](/docs/network/latency/regions/europe-to-asia-pacific) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Singapore to Frankfurt latency and RTT](/docs/network/latency/pairs/sin-fra-rtt): 151.6 ms **Fastest routes departing Frankfurt (FRA)** * [Frankfurt to Amsterdam latency and RTT](/docs/network/latency/pairs/fra-ams-rtt): 5.9 ms * [Frankfurt to Berlin latency and RTT](/docs/network/latency/pairs/fra-ber-rtt): 6.1 ms * [Frankfurt to Paris latency and RTT](/docs/network/latency/pairs/fra-par-rtt): 7.6 ms * [Frankfurt to London latency and RTT](/docs/network/latency/pairs/fra-lon-rtt): 12.9 ms * [Frankfurt to Marseille latency and RTT](/docs/network/latency/pairs/fra-mrs-rtt): 16.4 ms **Fastest routes arriving at Singapore (SIN)** * [Hong Kong to Singapore latency and RTT](/docs/network/latency/pairs/hkg-sin-rtt): 31.5 ms * [Taipei to Singapore latency and RTT](/docs/network/latency/pairs/tpe-sin-rtt): 44.3 ms * [Tokyo to Singapore latency and RTT](/docs/network/latency/pairs/tyo-sin-rtt): 69.4 ms * [Melbourne to Singapore latency and RTT](/docs/network/latency/pairs/mel-sin-rtt): 88.4 ms * [Sydney to Singapore latency and RTT](/docs/network/latency/pairs/syd-sin-rtt): 94.5 ms **Same corridor (Europe → Asia Pacific)** * [Moscow to Hong Kong latency and RTT](/docs/network/latency/pairs/mow-hkg-rtt): 118.6 ms * [Moscow to Taipei latency and RTT](/docs/network/latency/pairs/mow-tpe-rtt): 132.1 ms * [Marseille to Singapore latency and RTT](/docs/network/latency/pairs/mrs-sin-rtt): 139.6 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/fra-sin.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Frankfurt → Singapore round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/fra-sin-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Frankfurt, Germany → Sydney, Australia RTT 🇩🇪 **Frankfurt, Germany (FRA)** → 🇦🇺 **Sydney, Australia (SYD)** Measurement round: `2026-08-16T04:07:28Z`. Frankfurt to Sydney averaged 256 ms across 50 replies. The observed range was 251.25 to 274.41 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 4.51 ms, or 1.8% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------------------- | | RTT | **256 ms** | | Jitter | **3.47 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **161.37 ms** | | Fiber Efficiency | **63%** | | Latency Tier | High | | Source Region | [Europe](/docs/network/latency/regions/europe) | | Destination Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Frankfurt am Main](https://www.geonames.org/2925533) (50.11552, 8.68417) and [Sydney](https://www.geonames.org/2147714) (-33.86785, 151.20732). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | --------------- | | WGS-84 geodesic distance | **16,478.2 km** | | Vacuum RTT floor | **109.93 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **161.37 ms** | | Low-latency fiber material floor | **160.72 ms** | | Engineering floor (5% path allowance) | **169.46 ms** | | Research 1.33× mapped-fiber reference | **214.62 ms** | | Estimated unamplified path loss | **3460.4 dB** | | Transparent optical spans / inline amplifiers | **217 / 216** | | Published RTT inflation over fiber floor | **1.59×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **251.25 ms** | | Average RTT | **256 ms** | | Maximum RTT | **274.41 ms** | | Standard deviation | **4.51 ms** | | Stdev / average | **1.8%** | ## Route Context * Departure PoP: [Frankfurt (FRA)](/docs/network/latency/fra-frankfurt) * Destination PoP: [Sydney (SYD)](/docs/network/latency/syd-sydney) * Region overview: [Europe](/docs/network/latency/regions/europe) → [Asia Pacific](/docs/network/latency/regions/asia-pacific) * Corridor overview: [Europe to Asia Pacific](/docs/network/latency/regions/europe-to-asia-pacific) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Sydney to Frankfurt latency and RTT](/docs/network/latency/pairs/syd-fra-rtt): 254.4 ms **Fastest routes departing Frankfurt (FRA)** * [Frankfurt to Amsterdam latency and RTT](/docs/network/latency/pairs/fra-ams-rtt): 5.9 ms * [Frankfurt to Berlin latency and RTT](/docs/network/latency/pairs/fra-ber-rtt): 6.1 ms * [Frankfurt to Paris latency and RTT](/docs/network/latency/pairs/fra-par-rtt): 7.6 ms * [Frankfurt to London latency and RTT](/docs/network/latency/pairs/fra-lon-rtt): 12.9 ms * [Frankfurt to Marseille latency and RTT](/docs/network/latency/pairs/fra-mrs-rtt): 16.4 ms **Fastest routes arriving at Sydney (SYD)** * [Melbourne to Sydney latency and RTT](/docs/network/latency/pairs/mel-syd-rtt): 9.8 ms * [Singapore to Sydney latency and RTT](/docs/network/latency/pairs/sin-syd-rtt): 94.5 ms * [Tokyo to Sydney latency and RTT](/docs/network/latency/pairs/tyo-syd-rtt): 101.9 ms * [Taipei to Sydney latency and RTT](/docs/network/latency/pairs/tpe-syd-rtt): 133.2 ms * [Hong Kong to Sydney latency and RTT](/docs/network/latency/pairs/hkg-syd-rtt): 137.2 ms **Same corridor (Europe → Asia Pacific)** * [Moscow to Hong Kong latency and RTT](/docs/network/latency/pairs/mow-hkg-rtt): 118.6 ms * [Moscow to Taipei latency and RTT](/docs/network/latency/pairs/mow-tpe-rtt): 132.1 ms * [Marseille to Singapore latency and RTT](/docs/network/latency/pairs/mrs-sin-rtt): 139.6 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/fra-syd.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Frankfurt → Sydney round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/fra-syd-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Frankfurt, Germany → Taipei, Taiwan RTT 🇩🇪 **Frankfurt, Germany (FRA)** → 🇹🇼 **Taipei, Taiwan (TPE)** Measurement round: `2026-08-16T04:07:28Z`. Frankfurt to Taipei averaged 163.8 ms across 50 replies. The observed range was 159.37 to 176 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 3.63 ms, or 2.2% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------------------- | | RTT | **163.8 ms** | | Jitter | **3.43 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **91.96 ms** | | Fiber Efficiency | **56.1%** | | Latency Tier | Fair | | Source Region | [Europe](/docs/network/latency/regions/europe) | | Destination Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Frankfurt am Main](https://www.geonames.org/2925533) (50.11552, 8.68417) and [Taipei](https://www.geonames.org/1668341) (25.05306, 121.52639). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **9,390.8 km** | | Vacuum RTT floor | **62.65 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **91.96 ms** | | Low-latency fiber material floor | **91.59 ms** | | Engineering floor (5% path allowance) | **96.57 ms** | | Research 1.33× mapped-fiber reference | **122.31 ms** | | Estimated unamplified path loss | **1972.1 dB** | | Transparent optical spans / inline amplifiers | **124 / 123** | | Published RTT inflation over fiber floor | **1.78×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **159.37 ms** | | Average RTT | **163.8 ms** | | Maximum RTT | **176 ms** | | Standard deviation | **3.63 ms** | | Stdev / average | **2.2%** | ## Route Context * Departure PoP: [Frankfurt (FRA)](/docs/network/latency/fra-frankfurt) * Destination PoP: [Taipei (TPE)](/docs/network/latency/tpe-taipei) * Region overview: [Europe](/docs/network/latency/regions/europe) → [Asia Pacific](/docs/network/latency/regions/asia-pacific) * Corridor overview: [Europe to Asia Pacific](/docs/network/latency/regions/europe-to-asia-pacific) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Taipei to Frankfurt latency and RTT](/docs/network/latency/pairs/tpe-fra-rtt): 166 ms **Fastest routes departing Frankfurt (FRA)** * [Frankfurt to Amsterdam latency and RTT](/docs/network/latency/pairs/fra-ams-rtt): 5.9 ms * [Frankfurt to Berlin latency and RTT](/docs/network/latency/pairs/fra-ber-rtt): 6.1 ms * [Frankfurt to Paris latency and RTT](/docs/network/latency/pairs/fra-par-rtt): 7.6 ms * [Frankfurt to London latency and RTT](/docs/network/latency/pairs/fra-lon-rtt): 12.9 ms * [Frankfurt to Marseille latency and RTT](/docs/network/latency/pairs/fra-mrs-rtt): 16.4 ms **Fastest routes arriving at Taipei (TPE)** * [Hong Kong to Taipei latency and RTT](/docs/network/latency/pairs/hkg-tpe-rtt): 14.7 ms * [Tokyo to Taipei latency and RTT](/docs/network/latency/pairs/tyo-tpe-rtt): 32.3 ms * [Singapore to Taipei latency and RTT](/docs/network/latency/pairs/sin-tpe-rtt): 45.8 ms * [Seattle to Taipei latency and RTT](/docs/network/latency/pairs/sea-tpe-rtt): 114.7 ms * [Sydney to Taipei latency and RTT](/docs/network/latency/pairs/syd-tpe-rtt): 131.7 ms **Same corridor (Europe → Asia Pacific)** * [Moscow to Hong Kong latency and RTT](/docs/network/latency/pairs/mow-hkg-rtt): 118.6 ms * [Moscow to Taipei latency and RTT](/docs/network/latency/pairs/mow-tpe-rtt): 132.1 ms * [Marseille to Singapore latency and RTT](/docs/network/latency/pairs/mrs-sin-rtt): 139.6 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/fra-tpe.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Frankfurt → Taipei round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/fra-tpe-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Frankfurt, Germany → Tokyo, Japan RTT 🇩🇪 **Frankfurt, Germany (FRA)** → 🇯🇵 **Tokyo, Japan (TYO)** Measurement round: `2026-08-16T04:07:28Z`. Frankfurt to Tokyo averaged 194.5 ms, compared with a 91.61 ms city-centre fiber reference. The difference cannot be assigned to a particular cable or carrier from RTT alone. Standard deviation was 4.7 ms, or 2.4% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------------------- | | RTT | **194.5 ms** | | Jitter | **4.78 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **91.61 ms** | | Fiber Efficiency | **47.1%** | | Latency Tier | Fair | | Source Region | [Europe](/docs/network/latency/regions/europe) | | Destination Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Frankfurt am Main](https://www.geonames.org/2925533) (50.11552, 8.68417) and [Tokyo](https://www.geonames.org/1850147) (35.68950, 139.69171). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **9,354.4 km** | | Vacuum RTT floor | **62.41 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **91.61 ms** | | Low-latency fiber material floor | **91.24 ms** | | Engineering floor (5% path allowance) | **96.2 ms** | | Research 1.33× mapped-fiber reference | **121.84 ms** | | Estimated unamplified path loss | **1964.4 dB** | | Transparent optical spans / inline amplifiers | **123 / 122** | | Published RTT inflation over fiber floor | **2.12×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **189.69 ms** | | Average RTT | **194.5 ms** | | Maximum RTT | **207.25 ms** | | Standard deviation | **4.7 ms** | | Stdev / average | **2.4%** | ## Route Context * Departure PoP: [Frankfurt (FRA)](/docs/network/latency/fra-frankfurt) * Destination PoP: [Tokyo (TYO)](/docs/network/latency/tyo-tokyo) * Region overview: [Europe](/docs/network/latency/regions/europe) → [Asia Pacific](/docs/network/latency/regions/asia-pacific) * Corridor overview: [Europe to Asia Pacific](/docs/network/latency/regions/europe-to-asia-pacific) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Tokyo to Frankfurt latency and RTT](/docs/network/latency/pairs/tyo-fra-rtt): 193.2 ms **Fastest routes departing Frankfurt (FRA)** * [Frankfurt to Amsterdam latency and RTT](/docs/network/latency/pairs/fra-ams-rtt): 5.9 ms * [Frankfurt to Berlin latency and RTT](/docs/network/latency/pairs/fra-ber-rtt): 6.1 ms * [Frankfurt to Paris latency and RTT](/docs/network/latency/pairs/fra-par-rtt): 7.6 ms * [Frankfurt to London latency and RTT](/docs/network/latency/pairs/fra-lon-rtt): 12.9 ms * [Frankfurt to Marseille latency and RTT](/docs/network/latency/pairs/fra-mrs-rtt): 16.4 ms **Fastest routes arriving at Tokyo (TYO)** * [Taipei to Tokyo latency and RTT](/docs/network/latency/pairs/tpe-tyo-rtt): 31.9 ms * [Hong Kong to Tokyo latency and RTT](/docs/network/latency/pairs/hkg-tyo-rtt): 44.9 ms * [Singapore to Tokyo latency and RTT](/docs/network/latency/pairs/sin-tyo-rtt): 68.9 ms * [Seattle to Tokyo latency and RTT](/docs/network/latency/pairs/sea-tyo-rtt): 85 ms * [Los Angeles to Tokyo latency and RTT](/docs/network/latency/pairs/lax-tyo-rtt): 101.2 ms **Same corridor (Europe → Asia Pacific)** * [Moscow to Hong Kong latency and RTT](/docs/network/latency/pairs/mow-hkg-rtt): 118.6 ms * [Moscow to Taipei latency and RTT](/docs/network/latency/pairs/mow-tpe-rtt): 132.1 ms * [Marseille to Singapore latency and RTT](/docs/network/latency/pairs/mrs-sin-rtt): 139.6 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/fra-tyo.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Frankfurt → Tokyo round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/fra-tyo-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # São Paulo, Brazil → Amsterdam, Netherlands RTT 🇧🇷 **São Paulo, Brazil (GRU)** → 🇳🇱 **Amsterdam, Netherlands (AMS)** Measurement round: `2026-08-16T04:07:28Z`. São Paulo to Amsterdam averaged 180 ms across 50 replies. The observed range was 172.52 to 196.27 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 6.11 ms, or 3.4% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **180 ms** | | Jitter | **5.36 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **95.79 ms** | | Fiber Efficiency | **53.2%** | | Latency Tier | Fair | | Source Region | [South America](/docs/network/latency/regions/south-america) | | Destination Region | [Europe](/docs/network/latency/regions/europe) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [São Paulo](https://www.geonames.org/3448439) (-23.54750, -46.63611) and [Amsterdam](https://www.geonames.org/2759794) (52.37403, 4.88969). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **9,781.6 km** | | Vacuum RTT floor | **65.26 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **95.79 ms** | | Low-latency fiber material floor | **95.4 ms** | | Engineering floor (5% path allowance) | **100.59 ms** | | Research 1.33× mapped-fiber reference | **127.4 ms** | | Estimated unamplified path loss | **2054.1 dB** | | Transparent optical spans / inline amplifiers | **129 / 128** | | Published RTT inflation over fiber floor | **1.88×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **172.52 ms** | | Average RTT | **180 ms** | | Maximum RTT | **196.27 ms** | | Standard deviation | **6.11 ms** | | Stdev / average | **3.4%** | ## Route Context * Departure PoP: [São Paulo (GRU)](/docs/network/latency/gru-sao-paulo) * Destination PoP: [Amsterdam (AMS)](/docs/network/latency/ams-amsterdam) * Region overview: [South America](/docs/network/latency/regions/south-america) → [Europe](/docs/network/latency/regions/europe) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Amsterdam to São Paulo latency and RTT](/docs/network/latency/pairs/ams-gru-rtt): 175.5 ms **Fastest routes departing São Paulo (GRU)** * [São Paulo to Miami latency and RTT](/docs/network/latency/pairs/gru-mia-rtt): 74 ms * [São Paulo to New York latency and RTT](/docs/network/latency/pairs/gru-nyc-rtt): 122 ms * [São Paulo to Ashburn latency and RTT](/docs/network/latency/pairs/gru-iad-rtt): 128 ms * [São Paulo to Los Angeles latency and RTT](/docs/network/latency/pairs/gru-lax-rtt): 130.8 ms * [São Paulo to Seattle latency and RTT](/docs/network/latency/pairs/gru-sea-rtt): 155.9 ms **Fastest routes arriving at Amsterdam (AMS)** * [London to Amsterdam latency and RTT](/docs/network/latency/pairs/lon-ams-rtt): 5.2 ms * [Frankfurt to Amsterdam latency and RTT](/docs/network/latency/pairs/fra-ams-rtt): 5.9 ms * [Paris to Amsterdam latency and RTT](/docs/network/latency/pairs/par-ams-rtt): 7 ms * [Berlin to Amsterdam latency and RTT](/docs/network/latency/pairs/ber-ams-rtt): 7.2 ms * [Marseille to Amsterdam latency and RTT](/docs/network/latency/pairs/mrs-ams-rtt): 20.3 ms **Same corridor (South America → Europe)** * [São Paulo to London latency and RTT](/docs/network/latency/pairs/gru-lon-rtt): 175.8 ms * [São Paulo to Paris latency and RTT](/docs/network/latency/pairs/gru-par-rtt): 180.2 ms * [São Paulo to Frankfurt latency and RTT](/docs/network/latency/pairs/gru-fra-rtt): 186.2 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/gru-ams.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, São Paulo → Amsterdam round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/gru-ams-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # São Paulo, Brazil → Berlin, Germany RTT 🇧🇷 **São Paulo, Brazil (GRU)** → 🇩🇪 **Berlin, Germany (BER)** Measurement round: `2026-08-16T04:07:28Z`. São Paulo to Berlin averaged 191.3 ms across 50 replies. The observed range was 185.86 to 210.68 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 4.6 ms, or 2.4% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **191.3 ms** | | Jitter | **3.74 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **100.2 ms** | | Fiber Efficiency | **52.4%** | | Latency Tier | Fair | | Source Region | [South America](/docs/network/latency/regions/south-america) | | Destination Region | [Europe](/docs/network/latency/regions/europe) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [São Paulo](https://www.geonames.org/3448439) (-23.54750, -46.63611) and [Berlin](https://www.geonames.org/2950159) (52.52437, 13.41053). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | --------------- | | WGS-84 geodesic distance | **10,232.3 km** | | Vacuum RTT floor | **68.26 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **100.2 ms** | | Low-latency fiber material floor | **99.8 ms** | | Engineering floor (5% path allowance) | **105.23 ms** | | Research 1.33× mapped-fiber reference | **133.27 ms** | | Estimated unamplified path loss | **2148.8 dB** | | Transparent optical spans / inline amplifiers | **135 / 134** | | Published RTT inflation over fiber floor | **1.91×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **185.86 ms** | | Average RTT | **191.3 ms** | | Maximum RTT | **210.68 ms** | | Standard deviation | **4.6 ms** | | Stdev / average | **2.4%** | ## Route Context * Departure PoP: [São Paulo (GRU)](/docs/network/latency/gru-sao-paulo) * Destination PoP: [Berlin (BER)](/docs/network/latency/ber-berlin) * Region overview: [South America](/docs/network/latency/regions/south-america) → [Europe](/docs/network/latency/regions/europe) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Berlin to São Paulo latency and RTT](/docs/network/latency/pairs/ber-gru-rtt): 186.5 ms **Fastest routes departing São Paulo (GRU)** * [São Paulo to Miami latency and RTT](/docs/network/latency/pairs/gru-mia-rtt): 74 ms * [São Paulo to New York latency and RTT](/docs/network/latency/pairs/gru-nyc-rtt): 122 ms * [São Paulo to Ashburn latency and RTT](/docs/network/latency/pairs/gru-iad-rtt): 128 ms * [São Paulo to Los Angeles latency and RTT](/docs/network/latency/pairs/gru-lax-rtt): 130.8 ms * [São Paulo to Seattle latency and RTT](/docs/network/latency/pairs/gru-sea-rtt): 155.9 ms **Fastest routes arriving at Berlin (BER)** * [Frankfurt to Berlin latency and RTT](/docs/network/latency/pairs/fra-ber-rtt): 6.1 ms * [Amsterdam to Berlin latency and RTT](/docs/network/latency/pairs/ams-ber-rtt): 7.2 ms * [Paris to Berlin latency and RTT](/docs/network/latency/pairs/par-ber-rtt): 15.5 ms * [London to Berlin latency and RTT](/docs/network/latency/pairs/lon-ber-rtt): 16.2 ms * [Marseille to Berlin latency and RTT](/docs/network/latency/pairs/mrs-ber-rtt): 20.6 ms **Same corridor (South America → Europe)** * [São Paulo to London latency and RTT](/docs/network/latency/pairs/gru-lon-rtt): 175.8 ms * [São Paulo to Amsterdam latency and RTT](/docs/network/latency/pairs/gru-ams-rtt): 180 ms * [São Paulo to Paris latency and RTT](/docs/network/latency/pairs/gru-par-rtt): 180.2 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/gru-ber.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, São Paulo → Berlin round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/gru-ber-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # São Paulo, Brazil → Frankfurt, Germany RTT 🇧🇷 **São Paulo, Brazil (GRU)** → 🇩🇪 **Frankfurt, Germany (FRA)** Measurement round: `2026-08-16T04:07:28Z`. São Paulo to Frankfurt averaged 186.2 ms across 50 replies. The observed range was 180.59 to 202.2 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 5.01 ms, or 2.7% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **186.2 ms** | | Jitter | **4.54 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **96.04 ms** | | Fiber Efficiency | **51.6%** | | Latency Tier | Fair | | Source Region | [South America](/docs/network/latency/regions/south-america) | | Destination Region | [Europe](/docs/network/latency/regions/europe) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [São Paulo](https://www.geonames.org/3448439) (-23.54750, -46.63611) and [Frankfurt am Main](https://www.geonames.org/2925533) (50.11552, 8.68417). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **9,807.7 km** | | Vacuum RTT floor | **65.43 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **96.04 ms** | | Low-latency fiber material floor | **95.66 ms** | | Engineering floor (5% path allowance) | **100.86 ms** | | Research 1.33× mapped-fiber reference | **127.74 ms** | | Estimated unamplified path loss | **2059.6 dB** | | Transparent optical spans / inline amplifiers | **129 / 128** | | Published RTT inflation over fiber floor | **1.94×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **180.59 ms** | | Average RTT | **186.2 ms** | | Maximum RTT | **202.2 ms** | | Standard deviation | **5.01 ms** | | Stdev / average | **2.7%** | ## Route Context * Departure PoP: [São Paulo (GRU)](/docs/network/latency/gru-sao-paulo) * Destination PoP: [Frankfurt (FRA)](/docs/network/latency/fra-frankfurt) * Region overview: [South America](/docs/network/latency/regions/south-america) → [Europe](/docs/network/latency/regions/europe) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Frankfurt to São Paulo latency and RTT](/docs/network/latency/pairs/fra-gru-rtt): 181.3 ms **Fastest routes departing São Paulo (GRU)** * [São Paulo to Miami latency and RTT](/docs/network/latency/pairs/gru-mia-rtt): 74 ms * [São Paulo to New York latency and RTT](/docs/network/latency/pairs/gru-nyc-rtt): 122 ms * [São Paulo to Ashburn latency and RTT](/docs/network/latency/pairs/gru-iad-rtt): 128 ms * [São Paulo to Los Angeles latency and RTT](/docs/network/latency/pairs/gru-lax-rtt): 130.8 ms * [São Paulo to Seattle latency and RTT](/docs/network/latency/pairs/gru-sea-rtt): 155.9 ms **Fastest routes arriving at Frankfurt (FRA)** * [Amsterdam to Frankfurt latency and RTT](/docs/network/latency/pairs/ams-fra-rtt): 5.9 ms * [Berlin to Frankfurt latency and RTT](/docs/network/latency/pairs/ber-fra-rtt): 6.1 ms * [Paris to Frankfurt latency and RTT](/docs/network/latency/pairs/par-fra-rtt): 7.6 ms * [London to Frankfurt latency and RTT](/docs/network/latency/pairs/lon-fra-rtt): 13.6 ms * [Marseille to Frankfurt latency and RTT](/docs/network/latency/pairs/mrs-fra-rtt): 16 ms **Same corridor (South America → Europe)** * [São Paulo to London latency and RTT](/docs/network/latency/pairs/gru-lon-rtt): 175.8 ms * [São Paulo to Amsterdam latency and RTT](/docs/network/latency/pairs/gru-ams-rtt): 180 ms * [São Paulo to Paris latency and RTT](/docs/network/latency/pairs/gru-par-rtt): 180.2 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/gru-fra.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, São Paulo → Frankfurt round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/gru-fra-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # São Paulo, Brazil → Hong Kong RTT 🇧🇷 **São Paulo, Brazil (GRU)** → 🇭🇰 **Hong Kong (HKG)** Measurement round: `2026-08-16T04:07:28Z`. São Paulo to Hong Kong averaged 273.5 ms across 50 replies. The observed range was 267.18 to 286.64 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 4.57 ms, or 1.7% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **273.5 ms** | | Jitter | **4.53 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **176.86 ms** | | Fiber Efficiency | **64.7%** | | Latency Tier | High | | Source Region | [South America](/docs/network/latency/regions/south-america) | | Destination Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [São Paulo](https://www.geonames.org/3448439) (-23.54750, -46.63611) and [Hong Kong](https://www.geonames.org/1819729) (22.27832, 114.17469). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | ------------- | | WGS-84 geodesic distance | **18,060 km** | | Vacuum RTT floor | **120.48 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **176.86 ms** | | Low-latency fiber material floor | **176.15 ms** | | Engineering floor (5% path allowance) | **185.72 ms** | | Research 1.33× mapped-fiber reference | **235.22 ms** | | Estimated unamplified path loss | **3792.6 dB** | | Transparent optical spans / inline amplifiers | **238 / 237** | | Published RTT inflation over fiber floor | **1.55×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **267.18 ms** | | Average RTT | **273.5 ms** | | Maximum RTT | **286.64 ms** | | Standard deviation | **4.57 ms** | | Stdev / average | **1.7%** | ## Route Context * Departure PoP: [São Paulo (GRU)](/docs/network/latency/gru-sao-paulo) * Destination PoP: [Hong Kong (HKG)](/docs/network/latency/hkg-hong-kong) * Region overview: [South America](/docs/network/latency/regions/south-america) → [Asia Pacific](/docs/network/latency/regions/asia-pacific) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Hong Kong to São Paulo latency and RTT](/docs/network/latency/pairs/hkg-gru-rtt): 287.6 ms **Fastest routes departing São Paulo (GRU)** * [São Paulo to Miami latency and RTT](/docs/network/latency/pairs/gru-mia-rtt): 74 ms * [São Paulo to New York latency and RTT](/docs/network/latency/pairs/gru-nyc-rtt): 122 ms * [São Paulo to Ashburn latency and RTT](/docs/network/latency/pairs/gru-iad-rtt): 128 ms * [São Paulo to Los Angeles latency and RTT](/docs/network/latency/pairs/gru-lax-rtt): 130.8 ms * [São Paulo to Seattle latency and RTT](/docs/network/latency/pairs/gru-sea-rtt): 155.9 ms **Fastest routes arriving at Hong Kong (HKG)** * [Taipei to Hong Kong latency and RTT](/docs/network/latency/pairs/tpe-hkg-rtt): 15.5 ms * [Singapore to Hong Kong latency and RTT](/docs/network/latency/pairs/sin-hkg-rtt): 30.8 ms * [Tokyo to Hong Kong latency and RTT](/docs/network/latency/pairs/tyo-hkg-rtt): 44.8 ms * [Moscow to Hong Kong latency and RTT](/docs/network/latency/pairs/mow-hkg-rtt): 118.6 ms * [Seattle to Hong Kong latency and RTT](/docs/network/latency/pairs/sea-hkg-rtt): 130.4 ms **Same corridor (South America → Asia Pacific)** * [São Paulo to Tokyo latency and RTT](/docs/network/latency/pairs/gru-tyo-rtt): 230.3 ms * [São Paulo to Taipei latency and RTT](/docs/network/latency/pairs/gru-tpe-rtt): 259.1 ms * [São Paulo to Sydney latency and RTT](/docs/network/latency/pairs/gru-syd-rtt): 265.5 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/gru-hkg.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, São Paulo → Hong Kong round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/gru-hkg-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # São Paulo, Brazil → Ashburn, USA RTT 🇧🇷 **São Paulo, Brazil (GRU)** → 🇺🇸 **Ashburn, USA (IAD)** Measurement round: `2026-08-16T04:07:28Z`. São Paulo to Ashburn averaged 128 ms across 50 replies. The observed range was 121.55 to 142 ms; these figures describe one probe round, not a long-term service target. This is number 3 of 19 outgoing routes by average RTT. That ranking compares different destinations and is not adjusted for distance. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **128 ms** | | Jitter | **4.04 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **74.79 ms** | | Fiber Efficiency | **58.4%** | | Latency Tier | Good | | Source Region | [South America](/docs/network/latency/regions/south-america) | | Destination Region | [North America](/docs/network/latency/regions/north-america) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [São Paulo](https://www.geonames.org/3448439) (-23.54750, -46.63611) and [Ashburn](https://www.geonames.org/4744870) (39.04372, -77.48749). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **7,637.7 km** | | Vacuum RTT floor | **50.95 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **74.79 ms** | | Low-latency fiber material floor | **74.49 ms** | | Engineering floor (5% path allowance) | **78.54 ms** | | Research 1.33× mapped-fiber reference | **99.48 ms** | | Estimated unamplified path loss | **1603.9 dB** | | Transparent optical spans / inline amplifiers | **101 / 100** | | Published RTT inflation over fiber floor | **1.71×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **121.55 ms** | | Average RTT | **128 ms** | | Maximum RTT | **142 ms** | | Standard deviation | **4.73 ms** | | Stdev / average | **3.7%** | ## Route Context * Departure PoP: [São Paulo (GRU)](/docs/network/latency/gru-sao-paulo) * Destination PoP: [Ashburn (IAD)](/docs/network/latency/iad-ashburn) * Region overview: [South America](/docs/network/latency/regions/south-america) → [North America](/docs/network/latency/regions/north-america) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Ashburn to São Paulo latency and RTT](/docs/network/latency/pairs/iad-gru-rtt): 101.4 ms **Fastest routes departing São Paulo (GRU)** * [São Paulo to Miami latency and RTT](/docs/network/latency/pairs/gru-mia-rtt): 74 ms * [São Paulo to New York latency and RTT](/docs/network/latency/pairs/gru-nyc-rtt): 122 ms * [São Paulo to Los Angeles latency and RTT](/docs/network/latency/pairs/gru-lax-rtt): 130.8 ms * [São Paulo to Seattle latency and RTT](/docs/network/latency/pairs/gru-sea-rtt): 155.9 ms * [São Paulo to London latency and RTT](/docs/network/latency/pairs/gru-lon-rtt): 175.8 ms **Fastest routes arriving at Ashburn (IAD)** * [New York to Ashburn latency and RTT](/docs/network/latency/pairs/nyc-iad-rtt): 6 ms * [Miami to Ashburn latency and RTT](/docs/network/latency/pairs/mia-iad-rtt): 27.3 ms * [Los Angeles to Ashburn latency and RTT](/docs/network/latency/pairs/lax-iad-rtt): 60.4 ms * [Seattle to Ashburn latency and RTT](/docs/network/latency/pairs/sea-iad-rtt): 61.5 ms * [London to Ashburn latency and RTT](/docs/network/latency/pairs/lon-iad-rtt): 71 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/gru-iad.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, São Paulo → Ashburn round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/gru-iad-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # São Paulo, Brazil → Johannesburg, South Africa RTT 🇧🇷 **São Paulo, Brazil (GRU)** → 🇿🇦 **Johannesburg, South Africa (JNB)** Measurement round: `2026-08-16T04:07:28Z`. São Paulo to Johannesburg averaged 333.8 ms, compared with a 72.88 ms city-centre fiber reference. The difference cannot be assigned to a particular cable or carrier from RTT alone. Standard deviation was 10.62 ms, or 3.2% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **333.8 ms** | | Jitter | **9.28 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **72.88 ms** | | Fiber Efficiency | **21.8%** | | Latency Tier | High | | Source Region | [South America](/docs/network/latency/regions/south-america) | | Destination Region | [Africa](/docs/network/latency/regions/africa) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [São Paulo](https://www.geonames.org/3448439) (-23.54750, -46.63611) and [Johannesburg](https://www.geonames.org/993800) (-26.20227, 28.04363). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **7,441.9 km** | | Vacuum RTT floor | **49.65 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **72.88 ms** | | Low-latency fiber material floor | **72.58 ms** | | Engineering floor (5% path allowance) | **76.53 ms** | | Research 1.33× mapped-fiber reference | **96.93 ms** | | Estimated unamplified path loss | **1562.8 dB** | | Transparent optical spans / inline amplifiers | **98 / 97** | | Published RTT inflation over fiber floor | **4.58×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **321.14 ms** | | Average RTT | **333.8 ms** | | Maximum RTT | **369.06 ms** | | Standard deviation | **10.62 ms** | | Stdev / average | **3.2%** | ## Route Context * Departure PoP: [São Paulo (GRU)](/docs/network/latency/gru-sao-paulo) * Destination PoP: [Johannesburg (JNB)](/docs/network/latency/jnb-johannesburg) * Region overview: [South America](/docs/network/latency/regions/south-america) → [Africa](/docs/network/latency/regions/africa) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Johannesburg to São Paulo latency and RTT](/docs/network/latency/pairs/jnb-gru-rtt): 328.7 ms **Fastest routes departing São Paulo (GRU)** * [São Paulo to Miami latency and RTT](/docs/network/latency/pairs/gru-mia-rtt): 74 ms * [São Paulo to New York latency and RTT](/docs/network/latency/pairs/gru-nyc-rtt): 122 ms * [São Paulo to Ashburn latency and RTT](/docs/network/latency/pairs/gru-iad-rtt): 128 ms * [São Paulo to Los Angeles latency and RTT](/docs/network/latency/pairs/gru-lax-rtt): 130.8 ms * [São Paulo to Seattle latency and RTT](/docs/network/latency/pairs/gru-sea-rtt): 155.9 ms **Fastest routes arriving at Johannesburg (JNB)** * [Amsterdam to Johannesburg latency and RTT](/docs/network/latency/pairs/ams-jnb-rtt): 171.9 ms * [Berlin to Johannesburg latency and RTT](/docs/network/latency/pairs/ber-jnb-rtt): 172.1 ms * [London to Johannesburg latency and RTT](/docs/network/latency/pairs/lon-jnb-rtt): 179.6 ms * [Paris to Johannesburg latency and RTT](/docs/network/latency/pairs/par-jnb-rtt): 180.9 ms * [Frankfurt to Johannesburg latency and RTT](/docs/network/latency/pairs/fra-jnb-rtt): 188.4 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/gru-jnb.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, São Paulo → Johannesburg round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/gru-jnb-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # São Paulo, Brazil → Los Angeles, USA RTT 🇧🇷 **São Paulo, Brazil (GRU)** → 🇺🇸 **Los Angeles, USA (LAX)** Measurement round: `2026-08-16T04:07:28Z`. São Paulo to Los Angeles averaged 130.8 ms across 50 replies. The observed range was 125.52 to 141.14 ms; these figures describe one probe round, not a long-term service target. This is number 4 of 19 outgoing routes by average RTT. That ranking compares different destinations and is not adjusted for distance. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **130.8 ms** | | Jitter | **3.38 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **96.89 ms** | | Fiber Efficiency | **74.1%** | | Latency Tier | Good | | Source Region | [South America](/docs/network/latency/regions/south-america) | | Destination Region | [North America](/docs/network/latency/regions/north-america) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [São Paulo](https://www.geonames.org/3448439) (-23.54750, -46.63611) and [Los Angeles](https://www.geonames.org/5368361) (34.05223, -118.24368). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **9,894.5 km** | | Vacuum RTT floor | **66.01 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **96.89 ms** | | Low-latency fiber material floor | **96.51 ms** | | Engineering floor (5% path allowance) | **101.75 ms** | | Research 1.33× mapped-fiber reference | **128.87 ms** | | Estimated unamplified path loss | **2077.8 dB** | | Transparent optical spans / inline amplifiers | **130 / 129** | | Published RTT inflation over fiber floor | **1.35×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **125.52 ms** | | Average RTT | **130.8 ms** | | Maximum RTT | **141.14 ms** | | Standard deviation | **3.75 ms** | | Stdev / average | **2.9%** | ## Route Context * Departure PoP: [São Paulo (GRU)](/docs/network/latency/gru-sao-paulo) * Destination PoP: [Los Angeles (LAX)](/docs/network/latency/lax-los-angeles) * Region overview: [South America](/docs/network/latency/regions/south-america) → [North America](/docs/network/latency/regions/north-america) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Los Angeles to São Paulo latency and RTT](/docs/network/latency/pairs/lax-gru-rtt): 131.3 ms **Fastest routes departing São Paulo (GRU)** * [São Paulo to Miami latency and RTT](/docs/network/latency/pairs/gru-mia-rtt): 74 ms * [São Paulo to New York latency and RTT](/docs/network/latency/pairs/gru-nyc-rtt): 122 ms * [São Paulo to Ashburn latency and RTT](/docs/network/latency/pairs/gru-iad-rtt): 128 ms * [São Paulo to Seattle latency and RTT](/docs/network/latency/pairs/gru-sea-rtt): 155.9 ms * [São Paulo to London latency and RTT](/docs/network/latency/pairs/gru-lon-rtt): 175.8 ms **Fastest routes arriving at Los Angeles (LAX)** * [Seattle to Los Angeles latency and RTT](/docs/network/latency/pairs/sea-lax-rtt): 25.9 ms * [Miami to Los Angeles latency and RTT](/docs/network/latency/pairs/mia-lax-rtt): 56.8 ms * [New York to Los Angeles latency and RTT](/docs/network/latency/pairs/nyc-lax-rtt): 58.9 ms * [Ashburn to Los Angeles latency and RTT](/docs/network/latency/pairs/iad-lax-rtt): 60.7 ms * [Tokyo to Los Angeles latency and RTT](/docs/network/latency/pairs/tyo-lax-rtt): 101.2 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/gru-lax.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, São Paulo → Los Angeles round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/gru-lax-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # São Paulo, Brazil → London, UK RTT 🇧🇷 **São Paulo, Brazil (GRU)** → 🇬🇧 **London, UK (LON)** Measurement round: `2026-08-16T04:07:28Z`. São Paulo to London averaged 175.8 ms across 50 replies. The observed range was 170.28 to 196 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 5.34 ms, or 3.0% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **175.8 ms** | | Jitter | **3.82 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **92.76 ms** | | Fiber Efficiency | **52.8%** | | Latency Tier | Fair | | Source Region | [South America](/docs/network/latency/regions/south-america) | | Destination Region | [Europe](/docs/network/latency/regions/europe) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [São Paulo](https://www.geonames.org/3448439) (-23.54750, -46.63611) and [London](https://www.geonames.org/2643743) (51.50853, -0.12574). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **9,472.8 km** | | Vacuum RTT floor | **63.2 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **92.76 ms** | | Low-latency fiber material floor | **92.39 ms** | | Engineering floor (5% path allowance) | **97.42 ms** | | Research 1.33× mapped-fiber reference | **123.38 ms** | | Estimated unamplified path loss | **1989.3 dB** | | Transparent optical spans / inline amplifiers | **125 / 124** | | Published RTT inflation over fiber floor | **1.9×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **170.28 ms** | | Average RTT | **175.8 ms** | | Maximum RTT | **196 ms** | | Standard deviation | **5.34 ms** | | Stdev / average | **3.0%** | ## Route Context * Departure PoP: [São Paulo (GRU)](/docs/network/latency/gru-sao-paulo) * Destination PoP: [London (LON)](/docs/network/latency/lon-london) * Region overview: [South America](/docs/network/latency/regions/south-america) → [Europe](/docs/network/latency/regions/europe) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [London to São Paulo latency and RTT](/docs/network/latency/pairs/lon-gru-rtt): 170.7 ms **Fastest routes departing São Paulo (GRU)** * [São Paulo to Miami latency and RTT](/docs/network/latency/pairs/gru-mia-rtt): 74 ms * [São Paulo to New York latency and RTT](/docs/network/latency/pairs/gru-nyc-rtt): 122 ms * [São Paulo to Ashburn latency and RTT](/docs/network/latency/pairs/gru-iad-rtt): 128 ms * [São Paulo to Los Angeles latency and RTT](/docs/network/latency/pairs/gru-lax-rtt): 130.8 ms * [São Paulo to Seattle latency and RTT](/docs/network/latency/pairs/gru-sea-rtt): 155.9 ms **Fastest routes arriving at London (LON)** * [Amsterdam to London latency and RTT](/docs/network/latency/pairs/ams-lon-rtt): 5.2 ms * [Paris to London latency and RTT](/docs/network/latency/pairs/par-lon-rtt): 6.4 ms * [Frankfurt to London latency and RTT](/docs/network/latency/pairs/fra-lon-rtt): 12.9 ms * [Berlin to London latency and RTT](/docs/network/latency/pairs/ber-lon-rtt): 15.2 ms * [Marseille to London latency and RTT](/docs/network/latency/pairs/mrs-lon-rtt): 18.9 ms **Same corridor (South America → Europe)** * [São Paulo to Amsterdam latency and RTT](/docs/network/latency/pairs/gru-ams-rtt): 180 ms * [São Paulo to Paris latency and RTT](/docs/network/latency/pairs/gru-par-rtt): 180.2 ms * [São Paulo to Frankfurt latency and RTT](/docs/network/latency/pairs/gru-fra-rtt): 186.2 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/gru-lon.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, São Paulo → London round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/gru-lon-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # São Paulo, Brazil → Melbourne, Australia RTT 🇧🇷 **São Paulo, Brazil (GRU)** → 🇦🇺 **Melbourne, Australia (MEL)** Measurement round: `2026-08-16T04:07:28Z`. São Paulo to Melbourne averaged 274.7 ms, compared with a 128.34 ms city-centre fiber reference. The difference cannot be assigned to a particular cable or carrier from RTT alone. Standard deviation was 5.82 ms, or 2.1% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **274.7 ms** | | Jitter | **4.86 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **128.34 ms** | | Fiber Efficiency | **46.7%** | | Latency Tier | High | | Source Region | [South America](/docs/network/latency/regions/south-america) | | Destination Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [São Paulo](https://www.geonames.org/3448439) (-23.54750, -46.63611) and [Melbourne](https://www.geonames.org/2158177) (-37.81400, 144.96332). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | --------------- | | WGS-84 geodesic distance | **13,105.1 km** | | Vacuum RTT floor | **87.43 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **128.34 ms** | | Low-latency fiber material floor | **127.82 ms** | | Engineering floor (5% path allowance) | **134.77 ms** | | Research 1.33× mapped-fiber reference | **170.69 ms** | | Estimated unamplified path loss | **2752.1 dB** | | Transparent optical spans / inline amplifiers | **173 / 172** | | Published RTT inflation over fiber floor | **2.14×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **268.04 ms** | | Average RTT | **274.7 ms** | | Maximum RTT | **300.54 ms** | | Standard deviation | **5.82 ms** | | Stdev / average | **2.1%** | ## Route Context * Departure PoP: [São Paulo (GRU)](/docs/network/latency/gru-sao-paulo) * Destination PoP: [Melbourne (MEL)](/docs/network/latency/mel-melbourne) * Region overview: [South America](/docs/network/latency/regions/south-america) → [Asia Pacific](/docs/network/latency/regions/asia-pacific) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Melbourne to São Paulo latency and RTT](/docs/network/latency/pairs/mel-gru-rtt): 305.3 ms **Fastest routes departing São Paulo (GRU)** * [São Paulo to Miami latency and RTT](/docs/network/latency/pairs/gru-mia-rtt): 74 ms * [São Paulo to New York latency and RTT](/docs/network/latency/pairs/gru-nyc-rtt): 122 ms * [São Paulo to Ashburn latency and RTT](/docs/network/latency/pairs/gru-iad-rtt): 128 ms * [São Paulo to Los Angeles latency and RTT](/docs/network/latency/pairs/gru-lax-rtt): 130.8 ms * [São Paulo to Seattle latency and RTT](/docs/network/latency/pairs/gru-sea-rtt): 155.9 ms **Fastest routes arriving at Melbourne (MEL)** * [Sydney to Melbourne latency and RTT](/docs/network/latency/pairs/syd-mel-rtt): 9.8 ms * [Singapore to Melbourne latency and RTT](/docs/network/latency/pairs/sin-mel-rtt): 88.6 ms * [Tokyo to Melbourne latency and RTT](/docs/network/latency/pairs/tyo-mel-rtt): 113.9 ms * [Hong Kong to Melbourne latency and RTT](/docs/network/latency/pairs/hkg-mel-rtt): 138.4 ms * [Taipei to Melbourne latency and RTT](/docs/network/latency/pairs/tpe-mel-rtt): 142 ms **Same corridor (South America → Asia Pacific)** * [São Paulo to Tokyo latency and RTT](/docs/network/latency/pairs/gru-tyo-rtt): 230.3 ms * [São Paulo to Taipei latency and RTT](/docs/network/latency/pairs/gru-tpe-rtt): 259.1 ms * [São Paulo to Sydney latency and RTT](/docs/network/latency/pairs/gru-syd-rtt): 265.5 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/gru-mel.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, São Paulo → Melbourne round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/gru-mel-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # São Paulo, Brazil → Miami, USA RTT 🇧🇷 **São Paulo, Brazil (GRU)** → 🇺🇸 **Miami, USA (MIA)** Measurement round: `2026-08-16T04:07:28Z`. São Paulo to Miami averaged 74 ms across 50 replies. The observed range was 70.55 to 83.79 ms; these figures describe one probe round, not a long-term service target. This is number 1 of 19 outgoing routes by average RTT. That ranking compares different destinations and is not adjusted for distance. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **74 ms** | | Jitter | **2.29 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **64.11 ms** | | Fiber Efficiency | **86.6%** | | Latency Tier | Excellent | | Source Region | [South America](/docs/network/latency/regions/south-america) | | Destination Region | [North America](/docs/network/latency/regions/north-america) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [São Paulo](https://www.geonames.org/3448439) (-23.54750, -46.63611) and [Miami](https://www.geonames.org/4164138) (25.77427, -80.19366). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **6,546.4 km** | | Vacuum RTT floor | **43.67 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **64.11 ms** | | Low-latency fiber material floor | **63.85 ms** | | Engineering floor (5% path allowance) | **67.32 ms** | | Research 1.33× mapped-fiber reference | **85.26 ms** | | Estimated unamplified path loss | **1374.7 dB** | | Transparent optical spans / inline amplifiers | **86 / 85** | | Published RTT inflation over fiber floor | **1.15×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------ | | Minimum RTT | **70.55 ms** | | Average RTT | **74 ms** | | Maximum RTT | **83.79 ms** | | Standard deviation | **3.07 ms** | | Stdev / average | **4.1%** | ## Route Context * Departure PoP: [São Paulo (GRU)](/docs/network/latency/gru-sao-paulo) * Destination PoP: [Miami (MIA)](/docs/network/latency/mia-miami) * Region overview: [South America](/docs/network/latency/regions/south-america) → [North America](/docs/network/latency/regions/north-america) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Miami to São Paulo latency and RTT](/docs/network/latency/pairs/mia-gru-rtt): 128.7 ms **Fastest routes departing São Paulo (GRU)** * [São Paulo to New York latency and RTT](/docs/network/latency/pairs/gru-nyc-rtt): 122 ms * [São Paulo to Ashburn latency and RTT](/docs/network/latency/pairs/gru-iad-rtt): 128 ms * [São Paulo to Los Angeles latency and RTT](/docs/network/latency/pairs/gru-lax-rtt): 130.8 ms * [São Paulo to Seattle latency and RTT](/docs/network/latency/pairs/gru-sea-rtt): 155.9 ms * [São Paulo to London latency and RTT](/docs/network/latency/pairs/gru-lon-rtt): 175.8 ms **Fastest routes arriving at Miami (MIA)** * [Ashburn to Miami latency and RTT](/docs/network/latency/pairs/iad-mia-rtt): 27.4 ms * [New York to Miami latency and RTT](/docs/network/latency/pairs/nyc-mia-rtt): 32.9 ms * [Los Angeles to Miami latency and RTT](/docs/network/latency/pairs/lax-mia-rtt): 57.3 ms * [Seattle to Miami latency and RTT](/docs/network/latency/pairs/sea-mia-rtt): 81.7 ms * [London to Miami latency and RTT](/docs/network/latency/pairs/lon-mia-rtt): 101.9 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/gru-mia.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, São Paulo → Miami round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/gru-mia-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # São Paulo, Brazil → Moscow, Russia RTT 🇧🇷 **São Paulo, Brazil (GRU)** → 🇷🇺 **Moscow, Russia (MOW)** Measurement round: `2026-08-16T04:07:28Z`. São Paulo to Moscow averaged 219.5 ms across 50 replies. The observed range was 208.05 to 246.05 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 7.63 ms, or 3.5% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **219.5 ms** | | Jitter | **6.64 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **115.47 ms** | | Fiber Efficiency | **52.6%** | | Latency Tier | Fair | | Source Region | [South America](/docs/network/latency/regions/south-america) | | Destination Region | [Europe](/docs/network/latency/regions/europe) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [São Paulo](https://www.geonames.org/3448439) (-23.54750, -46.63611) and [Moscow](https://www.geonames.org/524901) (55.75204, 37.61781). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | --------------- | | WGS-84 geodesic distance | **11,791.1 km** | | Vacuum RTT floor | **78.66 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **115.47 ms** | | Low-latency fiber material floor | **115 ms** | | Engineering floor (5% path allowance) | **121.26 ms** | | Research 1.33× mapped-fiber reference | **153.57 ms** | | Estimated unamplified path loss | **2476.1 dB** | | Transparent optical spans / inline amplifiers | **155 / 154** | | Published RTT inflation over fiber floor | **1.9×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **208.05 ms** | | Average RTT | **219.5 ms** | | Maximum RTT | **246.05 ms** | | Standard deviation | **7.63 ms** | | Stdev / average | **3.5%** | ## Route Context * Departure PoP: [São Paulo (GRU)](/docs/network/latency/gru-sao-paulo) * Destination PoP: [Moscow (MOW)](/docs/network/latency/mow-moscow) * Region overview: [South America](/docs/network/latency/regions/south-america) → [Europe](/docs/network/latency/regions/europe) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Moscow to São Paulo latency and RTT](/docs/network/latency/pairs/mow-gru-rtt): 214.4 ms **Fastest routes departing São Paulo (GRU)** * [São Paulo to Miami latency and RTT](/docs/network/latency/pairs/gru-mia-rtt): 74 ms * [São Paulo to New York latency and RTT](/docs/network/latency/pairs/gru-nyc-rtt): 122 ms * [São Paulo to Ashburn latency and RTT](/docs/network/latency/pairs/gru-iad-rtt): 128 ms * [São Paulo to Los Angeles latency and RTT](/docs/network/latency/pairs/gru-lax-rtt): 130.8 ms * [São Paulo to Seattle latency and RTT](/docs/network/latency/pairs/gru-sea-rtt): 155.9 ms **Fastest routes arriving at Moscow (MOW)** * [Berlin to Moscow latency and RTT](/docs/network/latency/pairs/ber-mow-rtt): 28.6 ms * [Frankfurt to Moscow latency and RTT](/docs/network/latency/pairs/fra-mow-rtt): 34.7 ms * [Amsterdam to Moscow latency and RTT](/docs/network/latency/pairs/ams-mow-rtt): 38.1 ms * [London to Moscow latency and RTT](/docs/network/latency/pairs/lon-mow-rtt): 43.5 ms * [Paris to Moscow latency and RTT](/docs/network/latency/pairs/par-mow-rtt): 44.7 ms **Same corridor (South America → Europe)** * [São Paulo to London latency and RTT](/docs/network/latency/pairs/gru-lon-rtt): 175.8 ms * [São Paulo to Amsterdam latency and RTT](/docs/network/latency/pairs/gru-ams-rtt): 180 ms * [São Paulo to Paris latency and RTT](/docs/network/latency/pairs/gru-par-rtt): 180.2 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/gru-mow.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, São Paulo → Moscow round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/gru-mow-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # São Paulo, Brazil → Marseille, France RTT 🇧🇷 **São Paulo, Brazil (GRU)** → 🇫🇷 **Marseille, France (MRS)** Measurement round: `2026-08-16T04:07:28Z`. São Paulo to Marseille averaged 187.7 ms, compared with a 89.24 ms city-centre fiber reference. The difference cannot be assigned to a particular cable or carrier from RTT alone. Standard deviation was 8.46 ms, or 4.5% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **187.7 ms** | | Jitter | **7.15 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **89.24 ms** | | Fiber Efficiency | **47.5%** | | Latency Tier | Fair | | Source Region | [South America](/docs/network/latency/regions/south-america) | | Destination Region | [Europe](/docs/network/latency/regions/europe) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [São Paulo](https://www.geonames.org/3448439) (-23.54750, -46.63611) and [Marseille](https://www.geonames.org/2995469) (43.29695, 5.38107). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | ------------- | | WGS-84 geodesic distance | **9,113 km** | | Vacuum RTT floor | **60.8 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **89.24 ms** | | Low-latency fiber material floor | **88.88 ms** | | Engineering floor (5% path allowance) | **93.72 ms** | | Research 1.33× mapped-fiber reference | **118.69 ms** | | Estimated unamplified path loss | **1913.7 dB** | | Transparent optical spans / inline amplifiers | **120 / 119** | | Published RTT inflation over fiber floor | **2.1×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **176.81 ms** | | Average RTT | **187.7 ms** | | Maximum RTT | **212.03 ms** | | Standard deviation | **8.46 ms** | | Stdev / average | **4.5%** | ## Route Context * Departure PoP: [São Paulo (GRU)](/docs/network/latency/gru-sao-paulo) * Destination PoP: [Marseille (MRS)](/docs/network/latency/mrs-marseille) * Region overview: [South America](/docs/network/latency/regions/south-america) → [Europe](/docs/network/latency/regions/europe) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Marseille to São Paulo latency and RTT](/docs/network/latency/pairs/mrs-gru-rtt): 186.7 ms **Fastest routes departing São Paulo (GRU)** * [São Paulo to Miami latency and RTT](/docs/network/latency/pairs/gru-mia-rtt): 74 ms * [São Paulo to New York latency and RTT](/docs/network/latency/pairs/gru-nyc-rtt): 122 ms * [São Paulo to Ashburn latency and RTT](/docs/network/latency/pairs/gru-iad-rtt): 128 ms * [São Paulo to Los Angeles latency and RTT](/docs/network/latency/pairs/gru-lax-rtt): 130.8 ms * [São Paulo to Seattle latency and RTT](/docs/network/latency/pairs/gru-sea-rtt): 155.9 ms **Fastest routes arriving at Marseille (MRS)** * [Paris to Marseille latency and RTT](/docs/network/latency/pairs/par-mrs-rtt): 8.9 ms * [Frankfurt to Marseille latency and RTT](/docs/network/latency/pairs/fra-mrs-rtt): 16.4 ms * [London to Marseille latency and RTT](/docs/network/latency/pairs/lon-mrs-rtt): 17.9 ms * [Amsterdam to Marseille latency and RTT](/docs/network/latency/pairs/ams-mrs-rtt): 19.6 ms * [Berlin to Marseille latency and RTT](/docs/network/latency/pairs/ber-mrs-rtt): 20.2 ms **Same corridor (South America → Europe)** * [São Paulo to London latency and RTT](/docs/network/latency/pairs/gru-lon-rtt): 175.8 ms * [São Paulo to Amsterdam latency and RTT](/docs/network/latency/pairs/gru-ams-rtt): 180 ms * [São Paulo to Paris latency and RTT](/docs/network/latency/pairs/gru-par-rtt): 180.2 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/gru-mrs.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, São Paulo → Marseille round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/gru-mrs-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # São Paulo, Brazil → New York, USA RTT 🇧🇷 **São Paulo, Brazil (GRU)** → 🇺🇸 **New York, USA (NYC)** Measurement round: `2026-08-16T04:07:28Z`. São Paulo to New York averaged 122 ms across 50 replies. The observed range was 116.85 to 132.54 ms; these figures describe one probe round, not a long-term service target. This is number 2 of 19 outgoing routes by average RTT. That ranking compares different destinations and is not adjusted for distance. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **122 ms** | | Jitter | **3.62 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **74.99 ms** | | Fiber Efficiency | **61.5%** | | Latency Tier | Good | | Source Region | [South America](/docs/network/latency/regions/south-america) | | Destination Region | [North America](/docs/network/latency/regions/north-america) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [São Paulo](https://www.geonames.org/3448439) (-23.54750, -46.63611) and [New York City](https://www.geonames.org/5128581) (40.71427, -74.00597). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **7,657.6 km** | | Vacuum RTT floor | **51.09 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **74.99 ms** | | Low-latency fiber material floor | **74.69 ms** | | Engineering floor (5% path allowance) | **78.75 ms** | | Research 1.33× mapped-fiber reference | **99.74 ms** | | Estimated unamplified path loss | **1608.1 dB** | | Transparent optical spans / inline amplifiers | **101 / 100** | | Published RTT inflation over fiber floor | **1.63×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **116.85 ms** | | Average RTT | **122 ms** | | Maximum RTT | **132.54 ms** | | Standard deviation | **3.8 ms** | | Stdev / average | **3.1%** | ## Route Context * Departure PoP: [São Paulo (GRU)](/docs/network/latency/gru-sao-paulo) * Destination PoP: [New York (NYC)](/docs/network/latency/nyc-new-york) * Region overview: [South America](/docs/network/latency/regions/south-america) → [North America](/docs/network/latency/regions/north-america) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [New York to São Paulo latency and RTT](/docs/network/latency/pairs/nyc-gru-rtt): 106.9 ms **Fastest routes departing São Paulo (GRU)** * [São Paulo to Miami latency and RTT](/docs/network/latency/pairs/gru-mia-rtt): 74 ms * [São Paulo to Ashburn latency and RTT](/docs/network/latency/pairs/gru-iad-rtt): 128 ms * [São Paulo to Los Angeles latency and RTT](/docs/network/latency/pairs/gru-lax-rtt): 130.8 ms * [São Paulo to Seattle latency and RTT](/docs/network/latency/pairs/gru-sea-rtt): 155.9 ms * [São Paulo to London latency and RTT](/docs/network/latency/pairs/gru-lon-rtt): 175.8 ms **Fastest routes arriving at New York (NYC)** * [Ashburn to New York latency and RTT](/docs/network/latency/pairs/iad-nyc-rtt): 6 ms * [Miami to New York latency and RTT](/docs/network/latency/pairs/mia-nyc-rtt): 33.3 ms * [Los Angeles to New York latency and RTT](/docs/network/latency/pairs/lax-nyc-rtt): 57.9 ms * [Seattle to New York latency and RTT](/docs/network/latency/pairs/sea-nyc-rtt): 59.5 ms * [London to New York latency and RTT](/docs/network/latency/pairs/lon-nyc-rtt): 63.8 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/gru-nyc.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, São Paulo → New York round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/gru-nyc-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # São Paulo, Brazil → Paris, France RTT 🇧🇷 **São Paulo, Brazil (GRU)** → 🇫🇷 **Paris, France (PAR)** Measurement round: `2026-08-16T04:07:28Z`. São Paulo to Paris averaged 180.2 ms across 50 replies. The observed range was 169.53 to 210.06 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 8.35 ms, or 4.6% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **180.2 ms** | | Jitter | **7.23 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **91.83 ms** | | Fiber Efficiency | **51%** | | Latency Tier | Fair | | Source Region | [South America](/docs/network/latency/regions/south-america) | | Destination Region | [Europe](/docs/network/latency/regions/europe) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [São Paulo](https://www.geonames.org/3448439) (-23.54750, -46.63611) and [Paris](https://www.geonames.org/2988507) (48.85341, 2.34880). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **9,377.8 km** | | Vacuum RTT floor | **62.56 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **91.83 ms** | | Low-latency fiber material floor | **91.47 ms** | | Engineering floor (5% path allowance) | **96.44 ms** | | Research 1.33× mapped-fiber reference | **122.14 ms** | | Estimated unamplified path loss | **1969.3 dB** | | Transparent optical spans / inline amplifiers | **124 / 123** | | Published RTT inflation over fiber floor | **1.96×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **169.53 ms** | | Average RTT | **180.2 ms** | | Maximum RTT | **210.06 ms** | | Standard deviation | **8.35 ms** | | Stdev / average | **4.6%** | ## Route Context * Departure PoP: [São Paulo (GRU)](/docs/network/latency/gru-sao-paulo) * Destination PoP: [Paris (PAR)](/docs/network/latency/par-paris) * Region overview: [South America](/docs/network/latency/regions/south-america) → [Europe](/docs/network/latency/regions/europe) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Paris to São Paulo latency and RTT](/docs/network/latency/pairs/par-gru-rtt): 175.8 ms **Fastest routes departing São Paulo (GRU)** * [São Paulo to Miami latency and RTT](/docs/network/latency/pairs/gru-mia-rtt): 74 ms * [São Paulo to New York latency and RTT](/docs/network/latency/pairs/gru-nyc-rtt): 122 ms * [São Paulo to Ashburn latency and RTT](/docs/network/latency/pairs/gru-iad-rtt): 128 ms * [São Paulo to Los Angeles latency and RTT](/docs/network/latency/pairs/gru-lax-rtt): 130.8 ms * [São Paulo to Seattle latency and RTT](/docs/network/latency/pairs/gru-sea-rtt): 155.9 ms **Fastest routes arriving at Paris (PAR)** * [London to Paris latency and RTT](/docs/network/latency/pairs/lon-par-rtt): 6.4 ms * [Amsterdam to Paris latency and RTT](/docs/network/latency/pairs/ams-par-rtt): 7 ms * [Frankfurt to Paris latency and RTT](/docs/network/latency/pairs/fra-par-rtt): 7.6 ms * [Marseille to Paris latency and RTT](/docs/network/latency/pairs/mrs-par-rtt): 8.9 ms * [Berlin to Paris latency and RTT](/docs/network/latency/pairs/ber-par-rtt): 15.6 ms **Same corridor (South America → Europe)** * [São Paulo to London latency and RTT](/docs/network/latency/pairs/gru-lon-rtt): 175.8 ms * [São Paulo to Amsterdam latency and RTT](/docs/network/latency/pairs/gru-ams-rtt): 180 ms * [São Paulo to Frankfurt latency and RTT](/docs/network/latency/pairs/gru-fra-rtt): 186.2 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/gru-par.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, São Paulo → Paris round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/gru-par-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # São Paulo, Brazil → Seattle, USA RTT 🇧🇷 **São Paulo, Brazil (GRU)** → 🇺🇸 **Seattle, USA (SEA)** Measurement round: `2026-08-16T04:07:28Z`. São Paulo to Seattle averaged 155.9 ms across 50 replies. The observed range was 151.12 to 166.61 ms; these figures describe one probe round, not a long-term service target. This is number 5 of 19 outgoing routes by average RTT. That ranking compares different destinations and is not adjusted for distance. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **155.9 ms** | | Jitter | **3.4 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **106.76 ms** | | Fiber Efficiency | **68.5%** | | Latency Tier | Fair | | Source Region | [South America](/docs/network/latency/regions/south-america) | | Destination Region | [North America](/docs/network/latency/regions/north-america) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [São Paulo](https://www.geonames.org/3448439) (-23.54750, -46.63611) and [Seattle](https://www.geonames.org/5809844) (47.60621, -122.33207). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | --------------- | | WGS-84 geodesic distance | **10,902.3 km** | | Vacuum RTT floor | **72.73 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **106.76 ms** | | Low-latency fiber material floor | **106.33 ms** | | Engineering floor (5% path allowance) | **112.12 ms** | | Research 1.33× mapped-fiber reference | **142 ms** | | Estimated unamplified path loss | **2289.5 dB** | | Transparent optical spans / inline amplifiers | **144 / 143** | | Published RTT inflation over fiber floor | **1.46×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **151.12 ms** | | Average RTT | **155.9 ms** | | Maximum RTT | **166.61 ms** | | Standard deviation | **3.7 ms** | | Stdev / average | **2.4%** | ## Route Context * Departure PoP: [São Paulo (GRU)](/docs/network/latency/gru-sao-paulo) * Destination PoP: [Seattle (SEA)](/docs/network/latency/sea-seattle) * Region overview: [South America](/docs/network/latency/regions/south-america) → [North America](/docs/network/latency/regions/north-america) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Seattle to São Paulo latency and RTT](/docs/network/latency/pairs/sea-gru-rtt): 155.7 ms **Fastest routes departing São Paulo (GRU)** * [São Paulo to Miami latency and RTT](/docs/network/latency/pairs/gru-mia-rtt): 74 ms * [São Paulo to New York latency and RTT](/docs/network/latency/pairs/gru-nyc-rtt): 122 ms * [São Paulo to Ashburn latency and RTT](/docs/network/latency/pairs/gru-iad-rtt): 128 ms * [São Paulo to Los Angeles latency and RTT](/docs/network/latency/pairs/gru-lax-rtt): 130.8 ms * [São Paulo to London latency and RTT](/docs/network/latency/pairs/gru-lon-rtt): 175.8 ms **Fastest routes arriving at Seattle (SEA)** * [Los Angeles to Seattle latency and RTT](/docs/network/latency/pairs/lax-sea-rtt): 26.9 ms * [New York to Seattle latency and RTT](/docs/network/latency/pairs/nyc-sea-rtt): 58.7 ms * [Ashburn to Seattle latency and RTT](/docs/network/latency/pairs/iad-sea-rtt): 62.3 ms * [Miami to Seattle latency and RTT](/docs/network/latency/pairs/mia-sea-rtt): 81.9 ms * [Tokyo to Seattle latency and RTT](/docs/network/latency/pairs/tyo-sea-rtt): 84.8 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/gru-sea.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, São Paulo → Seattle round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/gru-sea-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # São Paulo, Brazil → Singapore RTT 🇧🇷 **São Paulo, Brazil (GRU)** → 🇸🇬 **Singapore (SIN)** Measurement round: `2026-08-16T04:07:28Z`. São Paulo to Singapore averaged 297.1 ms across 50 replies. The observed range was 280.43 to 326.59 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 11.31 ms, or 3.8% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **297.1 ms** | | Jitter | **10.86 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **156.67 ms** | | Fiber Efficiency | **52.7%** | | Latency Tier | High | | Source Region | [South America](/docs/network/latency/regions/south-america) | | Destination Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [São Paulo](https://www.geonames.org/3448439) (-23.54750, -46.63611) and [Singapore](https://www.geonames.org/1880252) (1.28967, 103.85007). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | --------------- | | WGS-84 geodesic distance | **15,998.2 km** | | Vacuum RTT floor | **106.73 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **156.67 ms** | | Low-latency fiber material floor | **156.04 ms** | | Engineering floor (5% path allowance) | **164.52 ms** | | Research 1.33× mapped-fiber reference | **208.37 ms** | | Estimated unamplified path loss | **3359.6 dB** | | Transparent optical spans / inline amplifiers | **210 / 209** | | Published RTT inflation over fiber floor | **1.9×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **280.43 ms** | | Average RTT | **297.1 ms** | | Maximum RTT | **326.59 ms** | | Standard deviation | **11.31 ms** | | Stdev / average | **3.8%** | ## Route Context * Departure PoP: [São Paulo (GRU)](/docs/network/latency/gru-sao-paulo) * Destination PoP: [Singapore (SIN)](/docs/network/latency/sin-singapore) * Region overview: [South America](/docs/network/latency/regions/south-america) → [Asia Pacific](/docs/network/latency/regions/asia-pacific) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Singapore to São Paulo latency and RTT](/docs/network/latency/pairs/sin-gru-rtt): 311.4 ms **Fastest routes departing São Paulo (GRU)** * [São Paulo to Miami latency and RTT](/docs/network/latency/pairs/gru-mia-rtt): 74 ms * [São Paulo to New York latency and RTT](/docs/network/latency/pairs/gru-nyc-rtt): 122 ms * [São Paulo to Ashburn latency and RTT](/docs/network/latency/pairs/gru-iad-rtt): 128 ms * [São Paulo to Los Angeles latency and RTT](/docs/network/latency/pairs/gru-lax-rtt): 130.8 ms * [São Paulo to Seattle latency and RTT](/docs/network/latency/pairs/gru-sea-rtt): 155.9 ms **Fastest routes arriving at Singapore (SIN)** * [Hong Kong to Singapore latency and RTT](/docs/network/latency/pairs/hkg-sin-rtt): 31.5 ms * [Taipei to Singapore latency and RTT](/docs/network/latency/pairs/tpe-sin-rtt): 44.3 ms * [Tokyo to Singapore latency and RTT](/docs/network/latency/pairs/tyo-sin-rtt): 69.4 ms * [Melbourne to Singapore latency and RTT](/docs/network/latency/pairs/mel-sin-rtt): 88.4 ms * [Sydney to Singapore latency and RTT](/docs/network/latency/pairs/syd-sin-rtt): 94.5 ms **Same corridor (South America → Asia Pacific)** * [São Paulo to Tokyo latency and RTT](/docs/network/latency/pairs/gru-tyo-rtt): 230.3 ms * [São Paulo to Taipei latency and RTT](/docs/network/latency/pairs/gru-tpe-rtt): 259.1 ms * [São Paulo to Sydney latency and RTT](/docs/network/latency/pairs/gru-syd-rtt): 265.5 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/gru-sin.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, São Paulo → Singapore round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/gru-sin-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # São Paulo, Brazil → Sydney, Australia RTT 🇧🇷 **São Paulo, Brazil (GRU)** → 🇦🇺 **Sydney, Australia (SYD)** Measurement round: `2026-08-16T04:07:28Z`. São Paulo to Sydney averaged 265.5 ms, compared with a 131 ms city-centre fiber reference. The difference cannot be assigned to a particular cable or carrier from RTT alone. Standard deviation was 9.22 ms, or 3.5% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **265.5 ms** | | Jitter | **8.27 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **131 ms** | | Fiber Efficiency | **49.3%** | | Latency Tier | High | | Source Region | [South America](/docs/network/latency/regions/south-america) | | Destination Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [São Paulo](https://www.geonames.org/3448439) (-23.54750, -46.63611) and [Sydney](https://www.geonames.org/2147714) (-33.86785, 151.20732). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | ------------- | | WGS-84 geodesic distance | **13,377 km** | | Vacuum RTT floor | **89.24 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **131 ms** | | Low-latency fiber material floor | **130.47 ms** | | Engineering floor (5% path allowance) | **137.57 ms** | | Research 1.33× mapped-fiber reference | **174.23 ms** | | Estimated unamplified path loss | **2809.2 dB** | | Transparent optical spans / inline amplifiers | **176 / 175** | | Published RTT inflation over fiber floor | **2.03×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **255.6 ms** | | Average RTT | **265.5 ms** | | Maximum RTT | **295.52 ms** | | Standard deviation | **9.22 ms** | | Stdev / average | **3.5%** | ## Route Context * Departure PoP: [São Paulo (GRU)](/docs/network/latency/gru-sao-paulo) * Destination PoP: [Sydney (SYD)](/docs/network/latency/syd-sydney) * Region overview: [South America](/docs/network/latency/regions/south-america) → [Asia Pacific](/docs/network/latency/regions/asia-pacific) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Sydney to São Paulo latency and RTT](/docs/network/latency/pairs/syd-gru-rtt): 296.7 ms **Fastest routes departing São Paulo (GRU)** * [São Paulo to Miami latency and RTT](/docs/network/latency/pairs/gru-mia-rtt): 74 ms * [São Paulo to New York latency and RTT](/docs/network/latency/pairs/gru-nyc-rtt): 122 ms * [São Paulo to Ashburn latency and RTT](/docs/network/latency/pairs/gru-iad-rtt): 128 ms * [São Paulo to Los Angeles latency and RTT](/docs/network/latency/pairs/gru-lax-rtt): 130.8 ms * [São Paulo to Seattle latency and RTT](/docs/network/latency/pairs/gru-sea-rtt): 155.9 ms **Fastest routes arriving at Sydney (SYD)** * [Melbourne to Sydney latency and RTT](/docs/network/latency/pairs/mel-syd-rtt): 9.8 ms * [Singapore to Sydney latency and RTT](/docs/network/latency/pairs/sin-syd-rtt): 94.5 ms * [Tokyo to Sydney latency and RTT](/docs/network/latency/pairs/tyo-syd-rtt): 101.9 ms * [Taipei to Sydney latency and RTT](/docs/network/latency/pairs/tpe-syd-rtt): 133.2 ms * [Hong Kong to Sydney latency and RTT](/docs/network/latency/pairs/hkg-syd-rtt): 137.2 ms **Same corridor (South America → Asia Pacific)** * [São Paulo to Tokyo latency and RTT](/docs/network/latency/pairs/gru-tyo-rtt): 230.3 ms * [São Paulo to Taipei latency and RTT](/docs/network/latency/pairs/gru-tpe-rtt): 259.1 ms * [São Paulo to Hong Kong latency and RTT](/docs/network/latency/pairs/gru-hkg-rtt): 273.5 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/gru-syd.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, São Paulo → Sydney round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/gru-syd-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # São Paulo, Brazil → Taipei, Taiwan RTT 🇧🇷 **São Paulo, Brazil (GRU)** → 🇹🇼 **Taipei, Taiwan (TPE)** Measurement round: `2026-08-16T04:07:28Z`. São Paulo to Taipei averaged 259.1 ms across 50 replies. The observed range was 248.08 to 296.38 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 8.63 ms, or 3.3% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **259.1 ms** | | Jitter | **7.16 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **184.29 ms** | | Fiber Efficiency | **71.1%** | | Latency Tier | High | | Source Region | [South America](/docs/network/latency/regions/south-america) | | Destination Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [São Paulo](https://www.geonames.org/3448439) (-23.54750, -46.63611) and [Taipei](https://www.geonames.org/1668341) (25.05306, 121.52639). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | --------------- | | WGS-84 geodesic distance | **18,818.5 km** | | Vacuum RTT floor | **125.54 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **184.29 ms** | | Low-latency fiber material floor | **183.54 ms** | | Engineering floor (5% path allowance) | **193.52 ms** | | Research 1.33× mapped-fiber reference | **245.1 ms** | | Estimated unamplified path loss | **3951.9 dB** | | Transparent optical spans / inline amplifiers | **247 / 246** | | Published RTT inflation over fiber floor | **1.41×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **248.08 ms** | | Average RTT | **259.1 ms** | | Maximum RTT | **296.38 ms** | | Standard deviation | **8.63 ms** | | Stdev / average | **3.3%** | ## Route Context * Departure PoP: [São Paulo (GRU)](/docs/network/latency/gru-sao-paulo) * Destination PoP: [Taipei (TPE)](/docs/network/latency/tpe-taipei) * Region overview: [South America](/docs/network/latency/regions/south-america) → [Asia Pacific](/docs/network/latency/regions/asia-pacific) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Taipei to São Paulo latency and RTT](/docs/network/latency/pairs/tpe-gru-rtt): 273 ms **Fastest routes departing São Paulo (GRU)** * [São Paulo to Miami latency and RTT](/docs/network/latency/pairs/gru-mia-rtt): 74 ms * [São Paulo to New York latency and RTT](/docs/network/latency/pairs/gru-nyc-rtt): 122 ms * [São Paulo to Ashburn latency and RTT](/docs/network/latency/pairs/gru-iad-rtt): 128 ms * [São Paulo to Los Angeles latency and RTT](/docs/network/latency/pairs/gru-lax-rtt): 130.8 ms * [São Paulo to Seattle latency and RTT](/docs/network/latency/pairs/gru-sea-rtt): 155.9 ms **Fastest routes arriving at Taipei (TPE)** * [Hong Kong to Taipei latency and RTT](/docs/network/latency/pairs/hkg-tpe-rtt): 14.7 ms * [Tokyo to Taipei latency and RTT](/docs/network/latency/pairs/tyo-tpe-rtt): 32.3 ms * [Singapore to Taipei latency and RTT](/docs/network/latency/pairs/sin-tpe-rtt): 45.8 ms * [Seattle to Taipei latency and RTT](/docs/network/latency/pairs/sea-tpe-rtt): 114.7 ms * [Sydney to Taipei latency and RTT](/docs/network/latency/pairs/syd-tpe-rtt): 131.7 ms **Same corridor (South America → Asia Pacific)** * [São Paulo to Tokyo latency and RTT](/docs/network/latency/pairs/gru-tyo-rtt): 230.3 ms * [São Paulo to Sydney latency and RTT](/docs/network/latency/pairs/gru-syd-rtt): 265.5 ms * [São Paulo to Hong Kong latency and RTT](/docs/network/latency/pairs/gru-hkg-rtt): 273.5 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/gru-tpe.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, São Paulo → Taipei round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/gru-tpe-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # São Paulo, Brazil → Tokyo, Japan RTT 🇧🇷 **São Paulo, Brazil (GRU)** → 🇯🇵 **Tokyo, Japan (TYO)** Measurement round: `2026-08-16T04:07:28Z`. São Paulo to Tokyo averaged 230.3 ms across 50 replies. The observed range was 223.73 to 259.55 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 6.53 ms, or 2.8% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **230.3 ms** | | Jitter | **4.44 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **181.46 ms** | | Fiber Efficiency | **78.8%** | | Latency Tier | Fair | | Source Region | [South America](/docs/network/latency/regions/south-america) | | Destination Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [São Paulo](https://www.geonames.org/3448439) (-23.54750, -46.63611) and [Tokyo](https://www.geonames.org/1850147) (35.68950, 139.69171). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | --------------- | | WGS-84 geodesic distance | **18,530.2 km** | | Vacuum RTT floor | **123.62 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **181.46 ms** | | Low-latency fiber material floor | **180.73 ms** | | Engineering floor (5% path allowance) | **190.56 ms** | | Research 1.33× mapped-fiber reference | **241.34 ms** | | Estimated unamplified path loss | **3891.3 dB** | | Transparent optical spans / inline amplifiers | **244 / 243** | | Published RTT inflation over fiber floor | **1.27×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **223.73 ms** | | Average RTT | **230.3 ms** | | Maximum RTT | **259.55 ms** | | Standard deviation | **6.53 ms** | | Stdev / average | **2.8%** | ## Route Context * Departure PoP: [São Paulo (GRU)](/docs/network/latency/gru-sao-paulo) * Destination PoP: [Tokyo (TYO)](/docs/network/latency/tyo-tokyo) * Region overview: [South America](/docs/network/latency/regions/south-america) → [Asia Pacific](/docs/network/latency/regions/asia-pacific) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Tokyo to São Paulo latency and RTT](/docs/network/latency/pairs/tyo-gru-rtt): 230.5 ms **Fastest routes departing São Paulo (GRU)** * [São Paulo to Miami latency and RTT](/docs/network/latency/pairs/gru-mia-rtt): 74 ms * [São Paulo to New York latency and RTT](/docs/network/latency/pairs/gru-nyc-rtt): 122 ms * [São Paulo to Ashburn latency and RTT](/docs/network/latency/pairs/gru-iad-rtt): 128 ms * [São Paulo to Los Angeles latency and RTT](/docs/network/latency/pairs/gru-lax-rtt): 130.8 ms * [São Paulo to Seattle latency and RTT](/docs/network/latency/pairs/gru-sea-rtt): 155.9 ms **Fastest routes arriving at Tokyo (TYO)** * [Taipei to Tokyo latency and RTT](/docs/network/latency/pairs/tpe-tyo-rtt): 31.9 ms * [Hong Kong to Tokyo latency and RTT](/docs/network/latency/pairs/hkg-tyo-rtt): 44.9 ms * [Singapore to Tokyo latency and RTT](/docs/network/latency/pairs/sin-tyo-rtt): 68.9 ms * [Seattle to Tokyo latency and RTT](/docs/network/latency/pairs/sea-tyo-rtt): 85 ms * [Los Angeles to Tokyo latency and RTT](/docs/network/latency/pairs/lax-tyo-rtt): 101.2 ms **Same corridor (South America → Asia Pacific)** * [São Paulo to Taipei latency and RTT](/docs/network/latency/pairs/gru-tpe-rtt): 259.1 ms * [São Paulo to Sydney latency and RTT](/docs/network/latency/pairs/gru-syd-rtt): 265.5 ms * [São Paulo to Hong Kong latency and RTT](/docs/network/latency/pairs/gru-hkg-rtt): 273.5 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/gru-tyo.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, São Paulo → Tokyo round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/gru-tyo-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Hong Kong → Amsterdam, Netherlands RTT 🇭🇰 **Hong Kong (HKG)** → 🇳🇱 **Amsterdam, Netherlands (AMS)** Measurement round: `2026-08-16T04:07:28Z`. Hong Kong to Amsterdam averaged 154.4 ms across 50 replies. The observed range was 148.52 to 170.08 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 4.49 ms, or 2.9% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------------------- | | RTT | **154.4 ms** | | Jitter | **4.6 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **91.05 ms** | | Fiber Efficiency | **59%** | | Latency Tier | Fair | | Source Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | | Destination Region | [Europe](/docs/network/latency/regions/europe) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Hong Kong](https://www.geonames.org/1819729) (22.27832, 114.17469) and [Amsterdam](https://www.geonames.org/2759794) (52.37403, 4.88969). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **9,297.6 km** | | Vacuum RTT floor | **62.03 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **91.05 ms** | | Low-latency fiber material floor | **90.68 ms** | | Engineering floor (5% path allowance) | **95.61 ms** | | Research 1.33× mapped-fiber reference | **121.1 ms** | | Estimated unamplified path loss | **1952.5 dB** | | Transparent optical spans / inline amplifiers | **123 / 122** | | Published RTT inflation over fiber floor | **1.7×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **148.52 ms** | | Average RTT | **154.4 ms** | | Maximum RTT | **170.08 ms** | | Standard deviation | **4.49 ms** | | Stdev / average | **2.9%** | ## Route Context * Departure PoP: [Hong Kong (HKG)](/docs/network/latency/hkg-hong-kong) * Destination PoP: [Amsterdam (AMS)](/docs/network/latency/ams-amsterdam) * Region overview: [Asia Pacific](/docs/network/latency/regions/asia-pacific) → [Europe](/docs/network/latency/regions/europe) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Amsterdam to Hong Kong latency and RTT](/docs/network/latency/pairs/ams-hkg-rtt): 155.9 ms **Fastest routes departing Hong Kong (HKG)** * [Hong Kong to Taipei latency and RTT](/docs/network/latency/pairs/hkg-tpe-rtt): 14.7 ms * [Hong Kong to Singapore latency and RTT](/docs/network/latency/pairs/hkg-sin-rtt): 31.5 ms * [Hong Kong to Tokyo latency and RTT](/docs/network/latency/pairs/hkg-tyo-rtt): 44.9 ms * [Hong Kong to Moscow latency and RTT](/docs/network/latency/pairs/hkg-mow-rtt): 118.2 ms * [Hong Kong to Seattle latency and RTT](/docs/network/latency/pairs/hkg-sea-rtt): 131.6 ms **Fastest routes arriving at Amsterdam (AMS)** * [London to Amsterdam latency and RTT](/docs/network/latency/pairs/lon-ams-rtt): 5.2 ms * [Frankfurt to Amsterdam latency and RTT](/docs/network/latency/pairs/fra-ams-rtt): 5.9 ms * [Paris to Amsterdam latency and RTT](/docs/network/latency/pairs/par-ams-rtt): 7 ms * [Berlin to Amsterdam latency and RTT](/docs/network/latency/pairs/ber-ams-rtt): 7.2 ms * [Marseille to Amsterdam latency and RTT](/docs/network/latency/pairs/mrs-ams-rtt): 20.3 ms **Same corridor (Asia Pacific → Europe)** * [Taipei to Moscow latency and RTT](/docs/network/latency/pairs/tpe-mow-rtt): 132.7 ms * [Singapore to Marseille latency and RTT](/docs/network/latency/pairs/sin-mrs-rtt): 140.6 ms * [Hong Kong to Berlin latency and RTT](/docs/network/latency/pairs/hkg-ber-rtt): 145.1 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/hkg-ams.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Hong Kong → Amsterdam round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/hkg-ams-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Hong Kong → Berlin, Germany RTT 🇭🇰 **Hong Kong (HKG)** → 🇩🇪 **Berlin, Germany (BER)** Measurement round: `2026-08-16T04:07:28Z`. Hong Kong to Berlin averaged 145.1 ms across 50 replies. The observed range was 138.49 to 164.88 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 6.67 ms, or 4.6% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------------------- | | RTT | **145.1 ms** | | Jitter | **5.8 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **85.85 ms** | | Fiber Efficiency | **59.2%** | | Latency Tier | Good | | Source Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | | Destination Region | [Europe](/docs/network/latency/regions/europe) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Hong Kong](https://www.geonames.org/1819729) (22.27832, 114.17469) and [Berlin](https://www.geonames.org/2950159) (52.52437, 13.41053). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **8,767.1 km** | | Vacuum RTT floor | **58.49 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **85.85 ms** | | Low-latency fiber material floor | **85.51 ms** | | Engineering floor (5% path allowance) | **90.16 ms** | | Research 1.33× mapped-fiber reference | **114.19 ms** | | Estimated unamplified path loss | **1841.1 dB** | | Transparent optical spans / inline amplifiers | **116 / 115** | | Published RTT inflation over fiber floor | **1.69×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **138.49 ms** | | Average RTT | **145.1 ms** | | Maximum RTT | **164.88 ms** | | Standard deviation | **6.67 ms** | | Stdev / average | **4.6%** | ## Route Context * Departure PoP: [Hong Kong (HKG)](/docs/network/latency/hkg-hong-kong) * Destination PoP: [Berlin (BER)](/docs/network/latency/ber-berlin) * Region overview: [Asia Pacific](/docs/network/latency/regions/asia-pacific) → [Europe](/docs/network/latency/regions/europe) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Berlin to Hong Kong latency and RTT](/docs/network/latency/pairs/ber-hkg-rtt): 144.9 ms **Fastest routes departing Hong Kong (HKG)** * [Hong Kong to Taipei latency and RTT](/docs/network/latency/pairs/hkg-tpe-rtt): 14.7 ms * [Hong Kong to Singapore latency and RTT](/docs/network/latency/pairs/hkg-sin-rtt): 31.5 ms * [Hong Kong to Tokyo latency and RTT](/docs/network/latency/pairs/hkg-tyo-rtt): 44.9 ms * [Hong Kong to Moscow latency and RTT](/docs/network/latency/pairs/hkg-mow-rtt): 118.2 ms * [Hong Kong to Seattle latency and RTT](/docs/network/latency/pairs/hkg-sea-rtt): 131.6 ms **Fastest routes arriving at Berlin (BER)** * [Frankfurt to Berlin latency and RTT](/docs/network/latency/pairs/fra-ber-rtt): 6.1 ms * [Amsterdam to Berlin latency and RTT](/docs/network/latency/pairs/ams-ber-rtt): 7.2 ms * [Paris to Berlin latency and RTT](/docs/network/latency/pairs/par-ber-rtt): 15.5 ms * [London to Berlin latency and RTT](/docs/network/latency/pairs/lon-ber-rtt): 16.2 ms * [Marseille to Berlin latency and RTT](/docs/network/latency/pairs/mrs-ber-rtt): 20.6 ms **Same corridor (Asia Pacific → Europe)** * [Taipei to Moscow latency and RTT](/docs/network/latency/pairs/tpe-mow-rtt): 132.7 ms * [Singapore to Marseille latency and RTT](/docs/network/latency/pairs/sin-mrs-rtt): 140.6 ms * [Singapore to Moscow latency and RTT](/docs/network/latency/pairs/sin-mow-rtt): 147.7 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/hkg-ber.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Hong Kong → Berlin round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/hkg-ber-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Hong Kong → Frankfurt, Germany RTT 🇭🇰 **Hong Kong (HKG)** → 🇩🇪 **Frankfurt, Germany (FRA)** Measurement round: `2026-08-16T04:07:28Z`. Hong Kong to Frankfurt averaged 150.7 ms across 50 replies. The observed range was 147.02 to 163.83 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 3.53 ms, or 2.3% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------------------- | | RTT | **150.7 ms** | | Jitter | **2.66 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **89.87 ms** | | Fiber Efficiency | **59.6%** | | Latency Tier | Fair | | Source Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | | Destination Region | [Europe](/docs/network/latency/regions/europe) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Hong Kong](https://www.geonames.org/1819729) (22.27832, 114.17469) and [Frankfurt am Main](https://www.geonames.org/2925533) (50.11552, 8.68417). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **9,177.1 km** | | Vacuum RTT floor | **61.22 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **89.87 ms** | | Low-latency fiber material floor | **89.51 ms** | | Engineering floor (5% path allowance) | **94.37 ms** | | Research 1.33× mapped-fiber reference | **119.53 ms** | | Estimated unamplified path loss | **1927.2 dB** | | Transparent optical spans / inline amplifiers | **121 / 120** | | Published RTT inflation over fiber floor | **1.68×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **147.02 ms** | | Average RTT | **150.7 ms** | | Maximum RTT | **163.83 ms** | | Standard deviation | **3.53 ms** | | Stdev / average | **2.3%** | ## Route Context * Departure PoP: [Hong Kong (HKG)](/docs/network/latency/hkg-hong-kong) * Destination PoP: [Frankfurt (FRA)](/docs/network/latency/fra-frankfurt) * Region overview: [Asia Pacific](/docs/network/latency/regions/asia-pacific) → [Europe](/docs/network/latency/regions/europe) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Frankfurt to Hong Kong latency and RTT](/docs/network/latency/pairs/fra-hkg-rtt): 152 ms **Fastest routes departing Hong Kong (HKG)** * [Hong Kong to Taipei latency and RTT](/docs/network/latency/pairs/hkg-tpe-rtt): 14.7 ms * [Hong Kong to Singapore latency and RTT](/docs/network/latency/pairs/hkg-sin-rtt): 31.5 ms * [Hong Kong to Tokyo latency and RTT](/docs/network/latency/pairs/hkg-tyo-rtt): 44.9 ms * [Hong Kong to Moscow latency and RTT](/docs/network/latency/pairs/hkg-mow-rtt): 118.2 ms * [Hong Kong to Seattle latency and RTT](/docs/network/latency/pairs/hkg-sea-rtt): 131.6 ms **Fastest routes arriving at Frankfurt (FRA)** * [Amsterdam to Frankfurt latency and RTT](/docs/network/latency/pairs/ams-fra-rtt): 5.9 ms * [Berlin to Frankfurt latency and RTT](/docs/network/latency/pairs/ber-fra-rtt): 6.1 ms * [Paris to Frankfurt latency and RTT](/docs/network/latency/pairs/par-fra-rtt): 7.6 ms * [London to Frankfurt latency and RTT](/docs/network/latency/pairs/lon-fra-rtt): 13.6 ms * [Marseille to Frankfurt latency and RTT](/docs/network/latency/pairs/mrs-fra-rtt): 16 ms **Same corridor (Asia Pacific → Europe)** * [Taipei to Moscow latency and RTT](/docs/network/latency/pairs/tpe-mow-rtt): 132.7 ms * [Singapore to Marseille latency and RTT](/docs/network/latency/pairs/sin-mrs-rtt): 140.6 ms * [Hong Kong to Berlin latency and RTT](/docs/network/latency/pairs/hkg-ber-rtt): 145.1 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/hkg-fra.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Hong Kong → Frankfurt round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/hkg-fra-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Hong Kong → São Paulo, Brazil RTT 🇭🇰 **Hong Kong (HKG)** → 🇧🇷 **São Paulo, Brazil (GRU)** Measurement round: `2026-08-16T04:07:28Z`. Hong Kong to São Paulo averaged 287.6 ms across 50 replies. The observed range was 276.79 to 338.46 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 13.48 ms, or 4.7% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **287.6 ms** | | Jitter | **8.34 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **176.86 ms** | | Fiber Efficiency | **61.5%** | | Latency Tier | High | | Source Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | | Destination Region | [South America](/docs/network/latency/regions/south-america) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Hong Kong](https://www.geonames.org/1819729) (22.27832, 114.17469) and [São Paulo](https://www.geonames.org/3448439) (-23.54750, -46.63611). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | ------------- | | WGS-84 geodesic distance | **18,060 km** | | Vacuum RTT floor | **120.48 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **176.86 ms** | | Low-latency fiber material floor | **176.15 ms** | | Engineering floor (5% path allowance) | **185.72 ms** | | Research 1.33× mapped-fiber reference | **235.22 ms** | | Estimated unamplified path loss | **3792.6 dB** | | Transparent optical spans / inline amplifiers | **238 / 237** | | Published RTT inflation over fiber floor | **1.63×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **276.79 ms** | | Average RTT | **287.6 ms** | | Maximum RTT | **338.46 ms** | | Standard deviation | **13.48 ms** | | Stdev / average | **4.7%** | ## Route Context * Departure PoP: [Hong Kong (HKG)](/docs/network/latency/hkg-hong-kong) * Destination PoP: [São Paulo (GRU)](/docs/network/latency/gru-sao-paulo) * Region overview: [Asia Pacific](/docs/network/latency/regions/asia-pacific) → [South America](/docs/network/latency/regions/south-america) * Corridor overview: [Asia Pacific to South America](/docs/network/latency/regions/asia-pacific-to-south-america) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [São Paulo to Hong Kong latency and RTT](/docs/network/latency/pairs/gru-hkg-rtt): 273.5 ms **Fastest routes departing Hong Kong (HKG)** * [Hong Kong to Taipei latency and RTT](/docs/network/latency/pairs/hkg-tpe-rtt): 14.7 ms * [Hong Kong to Singapore latency and RTT](/docs/network/latency/pairs/hkg-sin-rtt): 31.5 ms * [Hong Kong to Tokyo latency and RTT](/docs/network/latency/pairs/hkg-tyo-rtt): 44.9 ms * [Hong Kong to Moscow latency and RTT](/docs/network/latency/pairs/hkg-mow-rtt): 118.2 ms * [Hong Kong to Seattle latency and RTT](/docs/network/latency/pairs/hkg-sea-rtt): 131.6 ms **Fastest routes arriving at São Paulo (GRU)** * [Ashburn to São Paulo latency and RTT](/docs/network/latency/pairs/iad-gru-rtt): 101.4 ms * [New York to São Paulo latency and RTT](/docs/network/latency/pairs/nyc-gru-rtt): 106.9 ms * [Miami to São Paulo latency and RTT](/docs/network/latency/pairs/mia-gru-rtt): 128.7 ms * [Los Angeles to São Paulo latency and RTT](/docs/network/latency/pairs/lax-gru-rtt): 131.3 ms * [Seattle to São Paulo latency and RTT](/docs/network/latency/pairs/sea-gru-rtt): 155.7 ms **Same corridor (Asia Pacific → South America)** * [Tokyo to São Paulo latency and RTT](/docs/network/latency/pairs/tyo-gru-rtt): 230.5 ms * [Taipei to São Paulo latency and RTT](/docs/network/latency/pairs/tpe-gru-rtt): 273 ms * [Sydney to São Paulo latency and RTT](/docs/network/latency/pairs/syd-gru-rtt): 296.7 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/hkg-gru.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Hong Kong → São Paulo round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/hkg-gru-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Hong Kong → Ashburn, USA RTT 🇭🇰 **Hong Kong (HKG)** → 🇺🇸 **Ashburn, USA (IAD)** Measurement round: `2026-08-16T04:07:28Z`. Hong Kong to Ashburn averaged 186.2 ms across 50 replies. The observed range was 178.2 to 209.61 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 8.04 ms, or 4.3% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **186.2 ms** | | Jitter | **5.28 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **128.37 ms** | | Fiber Efficiency | **68.9%** | | Latency Tier | Fair | | Source Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | | Destination Region | [North America](/docs/network/latency/regions/north-america) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Hong Kong](https://www.geonames.org/1819729) (22.27832, 114.17469) and [Ashburn](https://www.geonames.org/4744870) (39.04372, -77.48749). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | --------------- | | WGS-84 geodesic distance | **13,108.8 km** | | Vacuum RTT floor | **87.45 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **128.37 ms** | | Low-latency fiber material floor | **127.86 ms** | | Engineering floor (5% path allowance) | **134.81 ms** | | Research 1.33× mapped-fiber reference | **170.73 ms** | | Estimated unamplified path loss | **2752.9 dB** | | Transparent optical spans / inline amplifiers | **173 / 172** | | Published RTT inflation over fiber floor | **1.45×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **178.2 ms** | | Average RTT | **186.2 ms** | | Maximum RTT | **209.61 ms** | | Standard deviation | **8.04 ms** | | Stdev / average | **4.3%** | ## Route Context * Departure PoP: [Hong Kong (HKG)](/docs/network/latency/hkg-hong-kong) * Destination PoP: [Ashburn (IAD)](/docs/network/latency/iad-ashburn) * Region overview: [Asia Pacific](/docs/network/latency/regions/asia-pacific) → [North America](/docs/network/latency/regions/north-america) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Ashburn to Hong Kong latency and RTT](/docs/network/latency/pairs/iad-hkg-rtt): 187.2 ms **Fastest routes departing Hong Kong (HKG)** * [Hong Kong to Taipei latency and RTT](/docs/network/latency/pairs/hkg-tpe-rtt): 14.7 ms * [Hong Kong to Singapore latency and RTT](/docs/network/latency/pairs/hkg-sin-rtt): 31.5 ms * [Hong Kong to Tokyo latency and RTT](/docs/network/latency/pairs/hkg-tyo-rtt): 44.9 ms * [Hong Kong to Moscow latency and RTT](/docs/network/latency/pairs/hkg-mow-rtt): 118.2 ms * [Hong Kong to Seattle latency and RTT](/docs/network/latency/pairs/hkg-sea-rtt): 131.6 ms **Fastest routes arriving at Ashburn (IAD)** * [New York to Ashburn latency and RTT](/docs/network/latency/pairs/nyc-iad-rtt): 6 ms * [Miami to Ashburn latency and RTT](/docs/network/latency/pairs/mia-iad-rtt): 27.3 ms * [Los Angeles to Ashburn latency and RTT](/docs/network/latency/pairs/lax-iad-rtt): 60.4 ms * [Seattle to Ashburn latency and RTT](/docs/network/latency/pairs/sea-iad-rtt): 61.5 ms * [London to Ashburn latency and RTT](/docs/network/latency/pairs/lon-iad-rtt): 71 ms **Same corridor (Asia Pacific → North America)** * [Tokyo to Seattle latency and RTT](/docs/network/latency/pairs/tyo-sea-rtt): 84.8 ms * [Tokyo to Los Angeles latency and RTT](/docs/network/latency/pairs/tyo-lax-rtt): 101.2 ms * [Taipei to Seattle latency and RTT](/docs/network/latency/pairs/tpe-sea-rtt): 115.8 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/hkg-iad.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Hong Kong → Ashburn round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/hkg-iad-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Hong Kong → Johannesburg, South Africa RTT 🇭🇰 **Hong Kong (HKG)** → 🇿🇦 **Johannesburg, South Africa (JNB)** Measurement round: `2026-08-16T04:07:28Z`. Hong Kong to Johannesburg averaged 316.7 ms, compared with a 104.93 ms city-centre fiber reference. The difference cannot be assigned to a particular cable or carrier from RTT alone. Standard deviation was 11.99 ms, or 3.8% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------------------- | | RTT | **316.7 ms** | | Jitter | **10.88 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **104.93 ms** | | Fiber Efficiency | **33.1%** | | Latency Tier | High | | Source Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | | Destination Region | [Africa](/docs/network/latency/regions/africa) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Hong Kong](https://www.geonames.org/1819729) (22.27832, 114.17469) and [Johannesburg](https://www.geonames.org/993800) (-26.20227, 28.04363). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | --------------- | | WGS-84 geodesic distance | **10,715.5 km** | | Vacuum RTT floor | **71.49 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **104.93 ms** | | Low-latency fiber material floor | **104.51 ms** | | Engineering floor (5% path allowance) | **110.2 ms** | | Research 1.33× mapped-fiber reference | **139.56 ms** | | Estimated unamplified path loss | **2250.3 dB** | | Transparent optical spans / inline amplifiers | **141 / 140** | | Published RTT inflation over fiber floor | **3.02×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **301.28 ms** | | Average RTT | **316.7 ms** | | Maximum RTT | **350.25 ms** | | Standard deviation | **11.99 ms** | | Stdev / average | **3.8%** | ## Route Context * Departure PoP: [Hong Kong (HKG)](/docs/network/latency/hkg-hong-kong) * Destination PoP: [Johannesburg (JNB)](/docs/network/latency/jnb-johannesburg) * Region overview: [Asia Pacific](/docs/network/latency/regions/asia-pacific) → [Africa](/docs/network/latency/regions/africa) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Johannesburg to Hong Kong latency and RTT](/docs/network/latency/pairs/jnb-hkg-rtt): 318 ms **Fastest routes departing Hong Kong (HKG)** * [Hong Kong to Taipei latency and RTT](/docs/network/latency/pairs/hkg-tpe-rtt): 14.7 ms * [Hong Kong to Singapore latency and RTT](/docs/network/latency/pairs/hkg-sin-rtt): 31.5 ms * [Hong Kong to Tokyo latency and RTT](/docs/network/latency/pairs/hkg-tyo-rtt): 44.9 ms * [Hong Kong to Moscow latency and RTT](/docs/network/latency/pairs/hkg-mow-rtt): 118.2 ms * [Hong Kong to Seattle latency and RTT](/docs/network/latency/pairs/hkg-sea-rtt): 131.6 ms **Fastest routes arriving at Johannesburg (JNB)** * [Amsterdam to Johannesburg latency and RTT](/docs/network/latency/pairs/ams-jnb-rtt): 171.9 ms * [Berlin to Johannesburg latency and RTT](/docs/network/latency/pairs/ber-jnb-rtt): 172.1 ms * [London to Johannesburg latency and RTT](/docs/network/latency/pairs/lon-jnb-rtt): 179.6 ms * [Paris to Johannesburg latency and RTT](/docs/network/latency/pairs/par-jnb-rtt): 180.9 ms * [Frankfurt to Johannesburg latency and RTT](/docs/network/latency/pairs/fra-jnb-rtt): 188.4 ms **Same corridor (Asia Pacific → Africa)** * [Singapore to Johannesburg latency and RTT](/docs/network/latency/pairs/sin-jnb-rtt): 312.6 ms * [Taipei to Johannesburg latency and RTT](/docs/network/latency/pairs/tpe-jnb-rtt): 331 ms * [Tokyo to Johannesburg latency and RTT](/docs/network/latency/pairs/tyo-jnb-rtt): 359.2 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/hkg-jnb.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Hong Kong → Johannesburg round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/hkg-jnb-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Hong Kong → Los Angeles, USA RTT 🇭🇰 **Hong Kong (HKG)** → 🇺🇸 **Los Angeles, USA (LAX)** Measurement round: `2026-08-16T04:07:28Z`. Hong Kong to Los Angeles averaged 145 ms across 50 replies. The observed range was 139.67 to 160.38 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 4.08 ms, or 2.8% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **145 ms** | | Jitter | **3.79 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **114.3 ms** | | Fiber Efficiency | **78.8%** | | Latency Tier | Good | | Source Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | | Destination Region | [North America](/docs/network/latency/regions/north-america) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Hong Kong](https://www.geonames.org/1819729) (22.27832, 114.17469) and [Los Angeles](https://www.geonames.org/5368361) (34.05223, -118.24368). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | --------------- | | WGS-84 geodesic distance | **11,671.5 km** | | Vacuum RTT floor | **77.86 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **114.3 ms** | | Low-latency fiber material floor | **113.84 ms** | | Engineering floor (5% path allowance) | **120.03 ms** | | Research 1.33× mapped-fiber reference | **152.01 ms** | | Estimated unamplified path loss | **2451 dB** | | Transparent optical spans / inline amplifiers | **154 / 153** | | Published RTT inflation over fiber floor | **1.27×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **139.67 ms** | | Average RTT | **145 ms** | | Maximum RTT | **160.38 ms** | | Standard deviation | **4.08 ms** | | Stdev / average | **2.8%** | ## Route Context * Departure PoP: [Hong Kong (HKG)](/docs/network/latency/hkg-hong-kong) * Destination PoP: [Los Angeles (LAX)](/docs/network/latency/lax-los-angeles) * Region overview: [Asia Pacific](/docs/network/latency/regions/asia-pacific) → [North America](/docs/network/latency/regions/north-america) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Los Angeles to Hong Kong latency and RTT](/docs/network/latency/pairs/lax-hkg-rtt): 146.5 ms **Fastest routes departing Hong Kong (HKG)** * [Hong Kong to Taipei latency and RTT](/docs/network/latency/pairs/hkg-tpe-rtt): 14.7 ms * [Hong Kong to Singapore latency and RTT](/docs/network/latency/pairs/hkg-sin-rtt): 31.5 ms * [Hong Kong to Tokyo latency and RTT](/docs/network/latency/pairs/hkg-tyo-rtt): 44.9 ms * [Hong Kong to Moscow latency and RTT](/docs/network/latency/pairs/hkg-mow-rtt): 118.2 ms * [Hong Kong to Seattle latency and RTT](/docs/network/latency/pairs/hkg-sea-rtt): 131.6 ms **Fastest routes arriving at Los Angeles (LAX)** * [Seattle to Los Angeles latency and RTT](/docs/network/latency/pairs/sea-lax-rtt): 25.9 ms * [Miami to Los Angeles latency and RTT](/docs/network/latency/pairs/mia-lax-rtt): 56.8 ms * [New York to Los Angeles latency and RTT](/docs/network/latency/pairs/nyc-lax-rtt): 58.9 ms * [Ashburn to Los Angeles latency and RTT](/docs/network/latency/pairs/iad-lax-rtt): 60.7 ms * [Tokyo to Los Angeles latency and RTT](/docs/network/latency/pairs/tyo-lax-rtt): 101.2 ms **Same corridor (Asia Pacific → North America)** * [Tokyo to Seattle latency and RTT](/docs/network/latency/pairs/tyo-sea-rtt): 84.8 ms * [Taipei to Seattle latency and RTT](/docs/network/latency/pairs/tpe-sea-rtt): 115.8 ms * [Taipei to Los Angeles latency and RTT](/docs/network/latency/pairs/tpe-lax-rtt): 132 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/hkg-lax.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Hong Kong → Los Angeles round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/hkg-lax-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Hong Kong → London, UK RTT 🇭🇰 **Hong Kong (HKG)** → 🇬🇧 **London, UK (LON)** Measurement round: `2026-08-16T04:07:28Z`. Hong Kong to London averaged 158.7 ms across 50 replies. The observed range was 154.18 to 172.16 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 4.34 ms, or 2.7% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------------------- | | RTT | **158.7 ms** | | Jitter | **3.3 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **94.44 ms** | | Fiber Efficiency | **59.5%** | | Latency Tier | Fair | | Source Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | | Destination Region | [Europe](/docs/network/latency/regions/europe) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Hong Kong](https://www.geonames.org/1819729) (22.27832, 114.17469) and [London](https://www.geonames.org/2643743) (51.50853, -0.12574). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **9,643.6 km** | | Vacuum RTT floor | **64.34 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **94.44 ms** | | Low-latency fiber material floor | **94.06 ms** | | Engineering floor (5% path allowance) | **99.17 ms** | | Research 1.33× mapped-fiber reference | **125.6 ms** | | Estimated unamplified path loss | **2025.2 dB** | | Transparent optical spans / inline amplifiers | **127 / 126** | | Published RTT inflation over fiber floor | **1.68×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **154.18 ms** | | Average RTT | **158.7 ms** | | Maximum RTT | **172.16 ms** | | Standard deviation | **4.34 ms** | | Stdev / average | **2.7%** | ## Route Context * Departure PoP: [Hong Kong (HKG)](/docs/network/latency/hkg-hong-kong) * Destination PoP: [London (LON)](/docs/network/latency/lon-london) * Region overview: [Asia Pacific](/docs/network/latency/regions/asia-pacific) → [Europe](/docs/network/latency/regions/europe) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [London to Hong Kong latency and RTT](/docs/network/latency/pairs/lon-hkg-rtt): 160.6 ms **Fastest routes departing Hong Kong (HKG)** * [Hong Kong to Taipei latency and RTT](/docs/network/latency/pairs/hkg-tpe-rtt): 14.7 ms * [Hong Kong to Singapore latency and RTT](/docs/network/latency/pairs/hkg-sin-rtt): 31.5 ms * [Hong Kong to Tokyo latency and RTT](/docs/network/latency/pairs/hkg-tyo-rtt): 44.9 ms * [Hong Kong to Moscow latency and RTT](/docs/network/latency/pairs/hkg-mow-rtt): 118.2 ms * [Hong Kong to Seattle latency and RTT](/docs/network/latency/pairs/hkg-sea-rtt): 131.6 ms **Fastest routes arriving at London (LON)** * [Amsterdam to London latency and RTT](/docs/network/latency/pairs/ams-lon-rtt): 5.2 ms * [Paris to London latency and RTT](/docs/network/latency/pairs/par-lon-rtt): 6.4 ms * [Frankfurt to London latency and RTT](/docs/network/latency/pairs/fra-lon-rtt): 12.9 ms * [Berlin to London latency and RTT](/docs/network/latency/pairs/ber-lon-rtt): 15.2 ms * [Marseille to London latency and RTT](/docs/network/latency/pairs/mrs-lon-rtt): 18.9 ms **Same corridor (Asia Pacific → Europe)** * [Taipei to Moscow latency and RTT](/docs/network/latency/pairs/tpe-mow-rtt): 132.7 ms * [Singapore to Marseille latency and RTT](/docs/network/latency/pairs/sin-mrs-rtt): 140.6 ms * [Hong Kong to Berlin latency and RTT](/docs/network/latency/pairs/hkg-ber-rtt): 145.1 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/hkg-lon.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Hong Kong → London round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/hkg-lon-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Hong Kong → Melbourne, Australia RTT 🇭🇰 **Hong Kong (HKG)** → 🇦🇺 **Melbourne, Australia (MEL)** Measurement round: `2026-08-16T04:07:28Z`. Hong Kong to Melbourne averaged 138.4 ms across 50 replies. The observed range was 132.66 to 153.75 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 4.9 ms, or 3.5% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------------------- | | RTT | **138.4 ms** | | Jitter | **4.62 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **72.38 ms** | | Fiber Efficiency | **52.3%** | | Latency Tier | Good | | Source Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | | Destination Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Hong Kong](https://www.geonames.org/1819729) (22.27832, 114.17469) and [Melbourne](https://www.geonames.org/2158177) (-37.81400, 144.96332). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **7,391.1 km** | | Vacuum RTT floor | **49.31 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **72.38 ms** | | Low-latency fiber material floor | **72.09 ms** | | Engineering floor (5% path allowance) | **76.01 ms** | | Research 1.33× mapped-fiber reference | **96.26 ms** | | Estimated unamplified path loss | **1552.1 dB** | | Transparent optical spans / inline amplifiers | **98 / 97** | | Published RTT inflation over fiber floor | **1.91×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **132.66 ms** | | Average RTT | **138.4 ms** | | Maximum RTT | **153.75 ms** | | Standard deviation | **4.9 ms** | | Stdev / average | **3.5%** | ## Route Context * Departure PoP: [Hong Kong (HKG)](/docs/network/latency/hkg-hong-kong) * Destination PoP: [Melbourne (MEL)](/docs/network/latency/mel-melbourne) * Region overview: [Asia Pacific](/docs/network/latency/regions/asia-pacific) → [Asia Pacific](/docs/network/latency/regions/asia-pacific) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Melbourne to Hong Kong latency and RTT](/docs/network/latency/pairs/mel-hkg-rtt): 138.5 ms **Fastest routes departing Hong Kong (HKG)** * [Hong Kong to Taipei latency and RTT](/docs/network/latency/pairs/hkg-tpe-rtt): 14.7 ms * [Hong Kong to Singapore latency and RTT](/docs/network/latency/pairs/hkg-sin-rtt): 31.5 ms * [Hong Kong to Tokyo latency and RTT](/docs/network/latency/pairs/hkg-tyo-rtt): 44.9 ms * [Hong Kong to Moscow latency and RTT](/docs/network/latency/pairs/hkg-mow-rtt): 118.2 ms * [Hong Kong to Seattle latency and RTT](/docs/network/latency/pairs/hkg-sea-rtt): 131.6 ms **Fastest routes arriving at Melbourne (MEL)** * [Sydney to Melbourne latency and RTT](/docs/network/latency/pairs/syd-mel-rtt): 9.8 ms * [Singapore to Melbourne latency and RTT](/docs/network/latency/pairs/sin-mel-rtt): 88.6 ms * [Tokyo to Melbourne latency and RTT](/docs/network/latency/pairs/tyo-mel-rtt): 113.9 ms * [Taipei to Melbourne latency and RTT](/docs/network/latency/pairs/tpe-mel-rtt): 142 ms * [Los Angeles to Melbourne latency and RTT](/docs/network/latency/pairs/lax-mel-rtt): 148.2 ms **Same corridor (Asia Pacific → Asia Pacific)** * [Melbourne to Sydney latency and RTT](/docs/network/latency/pairs/mel-syd-rtt): 9.8 ms * [Taipei to Hong Kong latency and RTT](/docs/network/latency/pairs/tpe-hkg-rtt): 15.5 ms * [Singapore to Hong Kong latency and RTT](/docs/network/latency/pairs/sin-hkg-rtt): 30.8 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/hkg-mel.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Hong Kong → Melbourne round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/hkg-mel-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Hong Kong → Miami, USA RTT 🇭🇰 **Hong Kong (HKG)** → 🇺🇸 **Miami, USA (MIA)** Measurement round: `2026-08-16T04:07:28Z`. Hong Kong to Miami averaged 201.1 ms across 50 replies. The observed range was 194.49 to 226.67 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 6.05 ms, or 3.0% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **201.1 ms** | | Jitter | **4.97 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **141.69 ms** | | Fiber Efficiency | **70.5%** | | Latency Tier | Fair | | Source Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | | Destination Region | [North America](/docs/network/latency/regions/north-america) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Hong Kong](https://www.geonames.org/1819729) (22.27832, 114.17469) and [Miami](https://www.geonames.org/4164138) (25.77427, -80.19366). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | --------------- | | WGS-84 geodesic distance | **14,468.4 km** | | Vacuum RTT floor | **96.52 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **141.69 ms** | | Low-latency fiber material floor | **141.12 ms** | | Engineering floor (5% path allowance) | **148.79 ms** | | Research 1.33× mapped-fiber reference | **188.44 ms** | | Estimated unamplified path loss | **3038.4 dB** | | Transparent optical spans / inline amplifiers | **190 / 189** | | Published RTT inflation over fiber floor | **1.42×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **194.49 ms** | | Average RTT | **201.1 ms** | | Maximum RTT | **226.67 ms** | | Standard deviation | **6.05 ms** | | Stdev / average | **3.0%** | ## Route Context * Departure PoP: [Hong Kong (HKG)](/docs/network/latency/hkg-hong-kong) * Destination PoP: [Miami (MIA)](/docs/network/latency/mia-miami) * Region overview: [Asia Pacific](/docs/network/latency/regions/asia-pacific) → [North America](/docs/network/latency/regions/north-america) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Miami to Hong Kong latency and RTT](/docs/network/latency/pairs/mia-hkg-rtt): 199.5 ms **Fastest routes departing Hong Kong (HKG)** * [Hong Kong to Taipei latency and RTT](/docs/network/latency/pairs/hkg-tpe-rtt): 14.7 ms * [Hong Kong to Singapore latency and RTT](/docs/network/latency/pairs/hkg-sin-rtt): 31.5 ms * [Hong Kong to Tokyo latency and RTT](/docs/network/latency/pairs/hkg-tyo-rtt): 44.9 ms * [Hong Kong to Moscow latency and RTT](/docs/network/latency/pairs/hkg-mow-rtt): 118.2 ms * [Hong Kong to Seattle latency and RTT](/docs/network/latency/pairs/hkg-sea-rtt): 131.6 ms **Fastest routes arriving at Miami (MIA)** * [Ashburn to Miami latency and RTT](/docs/network/latency/pairs/iad-mia-rtt): 27.4 ms * [New York to Miami latency and RTT](/docs/network/latency/pairs/nyc-mia-rtt): 32.9 ms * [Los Angeles to Miami latency and RTT](/docs/network/latency/pairs/lax-mia-rtt): 57.3 ms * [São Paulo to Miami latency and RTT](/docs/network/latency/pairs/gru-mia-rtt): 74 ms * [Seattle to Miami latency and RTT](/docs/network/latency/pairs/sea-mia-rtt): 81.7 ms **Same corridor (Asia Pacific → North America)** * [Tokyo to Seattle latency and RTT](/docs/network/latency/pairs/tyo-sea-rtt): 84.8 ms * [Tokyo to Los Angeles latency and RTT](/docs/network/latency/pairs/tyo-lax-rtt): 101.2 ms * [Taipei to Seattle latency and RTT](/docs/network/latency/pairs/tpe-sea-rtt): 115.8 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/hkg-mia.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Hong Kong → Miami round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/hkg-mia-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Hong Kong → Moscow, Russia RTT 🇭🇰 **Hong Kong (HKG)** → 🇷🇺 **Moscow, Russia (MOW)** Measurement round: `2026-08-16T04:07:28Z`. Hong Kong to Moscow averaged 118.2 ms across 50 replies. The observed range was 115.72 to 128.25 ms; these figures describe one probe round, not a long-term service target. This is number 4 of 19 outgoing routes by average RTT. That ranking compares different destinations and is not adjusted for distance. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------------------- | | RTT | **118.2 ms** | | Jitter | **1.78 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **70.05 ms** | | Fiber Efficiency | **59.3%** | | Latency Tier | Good | | Source Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | | Destination Region | [Europe](/docs/network/latency/regions/europe) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Hong Kong](https://www.geonames.org/1819729) (22.27832, 114.17469) and [Moscow](https://www.geonames.org/524901) (55.75204, 37.61781). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **7,153.6 km** | | Vacuum RTT floor | **47.72 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **70.05 ms** | | Low-latency fiber material floor | **69.77 ms** | | Engineering floor (5% path allowance) | **73.57 ms** | | Research 1.33× mapped-fiber reference | **93.17 ms** | | Estimated unamplified path loss | **1502.3 dB** | | Transparent optical spans / inline amplifiers | **94 / 93** | | Published RTT inflation over fiber floor | **1.69×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **115.72 ms** | | Average RTT | **118.2 ms** | | Maximum RTT | **128.25 ms** | | Standard deviation | **2.23 ms** | | Stdev / average | **1.9%** | ## Route Context * Departure PoP: [Hong Kong (HKG)](/docs/network/latency/hkg-hong-kong) * Destination PoP: [Moscow (MOW)](/docs/network/latency/mow-moscow) * Region overview: [Asia Pacific](/docs/network/latency/regions/asia-pacific) → [Europe](/docs/network/latency/regions/europe) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Moscow to Hong Kong latency and RTT](/docs/network/latency/pairs/mow-hkg-rtt): 118.6 ms **Fastest routes departing Hong Kong (HKG)** * [Hong Kong to Taipei latency and RTT](/docs/network/latency/pairs/hkg-tpe-rtt): 14.7 ms * [Hong Kong to Singapore latency and RTT](/docs/network/latency/pairs/hkg-sin-rtt): 31.5 ms * [Hong Kong to Tokyo latency and RTT](/docs/network/latency/pairs/hkg-tyo-rtt): 44.9 ms * [Hong Kong to Seattle latency and RTT](/docs/network/latency/pairs/hkg-sea-rtt): 131.6 ms * [Hong Kong to Sydney latency and RTT](/docs/network/latency/pairs/hkg-syd-rtt): 137.2 ms **Fastest routes arriving at Moscow (MOW)** * [Berlin to Moscow latency and RTT](/docs/network/latency/pairs/ber-mow-rtt): 28.6 ms * [Frankfurt to Moscow latency and RTT](/docs/network/latency/pairs/fra-mow-rtt): 34.7 ms * [Amsterdam to Moscow latency and RTT](/docs/network/latency/pairs/ams-mow-rtt): 38.1 ms * [London to Moscow latency and RTT](/docs/network/latency/pairs/lon-mow-rtt): 43.5 ms * [Paris to Moscow latency and RTT](/docs/network/latency/pairs/par-mow-rtt): 44.7 ms **Same corridor (Asia Pacific → Europe)** * [Taipei to Moscow latency and RTT](/docs/network/latency/pairs/tpe-mow-rtt): 132.7 ms * [Singapore to Marseille latency and RTT](/docs/network/latency/pairs/sin-mrs-rtt): 140.6 ms * [Hong Kong to Berlin latency and RTT](/docs/network/latency/pairs/hkg-ber-rtt): 145.1 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/hkg-mow.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Hong Kong → Moscow round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/hkg-mow-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Hong Kong → Marseille, France RTT 🇭🇰 **Hong Kong (HKG)** → 🇫🇷 **Marseille, France (MRS)** Measurement round: `2026-08-16T04:07:28Z`. Hong Kong to Marseille averaged 165.4 ms across 50 replies. The observed range was 159.87 to 180.73 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 4.35 ms, or 2.6% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------------------- | | RTT | **165.4 ms** | | Jitter | **3.54 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **95.48 ms** | | Fiber Efficiency | **57.7%** | | Latency Tier | Fair | | Source Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | | Destination Region | [Europe](/docs/network/latency/regions/europe) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Hong Kong](https://www.geonames.org/1819729) (22.27832, 114.17469) and [Marseille](https://www.geonames.org/2995469) (43.29695, 5.38107). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **9,750.5 km** | | Vacuum RTT floor | **65.05 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **95.48 ms** | | Low-latency fiber material floor | **95.1 ms** | | Engineering floor (5% path allowance) | **100.27 ms** | | Research 1.33× mapped-fiber reference | **126.99 ms** | | Estimated unamplified path loss | **2047.6 dB** | | Transparent optical spans / inline amplifiers | **128 / 127** | | Published RTT inflation over fiber floor | **1.73×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **159.87 ms** | | Average RTT | **165.4 ms** | | Maximum RTT | **180.73 ms** | | Standard deviation | **4.35 ms** | | Stdev / average | **2.6%** | ## Route Context * Departure PoP: [Hong Kong (HKG)](/docs/network/latency/hkg-hong-kong) * Destination PoP: [Marseille (MRS)](/docs/network/latency/mrs-marseille) * Region overview: [Asia Pacific](/docs/network/latency/regions/asia-pacific) → [Europe](/docs/network/latency/regions/europe) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Marseille to Hong Kong latency and RTT](/docs/network/latency/pairs/mrs-hkg-rtt): 166.4 ms **Fastest routes departing Hong Kong (HKG)** * [Hong Kong to Taipei latency and RTT](/docs/network/latency/pairs/hkg-tpe-rtt): 14.7 ms * [Hong Kong to Singapore latency and RTT](/docs/network/latency/pairs/hkg-sin-rtt): 31.5 ms * [Hong Kong to Tokyo latency and RTT](/docs/network/latency/pairs/hkg-tyo-rtt): 44.9 ms * [Hong Kong to Moscow latency and RTT](/docs/network/latency/pairs/hkg-mow-rtt): 118.2 ms * [Hong Kong to Seattle latency and RTT](/docs/network/latency/pairs/hkg-sea-rtt): 131.6 ms **Fastest routes arriving at Marseille (MRS)** * [Paris to Marseille latency and RTT](/docs/network/latency/pairs/par-mrs-rtt): 8.9 ms * [Frankfurt to Marseille latency and RTT](/docs/network/latency/pairs/fra-mrs-rtt): 16.4 ms * [London to Marseille latency and RTT](/docs/network/latency/pairs/lon-mrs-rtt): 17.9 ms * [Amsterdam to Marseille latency and RTT](/docs/network/latency/pairs/ams-mrs-rtt): 19.6 ms * [Berlin to Marseille latency and RTT](/docs/network/latency/pairs/ber-mrs-rtt): 20.2 ms **Same corridor (Asia Pacific → Europe)** * [Taipei to Moscow latency and RTT](/docs/network/latency/pairs/tpe-mow-rtt): 132.7 ms * [Singapore to Marseille latency and RTT](/docs/network/latency/pairs/sin-mrs-rtt): 140.6 ms * [Hong Kong to Berlin latency and RTT](/docs/network/latency/pairs/hkg-ber-rtt): 145.1 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/hkg-mrs.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Hong Kong → Marseille round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/hkg-mrs-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Hong Kong → New York, USA RTT 🇭🇰 **Hong Kong (HKG)** → 🇺🇸 **New York, USA (NYC)** Measurement round: `2026-08-16T04:07:28Z`. Hong Kong to New York averaged 184.7 ms across 50 replies. The observed range was 181.35 to 196.05 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 3.49 ms, or 1.9% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **184.7 ms** | | Jitter | **2.83 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **127.11 ms** | | Fiber Efficiency | **68.8%** | | Latency Tier | Fair | | Source Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | | Destination Region | [North America](/docs/network/latency/regions/north-america) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Hong Kong](https://www.geonames.org/1819729) (22.27832, 114.17469) and [New York City](https://www.geonames.org/5128581) (40.71427, -74.00597). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | --------------- | | WGS-84 geodesic distance | **12,979.6 km** | | Vacuum RTT floor | **86.59 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **127.11 ms** | | Low-latency fiber material floor | **126.6 ms** | | Engineering floor (5% path allowance) | **133.48 ms** | | Research 1.33× mapped-fiber reference | **169.05 ms** | | Estimated unamplified path loss | **2725.7 dB** | | Transparent optical spans / inline amplifiers | **171 / 170** | | Published RTT inflation over fiber floor | **1.45×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **181.35 ms** | | Average RTT | **184.7 ms** | | Maximum RTT | **196.05 ms** | | Standard deviation | **3.49 ms** | | Stdev / average | **1.9%** | ## Route Context * Departure PoP: [Hong Kong (HKG)](/docs/network/latency/hkg-hong-kong) * Destination PoP: [New York (NYC)](/docs/network/latency/nyc-new-york) * Region overview: [Asia Pacific](/docs/network/latency/regions/asia-pacific) → [North America](/docs/network/latency/regions/north-america) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [New York to Hong Kong latency and RTT](/docs/network/latency/pairs/nyc-hkg-rtt): 187.1 ms **Fastest routes departing Hong Kong (HKG)** * [Hong Kong to Taipei latency and RTT](/docs/network/latency/pairs/hkg-tpe-rtt): 14.7 ms * [Hong Kong to Singapore latency and RTT](/docs/network/latency/pairs/hkg-sin-rtt): 31.5 ms * [Hong Kong to Tokyo latency and RTT](/docs/network/latency/pairs/hkg-tyo-rtt): 44.9 ms * [Hong Kong to Moscow latency and RTT](/docs/network/latency/pairs/hkg-mow-rtt): 118.2 ms * [Hong Kong to Seattle latency and RTT](/docs/network/latency/pairs/hkg-sea-rtt): 131.6 ms **Fastest routes arriving at New York (NYC)** * [Ashburn to New York latency and RTT](/docs/network/latency/pairs/iad-nyc-rtt): 6 ms * [Miami to New York latency and RTT](/docs/network/latency/pairs/mia-nyc-rtt): 33.3 ms * [Los Angeles to New York latency and RTT](/docs/network/latency/pairs/lax-nyc-rtt): 57.9 ms * [Seattle to New York latency and RTT](/docs/network/latency/pairs/sea-nyc-rtt): 59.5 ms * [London to New York latency and RTT](/docs/network/latency/pairs/lon-nyc-rtt): 63.8 ms **Same corridor (Asia Pacific → North America)** * [Tokyo to Seattle latency and RTT](/docs/network/latency/pairs/tyo-sea-rtt): 84.8 ms * [Tokyo to Los Angeles latency and RTT](/docs/network/latency/pairs/tyo-lax-rtt): 101.2 ms * [Taipei to Seattle latency and RTT](/docs/network/latency/pairs/tpe-sea-rtt): 115.8 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/hkg-nyc.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Hong Kong → New York round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/hkg-nyc-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Hong Kong → Paris, France RTT 🇭🇰 **Hong Kong (HKG)** → 🇫🇷 **Paris, France (PAR)** Measurement round: `2026-08-16T04:07:28Z`. Hong Kong to Paris averaged 159.4 ms across 50 replies. The observed range was 154.31 to 174.31 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 4.8 ms, or 3.0% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------------------- | | RTT | **159.4 ms** | | Jitter | **4.33 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **94.48 ms** | | Fiber Efficiency | **59.3%** | | Latency Tier | Fair | | Source Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | | Destination Region | [Europe](/docs/network/latency/regions/europe) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Hong Kong](https://www.geonames.org/1819729) (22.27832, 114.17469) and [Paris](https://www.geonames.org/2988507) (48.85341, 2.34880). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **9,647.6 km** | | Vacuum RTT floor | **64.36 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **94.48 ms** | | Low-latency fiber material floor | **94.1 ms** | | Engineering floor (5% path allowance) | **99.21 ms** | | Research 1.33× mapped-fiber reference | **125.65 ms** | | Estimated unamplified path loss | **2026 dB** | | Transparent optical spans / inline amplifiers | **127 / 126** | | Published RTT inflation over fiber floor | **1.69×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **154.31 ms** | | Average RTT | **159.4 ms** | | Maximum RTT | **174.31 ms** | | Standard deviation | **4.8 ms** | | Stdev / average | **3.0%** | ## Route Context * Departure PoP: [Hong Kong (HKG)](/docs/network/latency/hkg-hong-kong) * Destination PoP: [Paris (PAR)](/docs/network/latency/par-paris) * Region overview: [Asia Pacific](/docs/network/latency/regions/asia-pacific) → [Europe](/docs/network/latency/regions/europe) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Paris to Hong Kong latency and RTT](/docs/network/latency/pairs/par-hkg-rtt): 160.1 ms **Fastest routes departing Hong Kong (HKG)** * [Hong Kong to Taipei latency and RTT](/docs/network/latency/pairs/hkg-tpe-rtt): 14.7 ms * [Hong Kong to Singapore latency and RTT](/docs/network/latency/pairs/hkg-sin-rtt): 31.5 ms * [Hong Kong to Tokyo latency and RTT](/docs/network/latency/pairs/hkg-tyo-rtt): 44.9 ms * [Hong Kong to Moscow latency and RTT](/docs/network/latency/pairs/hkg-mow-rtt): 118.2 ms * [Hong Kong to Seattle latency and RTT](/docs/network/latency/pairs/hkg-sea-rtt): 131.6 ms **Fastest routes arriving at Paris (PAR)** * [London to Paris latency and RTT](/docs/network/latency/pairs/lon-par-rtt): 6.4 ms * [Amsterdam to Paris latency and RTT](/docs/network/latency/pairs/ams-par-rtt): 7 ms * [Frankfurt to Paris latency and RTT](/docs/network/latency/pairs/fra-par-rtt): 7.6 ms * [Marseille to Paris latency and RTT](/docs/network/latency/pairs/mrs-par-rtt): 8.9 ms * [Berlin to Paris latency and RTT](/docs/network/latency/pairs/ber-par-rtt): 15.6 ms **Same corridor (Asia Pacific → Europe)** * [Taipei to Moscow latency and RTT](/docs/network/latency/pairs/tpe-mow-rtt): 132.7 ms * [Singapore to Marseille latency and RTT](/docs/network/latency/pairs/sin-mrs-rtt): 140.6 ms * [Hong Kong to Berlin latency and RTT](/docs/network/latency/pairs/hkg-ber-rtt): 145.1 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/hkg-par.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Hong Kong → Paris round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/hkg-par-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Hong Kong → Seattle, USA RTT 🇭🇰 **Hong Kong (HKG)** → 🇺🇸 **Seattle, USA (SEA)** Measurement round: `2026-08-16T04:07:28Z`. Hong Kong to Seattle averaged 131.6 ms across 50 replies. The observed range was 125.78 to 162.34 ms; these figures describe one probe round, not a long-term service target. This is number 5 of 19 outgoing routes by average RTT. That ranking compares different destinations and is not adjusted for distance. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **131.6 ms** | | Jitter | **5.03 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **102.2 ms** | | Fiber Efficiency | **77.7%** | | Latency Tier | Good | | Source Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | | Destination Region | [North America](/docs/network/latency/regions/north-america) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Hong Kong](https://www.geonames.org/1819729) (22.27832, 114.17469) and [Seattle](https://www.geonames.org/5809844) (47.60621, -122.33207). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | --------------- | | WGS-84 geodesic distance | **10,435.9 km** | | Vacuum RTT floor | **69.62 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **102.2 ms** | | Low-latency fiber material floor | **101.79 ms** | | Engineering floor (5% path allowance) | **107.32 ms** | | Research 1.33× mapped-fiber reference | **135.92 ms** | | Estimated unamplified path loss | **2191.5 dB** | | Transparent optical spans / inline amplifiers | **137 / 136** | | Published RTT inflation over fiber floor | **1.29×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **125.78 ms** | | Average RTT | **131.6 ms** | | Maximum RTT | **162.34 ms** | | Standard deviation | **6.2 ms** | | Stdev / average | **4.7%** | ## Route Context * Departure PoP: [Hong Kong (HKG)](/docs/network/latency/hkg-hong-kong) * Destination PoP: [Seattle (SEA)](/docs/network/latency/sea-seattle) * Region overview: [Asia Pacific](/docs/network/latency/regions/asia-pacific) → [North America](/docs/network/latency/regions/north-america) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Seattle to Hong Kong latency and RTT](/docs/network/latency/pairs/sea-hkg-rtt): 130.4 ms **Fastest routes departing Hong Kong (HKG)** * [Hong Kong to Taipei latency and RTT](/docs/network/latency/pairs/hkg-tpe-rtt): 14.7 ms * [Hong Kong to Singapore latency and RTT](/docs/network/latency/pairs/hkg-sin-rtt): 31.5 ms * [Hong Kong to Tokyo latency and RTT](/docs/network/latency/pairs/hkg-tyo-rtt): 44.9 ms * [Hong Kong to Moscow latency and RTT](/docs/network/latency/pairs/hkg-mow-rtt): 118.2 ms * [Hong Kong to Sydney latency and RTT](/docs/network/latency/pairs/hkg-syd-rtt): 137.2 ms **Fastest routes arriving at Seattle (SEA)** * [Los Angeles to Seattle latency and RTT](/docs/network/latency/pairs/lax-sea-rtt): 26.9 ms * [New York to Seattle latency and RTT](/docs/network/latency/pairs/nyc-sea-rtt): 58.7 ms * [Ashburn to Seattle latency and RTT](/docs/network/latency/pairs/iad-sea-rtt): 62.3 ms * [Miami to Seattle latency and RTT](/docs/network/latency/pairs/mia-sea-rtt): 81.9 ms * [Tokyo to Seattle latency and RTT](/docs/network/latency/pairs/tyo-sea-rtt): 84.8 ms **Same corridor (Asia Pacific → North America)** * [Tokyo to Los Angeles latency and RTT](/docs/network/latency/pairs/tyo-lax-rtt): 101.2 ms * [Taipei to Seattle latency and RTT](/docs/network/latency/pairs/tpe-sea-rtt): 115.8 ms * [Taipei to Los Angeles latency and RTT](/docs/network/latency/pairs/tpe-lax-rtt): 132 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/hkg-sea.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Hong Kong → Seattle round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/hkg-sea-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Hong Kong → Singapore RTT 🇭🇰 **Hong Kong (HKG)** → 🇸🇬 **Singapore (SIN)** Measurement round: `2026-08-16T04:07:28Z`. Hong Kong to Singapore averaged 31.5 ms across 50 replies. The observed range was 29.73 to 35.06 ms; these figures describe one probe round, not a long-term service target. This is number 2 of 19 outgoing routes by average RTT. That ranking compares different destinations and is not adjusted for distance. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------------------- | | RTT | **31.5 ms** | | Jitter | **1.16 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **25.24 ms** | | Fiber Efficiency | **80.1%** | | Latency Tier | Excellent | | Source Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | | Destination Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Hong Kong](https://www.geonames.org/1819729) (22.27832, 114.17469) and [Singapore](https://www.geonames.org/1880252) (1.28967, 103.85007). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **2,577.2 km** | | Vacuum RTT floor | **17.19 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **25.24 ms** | | Low-latency fiber material floor | **25.14 ms** | | Engineering floor (5% path allowance) | **26.5 ms** | | Research 1.33× mapped-fiber reference | **33.57 ms** | | Estimated unamplified path loss | **541.2 dB** | | Transparent optical spans / inline amplifiers | **34 / 33** | | Published RTT inflation over fiber floor | **1.25×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------ | | Minimum RTT | **29.73 ms** | | Average RTT | **31.5 ms** | | Maximum RTT | **35.06 ms** | | Standard deviation | **1.22 ms** | | Stdev / average | **3.9%** | ## Route Context * Departure PoP: [Hong Kong (HKG)](/docs/network/latency/hkg-hong-kong) * Destination PoP: [Singapore (SIN)](/docs/network/latency/sin-singapore) * Region overview: [Asia Pacific](/docs/network/latency/regions/asia-pacific) → [Asia Pacific](/docs/network/latency/regions/asia-pacific) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Singapore to Hong Kong latency and RTT](/docs/network/latency/pairs/sin-hkg-rtt): 30.8 ms **Fastest routes departing Hong Kong (HKG)** * [Hong Kong to Taipei latency and RTT](/docs/network/latency/pairs/hkg-tpe-rtt): 14.7 ms * [Hong Kong to Tokyo latency and RTT](/docs/network/latency/pairs/hkg-tyo-rtt): 44.9 ms * [Hong Kong to Moscow latency and RTT](/docs/network/latency/pairs/hkg-mow-rtt): 118.2 ms * [Hong Kong to Seattle latency and RTT](/docs/network/latency/pairs/hkg-sea-rtt): 131.6 ms * [Hong Kong to Sydney latency and RTT](/docs/network/latency/pairs/hkg-syd-rtt): 137.2 ms **Fastest routes arriving at Singapore (SIN)** * [Taipei to Singapore latency and RTT](/docs/network/latency/pairs/tpe-sin-rtt): 44.3 ms * [Tokyo to Singapore latency and RTT](/docs/network/latency/pairs/tyo-sin-rtt): 69.4 ms * [Melbourne to Singapore latency and RTT](/docs/network/latency/pairs/mel-sin-rtt): 88.4 ms * [Sydney to Singapore latency and RTT](/docs/network/latency/pairs/syd-sin-rtt): 94.5 ms * [Marseille to Singapore latency and RTT](/docs/network/latency/pairs/mrs-sin-rtt): 139.6 ms **Same corridor (Asia Pacific → Asia Pacific)** * [Melbourne to Sydney latency and RTT](/docs/network/latency/pairs/mel-syd-rtt): 9.8 ms * [Sydney to Melbourne latency and RTT](/docs/network/latency/pairs/syd-mel-rtt): 9.8 ms * [Taipei to Hong Kong latency and RTT](/docs/network/latency/pairs/tpe-hkg-rtt): 15.5 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/hkg-sin.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Hong Kong → Singapore round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/hkg-sin-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Hong Kong → Sydney, Australia RTT 🇭🇰 **Hong Kong (HKG)** → 🇦🇺 **Sydney, Australia (SYD)** Measurement round: `2026-08-16T04:07:28Z`. Hong Kong to Sydney averaged 137.2 ms across 50 replies. The observed range was 129.36 to 162.96 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 6.53 ms, or 4.8% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------------------- | | RTT | **137.2 ms** | | Jitter | **5.69 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **71.97 ms** | | Fiber Efficiency | **52.5%** | | Latency Tier | Good | | Source Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | | Destination Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Hong Kong](https://www.geonames.org/1819729) (22.27832, 114.17469) and [Sydney](https://www.geonames.org/2147714) (-33.86785, 151.20732). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **7,349.1 km** | | Vacuum RTT floor | **49.03 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **71.97 ms** | | Low-latency fiber material floor | **71.68 ms** | | Engineering floor (5% path allowance) | **75.58 ms** | | Research 1.33× mapped-fiber reference | **95.72 ms** | | Estimated unamplified path loss | **1543.3 dB** | | Transparent optical spans / inline amplifiers | **97 / 96** | | Published RTT inflation over fiber floor | **1.91×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **129.36 ms** | | Average RTT | **137.2 ms** | | Maximum RTT | **162.96 ms** | | Standard deviation | **6.53 ms** | | Stdev / average | **4.8%** | ## Route Context * Departure PoP: [Hong Kong (HKG)](/docs/network/latency/hkg-hong-kong) * Destination PoP: [Sydney (SYD)](/docs/network/latency/syd-sydney) * Region overview: [Asia Pacific](/docs/network/latency/regions/asia-pacific) → [Asia Pacific](/docs/network/latency/regions/asia-pacific) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Sydney to Hong Kong latency and RTT](/docs/network/latency/pairs/syd-hkg-rtt): 135.7 ms **Fastest routes departing Hong Kong (HKG)** * [Hong Kong to Taipei latency and RTT](/docs/network/latency/pairs/hkg-tpe-rtt): 14.7 ms * [Hong Kong to Singapore latency and RTT](/docs/network/latency/pairs/hkg-sin-rtt): 31.5 ms * [Hong Kong to Tokyo latency and RTT](/docs/network/latency/pairs/hkg-tyo-rtt): 44.9 ms * [Hong Kong to Moscow latency and RTT](/docs/network/latency/pairs/hkg-mow-rtt): 118.2 ms * [Hong Kong to Seattle latency and RTT](/docs/network/latency/pairs/hkg-sea-rtt): 131.6 ms **Fastest routes arriving at Sydney (SYD)** * [Melbourne to Sydney latency and RTT](/docs/network/latency/pairs/mel-syd-rtt): 9.8 ms * [Singapore to Sydney latency and RTT](/docs/network/latency/pairs/sin-syd-rtt): 94.5 ms * [Tokyo to Sydney latency and RTT](/docs/network/latency/pairs/tyo-syd-rtt): 101.9 ms * [Taipei to Sydney latency and RTT](/docs/network/latency/pairs/tpe-syd-rtt): 133.2 ms * [Los Angeles to Sydney latency and RTT](/docs/network/latency/pairs/lax-syd-rtt): 137.6 ms **Same corridor (Asia Pacific → Asia Pacific)** * [Sydney to Melbourne latency and RTT](/docs/network/latency/pairs/syd-mel-rtt): 9.8 ms * [Taipei to Hong Kong latency and RTT](/docs/network/latency/pairs/tpe-hkg-rtt): 15.5 ms * [Singapore to Hong Kong latency and RTT](/docs/network/latency/pairs/sin-hkg-rtt): 30.8 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/hkg-syd.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Hong Kong → Sydney round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/hkg-syd-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Hong Kong → Taipei, Taiwan RTT 🇭🇰 **Hong Kong (HKG)** → 🇹🇼 **Taipei, Taiwan (TPE)** Measurement round: `2026-08-16T04:07:28Z`. Hong Kong to Taipei averaged 14.7 ms across 50 replies. The observed range was 13.95 to 17.24 ms; these figures describe one probe round, not a long-term service target. This is number 1 of 19 outgoing routes by average RTT. That ranking compares different destinations and is not adjusted for distance. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------------------- | | RTT | **14.7 ms** | | Jitter | **0.56 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **7.94 ms** | | Fiber Efficiency | **54%** | | Latency Tier | Ultra-Low | | Source Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | | Destination Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Hong Kong](https://www.geonames.org/1819729) (22.27832, 114.17469) and [Taipei](https://www.geonames.org/1668341) (25.05306, 121.52639). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | ------------ | | WGS-84 geodesic distance | **810.3 km** | | Vacuum RTT floor | **5.41 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **7.94 ms** | | Low-latency fiber material floor | **7.9 ms** | | Engineering floor (5% path allowance) | **8.33 ms** | | Research 1.33× mapped-fiber reference | **10.55 ms** | | Estimated unamplified path loss | **170.2 dB** | | Transparent optical spans / inline amplifiers | **11 / 10** | | Published RTT inflation over fiber floor | **1.85×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------ | | Minimum RTT | **13.95 ms** | | Average RTT | **14.7 ms** | | Maximum RTT | **17.24 ms** | | Standard deviation | **0.72 ms** | | Stdev / average | **4.9%** | ## Route Context * Departure PoP: [Hong Kong (HKG)](/docs/network/latency/hkg-hong-kong) * Destination PoP: [Taipei (TPE)](/docs/network/latency/tpe-taipei) * Region overview: [Asia Pacific](/docs/network/latency/regions/asia-pacific) → [Asia Pacific](/docs/network/latency/regions/asia-pacific) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Taipei to Hong Kong latency and RTT](/docs/network/latency/pairs/tpe-hkg-rtt): 15.5 ms **Fastest routes departing Hong Kong (HKG)** * [Hong Kong to Singapore latency and RTT](/docs/network/latency/pairs/hkg-sin-rtt): 31.5 ms * [Hong Kong to Tokyo latency and RTT](/docs/network/latency/pairs/hkg-tyo-rtt): 44.9 ms * [Hong Kong to Moscow latency and RTT](/docs/network/latency/pairs/hkg-mow-rtt): 118.2 ms * [Hong Kong to Seattle latency and RTT](/docs/network/latency/pairs/hkg-sea-rtt): 131.6 ms * [Hong Kong to Sydney latency and RTT](/docs/network/latency/pairs/hkg-syd-rtt): 137.2 ms **Fastest routes arriving at Taipei (TPE)** * [Tokyo to Taipei latency and RTT](/docs/network/latency/pairs/tyo-tpe-rtt): 32.3 ms * [Singapore to Taipei latency and RTT](/docs/network/latency/pairs/sin-tpe-rtt): 45.8 ms * [Seattle to Taipei latency and RTT](/docs/network/latency/pairs/sea-tpe-rtt): 114.7 ms * [Sydney to Taipei latency and RTT](/docs/network/latency/pairs/syd-tpe-rtt): 131.7 ms * [Moscow to Taipei latency and RTT](/docs/network/latency/pairs/mow-tpe-rtt): 132.1 ms **Same corridor (Asia Pacific → Asia Pacific)** * [Melbourne to Sydney latency and RTT](/docs/network/latency/pairs/mel-syd-rtt): 9.8 ms * [Sydney to Melbourne latency and RTT](/docs/network/latency/pairs/syd-mel-rtt): 9.8 ms * [Singapore to Hong Kong latency and RTT](/docs/network/latency/pairs/sin-hkg-rtt): 30.8 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/hkg-tpe.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Hong Kong → Taipei round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/hkg-tpe-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Hong Kong → Tokyo, Japan RTT 🇭🇰 **Hong Kong (HKG)** → 🇯🇵 **Tokyo, Japan (TYO)** Measurement round: `2026-08-16T04:07:28Z`. Hong Kong to Tokyo averaged 44.9 ms across 50 replies. The observed range was 42.32 to 50.91 ms; these figures describe one probe round, not a long-term service target. This is number 3 of 19 outgoing routes by average RTT. That ranking compares different destinations and is not adjusted for distance. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------------------- | | RTT | **44.9 ms** | | Jitter | **1.6 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **28.24 ms** | | Fiber Efficiency | **62.9%** | | Latency Tier | Excellent | | Source Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | | Destination Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Hong Kong](https://www.geonames.org/1819729) (22.27832, 114.17469) and [Tokyo](https://www.geonames.org/1850147) (35.68950, 139.69171). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **2,883.5 km** | | Vacuum RTT floor | **19.24 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **28.24 ms** | | Low-latency fiber material floor | **28.12 ms** | | Engineering floor (5% path allowance) | **29.65 ms** | | Research 1.33× mapped-fiber reference | **37.56 ms** | | Estimated unamplified path loss | **605.5 dB** | | Transparent optical spans / inline amplifiers | **38 / 37** | | Published RTT inflation over fiber floor | **1.59×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------ | | Minimum RTT | **42.32 ms** | | Average RTT | **44.9 ms** | | Maximum RTT | **50.91 ms** | | Standard deviation | **1.82 ms** | | Stdev / average | **4.1%** | ## Route Context * Departure PoP: [Hong Kong (HKG)](/docs/network/latency/hkg-hong-kong) * Destination PoP: [Tokyo (TYO)](/docs/network/latency/tyo-tokyo) * Region overview: [Asia Pacific](/docs/network/latency/regions/asia-pacific) → [Asia Pacific](/docs/network/latency/regions/asia-pacific) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Tokyo to Hong Kong latency and RTT](/docs/network/latency/pairs/tyo-hkg-rtt): 44.8 ms **Fastest routes departing Hong Kong (HKG)** * [Hong Kong to Taipei latency and RTT](/docs/network/latency/pairs/hkg-tpe-rtt): 14.7 ms * [Hong Kong to Singapore latency and RTT](/docs/network/latency/pairs/hkg-sin-rtt): 31.5 ms * [Hong Kong to Moscow latency and RTT](/docs/network/latency/pairs/hkg-mow-rtt): 118.2 ms * [Hong Kong to Seattle latency and RTT](/docs/network/latency/pairs/hkg-sea-rtt): 131.6 ms * [Hong Kong to Sydney latency and RTT](/docs/network/latency/pairs/hkg-syd-rtt): 137.2 ms **Fastest routes arriving at Tokyo (TYO)** * [Taipei to Tokyo latency and RTT](/docs/network/latency/pairs/tpe-tyo-rtt): 31.9 ms * [Singapore to Tokyo latency and RTT](/docs/network/latency/pairs/sin-tyo-rtt): 68.9 ms * [Seattle to Tokyo latency and RTT](/docs/network/latency/pairs/sea-tyo-rtt): 85 ms * [Los Angeles to Tokyo latency and RTT](/docs/network/latency/pairs/lax-tyo-rtt): 101.2 ms * [Sydney to Tokyo latency and RTT](/docs/network/latency/pairs/syd-tyo-rtt): 101.3 ms **Same corridor (Asia Pacific → Asia Pacific)** * [Melbourne to Sydney latency and RTT](/docs/network/latency/pairs/mel-syd-rtt): 9.8 ms * [Sydney to Melbourne latency and RTT](/docs/network/latency/pairs/syd-mel-rtt): 9.8 ms * [Taipei to Hong Kong latency and RTT](/docs/network/latency/pairs/tpe-hkg-rtt): 15.5 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/hkg-tyo.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Hong Kong → Tokyo round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/hkg-tyo-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Ashburn, USA → Amsterdam, Netherlands RTT 🇺🇸 **Ashburn, USA (IAD)** → 🇳🇱 **Amsterdam, Netherlands (AMS)** Measurement round: `2026-08-16T04:07:28Z`. Ashburn to Amsterdam averaged 75.2 ms across 50 replies. The observed range was 72.76 to 84.12 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 2.32 ms, or 3.1% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **75.2 ms** | | Jitter | **1.77 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **60.94 ms** | | Fiber Efficiency | **81%** | | Latency Tier | Excellent | | Source Region | [North America](/docs/network/latency/regions/north-america) | | Destination Region | [Europe](/docs/network/latency/regions/europe) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Ashburn](https://www.geonames.org/4744870) (39.04372, -77.48749) and [Amsterdam](https://www.geonames.org/2759794) (52.37403, 4.88969). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **6,223.2 km** | | Vacuum RTT floor | **41.52 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **60.94 ms** | | Low-latency fiber material floor | **60.7 ms** | | Engineering floor (5% path allowance) | **64 ms** | | Research 1.33× mapped-fiber reference | **81.05 ms** | | Estimated unamplified path loss | **1306.9 dB** | | Transparent optical spans / inline amplifiers | **82 / 81** | | Published RTT inflation over fiber floor | **1.23×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------ | | Minimum RTT | **72.76 ms** | | Average RTT | **75.2 ms** | | Maximum RTT | **84.12 ms** | | Standard deviation | **2.32 ms** | | Stdev / average | **3.1%** | ## Route Context * Departure PoP: [Ashburn (IAD)](/docs/network/latency/iad-ashburn) * Destination PoP: [Amsterdam (AMS)](/docs/network/latency/ams-amsterdam) * Region overview: [North America](/docs/network/latency/regions/north-america) → [Europe](/docs/network/latency/regions/europe) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Amsterdam to Ashburn latency and RTT](/docs/network/latency/pairs/ams-iad-rtt): 75.7 ms **Fastest routes departing Ashburn (IAD)** * [Ashburn to New York latency and RTT](/docs/network/latency/pairs/iad-nyc-rtt): 6 ms * [Ashburn to Miami latency and RTT](/docs/network/latency/pairs/iad-mia-rtt): 27.4 ms * [Ashburn to Los Angeles latency and RTT](/docs/network/latency/pairs/iad-lax-rtt): 60.7 ms * [Ashburn to Seattle latency and RTT](/docs/network/latency/pairs/iad-sea-rtt): 62.3 ms * [Ashburn to London latency and RTT](/docs/network/latency/pairs/iad-lon-rtt): 70.3 ms **Fastest routes arriving at Amsterdam (AMS)** * [London to Amsterdam latency and RTT](/docs/network/latency/pairs/lon-ams-rtt): 5.2 ms * [Frankfurt to Amsterdam latency and RTT](/docs/network/latency/pairs/fra-ams-rtt): 5.9 ms * [Paris to Amsterdam latency and RTT](/docs/network/latency/pairs/par-ams-rtt): 7 ms * [Berlin to Amsterdam latency and RTT](/docs/network/latency/pairs/ber-ams-rtt): 7.2 ms * [Marseille to Amsterdam latency and RTT](/docs/network/latency/pairs/mrs-ams-rtt): 20.3 ms **Same corridor (North America → Europe)** * [New York to London latency and RTT](/docs/network/latency/pairs/nyc-lon-rtt): 63.5 ms * [New York to Amsterdam latency and RTT](/docs/network/latency/pairs/nyc-ams-rtt): 68.7 ms * [New York to Paris latency and RTT](/docs/network/latency/pairs/nyc-par-rtt): 68.9 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/iad-ams.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Ashburn → Amsterdam round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/iad-ams-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Ashburn, USA → Berlin, Germany RTT 🇺🇸 **Ashburn, USA (IAD)** → 🇩🇪 **Berlin, Germany (BER)** Measurement round: `2026-08-16T04:07:28Z`. Ashburn to Berlin averaged 86.1 ms across 50 replies. The observed range was 84.07 to 94.45 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 2.05 ms, or 2.4% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **86.1 ms** | | Jitter | **1.6 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **66.05 ms** | | Fiber Efficiency | **76.7%** | | Latency Tier | Good | | Source Region | [North America](/docs/network/latency/regions/north-america) | | Destination Region | [Europe](/docs/network/latency/regions/europe) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Ashburn](https://www.geonames.org/4744870) (39.04372, -77.48749) and [Berlin](https://www.geonames.org/2950159) (52.52437, 13.41053). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **6,744.5 km** | | Vacuum RTT floor | **44.99 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **66.05 ms** | | Low-latency fiber material floor | **65.78 ms** | | Engineering floor (5% path allowance) | **69.36 ms** | | Research 1.33× mapped-fiber reference | **87.84 ms** | | Estimated unamplified path loss | **1416.4 dB** | | Transparent optical spans / inline amplifiers | **89 / 88** | | Published RTT inflation over fiber floor | **1.3×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------ | | Minimum RTT | **84.07 ms** | | Average RTT | **86.1 ms** | | Maximum RTT | **94.45 ms** | | Standard deviation | **2.05 ms** | | Stdev / average | **2.4%** | ## Route Context * Departure PoP: [Ashburn (IAD)](/docs/network/latency/iad-ashburn) * Destination PoP: [Berlin (BER)](/docs/network/latency/ber-berlin) * Region overview: [North America](/docs/network/latency/regions/north-america) → [Europe](/docs/network/latency/regions/europe) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Berlin to Ashburn latency and RTT](/docs/network/latency/pairs/ber-iad-rtt): 85.7 ms **Fastest routes departing Ashburn (IAD)** * [Ashburn to New York latency and RTT](/docs/network/latency/pairs/iad-nyc-rtt): 6 ms * [Ashburn to Miami latency and RTT](/docs/network/latency/pairs/iad-mia-rtt): 27.4 ms * [Ashburn to Los Angeles latency and RTT](/docs/network/latency/pairs/iad-lax-rtt): 60.7 ms * [Ashburn to Seattle latency and RTT](/docs/network/latency/pairs/iad-sea-rtt): 62.3 ms * [Ashburn to London latency and RTT](/docs/network/latency/pairs/iad-lon-rtt): 70.3 ms **Fastest routes arriving at Berlin (BER)** * [Frankfurt to Berlin latency and RTT](/docs/network/latency/pairs/fra-ber-rtt): 6.1 ms * [Amsterdam to Berlin latency and RTT](/docs/network/latency/pairs/ams-ber-rtt): 7.2 ms * [Paris to Berlin latency and RTT](/docs/network/latency/pairs/par-ber-rtt): 15.5 ms * [London to Berlin latency and RTT](/docs/network/latency/pairs/lon-ber-rtt): 16.2 ms * [Marseille to Berlin latency and RTT](/docs/network/latency/pairs/mrs-ber-rtt): 20.6 ms **Same corridor (North America → Europe)** * [New York to London latency and RTT](/docs/network/latency/pairs/nyc-lon-rtt): 63.5 ms * [New York to Amsterdam latency and RTT](/docs/network/latency/pairs/nyc-ams-rtt): 68.7 ms * [New York to Paris latency and RTT](/docs/network/latency/pairs/nyc-par-rtt): 68.9 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/iad-ber.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Ashburn → Berlin round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/iad-ber-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Ashburn, USA → Frankfurt, Germany RTT 🇺🇸 **Ashburn, USA (IAD)** → 🇩🇪 **Frankfurt, Germany (FRA)** Measurement round: `2026-08-16T04:07:28Z`. Ashburn to Frankfurt averaged 81.2 ms across 50 replies. The observed range was 79.1 to 86.57 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 1.81 ms, or 2.2% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **81.2 ms** | | Jitter | **1.11 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **64.3 ms** | | Fiber Efficiency | **79.2%** | | Latency Tier | Good | | Source Region | [North America](/docs/network/latency/regions/north-america) | | Destination Region | [Europe](/docs/network/latency/regions/europe) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Ashburn](https://www.geonames.org/4744870) (39.04372, -77.48749) and [Frankfurt am Main](https://www.geonames.org/2925533) (50.11552, 8.68417). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | ------------- | | WGS-84 geodesic distance | **6,566 km** | | Vacuum RTT floor | **43.8 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **64.3 ms** | | Low-latency fiber material floor | **64.04 ms** | | Engineering floor (5% path allowance) | **67.52 ms** | | Research 1.33× mapped-fiber reference | **85.52 ms** | | Estimated unamplified path loss | **1378.9 dB** | | Transparent optical spans / inline amplifiers | **87 / 86** | | Published RTT inflation over fiber floor | **1.26×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------ | | Minimum RTT | **79.1 ms** | | Average RTT | **81.2 ms** | | Maximum RTT | **86.57 ms** | | Standard deviation | **1.81 ms** | | Stdev / average | **2.2%** | ## Route Context * Departure PoP: [Ashburn (IAD)](/docs/network/latency/iad-ashburn) * Destination PoP: [Frankfurt (FRA)](/docs/network/latency/fra-frankfurt) * Region overview: [North America](/docs/network/latency/regions/north-america) → [Europe](/docs/network/latency/regions/europe) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Frankfurt to Ashburn latency and RTT](/docs/network/latency/pairs/fra-iad-rtt): 80.7 ms **Fastest routes departing Ashburn (IAD)** * [Ashburn to New York latency and RTT](/docs/network/latency/pairs/iad-nyc-rtt): 6 ms * [Ashburn to Miami latency and RTT](/docs/network/latency/pairs/iad-mia-rtt): 27.4 ms * [Ashburn to Los Angeles latency and RTT](/docs/network/latency/pairs/iad-lax-rtt): 60.7 ms * [Ashburn to Seattle latency and RTT](/docs/network/latency/pairs/iad-sea-rtt): 62.3 ms * [Ashburn to London latency and RTT](/docs/network/latency/pairs/iad-lon-rtt): 70.3 ms **Fastest routes arriving at Frankfurt (FRA)** * [Amsterdam to Frankfurt latency and RTT](/docs/network/latency/pairs/ams-fra-rtt): 5.9 ms * [Berlin to Frankfurt latency and RTT](/docs/network/latency/pairs/ber-fra-rtt): 6.1 ms * [Paris to Frankfurt latency and RTT](/docs/network/latency/pairs/par-fra-rtt): 7.6 ms * [London to Frankfurt latency and RTT](/docs/network/latency/pairs/lon-fra-rtt): 13.6 ms * [Marseille to Frankfurt latency and RTT](/docs/network/latency/pairs/mrs-fra-rtt): 16 ms **Same corridor (North America → Europe)** * [New York to London latency and RTT](/docs/network/latency/pairs/nyc-lon-rtt): 63.5 ms * [New York to Amsterdam latency and RTT](/docs/network/latency/pairs/nyc-ams-rtt): 68.7 ms * [New York to Paris latency and RTT](/docs/network/latency/pairs/nyc-par-rtt): 68.9 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/iad-fra.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Ashburn → Frankfurt round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/iad-fra-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Ashburn, USA → São Paulo, Brazil RTT 🇺🇸 **Ashburn, USA (IAD)** → 🇧🇷 **São Paulo, Brazil (GRU)** Measurement round: `2026-08-16T04:07:28Z`. Ashburn to São Paulo averaged 101.4 ms across 50 replies. The observed range was 98.16 to 107.74 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 2.31 ms, or 2.3% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **101.4 ms** | | Jitter | **2.06 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **74.79 ms** | | Fiber Efficiency | **73.8%** | | Latency Tier | Good | | Source Region | [North America](/docs/network/latency/regions/north-america) | | Destination Region | [South America](/docs/network/latency/regions/south-america) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Ashburn](https://www.geonames.org/4744870) (39.04372, -77.48749) and [São Paulo](https://www.geonames.org/3448439) (-23.54750, -46.63611). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **7,637.7 km** | | Vacuum RTT floor | **50.95 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **74.79 ms** | | Low-latency fiber material floor | **74.49 ms** | | Engineering floor (5% path allowance) | **78.54 ms** | | Research 1.33× mapped-fiber reference | **99.48 ms** | | Estimated unamplified path loss | **1603.9 dB** | | Transparent optical spans / inline amplifiers | **101 / 100** | | Published RTT inflation over fiber floor | **1.36×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **98.16 ms** | | Average RTT | **101.4 ms** | | Maximum RTT | **107.74 ms** | | Standard deviation | **2.31 ms** | | Stdev / average | **2.3%** | ## Route Context * Departure PoP: [Ashburn (IAD)](/docs/network/latency/iad-ashburn) * Destination PoP: [São Paulo (GRU)](/docs/network/latency/gru-sao-paulo) * Region overview: [North America](/docs/network/latency/regions/north-america) → [South America](/docs/network/latency/regions/south-america) * Corridor overview: [North America to South America](/docs/network/latency/regions/north-america-to-south-america) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [São Paulo to Ashburn latency and RTT](/docs/network/latency/pairs/gru-iad-rtt): 128 ms **Fastest routes departing Ashburn (IAD)** * [Ashburn to New York latency and RTT](/docs/network/latency/pairs/iad-nyc-rtt): 6 ms * [Ashburn to Miami latency and RTT](/docs/network/latency/pairs/iad-mia-rtt): 27.4 ms * [Ashburn to Los Angeles latency and RTT](/docs/network/latency/pairs/iad-lax-rtt): 60.7 ms * [Ashburn to Seattle latency and RTT](/docs/network/latency/pairs/iad-sea-rtt): 62.3 ms * [Ashburn to London latency and RTT](/docs/network/latency/pairs/iad-lon-rtt): 70.3 ms **Fastest routes arriving at São Paulo (GRU)** * [New York to São Paulo latency and RTT](/docs/network/latency/pairs/nyc-gru-rtt): 106.9 ms * [Miami to São Paulo latency and RTT](/docs/network/latency/pairs/mia-gru-rtt): 128.7 ms * [Los Angeles to São Paulo latency and RTT](/docs/network/latency/pairs/lax-gru-rtt): 131.3 ms * [Seattle to São Paulo latency and RTT](/docs/network/latency/pairs/sea-gru-rtt): 155.7 ms * [London to São Paulo latency and RTT](/docs/network/latency/pairs/lon-gru-rtt): 170.7 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/iad-gru.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Ashburn → São Paulo round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/iad-gru-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Ashburn, USA → Hong Kong RTT 🇺🇸 **Ashburn, USA (IAD)** → 🇭🇰 **Hong Kong (HKG)** Measurement round: `2026-08-16T04:07:28Z`. Ashburn to Hong Kong averaged 187.2 ms across 50 replies. The observed range was 178.13 to 213.4 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 8.01 ms, or 4.3% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **187.2 ms** | | Jitter | **6.84 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **128.37 ms** | | Fiber Efficiency | **68.6%** | | Latency Tier | Fair | | Source Region | [North America](/docs/network/latency/regions/north-america) | | Destination Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Ashburn](https://www.geonames.org/4744870) (39.04372, -77.48749) and [Hong Kong](https://www.geonames.org/1819729) (22.27832, 114.17469). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | --------------- | | WGS-84 geodesic distance | **13,108.8 km** | | Vacuum RTT floor | **87.45 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **128.37 ms** | | Low-latency fiber material floor | **127.86 ms** | | Engineering floor (5% path allowance) | **134.81 ms** | | Research 1.33× mapped-fiber reference | **170.73 ms** | | Estimated unamplified path loss | **2752.9 dB** | | Transparent optical spans / inline amplifiers | **173 / 172** | | Published RTT inflation over fiber floor | **1.46×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **178.13 ms** | | Average RTT | **187.2 ms** | | Maximum RTT | **213.4 ms** | | Standard deviation | **8.01 ms** | | Stdev / average | **4.3%** | ## Route Context * Departure PoP: [Ashburn (IAD)](/docs/network/latency/iad-ashburn) * Destination PoP: [Hong Kong (HKG)](/docs/network/latency/hkg-hong-kong) * Region overview: [North America](/docs/network/latency/regions/north-america) → [Asia Pacific](/docs/network/latency/regions/asia-pacific) * Corridor overview: [North America to Asia Pacific](/docs/network/latency/regions/north-america-to-asia-pacific) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Hong Kong to Ashburn latency and RTT](/docs/network/latency/pairs/hkg-iad-rtt): 186.2 ms **Fastest routes departing Ashburn (IAD)** * [Ashburn to New York latency and RTT](/docs/network/latency/pairs/iad-nyc-rtt): 6 ms * [Ashburn to Miami latency and RTT](/docs/network/latency/pairs/iad-mia-rtt): 27.4 ms * [Ashburn to Los Angeles latency and RTT](/docs/network/latency/pairs/iad-lax-rtt): 60.7 ms * [Ashburn to Seattle latency and RTT](/docs/network/latency/pairs/iad-sea-rtt): 62.3 ms * [Ashburn to London latency and RTT](/docs/network/latency/pairs/iad-lon-rtt): 70.3 ms **Fastest routes arriving at Hong Kong (HKG)** * [Taipei to Hong Kong latency and RTT](/docs/network/latency/pairs/tpe-hkg-rtt): 15.5 ms * [Singapore to Hong Kong latency and RTT](/docs/network/latency/pairs/sin-hkg-rtt): 30.8 ms * [Tokyo to Hong Kong latency and RTT](/docs/network/latency/pairs/tyo-hkg-rtt): 44.8 ms * [Moscow to Hong Kong latency and RTT](/docs/network/latency/pairs/mow-hkg-rtt): 118.6 ms * [Seattle to Hong Kong latency and RTT](/docs/network/latency/pairs/sea-hkg-rtt): 130.4 ms **Same corridor (North America → Asia Pacific)** * [Seattle to Tokyo latency and RTT](/docs/network/latency/pairs/sea-tyo-rtt): 85 ms * [Los Angeles to Tokyo latency and RTT](/docs/network/latency/pairs/lax-tyo-rtt): 101.2 ms * [Seattle to Taipei latency and RTT](/docs/network/latency/pairs/sea-tpe-rtt): 114.7 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/iad-hkg.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Ashburn → Hong Kong round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/iad-hkg-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Ashburn, USA → Johannesburg, South Africa RTT 🇺🇸 **Ashburn, USA (IAD)** → 🇿🇦 **Johannesburg, South Africa (JNB)** Measurement round: `2026-08-16T04:07:28Z`. Ashburn to Johannesburg averaged 228.3 ms across 50 replies. The observed range was 216.38 to 264.39 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 10.74 ms, or 4.7% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **228.3 ms** | | Jitter | **7.98 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **128.11 ms** | | Fiber Efficiency | **56.1%** | | Latency Tier | Fair | | Source Region | [North America](/docs/network/latency/regions/north-america) | | Destination Region | [Africa](/docs/network/latency/regions/africa) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Ashburn](https://www.geonames.org/4744870) (39.04372, -77.48749) and [Johannesburg](https://www.geonames.org/993800) (-26.20227, 28.04363). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | ------------- | | WGS-84 geodesic distance | **13,082 km** | | Vacuum RTT floor | **87.27 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **128.11 ms** | | Low-latency fiber material floor | **127.59 ms** | | Engineering floor (5% path allowance) | **134.53 ms** | | Research 1.33× mapped-fiber reference | **170.38 ms** | | Estimated unamplified path loss | **2747.2 dB** | | Transparent optical spans / inline amplifiers | **172 / 171** | | Published RTT inflation over fiber floor | **1.78×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **216.38 ms** | | Average RTT | **228.3 ms** | | Maximum RTT | **264.39 ms** | | Standard deviation | **10.74 ms** | | Stdev / average | **4.7%** | ## Route Context * Departure PoP: [Ashburn (IAD)](/docs/network/latency/iad-ashburn) * Destination PoP: [Johannesburg (JNB)](/docs/network/latency/jnb-johannesburg) * Region overview: [North America](/docs/network/latency/regions/north-america) → [Africa](/docs/network/latency/regions/africa) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Johannesburg to Ashburn latency and RTT](/docs/network/latency/pairs/jnb-iad-rtt): 229 ms **Fastest routes departing Ashburn (IAD)** * [Ashburn to New York latency and RTT](/docs/network/latency/pairs/iad-nyc-rtt): 6 ms * [Ashburn to Miami latency and RTT](/docs/network/latency/pairs/iad-mia-rtt): 27.4 ms * [Ashburn to Los Angeles latency and RTT](/docs/network/latency/pairs/iad-lax-rtt): 60.7 ms * [Ashburn to Seattle latency and RTT](/docs/network/latency/pairs/iad-sea-rtt): 62.3 ms * [Ashburn to London latency and RTT](/docs/network/latency/pairs/iad-lon-rtt): 70.3 ms **Fastest routes arriving at Johannesburg (JNB)** * [Amsterdam to Johannesburg latency and RTT](/docs/network/latency/pairs/ams-jnb-rtt): 171.9 ms * [Berlin to Johannesburg latency and RTT](/docs/network/latency/pairs/ber-jnb-rtt): 172.1 ms * [London to Johannesburg latency and RTT](/docs/network/latency/pairs/lon-jnb-rtt): 179.6 ms * [Paris to Johannesburg latency and RTT](/docs/network/latency/pairs/par-jnb-rtt): 180.9 ms * [Frankfurt to Johannesburg latency and RTT](/docs/network/latency/pairs/fra-jnb-rtt): 188.4 ms **Same corridor (North America → Africa)** * [New York to Johannesburg latency and RTT](/docs/network/latency/pairs/nyc-jnb-rtt): 221.5 ms * [Miami to Johannesburg latency and RTT](/docs/network/latency/pairs/mia-jnb-rtt): 259.8 ms * [Seattle to Johannesburg latency and RTT](/docs/network/latency/pairs/sea-jnb-rtt): 282.1 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/iad-jnb.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Ashburn → Johannesburg round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/iad-jnb-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Ashburn, USA → Los Angeles, USA RTT 🇺🇸 **Ashburn, USA (IAD)** → 🇺🇸 **Los Angeles, USA (LAX)** Measurement round: `2026-08-16T04:07:28Z`. Ashburn to Los Angeles averaged 60.7 ms across 50 replies. The observed range was 58.56 to 67.73 ms; these figures describe one probe round, not a long-term service target. This is number 3 of 19 outgoing routes by average RTT. That ranking compares different destinations and is not adjusted for distance. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **60.7 ms** | | Jitter | **1.64 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **35.84 ms** | | Fiber Efficiency | **59%** | | Latency Tier | Excellent | | Source Region | [North America](/docs/network/latency/regions/north-america) | | Destination Region | [North America](/docs/network/latency/regions/north-america) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Ashburn](https://www.geonames.org/4744870) (39.04372, -77.48749) and [Los Angeles](https://www.geonames.org/5368361) (34.05223, -118.24368). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **3,660.1 km** | | Vacuum RTT floor | **24.42 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **35.84 ms** | | Low-latency fiber material floor | **35.7 ms** | | Engineering floor (5% path allowance) | **37.64 ms** | | Research 1.33× mapped-fiber reference | **47.67 ms** | | Estimated unamplified path loss | **768.6 dB** | | Transparent optical spans / inline amplifiers | **49 / 48** | | Published RTT inflation over fiber floor | **1.69×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------ | | Minimum RTT | **58.56 ms** | | Average RTT | **60.7 ms** | | Maximum RTT | **67.73 ms** | | Standard deviation | **1.84 ms** | | Stdev / average | **3.0%** | ## Route Context * Departure PoP: [Ashburn (IAD)](/docs/network/latency/iad-ashburn) * Destination PoP: [Los Angeles (LAX)](/docs/network/latency/lax-los-angeles) * Region overview: [North America](/docs/network/latency/regions/north-america) → [North America](/docs/network/latency/regions/north-america) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Los Angeles to Ashburn latency and RTT](/docs/network/latency/pairs/lax-iad-rtt): 60.4 ms **Fastest routes departing Ashburn (IAD)** * [Ashburn to New York latency and RTT](/docs/network/latency/pairs/iad-nyc-rtt): 6 ms * [Ashburn to Miami latency and RTT](/docs/network/latency/pairs/iad-mia-rtt): 27.4 ms * [Ashburn to Seattle latency and RTT](/docs/network/latency/pairs/iad-sea-rtt): 62.3 ms * [Ashburn to London latency and RTT](/docs/network/latency/pairs/iad-lon-rtt): 70.3 ms * [Ashburn to Amsterdam latency and RTT](/docs/network/latency/pairs/iad-ams-rtt): 75.2 ms **Fastest routes arriving at Los Angeles (LAX)** * [Seattle to Los Angeles latency and RTT](/docs/network/latency/pairs/sea-lax-rtt): 25.9 ms * [Miami to Los Angeles latency and RTT](/docs/network/latency/pairs/mia-lax-rtt): 56.8 ms * [New York to Los Angeles latency and RTT](/docs/network/latency/pairs/nyc-lax-rtt): 58.9 ms * [Tokyo to Los Angeles latency and RTT](/docs/network/latency/pairs/tyo-lax-rtt): 101.2 ms * [London to Los Angeles latency and RTT](/docs/network/latency/pairs/lon-lax-rtt): 127.5 ms **Same corridor (North America → North America)** * [New York to Ashburn latency and RTT](/docs/network/latency/pairs/nyc-iad-rtt): 6 ms * [Los Angeles to Seattle latency and RTT](/docs/network/latency/pairs/lax-sea-rtt): 26.9 ms * [Miami to Ashburn latency and RTT](/docs/network/latency/pairs/mia-iad-rtt): 27.3 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/iad-lax.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Ashburn → Los Angeles round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/iad-lax-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Ashburn, USA → London, UK RTT 🇺🇸 **Ashburn, USA (IAD)** → 🇬🇧 **London, UK (LON)** Measurement round: `2026-08-16T04:07:28Z`. Ashburn to London averaged 70.3 ms across 50 replies. The observed range was 68.79 to 73.78 ms; these figures describe one probe round, not a long-term service target. This is number 5 of 19 outgoing routes by average RTT. That ranking compares different destinations and is not adjusted for distance. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **70.3 ms** | | Jitter | **1.06 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **58.1 ms** | | Fiber Efficiency | **82.6%** | | Latency Tier | Excellent | | Source Region | [North America](/docs/network/latency/regions/north-america) | | Destination Region | [Europe](/docs/network/latency/regions/europe) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Ashburn](https://www.geonames.org/4744870) (39.04372, -77.48749) and [London](https://www.geonames.org/2643743) (51.50853, -0.12574). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **5,932.8 km** | | Vacuum RTT floor | **39.58 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **58.1 ms** | | Low-latency fiber material floor | **57.87 ms** | | Engineering floor (5% path allowance) | **61.01 ms** | | Research 1.33× mapped-fiber reference | **77.27 ms** | | Estimated unamplified path loss | **1245.9 dB** | | Transparent optical spans / inline amplifiers | **78 / 77** | | Published RTT inflation over fiber floor | **1.21×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------ | | Minimum RTT | **68.79 ms** | | Average RTT | **70.3 ms** | | Maximum RTT | **73.78 ms** | | Standard deviation | **1.22 ms** | | Stdev / average | **1.7%** | ## Route Context * Departure PoP: [Ashburn (IAD)](/docs/network/latency/iad-ashburn) * Destination PoP: [London (LON)](/docs/network/latency/lon-london) * Region overview: [North America](/docs/network/latency/regions/north-america) → [Europe](/docs/network/latency/regions/europe) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [London to Ashburn latency and RTT](/docs/network/latency/pairs/lon-iad-rtt): 71 ms **Fastest routes departing Ashburn (IAD)** * [Ashburn to New York latency and RTT](/docs/network/latency/pairs/iad-nyc-rtt): 6 ms * [Ashburn to Miami latency and RTT](/docs/network/latency/pairs/iad-mia-rtt): 27.4 ms * [Ashburn to Los Angeles latency and RTT](/docs/network/latency/pairs/iad-lax-rtt): 60.7 ms * [Ashburn to Seattle latency and RTT](/docs/network/latency/pairs/iad-sea-rtt): 62.3 ms * [Ashburn to Amsterdam latency and RTT](/docs/network/latency/pairs/iad-ams-rtt): 75.2 ms **Fastest routes arriving at London (LON)** * [Amsterdam to London latency and RTT](/docs/network/latency/pairs/ams-lon-rtt): 5.2 ms * [Paris to London latency and RTT](/docs/network/latency/pairs/par-lon-rtt): 6.4 ms * [Frankfurt to London latency and RTT](/docs/network/latency/pairs/fra-lon-rtt): 12.9 ms * [Berlin to London latency and RTT](/docs/network/latency/pairs/ber-lon-rtt): 15.2 ms * [Marseille to London latency and RTT](/docs/network/latency/pairs/mrs-lon-rtt): 18.9 ms **Same corridor (North America → Europe)** * [New York to London latency and RTT](/docs/network/latency/pairs/nyc-lon-rtt): 63.5 ms * [New York to Amsterdam latency and RTT](/docs/network/latency/pairs/nyc-ams-rtt): 68.7 ms * [New York to Paris latency and RTT](/docs/network/latency/pairs/nyc-par-rtt): 68.9 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/iad-lon.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Ashburn → London round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/iad-lon-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Ashburn, USA → Melbourne, Australia RTT 🇺🇸 **Ashburn, USA (IAD)** → 🇦🇺 **Melbourne, Australia (MEL)** Measurement round: `2026-08-16T04:07:28Z`. Ashburn to Melbourne averaged 203.9 ms across 50 replies. The observed range was 200.88 to 217.16 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 3.32 ms, or 1.6% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **203.9 ms** | | Jitter | **2.28 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **160.09 ms** | | Fiber Efficiency | **78.5%** | | Latency Tier | Fair | | Source Region | [North America](/docs/network/latency/regions/north-america) | | Destination Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Ashburn](https://www.geonames.org/4744870) (39.04372, -77.48749) and [Melbourne](https://www.geonames.org/2158177) (-37.81400, 144.96332). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | --------------- | | WGS-84 geodesic distance | **16,348.2 km** | | Vacuum RTT floor | **109.06 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **160.09 ms** | | Low-latency fiber material floor | **159.45 ms** | | Engineering floor (5% path allowance) | **168.12 ms** | | Research 1.33× mapped-fiber reference | **212.92 ms** | | Estimated unamplified path loss | **3433.1 dB** | | Transparent optical spans / inline amplifiers | **215 / 214** | | Published RTT inflation over fiber floor | **1.27×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **200.88 ms** | | Average RTT | **203.9 ms** | | Maximum RTT | **217.16 ms** | | Standard deviation | **3.32 ms** | | Stdev / average | **1.6%** | ## Route Context * Departure PoP: [Ashburn (IAD)](/docs/network/latency/iad-ashburn) * Destination PoP: [Melbourne (MEL)](/docs/network/latency/mel-melbourne) * Region overview: [North America](/docs/network/latency/regions/north-america) → [Asia Pacific](/docs/network/latency/regions/asia-pacific) * Corridor overview: [North America to Asia Pacific](/docs/network/latency/regions/north-america-to-asia-pacific) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Melbourne to Ashburn latency and RTT](/docs/network/latency/pairs/mel-iad-rtt): 203.9 ms **Fastest routes departing Ashburn (IAD)** * [Ashburn to New York latency and RTT](/docs/network/latency/pairs/iad-nyc-rtt): 6 ms * [Ashburn to Miami latency and RTT](/docs/network/latency/pairs/iad-mia-rtt): 27.4 ms * [Ashburn to Los Angeles latency and RTT](/docs/network/latency/pairs/iad-lax-rtt): 60.7 ms * [Ashburn to Seattle latency and RTT](/docs/network/latency/pairs/iad-sea-rtt): 62.3 ms * [Ashburn to London latency and RTT](/docs/network/latency/pairs/iad-lon-rtt): 70.3 ms **Fastest routes arriving at Melbourne (MEL)** * [Sydney to Melbourne latency and RTT](/docs/network/latency/pairs/syd-mel-rtt): 9.8 ms * [Singapore to Melbourne latency and RTT](/docs/network/latency/pairs/sin-mel-rtt): 88.6 ms * [Tokyo to Melbourne latency and RTT](/docs/network/latency/pairs/tyo-mel-rtt): 113.9 ms * [Hong Kong to Melbourne latency and RTT](/docs/network/latency/pairs/hkg-mel-rtt): 138.4 ms * [Taipei to Melbourne latency and RTT](/docs/network/latency/pairs/tpe-mel-rtt): 142 ms **Same corridor (North America → Asia Pacific)** * [Seattle to Tokyo latency and RTT](/docs/network/latency/pairs/sea-tyo-rtt): 85 ms * [Los Angeles to Tokyo latency and RTT](/docs/network/latency/pairs/lax-tyo-rtt): 101.2 ms * [Seattle to Taipei latency and RTT](/docs/network/latency/pairs/sea-tpe-rtt): 114.7 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/iad-mel.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Ashburn → Melbourne round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/iad-mel-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Ashburn, USA → Miami, USA RTT 🇺🇸 **Ashburn, USA (IAD)** → 🇺🇸 **Miami, USA (MIA)** Measurement round: `2026-08-16T04:07:28Z`. Ashburn to Miami averaged 27.4 ms across 50 replies. The observed range was 26.09 to 30.01 ms; these figures describe one probe round, not a long-term service target. This is number 2 of 19 outgoing routes by average RTT. That ranking compares different destinations and is not adjusted for distance. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **27.4 ms** | | Jitter | **0.98 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **14.62 ms** | | Fiber Efficiency | **53.4%** | | Latency Tier | Ultra-Low | | Source Region | [North America](/docs/network/latency/regions/north-america) | | Destination Region | [North America](/docs/network/latency/regions/north-america) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Ashburn](https://www.geonames.org/4744870) (39.04372, -77.48749) and [Miami](https://www.geonames.org/4164138) (25.77427, -80.19366). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **1,493.2 km** | | Vacuum RTT floor | **9.96 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **14.62 ms** | | Low-latency fiber material floor | **14.56 ms** | | Engineering floor (5% path allowance) | **15.36 ms** | | Research 1.33× mapped-fiber reference | **19.45 ms** | | Estimated unamplified path loss | **313.6 dB** | | Transparent optical spans / inline amplifiers | **20 / 19** | | Published RTT inflation over fiber floor | **1.87×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------ | | Minimum RTT | **26.09 ms** | | Average RTT | **27.4 ms** | | Maximum RTT | **30.01 ms** | | Standard deviation | **1.06 ms** | | Stdev / average | **3.9%** | ## Route Context * Departure PoP: [Ashburn (IAD)](/docs/network/latency/iad-ashburn) * Destination PoP: [Miami (MIA)](/docs/network/latency/mia-miami) * Region overview: [North America](/docs/network/latency/regions/north-america) → [North America](/docs/network/latency/regions/north-america) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Miami to Ashburn latency and RTT](/docs/network/latency/pairs/mia-iad-rtt): 27.3 ms **Fastest routes departing Ashburn (IAD)** * [Ashburn to New York latency and RTT](/docs/network/latency/pairs/iad-nyc-rtt): 6 ms * [Ashburn to Los Angeles latency and RTT](/docs/network/latency/pairs/iad-lax-rtt): 60.7 ms * [Ashburn to Seattle latency and RTT](/docs/network/latency/pairs/iad-sea-rtt): 62.3 ms * [Ashburn to London latency and RTT](/docs/network/latency/pairs/iad-lon-rtt): 70.3 ms * [Ashburn to Amsterdam latency and RTT](/docs/network/latency/pairs/iad-ams-rtt): 75.2 ms **Fastest routes arriving at Miami (MIA)** * [New York to Miami latency and RTT](/docs/network/latency/pairs/nyc-mia-rtt): 32.9 ms * [Los Angeles to Miami latency and RTT](/docs/network/latency/pairs/lax-mia-rtt): 57.3 ms * [São Paulo to Miami latency and RTT](/docs/network/latency/pairs/gru-mia-rtt): 74 ms * [Seattle to Miami latency and RTT](/docs/network/latency/pairs/sea-mia-rtt): 81.7 ms * [London to Miami latency and RTT](/docs/network/latency/pairs/lon-mia-rtt): 101.9 ms **Same corridor (North America → North America)** * [New York to Ashburn latency and RTT](/docs/network/latency/pairs/nyc-iad-rtt): 6 ms * [Seattle to Los Angeles latency and RTT](/docs/network/latency/pairs/sea-lax-rtt): 25.9 ms * [Los Angeles to Seattle latency and RTT](/docs/network/latency/pairs/lax-sea-rtt): 26.9 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/iad-mia.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Ashburn → Miami round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/iad-mia-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Ashburn, USA → Moscow, Russia RTT 🇺🇸 **Ashburn, USA (IAD)** → 🇷🇺 **Moscow, Russia (MOW)** Measurement round: `2026-08-16T04:07:28Z`. Ashburn to Moscow averaged 112.8 ms across 50 replies. The observed range was 109.18 to 122.64 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 3.63 ms, or 3.2% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **112.8 ms** | | Jitter | **3.16 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **76.88 ms** | | Fiber Efficiency | **68.2%** | | Latency Tier | Good | | Source Region | [North America](/docs/network/latency/regions/north-america) | | Destination Region | [Europe](/docs/network/latency/regions/europe) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Ashburn](https://www.geonames.org/4744870) (39.04372, -77.48749) and [Moscow](https://www.geonames.org/524901) (55.75204, 37.61781). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **7,850.6 km** | | Vacuum RTT floor | **52.37 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **76.88 ms** | | Low-latency fiber material floor | **76.57 ms** | | Engineering floor (5% path allowance) | **80.73 ms** | | Research 1.33× mapped-fiber reference | **102.25 ms** | | Estimated unamplified path loss | **1648.6 dB** | | Transparent optical spans / inline amplifiers | **104 / 103** | | Published RTT inflation over fiber floor | **1.47×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **109.18 ms** | | Average RTT | **112.8 ms** | | Maximum RTT | **122.64 ms** | | Standard deviation | **3.63 ms** | | Stdev / average | **3.2%** | ## Route Context * Departure PoP: [Ashburn (IAD)](/docs/network/latency/iad-ashburn) * Destination PoP: [Moscow (MOW)](/docs/network/latency/mow-moscow) * Region overview: [North America](/docs/network/latency/regions/north-america) → [Europe](/docs/network/latency/regions/europe) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Moscow to Ashburn latency and RTT](/docs/network/latency/pairs/mow-iad-rtt): 113.9 ms **Fastest routes departing Ashburn (IAD)** * [Ashburn to New York latency and RTT](/docs/network/latency/pairs/iad-nyc-rtt): 6 ms * [Ashburn to Miami latency and RTT](/docs/network/latency/pairs/iad-mia-rtt): 27.4 ms * [Ashburn to Los Angeles latency and RTT](/docs/network/latency/pairs/iad-lax-rtt): 60.7 ms * [Ashburn to Seattle latency and RTT](/docs/network/latency/pairs/iad-sea-rtt): 62.3 ms * [Ashburn to London latency and RTT](/docs/network/latency/pairs/iad-lon-rtt): 70.3 ms **Fastest routes arriving at Moscow (MOW)** * [Berlin to Moscow latency and RTT](/docs/network/latency/pairs/ber-mow-rtt): 28.6 ms * [Frankfurt to Moscow latency and RTT](/docs/network/latency/pairs/fra-mow-rtt): 34.7 ms * [Amsterdam to Moscow latency and RTT](/docs/network/latency/pairs/ams-mow-rtt): 38.1 ms * [London to Moscow latency and RTT](/docs/network/latency/pairs/lon-mow-rtt): 43.5 ms * [Paris to Moscow latency and RTT](/docs/network/latency/pairs/par-mow-rtt): 44.7 ms **Same corridor (North America → Europe)** * [New York to London latency and RTT](/docs/network/latency/pairs/nyc-lon-rtt): 63.5 ms * [New York to Amsterdam latency and RTT](/docs/network/latency/pairs/nyc-ams-rtt): 68.7 ms * [New York to Paris latency and RTT](/docs/network/latency/pairs/nyc-par-rtt): 68.9 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/iad-mow.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Ashburn → Moscow round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/iad-mow-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Ashburn, USA → Marseille, France RTT 🇺🇸 **Ashburn, USA (IAD)** → 🇫🇷 **Marseille, France (MRS)** Measurement round: `2026-08-16T04:07:28Z`. Ashburn to Marseille averaged 86.5 ms across 50 replies. The observed range was 81.69 to 97.62 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 3.63 ms, or 4.2% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **86.5 ms** | | Jitter | **3.41 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **65.34 ms** | | Fiber Efficiency | **75.5%** | | Latency Tier | Good | | Source Region | [North America](/docs/network/latency/regions/north-america) | | Destination Region | [Europe](/docs/network/latency/regions/europe) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Ashburn](https://www.geonames.org/4744870) (39.04372, -77.48749) and [Marseille](https://www.geonames.org/2995469) (43.29695, 5.38107). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **6,672.7 km** | | Vacuum RTT floor | **44.52 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **65.34 ms** | | Low-latency fiber material floor | **65.08 ms** | | Engineering floor (5% path allowance) | **68.62 ms** | | Research 1.33× mapped-fiber reference | **86.91 ms** | | Estimated unamplified path loss | **1401.3 dB** | | Transparent optical spans / inline amplifiers | **88 / 87** | | Published RTT inflation over fiber floor | **1.32×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------ | | Minimum RTT | **81.69 ms** | | Average RTT | **86.5 ms** | | Maximum RTT | **97.62 ms** | | Standard deviation | **3.63 ms** | | Stdev / average | **4.2%** | ## Route Context * Departure PoP: [Ashburn (IAD)](/docs/network/latency/iad-ashburn) * Destination PoP: [Marseille (MRS)](/docs/network/latency/mrs-marseille) * Region overview: [North America](/docs/network/latency/regions/north-america) → [Europe](/docs/network/latency/regions/europe) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Marseille to Ashburn latency and RTT](/docs/network/latency/pairs/mrs-iad-rtt): 85.7 ms **Fastest routes departing Ashburn (IAD)** * [Ashburn to New York latency and RTT](/docs/network/latency/pairs/iad-nyc-rtt): 6 ms * [Ashburn to Miami latency and RTT](/docs/network/latency/pairs/iad-mia-rtt): 27.4 ms * [Ashburn to Los Angeles latency and RTT](/docs/network/latency/pairs/iad-lax-rtt): 60.7 ms * [Ashburn to Seattle latency and RTT](/docs/network/latency/pairs/iad-sea-rtt): 62.3 ms * [Ashburn to London latency and RTT](/docs/network/latency/pairs/iad-lon-rtt): 70.3 ms **Fastest routes arriving at Marseille (MRS)** * [Paris to Marseille latency and RTT](/docs/network/latency/pairs/par-mrs-rtt): 8.9 ms * [Frankfurt to Marseille latency and RTT](/docs/network/latency/pairs/fra-mrs-rtt): 16.4 ms * [London to Marseille latency and RTT](/docs/network/latency/pairs/lon-mrs-rtt): 17.9 ms * [Amsterdam to Marseille latency and RTT](/docs/network/latency/pairs/ams-mrs-rtt): 19.6 ms * [Berlin to Marseille latency and RTT](/docs/network/latency/pairs/ber-mrs-rtt): 20.2 ms **Same corridor (North America → Europe)** * [New York to London latency and RTT](/docs/network/latency/pairs/nyc-lon-rtt): 63.5 ms * [New York to Amsterdam latency and RTT](/docs/network/latency/pairs/nyc-ams-rtt): 68.7 ms * [New York to Paris latency and RTT](/docs/network/latency/pairs/nyc-par-rtt): 68.9 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/iad-mrs.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Ashburn → Marseille round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/iad-mrs-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Ashburn, USA → New York, USA RTT 🇺🇸 **Ashburn, USA (IAD)** → 🇺🇸 **New York, USA (NYC)** Measurement round: `2026-08-16T04:07:28Z`. Ashburn to New York averaged 6 ms across 50 replies. The observed range was 5.77 to 6.79 ms; these figures describe one probe round, not a long-term service target. This is number 1 of 19 outgoing routes by average RTT. That ranking compares different destinations and is not adjusted for distance. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **6 ms** | | Jitter | **0.19 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **3.44 ms** | | Fiber Efficiency | **57.3%** | | Latency Tier | Ultra-Low | | Source Region | [North America](/docs/network/latency/regions/north-america) | | Destination Region | [North America](/docs/network/latency/regions/north-america) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Ashburn](https://www.geonames.org/4744870) (39.04372, -77.48749) and [New York City](https://www.geonames.org/5128581) (40.71427, -74.00597). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | ------------ | | WGS-84 geodesic distance | **350.8 km** | | Vacuum RTT floor | **2.34 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **3.44 ms** | | Low-latency fiber material floor | **3.42 ms** | | Engineering floor (5% path allowance) | **3.61 ms** | | Research 1.33× mapped-fiber reference | **4.57 ms** | | Estimated unamplified path loss | **73.7 dB** | | Transparent optical spans / inline amplifiers | **5 / 4** | | Published RTT inflation over fiber floor | **1.75×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ----------- | | Minimum RTT | **5.77 ms** | | Average RTT | **6 ms** | | Maximum RTT | **6.79 ms** | | Standard deviation | **0.2 ms** | | Stdev / average | **3.3%** | ## Route Context * Departure PoP: [Ashburn (IAD)](/docs/network/latency/iad-ashburn) * Destination PoP: [New York (NYC)](/docs/network/latency/nyc-new-york) * Region overview: [North America](/docs/network/latency/regions/north-america) → [North America](/docs/network/latency/regions/north-america) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [New York to Ashburn latency and RTT](/docs/network/latency/pairs/nyc-iad-rtt): 6 ms **Fastest routes departing Ashburn (IAD)** * [Ashburn to Miami latency and RTT](/docs/network/latency/pairs/iad-mia-rtt): 27.4 ms * [Ashburn to Los Angeles latency and RTT](/docs/network/latency/pairs/iad-lax-rtt): 60.7 ms * [Ashburn to Seattle latency and RTT](/docs/network/latency/pairs/iad-sea-rtt): 62.3 ms * [Ashburn to London latency and RTT](/docs/network/latency/pairs/iad-lon-rtt): 70.3 ms * [Ashburn to Amsterdam latency and RTT](/docs/network/latency/pairs/iad-ams-rtt): 75.2 ms **Fastest routes arriving at New York (NYC)** * [Miami to New York latency and RTT](/docs/network/latency/pairs/mia-nyc-rtt): 33.3 ms * [Los Angeles to New York latency and RTT](/docs/network/latency/pairs/lax-nyc-rtt): 57.9 ms * [Seattle to New York latency and RTT](/docs/network/latency/pairs/sea-nyc-rtt): 59.5 ms * [London to New York latency and RTT](/docs/network/latency/pairs/lon-nyc-rtt): 63.8 ms * [Amsterdam to New York latency and RTT](/docs/network/latency/pairs/ams-nyc-rtt): 68.6 ms **Same corridor (North America → North America)** * [Seattle to Los Angeles latency and RTT](/docs/network/latency/pairs/sea-lax-rtt): 25.9 ms * [Los Angeles to Seattle latency and RTT](/docs/network/latency/pairs/lax-sea-rtt): 26.9 ms * [Miami to Ashburn latency and RTT](/docs/network/latency/pairs/mia-iad-rtt): 27.3 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/iad-nyc.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Ashburn → New York round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/iad-nyc-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Ashburn, USA → Paris, France RTT 🇺🇸 **Ashburn, USA (IAD)** → 🇫🇷 **Paris, France (PAR)** Measurement round: `2026-08-16T04:07:28Z`. Ashburn to Paris averaged 76.1 ms across 50 replies. The observed range was 74.36 to 85 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 1.76 ms, or 2.3% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **76.1 ms** | | Jitter | **1.12 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **60.73 ms** | | Fiber Efficiency | **79.8%** | | Latency Tier | Excellent | | Source Region | [North America](/docs/network/latency/regions/north-america) | | Destination Region | [Europe](/docs/network/latency/regions/europe) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Ashburn](https://www.geonames.org/4744870) (39.04372, -77.48749) and [Paris](https://www.geonames.org/2988507) (48.85341, 2.34880). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | ------------- | | WGS-84 geodesic distance | **6,202 km** | | Vacuum RTT floor | **41.38 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **60.73 ms** | | Low-latency fiber material floor | **60.49 ms** | | Engineering floor (5% path allowance) | **63.78 ms** | | Research 1.33× mapped-fiber reference | **80.78 ms** | | Estimated unamplified path loss | **1302.4 dB** | | Transparent optical spans / inline amplifiers | **82 / 81** | | Published RTT inflation over fiber floor | **1.25×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------ | | Minimum RTT | **74.36 ms** | | Average RTT | **76.1 ms** | | Maximum RTT | **85 ms** | | Standard deviation | **1.76 ms** | | Stdev / average | **2.3%** | ## Route Context * Departure PoP: [Ashburn (IAD)](/docs/network/latency/iad-ashburn) * Destination PoP: [Paris (PAR)](/docs/network/latency/par-paris) * Region overview: [North America](/docs/network/latency/regions/north-america) → [Europe](/docs/network/latency/regions/europe) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Paris to Ashburn latency and RTT](/docs/network/latency/pairs/par-iad-rtt): 75.5 ms **Fastest routes departing Ashburn (IAD)** * [Ashburn to New York latency and RTT](/docs/network/latency/pairs/iad-nyc-rtt): 6 ms * [Ashburn to Miami latency and RTT](/docs/network/latency/pairs/iad-mia-rtt): 27.4 ms * [Ashburn to Los Angeles latency and RTT](/docs/network/latency/pairs/iad-lax-rtt): 60.7 ms * [Ashburn to Seattle latency and RTT](/docs/network/latency/pairs/iad-sea-rtt): 62.3 ms * [Ashburn to London latency and RTT](/docs/network/latency/pairs/iad-lon-rtt): 70.3 ms **Fastest routes arriving at Paris (PAR)** * [London to Paris latency and RTT](/docs/network/latency/pairs/lon-par-rtt): 6.4 ms * [Amsterdam to Paris latency and RTT](/docs/network/latency/pairs/ams-par-rtt): 7 ms * [Frankfurt to Paris latency and RTT](/docs/network/latency/pairs/fra-par-rtt): 7.6 ms * [Marseille to Paris latency and RTT](/docs/network/latency/pairs/mrs-par-rtt): 8.9 ms * [Berlin to Paris latency and RTT](/docs/network/latency/pairs/ber-par-rtt): 15.6 ms **Same corridor (North America → Europe)** * [New York to London latency and RTT](/docs/network/latency/pairs/nyc-lon-rtt): 63.5 ms * [New York to Amsterdam latency and RTT](/docs/network/latency/pairs/nyc-ams-rtt): 68.7 ms * [New York to Paris latency and RTT](/docs/network/latency/pairs/nyc-par-rtt): 68.9 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/iad-par.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Ashburn → Paris round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/iad-par-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Ashburn, USA → Seattle, USA RTT 🇺🇸 **Ashburn, USA (IAD)** → 🇺🇸 **Seattle, USA (SEA)** Measurement round: `2026-08-16T04:07:28Z`. Ashburn to Seattle averaged 62.3 ms across 50 replies. The observed range was 59.37 to 75.24 ms; these figures describe one probe round, not a long-term service target. This is number 4 of 19 outgoing routes by average RTT. That ranking compares different destinations and is not adjusted for distance. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **62.3 ms** | | Jitter | **1.89 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **36.27 ms** | | Fiber Efficiency | **58.2%** | | Latency Tier | Excellent | | Source Region | [North America](/docs/network/latency/regions/north-america) | | Destination Region | [North America](/docs/network/latency/regions/north-america) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Ashburn](https://www.geonames.org/4744870) (39.04372, -77.48749) and [Seattle](https://www.geonames.org/5809844) (47.60621, -122.33207). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | ------------ | | WGS-84 geodesic distance | **3,704 km** | | Vacuum RTT floor | **24.71 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **36.27 ms** | | Low-latency fiber material floor | **36.13 ms** | | Engineering floor (5% path allowance) | **38.09 ms** | | Research 1.33× mapped-fiber reference | **48.24 ms** | | Estimated unamplified path loss | **777.8 dB** | | Transparent optical spans / inline amplifiers | **49 / 48** | | Published RTT inflation over fiber floor | **1.72×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------ | | Minimum RTT | **59.37 ms** | | Average RTT | **62.3 ms** | | Maximum RTT | **75.24 ms** | | Standard deviation | **2.67 ms** | | Stdev / average | **4.3%** | ## Route Context * Departure PoP: [Ashburn (IAD)](/docs/network/latency/iad-ashburn) * Destination PoP: [Seattle (SEA)](/docs/network/latency/sea-seattle) * Region overview: [North America](/docs/network/latency/regions/north-america) → [North America](/docs/network/latency/regions/north-america) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Seattle to Ashburn latency and RTT](/docs/network/latency/pairs/sea-iad-rtt): 61.5 ms **Fastest routes departing Ashburn (IAD)** * [Ashburn to New York latency and RTT](/docs/network/latency/pairs/iad-nyc-rtt): 6 ms * [Ashburn to Miami latency and RTT](/docs/network/latency/pairs/iad-mia-rtt): 27.4 ms * [Ashburn to Los Angeles latency and RTT](/docs/network/latency/pairs/iad-lax-rtt): 60.7 ms * [Ashburn to London latency and RTT](/docs/network/latency/pairs/iad-lon-rtt): 70.3 ms * [Ashburn to Amsterdam latency and RTT](/docs/network/latency/pairs/iad-ams-rtt): 75.2 ms **Fastest routes arriving at Seattle (SEA)** * [Los Angeles to Seattle latency and RTT](/docs/network/latency/pairs/lax-sea-rtt): 26.9 ms * [New York to Seattle latency and RTT](/docs/network/latency/pairs/nyc-sea-rtt): 58.7 ms * [Miami to Seattle latency and RTT](/docs/network/latency/pairs/mia-sea-rtt): 81.9 ms * [Tokyo to Seattle latency and RTT](/docs/network/latency/pairs/tyo-sea-rtt): 84.8 ms * [Taipei to Seattle latency and RTT](/docs/network/latency/pairs/tpe-sea-rtt): 115.8 ms **Same corridor (North America → North America)** * [New York to Ashburn latency and RTT](/docs/network/latency/pairs/nyc-iad-rtt): 6 ms * [Seattle to Los Angeles latency and RTT](/docs/network/latency/pairs/sea-lax-rtt): 25.9 ms * [Miami to Ashburn latency and RTT](/docs/network/latency/pairs/mia-iad-rtt): 27.3 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/iad-sea.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Ashburn → Seattle round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/iad-sea-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Ashburn, USA → Singapore RTT 🇺🇸 **Ashburn, USA (IAD)** → 🇸🇬 **Singapore (SIN)** Measurement round: `2026-08-16T04:07:28Z`. Ashburn to Singapore averaged 211.5 ms across 50 replies. The observed range was 203.83 to 243.56 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 6.36 ms, or 3.0% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **211.5 ms** | | Jitter | **5.45 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **152.14 ms** | | Fiber Efficiency | **71.9%** | | Latency Tier | Fair | | Source Region | [North America](/docs/network/latency/regions/north-america) | | Destination Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Ashburn](https://www.geonames.org/4744870) (39.04372, -77.48749) and [Singapore](https://www.geonames.org/1880252) (1.28967, 103.85007). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | --------------- | | WGS-84 geodesic distance | **15,535.9 km** | | Vacuum RTT floor | **103.64 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **152.14 ms** | | Low-latency fiber material floor | **151.53 ms** | | Engineering floor (5% path allowance) | **159.77 ms** | | Research 1.33× mapped-fiber reference | **202.35 ms** | | Estimated unamplified path loss | **3262.5 dB** | | Transparent optical spans / inline amplifiers | **204 / 203** | | Published RTT inflation over fiber floor | **1.39×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **203.83 ms** | | Average RTT | **211.5 ms** | | Maximum RTT | **243.56 ms** | | Standard deviation | **6.36 ms** | | Stdev / average | **3.0%** | ## Route Context * Departure PoP: [Ashburn (IAD)](/docs/network/latency/iad-ashburn) * Destination PoP: [Singapore (SIN)](/docs/network/latency/sin-singapore) * Region overview: [North America](/docs/network/latency/regions/north-america) → [Asia Pacific](/docs/network/latency/regions/asia-pacific) * Corridor overview: [North America to Asia Pacific](/docs/network/latency/regions/north-america-to-asia-pacific) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Singapore to Ashburn latency and RTT](/docs/network/latency/pairs/sin-iad-rtt): 210 ms **Fastest routes departing Ashburn (IAD)** * [Ashburn to New York latency and RTT](/docs/network/latency/pairs/iad-nyc-rtt): 6 ms * [Ashburn to Miami latency and RTT](/docs/network/latency/pairs/iad-mia-rtt): 27.4 ms * [Ashburn to Los Angeles latency and RTT](/docs/network/latency/pairs/iad-lax-rtt): 60.7 ms * [Ashburn to Seattle latency and RTT](/docs/network/latency/pairs/iad-sea-rtt): 62.3 ms * [Ashburn to London latency and RTT](/docs/network/latency/pairs/iad-lon-rtt): 70.3 ms **Fastest routes arriving at Singapore (SIN)** * [Hong Kong to Singapore latency and RTT](/docs/network/latency/pairs/hkg-sin-rtt): 31.5 ms * [Taipei to Singapore latency and RTT](/docs/network/latency/pairs/tpe-sin-rtt): 44.3 ms * [Tokyo to Singapore latency and RTT](/docs/network/latency/pairs/tyo-sin-rtt): 69.4 ms * [Melbourne to Singapore latency and RTT](/docs/network/latency/pairs/mel-sin-rtt): 88.4 ms * [Sydney to Singapore latency and RTT](/docs/network/latency/pairs/syd-sin-rtt): 94.5 ms **Same corridor (North America → Asia Pacific)** * [Seattle to Tokyo latency and RTT](/docs/network/latency/pairs/sea-tyo-rtt): 85 ms * [Los Angeles to Tokyo latency and RTT](/docs/network/latency/pairs/lax-tyo-rtt): 101.2 ms * [Seattle to Taipei latency and RTT](/docs/network/latency/pairs/sea-tpe-rtt): 114.7 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/iad-sin.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Ashburn → Singapore round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/iad-sin-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Ashburn, USA → Sydney, Australia RTT 🇺🇸 **Ashburn, USA (IAD)** → 🇦🇺 **Sydney, Australia (SYD)** Measurement round: `2026-08-16T04:07:28Z`. Ashburn to Sydney averaged 193.6 ms across 50 replies. The observed range was 185.3 to 213.55 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 6.56 ms, or 3.4% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **193.6 ms** | | Jitter | **5.83 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **153.48 ms** | | Fiber Efficiency | **79.3%** | | Latency Tier | Fair | | Source Region | [North America](/docs/network/latency/regions/north-america) | | Destination Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Ashburn](https://www.geonames.org/4744870) (39.04372, -77.48749) and [Sydney](https://www.geonames.org/2147714) (-33.86785, 151.20732). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | --------------- | | WGS-84 geodesic distance | **15,673.1 km** | | Vacuum RTT floor | **104.56 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **153.48 ms** | | Low-latency fiber material floor | **152.87 ms** | | Engineering floor (5% path allowance) | **161.18 ms** | | Research 1.33× mapped-fiber reference | **204.13 ms** | | Estimated unamplified path loss | **3291.4 dB** | | Transparent optical spans / inline amplifiers | **206 / 205** | | Published RTT inflation over fiber floor | **1.26×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **185.3 ms** | | Average RTT | **193.6 ms** | | Maximum RTT | **213.55 ms** | | Standard deviation | **6.56 ms** | | Stdev / average | **3.4%** | ## Route Context * Departure PoP: [Ashburn (IAD)](/docs/network/latency/iad-ashburn) * Destination PoP: [Sydney (SYD)](/docs/network/latency/syd-sydney) * Region overview: [North America](/docs/network/latency/regions/north-america) → [Asia Pacific](/docs/network/latency/regions/asia-pacific) * Corridor overview: [North America to Asia Pacific](/docs/network/latency/regions/north-america-to-asia-pacific) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Sydney to Ashburn latency and RTT](/docs/network/latency/pairs/syd-iad-rtt): 195.3 ms **Fastest routes departing Ashburn (IAD)** * [Ashburn to New York latency and RTT](/docs/network/latency/pairs/iad-nyc-rtt): 6 ms * [Ashburn to Miami latency and RTT](/docs/network/latency/pairs/iad-mia-rtt): 27.4 ms * [Ashburn to Los Angeles latency and RTT](/docs/network/latency/pairs/iad-lax-rtt): 60.7 ms * [Ashburn to Seattle latency and RTT](/docs/network/latency/pairs/iad-sea-rtt): 62.3 ms * [Ashburn to London latency and RTT](/docs/network/latency/pairs/iad-lon-rtt): 70.3 ms **Fastest routes arriving at Sydney (SYD)** * [Melbourne to Sydney latency and RTT](/docs/network/latency/pairs/mel-syd-rtt): 9.8 ms * [Singapore to Sydney latency and RTT](/docs/network/latency/pairs/sin-syd-rtt): 94.5 ms * [Tokyo to Sydney latency and RTT](/docs/network/latency/pairs/tyo-syd-rtt): 101.9 ms * [Taipei to Sydney latency and RTT](/docs/network/latency/pairs/tpe-syd-rtt): 133.2 ms * [Hong Kong to Sydney latency and RTT](/docs/network/latency/pairs/hkg-syd-rtt): 137.2 ms **Same corridor (North America → Asia Pacific)** * [Seattle to Tokyo latency and RTT](/docs/network/latency/pairs/sea-tyo-rtt): 85 ms * [Los Angeles to Tokyo latency and RTT](/docs/network/latency/pairs/lax-tyo-rtt): 101.2 ms * [Seattle to Taipei latency and RTT](/docs/network/latency/pairs/sea-tpe-rtt): 114.7 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/iad-syd.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Ashburn → Sydney round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/iad-syd-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Ashburn, USA → Taipei, Taiwan RTT 🇺🇸 **Ashburn, USA (IAD)** → 🇹🇼 **Taipei, Taiwan (TPE)** Measurement round: `2026-08-16T04:07:28Z`. Ashburn to Taipei averaged 172.3 ms across 50 replies. The observed range was 164.68 to 192.64 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 7.13 ms, or 4.1% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **172.3 ms** | | Jitter | **6.51 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **123.78 ms** | | Fiber Efficiency | **71.8%** | | Latency Tier | Fair | | Source Region | [North America](/docs/network/latency/regions/north-america) | | Destination Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Ashburn](https://www.geonames.org/4744870) (39.04372, -77.48749) and [Taipei](https://www.geonames.org/1668341) (25.05306, 121.52639). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | --------------- | | WGS-84 geodesic distance | **12,639.6 km** | | Vacuum RTT floor | **84.32 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **123.78 ms** | | Low-latency fiber material floor | **123.28 ms** | | Engineering floor (5% path allowance) | **129.98 ms** | | Research 1.33× mapped-fiber reference | **164.62 ms** | | Estimated unamplified path loss | **2654.3 dB** | | Transparent optical spans / inline amplifiers | **166 / 165** | | Published RTT inflation over fiber floor | **1.39×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **164.68 ms** | | Average RTT | **172.3 ms** | | Maximum RTT | **192.64 ms** | | Standard deviation | **7.13 ms** | | Stdev / average | **4.1%** | ## Route Context * Departure PoP: [Ashburn (IAD)](/docs/network/latency/iad-ashburn) * Destination PoP: [Taipei (TPE)](/docs/network/latency/tpe-taipei) * Region overview: [North America](/docs/network/latency/regions/north-america) → [Asia Pacific](/docs/network/latency/regions/asia-pacific) * Corridor overview: [North America to Asia Pacific](/docs/network/latency/regions/north-america-to-asia-pacific) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Taipei to Ashburn latency and RTT](/docs/network/latency/pairs/tpe-iad-rtt): 171.6 ms **Fastest routes departing Ashburn (IAD)** * [Ashburn to New York latency and RTT](/docs/network/latency/pairs/iad-nyc-rtt): 6 ms * [Ashburn to Miami latency and RTT](/docs/network/latency/pairs/iad-mia-rtt): 27.4 ms * [Ashburn to Los Angeles latency and RTT](/docs/network/latency/pairs/iad-lax-rtt): 60.7 ms * [Ashburn to Seattle latency and RTT](/docs/network/latency/pairs/iad-sea-rtt): 62.3 ms * [Ashburn to London latency and RTT](/docs/network/latency/pairs/iad-lon-rtt): 70.3 ms **Fastest routes arriving at Taipei (TPE)** * [Hong Kong to Taipei latency and RTT](/docs/network/latency/pairs/hkg-tpe-rtt): 14.7 ms * [Tokyo to Taipei latency and RTT](/docs/network/latency/pairs/tyo-tpe-rtt): 32.3 ms * [Singapore to Taipei latency and RTT](/docs/network/latency/pairs/sin-tpe-rtt): 45.8 ms * [Seattle to Taipei latency and RTT](/docs/network/latency/pairs/sea-tpe-rtt): 114.7 ms * [Sydney to Taipei latency and RTT](/docs/network/latency/pairs/syd-tpe-rtt): 131.7 ms **Same corridor (North America → Asia Pacific)** * [Seattle to Tokyo latency and RTT](/docs/network/latency/pairs/sea-tyo-rtt): 85 ms * [Los Angeles to Tokyo latency and RTT](/docs/network/latency/pairs/lax-tyo-rtt): 101.2 ms * [Seattle to Hong Kong latency and RTT](/docs/network/latency/pairs/sea-hkg-rtt): 130.4 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/iad-tpe.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Ashburn → Taipei round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/iad-tpe-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Ashburn, USA → Tokyo, Japan RTT 🇺🇸 **Ashburn, USA (IAD)** → 🇯🇵 **Tokyo, Japan (TYO)** Measurement round: `2026-08-16T04:07:28Z`. Ashburn to Tokyo averaged 141.9 ms across 50 replies. The observed range was 136.1 to 154.55 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 4.32 ms, or 3.0% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **141.9 ms** | | Jitter | **3.99 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **106.69 ms** | | Fiber Efficiency | **75.2%** | | Latency Tier | Good | | Source Region | [North America](/docs/network/latency/regions/north-america) | | Destination Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Ashburn](https://www.geonames.org/4744870) (39.04372, -77.48749) and [Tokyo](https://www.geonames.org/1850147) (35.68950, 139.69171). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | --------------- | | WGS-84 geodesic distance | **10,894.6 km** | | Vacuum RTT floor | **72.68 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **106.69 ms** | | Low-latency fiber material floor | **106.26 ms** | | Engineering floor (5% path allowance) | **112.04 ms** | | Research 1.33× mapped-fiber reference | **141.9 ms** | | Estimated unamplified path loss | **2287.9 dB** | | Transparent optical spans / inline amplifiers | **143 / 142** | | Published RTT inflation over fiber floor | **1.33×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **136.1 ms** | | Average RTT | **141.9 ms** | | Maximum RTT | **154.55 ms** | | Standard deviation | **4.32 ms** | | Stdev / average | **3.0%** | ## Route Context * Departure PoP: [Ashburn (IAD)](/docs/network/latency/iad-ashburn) * Destination PoP: [Tokyo (TYO)](/docs/network/latency/tyo-tokyo) * Region overview: [North America](/docs/network/latency/regions/north-america) → [Asia Pacific](/docs/network/latency/regions/asia-pacific) * Corridor overview: [North America to Asia Pacific](/docs/network/latency/regions/north-america-to-asia-pacific) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Tokyo to Ashburn latency and RTT](/docs/network/latency/pairs/tyo-iad-rtt): 142.9 ms **Fastest routes departing Ashburn (IAD)** * [Ashburn to New York latency and RTT](/docs/network/latency/pairs/iad-nyc-rtt): 6 ms * [Ashburn to Miami latency and RTT](/docs/network/latency/pairs/iad-mia-rtt): 27.4 ms * [Ashburn to Los Angeles latency and RTT](/docs/network/latency/pairs/iad-lax-rtt): 60.7 ms * [Ashburn to Seattle latency and RTT](/docs/network/latency/pairs/iad-sea-rtt): 62.3 ms * [Ashburn to London latency and RTT](/docs/network/latency/pairs/iad-lon-rtt): 70.3 ms **Fastest routes arriving at Tokyo (TYO)** * [Taipei to Tokyo latency and RTT](/docs/network/latency/pairs/tpe-tyo-rtt): 31.9 ms * [Hong Kong to Tokyo latency and RTT](/docs/network/latency/pairs/hkg-tyo-rtt): 44.9 ms * [Singapore to Tokyo latency and RTT](/docs/network/latency/pairs/sin-tyo-rtt): 68.9 ms * [Seattle to Tokyo latency and RTT](/docs/network/latency/pairs/sea-tyo-rtt): 85 ms * [Los Angeles to Tokyo latency and RTT](/docs/network/latency/pairs/lax-tyo-rtt): 101.2 ms **Same corridor (North America → Asia Pacific)** * [Seattle to Taipei latency and RTT](/docs/network/latency/pairs/sea-tpe-rtt): 114.7 ms * [Seattle to Hong Kong latency and RTT](/docs/network/latency/pairs/sea-hkg-rtt): 130.4 ms * [Los Angeles to Taipei latency and RTT](/docs/network/latency/pairs/lax-tpe-rtt): 132.9 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/iad-tyo.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Ashburn → Tokyo round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/iad-tyo-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # City-to-City Ping & RTT Latency | AS203314 The directory contains 380 directed measurements between Hats Network PoPs. Each route has its own probe round, propagation reference and downloadable data. The reverse direction is a separate measurement. Measurement round: `2026-08-16T04:07:28Z`. ## Browse by Source City City pages compare incoming and outgoing routes for one PoP. | Source PoP | Region | Routes | | ----------------------------------------------------------------- | ------------- | ----------- | | [🇳🇱 Amsterdam (AMS)](/docs/network/latency/ams-amsterdam) | Europe | 19 outbound | | [🇩🇪 Berlin (BER)](/docs/network/latency/ber-berlin) | Europe | 19 outbound | | [🇩🇪 Frankfurt (FRA)](/docs/network/latency/fra-frankfurt) | Europe | 19 outbound | | [🇬🇧 London (LON)](/docs/network/latency/lon-london) | Europe | 19 outbound | | [🇫🇷 Marseille (MRS)](/docs/network/latency/mrs-marseille) | Europe | 19 outbound | | [🇷🇺 Moscow (MOW)](/docs/network/latency/mow-moscow) | Europe | 19 outbound | | [🇫🇷 Paris (PAR)](/docs/network/latency/par-paris) | Europe | 19 outbound | | [🇺🇸 Ashburn (IAD)](/docs/network/latency/iad-ashburn) | North America | 19 outbound | | [🇺🇸 Los Angeles (LAX)](/docs/network/latency/lax-los-angeles) | North America | 19 outbound | | [🇺🇸 Miami (MIA)](/docs/network/latency/mia-miami) | North America | 19 outbound | | [🇺🇸 New York (NYC)](/docs/network/latency/nyc-new-york) | North America | 19 outbound | | [🇺🇸 Seattle (SEA)](/docs/network/latency/sea-seattle) | North America | 19 outbound | | [🇧🇷 São Paulo (GRU)](/docs/network/latency/gru-sao-paulo) | South America | 19 outbound | | [🇭🇰 Hong Kong (HKG)](/docs/network/latency/hkg-hong-kong) | Asia Pacific | 19 outbound | | [🇦🇺 Melbourne (MEL)](/docs/network/latency/mel-melbourne) | Asia Pacific | 19 outbound | | [🇸🇬 Singapore (SIN)](/docs/network/latency/sin-singapore) | Asia Pacific | 19 outbound | | [🇦🇺 Sydney (SYD)](/docs/network/latency/syd-sydney) | Asia Pacific | 19 outbound | | [🇹🇼 Taipei (TPE)](/docs/network/latency/tpe-taipei) | Asia Pacific | 19 outbound | | [🇯🇵 Tokyo (TYO)](/docs/network/latency/tyo-tokyo) | Asia Pacific | 19 outbound | | [🇿🇦 Johannesburg (JNB)](/docs/network/latency/jnb-johannesburg) | Africa | 19 outbound | ## Fastest City-Pair Routes | Route | RTT (ms) | Tier | | ------------------------------------------- | -------------------------------------------------- | --------- | | 🇳🇱 Amsterdam (AMS) → 🇬🇧 London (LON) | **[5.2](/docs/network/latency/pairs/ams-lon-rtt)** | Ultra-Low | | 🇬🇧 London (LON) → 🇳🇱 Amsterdam (AMS) | **[5.2](/docs/network/latency/pairs/lon-ams-rtt)** | Ultra-Low | | 🇳🇱 Amsterdam (AMS) → 🇩🇪 Frankfurt (FRA) | **[5.9](/docs/network/latency/pairs/ams-fra-rtt)** | Ultra-Low | | 🇩🇪 Frankfurt (FRA) → 🇳🇱 Amsterdam (AMS) | **[5.9](/docs/network/latency/pairs/fra-ams-rtt)** | Ultra-Low | | 🇺🇸 Ashburn (IAD) → 🇺🇸 New York (NYC) | **[6](/docs/network/latency/pairs/iad-nyc-rtt)** | Ultra-Low | | 🇺🇸 New York (NYC) → 🇺🇸 Ashburn (IAD) | **[6](/docs/network/latency/pairs/nyc-iad-rtt)** | Ultra-Low | | 🇩🇪 Berlin (BER) → 🇩🇪 Frankfurt (FRA) | **[6.1](/docs/network/latency/pairs/ber-fra-rtt)** | Ultra-Low | | 🇩🇪 Frankfurt (FRA) → 🇩🇪 Berlin (BER) | **[6.1](/docs/network/latency/pairs/fra-ber-rtt)** | Ultra-Low | | 🇬🇧 London (LON) → 🇫🇷 Paris (PAR) | **[6.4](/docs/network/latency/pairs/lon-par-rtt)** | Ultra-Low | | 🇫🇷 Paris (PAR) → 🇬🇧 London (LON) | **[6.4](/docs/network/latency/pairs/par-lon-rtt)** | Ultra-Low | ## All 380 City-Pair Routes Routes are grouped by origin. Choose a destination to inspect the underlying replies. ### Routes from Europe (133) **🇳🇱 Amsterdam (AMS): 19 outbound routes** | Destination | RTT | Tier | | ------------------------------------------------------------------ | -----------: | --------- | | [🇩🇪 Berlin (BER)](/docs/network/latency/pairs/ams-ber-rtt) | **7.2 ms** | Ultra-Low | | [🇩🇪 Frankfurt (FRA)](/docs/network/latency/pairs/ams-fra-rtt) | **5.9 ms** | Ultra-Low | | [🇬🇧 London (LON)](/docs/network/latency/pairs/ams-lon-rtt) | **5.2 ms** | Ultra-Low | | [🇫🇷 Marseille (MRS)](/docs/network/latency/pairs/ams-mrs-rtt) | **19.6 ms** | Ultra-Low | | [🇷🇺 Moscow (MOW)](/docs/network/latency/pairs/ams-mow-rtt) | **38.1 ms** | Excellent | | [🇫🇷 Paris (PAR)](/docs/network/latency/pairs/ams-par-rtt) | **7 ms** | Ultra-Low | | [🇺🇸 Ashburn (IAD)](/docs/network/latency/pairs/ams-iad-rtt) | **75.7 ms** | Excellent | | [🇺🇸 Los Angeles (LAX)](/docs/network/latency/pairs/ams-lax-rtt) | **133.7 ms** | Good | | [🇺🇸 Miami (MIA)](/docs/network/latency/pairs/ams-mia-rtt) | **106 ms** | Good | | [🇺🇸 New York (NYC)](/docs/network/latency/pairs/ams-nyc-rtt) | **68.6 ms** | Excellent | | [🇺🇸 Seattle (SEA)](/docs/network/latency/pairs/ams-sea-rtt) | **130.2 ms** | Good | | [🇧🇷 São Paulo (GRU)](/docs/network/latency/pairs/ams-gru-rtt) | **175.5 ms** | Fair | | [🇭🇰 Hong Kong (HKG)](/docs/network/latency/pairs/ams-hkg-rtt) | **155.9 ms** | Fair | | [🇦🇺 Melbourne (MEL)](/docs/network/latency/pairs/ams-mel-rtt) | **252.5 ms** | High | | [🇸🇬 Singapore (SIN)](/docs/network/latency/pairs/ams-sin-rtt) | **157.6 ms** | Fair | | [🇦🇺 Sydney (SYD)](/docs/network/latency/pairs/ams-syd-rtt) | **258 ms** | High | | [🇹🇼 Taipei (TPE)](/docs/network/latency/pairs/ams-tpe-rtt) | **168.7 ms** | Fair | | [🇯🇵 Tokyo (TYO)](/docs/network/latency/pairs/ams-tyo-rtt) | **198.2 ms** | Fair | | [🇿🇦 Johannesburg (JNB)](/docs/network/latency/pairs/ams-jnb-rtt) | **171.9 ms** | Fair | **🇩🇪 Berlin (BER): 19 outbound routes** | Destination | RTT | Tier | | ------------------------------------------------------------------ | -----------: | --------- | | [🇳🇱 Amsterdam (AMS)](/docs/network/latency/pairs/ber-ams-rtt) | **7.2 ms** | Ultra-Low | | [🇩🇪 Frankfurt (FRA)](/docs/network/latency/pairs/ber-fra-rtt) | **6.1 ms** | Ultra-Low | | [🇬🇧 London (LON)](/docs/network/latency/pairs/ber-lon-rtt) | **15.2 ms** | Ultra-Low | | [🇫🇷 Marseille (MRS)](/docs/network/latency/pairs/ber-mrs-rtt) | **20.2 ms** | Ultra-Low | | [🇷🇺 Moscow (MOW)](/docs/network/latency/pairs/ber-mow-rtt) | **28.6 ms** | Ultra-Low | | [🇫🇷 Paris (PAR)](/docs/network/latency/pairs/ber-par-rtt) | **15.6 ms** | Ultra-Low | | [🇺🇸 Ashburn (IAD)](/docs/network/latency/pairs/ber-iad-rtt) | **85.7 ms** | Good | | [🇺🇸 Los Angeles (LAX)](/docs/network/latency/pairs/ber-lax-rtt) | **140.9 ms** | Good | | [🇺🇸 Miami (MIA)](/docs/network/latency/pairs/ber-mia-rtt) | **118.3 ms** | Good | | [🇺🇸 New York (NYC)](/docs/network/latency/pairs/ber-nyc-rtt) | **79.6 ms** | Excellent | | [🇺🇸 Seattle (SEA)](/docs/network/latency/pairs/ber-sea-rtt) | **139.7 ms** | Good | | [🇧🇷 São Paulo (GRU)](/docs/network/latency/pairs/ber-gru-rtt) | **186.5 ms** | Fair | | [🇭🇰 Hong Kong (HKG)](/docs/network/latency/pairs/ber-hkg-rtt) | **144.9 ms** | Good | | [🇦🇺 Melbourne (MEL)](/docs/network/latency/pairs/ber-mel-rtt) | **256.2 ms** | High | | [🇸🇬 Singapore (SIN)](/docs/network/latency/pairs/ber-sin-rtt) | **158.3 ms** | Fair | | [🇦🇺 Sydney (SYD)](/docs/network/latency/pairs/ber-syd-rtt) | **262.1 ms** | High | | [🇹🇼 Taipei (TPE)](/docs/network/latency/pairs/ber-tpe-rtt) | **156.8 ms** | Fair | | [🇯🇵 Tokyo (TYO)](/docs/network/latency/pairs/ber-tyo-rtt) | **187.5 ms** | Fair | | [🇿🇦 Johannesburg (JNB)](/docs/network/latency/pairs/ber-jnb-rtt) | **172.1 ms** | Fair | **🇩🇪 Frankfurt (FRA): 19 outbound routes** | Destination | RTT | Tier | | ------------------------------------------------------------------ | -----------: | --------- | | [🇳🇱 Amsterdam (AMS)](/docs/network/latency/pairs/fra-ams-rtt) | **5.9 ms** | Ultra-Low | | [🇩🇪 Berlin (BER)](/docs/network/latency/pairs/fra-ber-rtt) | **6.1 ms** | Ultra-Low | | [🇬🇧 London (LON)](/docs/network/latency/pairs/fra-lon-rtt) | **12.9 ms** | Ultra-Low | | [🇫🇷 Marseille (MRS)](/docs/network/latency/pairs/fra-mrs-rtt) | **16.4 ms** | Ultra-Low | | [🇷🇺 Moscow (MOW)](/docs/network/latency/pairs/fra-mow-rtt) | **34.7 ms** | Excellent | | [🇫🇷 Paris (PAR)](/docs/network/latency/pairs/fra-par-rtt) | **7.6 ms** | Ultra-Low | | [🇺🇸 Ashburn (IAD)](/docs/network/latency/pairs/fra-iad-rtt) | **80.7 ms** | Good | | [🇺🇸 Los Angeles (LAX)](/docs/network/latency/pairs/fra-lax-rtt) | **141.8 ms** | Good | | [🇺🇸 Miami (MIA)](/docs/network/latency/pairs/fra-mia-rtt) | **112 ms** | Good | | [🇺🇸 New York (NYC)](/docs/network/latency/pairs/fra-nyc-rtt) | **74.4 ms** | Excellent | | [🇺🇸 Seattle (SEA)](/docs/network/latency/pairs/fra-sea-rtt) | **135.2 ms** | Good | | [🇧🇷 São Paulo (GRU)](/docs/network/latency/pairs/fra-gru-rtt) | **181.3 ms** | Fair | | [🇭🇰 Hong Kong (HKG)](/docs/network/latency/pairs/fra-hkg-rtt) | **152 ms** | Fair | | [🇦🇺 Melbourne (MEL)](/docs/network/latency/pairs/fra-mel-rtt) | **250.1 ms** | High | | [🇸🇬 Singapore (SIN)](/docs/network/latency/pairs/fra-sin-rtt) | **151.4 ms** | Fair | | [🇦🇺 Sydney (SYD)](/docs/network/latency/pairs/fra-syd-rtt) | **256 ms** | High | | [🇹🇼 Taipei (TPE)](/docs/network/latency/pairs/fra-tpe-rtt) | **163.8 ms** | Fair | | [🇯🇵 Tokyo (TYO)](/docs/network/latency/pairs/fra-tyo-rtt) | **194.5 ms** | Fair | | [🇿🇦 Johannesburg (JNB)](/docs/network/latency/pairs/fra-jnb-rtt) | **188.4 ms** | Fair | **🇬🇧 London (LON): 19 outbound routes** | Destination | RTT | Tier | | ------------------------------------------------------------------ | -----------: | --------- | | [🇳🇱 Amsterdam (AMS)](/docs/network/latency/pairs/lon-ams-rtt) | **5.2 ms** | Ultra-Low | | [🇩🇪 Berlin (BER)](/docs/network/latency/pairs/lon-ber-rtt) | **16.2 ms** | Ultra-Low | | [🇩🇪 Frankfurt (FRA)](/docs/network/latency/pairs/lon-fra-rtt) | **13.6 ms** | Ultra-Low | | [🇫🇷 Marseille (MRS)](/docs/network/latency/pairs/lon-mrs-rtt) | **17.9 ms** | Ultra-Low | | [🇷🇺 Moscow (MOW)](/docs/network/latency/pairs/lon-mow-rtt) | **43.5 ms** | Excellent | | [🇫🇷 Paris (PAR)](/docs/network/latency/pairs/lon-par-rtt) | **6.4 ms** | Ultra-Low | | [🇺🇸 Ashburn (IAD)](/docs/network/latency/pairs/lon-iad-rtt) | **71 ms** | Excellent | | [🇺🇸 Los Angeles (LAX)](/docs/network/latency/pairs/lon-lax-rtt) | **127.5 ms** | Good | | [🇺🇸 Miami (MIA)](/docs/network/latency/pairs/lon-mia-rtt) | **101.9 ms** | Good | | [🇺🇸 New York (NYC)](/docs/network/latency/pairs/lon-nyc-rtt) | **63.8 ms** | Excellent | | [🇺🇸 Seattle (SEA)](/docs/network/latency/pairs/lon-sea-rtt) | **123.2 ms** | Good | | [🇧🇷 São Paulo (GRU)](/docs/network/latency/pairs/lon-gru-rtt) | **170.7 ms** | Fair | | [🇭🇰 Hong Kong (HKG)](/docs/network/latency/pairs/lon-hkg-rtt) | **160.6 ms** | Fair | | [🇦🇺 Melbourne (MEL)](/docs/network/latency/pairs/lon-mel-rtt) | **249.6 ms** | Fair | | [🇸🇬 Singapore (SIN)](/docs/network/latency/pairs/lon-sin-rtt) | **156.5 ms** | Fair | | [🇦🇺 Sydney (SYD)](/docs/network/latency/pairs/lon-syd-rtt) | **258.1 ms** | High | | [🇹🇼 Taipei (TPE)](/docs/network/latency/pairs/lon-tpe-rtt) | **173.8 ms** | Fair | | [🇯🇵 Tokyo (TYO)](/docs/network/latency/pairs/lon-tyo-rtt) | **202.9 ms** | Fair | | [🇿🇦 Johannesburg (JNB)](/docs/network/latency/pairs/lon-jnb-rtt) | **179.6 ms** | Fair | **🇫🇷 Marseille (MRS): 19 outbound routes** | Destination | RTT | Tier | | ------------------------------------------------------------------ | -----------: | --------- | | [🇳🇱 Amsterdam (AMS)](/docs/network/latency/pairs/mrs-ams-rtt) | **20.3 ms** | Ultra-Low | | [🇩🇪 Berlin (BER)](/docs/network/latency/pairs/mrs-ber-rtt) | **20.6 ms** | Ultra-Low | | [🇩🇪 Frankfurt (FRA)](/docs/network/latency/pairs/mrs-fra-rtt) | **16 ms** | Ultra-Low | | [🇬🇧 London (LON)](/docs/network/latency/pairs/mrs-lon-rtt) | **18.9 ms** | Ultra-Low | | [🇷🇺 Moscow (MOW)](/docs/network/latency/pairs/mrs-mow-rtt) | **49.6 ms** | Excellent | | [🇫🇷 Paris (PAR)](/docs/network/latency/pairs/mrs-par-rtt) | **8.9 ms** | Ultra-Low | | [🇺🇸 Ashburn (IAD)](/docs/network/latency/pairs/mrs-iad-rtt) | **85.7 ms** | Good | | [🇺🇸 Los Angeles (LAX)](/docs/network/latency/pairs/mrs-lax-rtt) | **144.5 ms** | Good | | [🇺🇸 Miami (MIA)](/docs/network/latency/pairs/mrs-mia-rtt) | **112.4 ms** | Good | | [🇺🇸 New York (NYC)](/docs/network/latency/pairs/mrs-nyc-rtt) | **79.8 ms** | Excellent | | [🇺🇸 Seattle (SEA)](/docs/network/latency/pairs/mrs-sea-rtt) | **141.7 ms** | Good | | [🇧🇷 São Paulo (GRU)](/docs/network/latency/pairs/mrs-gru-rtt) | **186.7 ms** | Fair | | [🇭🇰 Hong Kong (HKG)](/docs/network/latency/pairs/mrs-hkg-rtt) | **166.4 ms** | Fair | | [🇦🇺 Melbourne (MEL)](/docs/network/latency/pairs/mrs-mel-rtt) | **234.2 ms** | Fair | | [🇸🇬 Singapore (SIN)](/docs/network/latency/pairs/mrs-sin-rtt) | **139.6 ms** | Good | | [🇦🇺 Sydney (SYD)](/docs/network/latency/pairs/mrs-syd-rtt) | **240.6 ms** | Fair | | [🇹🇼 Taipei (TPE)](/docs/network/latency/pairs/mrs-tpe-rtt) | **180.9 ms** | Fair | | [🇯🇵 Tokyo (TYO)](/docs/network/latency/pairs/mrs-tyo-rtt) | **203.8 ms** | Fair | | [🇿🇦 Johannesburg (JNB)](/docs/network/latency/pairs/mrs-jnb-rtt) | **198.5 ms** | Fair | **🇷🇺 Moscow (MOW): 19 outbound routes** | Destination | RTT | Tier | | ------------------------------------------------------------------ | -----------: | --------- | | [🇳🇱 Amsterdam (AMS)](/docs/network/latency/pairs/mow-ams-rtt) | **37.6 ms** | Excellent | | [🇩🇪 Berlin (BER)](/docs/network/latency/pairs/mow-ber-rtt) | **27.5 ms** | Ultra-Low | | [🇩🇪 Frankfurt (FRA)](/docs/network/latency/pairs/mow-fra-rtt) | **35.9 ms** | Excellent | | [🇬🇧 London (LON)](/docs/network/latency/pairs/mow-lon-rtt) | **42.4 ms** | Excellent | | [🇫🇷 Marseille (MRS)](/docs/network/latency/pairs/mow-mrs-rtt) | **49.9 ms** | Excellent | | [🇫🇷 Paris (PAR)](/docs/network/latency/pairs/mow-par-rtt) | **44 ms** | Excellent | | [🇺🇸 Ashburn (IAD)](/docs/network/latency/pairs/mow-iad-rtt) | **113.9 ms** | Good | | [🇺🇸 Los Angeles (LAX)](/docs/network/latency/pairs/mow-lax-rtt) | **177.7 ms** | Fair | | [🇺🇸 Miami (MIA)](/docs/network/latency/pairs/mow-mia-rtt) | **144.5 ms** | Good | | [🇺🇸 New York (NYC)](/docs/network/latency/pairs/mow-nyc-rtt) | **107.5 ms** | Good | | [🇺🇸 Seattle (SEA)](/docs/network/latency/pairs/mow-sea-rtt) | **164.8 ms** | Fair | | [🇧🇷 São Paulo (GRU)](/docs/network/latency/pairs/mow-gru-rtt) | **214.4 ms** | Fair | | [🇭🇰 Hong Kong (HKG)](/docs/network/latency/pairs/mow-hkg-rtt) | **118.6 ms** | Good | | [🇦🇺 Melbourne (MEL)](/docs/network/latency/pairs/mow-mel-rtt) | **235.9 ms** | Fair | | [🇸🇬 Singapore (SIN)](/docs/network/latency/pairs/mow-sin-rtt) | **147.3 ms** | Good | | [🇦🇺 Sydney (SYD)](/docs/network/latency/pairs/mow-syd-rtt) | **241.8 ms** | Fair | | [🇹🇼 Taipei (TPE)](/docs/network/latency/pairs/mow-tpe-rtt) | **132.1 ms** | Good | | [🇯🇵 Tokyo (TYO)](/docs/network/latency/pairs/mow-tyo-rtt) | **162.2 ms** | Fair | | [🇿🇦 Johannesburg (JNB)](/docs/network/latency/pairs/mow-jnb-rtt) | **201.9 ms** | Fair | **🇫🇷 Paris (PAR): 19 outbound routes** | Destination | RTT | Tier | | ------------------------------------------------------------------ | -----------: | --------- | | [🇳🇱 Amsterdam (AMS)](/docs/network/latency/pairs/par-ams-rtt) | **7 ms** | Ultra-Low | | [🇩🇪 Berlin (BER)](/docs/network/latency/pairs/par-ber-rtt) | **15.5 ms** | Ultra-Low | | [🇩🇪 Frankfurt (FRA)](/docs/network/latency/pairs/par-fra-rtt) | **7.6 ms** | Ultra-Low | | [🇬🇧 London (LON)](/docs/network/latency/pairs/par-lon-rtt) | **6.4 ms** | Ultra-Low | | [🇫🇷 Marseille (MRS)](/docs/network/latency/pairs/par-mrs-rtt) | **8.9 ms** | Ultra-Low | | [🇷🇺 Moscow (MOW)](/docs/network/latency/pairs/par-mow-rtt) | **44.7 ms** | Excellent | | [🇺🇸 Ashburn (IAD)](/docs/network/latency/pairs/par-iad-rtt) | **75.5 ms** | Excellent | | [🇺🇸 Los Angeles (LAX)](/docs/network/latency/pairs/par-lax-rtt) | **136.5 ms** | Good | | [🇺🇸 Miami (MIA)](/docs/network/latency/pairs/par-mia-rtt) | **106.8 ms** | Good | | [🇺🇸 New York (NYC)](/docs/network/latency/pairs/par-nyc-rtt) | **68.9 ms** | Excellent | | [🇺🇸 Seattle (SEA)](/docs/network/latency/pairs/par-sea-rtt) | **128 ms** | Good | | [🇧🇷 São Paulo (GRU)](/docs/network/latency/pairs/par-gru-rtt) | **175.8 ms** | Fair | | [🇭🇰 Hong Kong (HKG)](/docs/network/latency/pairs/par-hkg-rtt) | **160.1 ms** | Fair | | [🇦🇺 Melbourne (MEL)](/docs/network/latency/pairs/par-mel-rtt) | **246.4 ms** | Fair | | [🇸🇬 Singapore (SIN)](/docs/network/latency/pairs/par-sin-rtt) | **150.3 ms** | Fair | | [🇦🇺 Sydney (SYD)](/docs/network/latency/pairs/par-syd-rtt) | **249.7 ms** | Fair | | [🇹🇼 Taipei (TPE)](/docs/network/latency/pairs/par-tpe-rtt) | **171.9 ms** | Fair | | [🇯🇵 Tokyo (TYO)](/docs/network/latency/pairs/par-tyo-rtt) | **204.7 ms** | Fair | | [🇿🇦 Johannesburg (JNB)](/docs/network/latency/pairs/par-jnb-rtt) | **180.9 ms** | Fair | ### Routes from North America (95) **🇺🇸 Ashburn (IAD): 19 outbound routes** | Destination | RTT | Tier | | ------------------------------------------------------------------ | -----------: | --------- | | [🇳🇱 Amsterdam (AMS)](/docs/network/latency/pairs/iad-ams-rtt) | **75.2 ms** | Excellent | | [🇩🇪 Berlin (BER)](/docs/network/latency/pairs/iad-ber-rtt) | **86.1 ms** | Good | | [🇩🇪 Frankfurt (FRA)](/docs/network/latency/pairs/iad-fra-rtt) | **81.2 ms** | Good | | [🇬🇧 London (LON)](/docs/network/latency/pairs/iad-lon-rtt) | **70.3 ms** | Excellent | | [🇫🇷 Marseille (MRS)](/docs/network/latency/pairs/iad-mrs-rtt) | **86.5 ms** | Good | | [🇷🇺 Moscow (MOW)](/docs/network/latency/pairs/iad-mow-rtt) | **112.8 ms** | Good | | [🇫🇷 Paris (PAR)](/docs/network/latency/pairs/iad-par-rtt) | **76.1 ms** | Excellent | | [🇺🇸 Los Angeles (LAX)](/docs/network/latency/pairs/iad-lax-rtt) | **60.7 ms** | Excellent | | [🇺🇸 Miami (MIA)](/docs/network/latency/pairs/iad-mia-rtt) | **27.4 ms** | Ultra-Low | | [🇺🇸 New York (NYC)](/docs/network/latency/pairs/iad-nyc-rtt) | **6 ms** | Ultra-Low | | [🇺🇸 Seattle (SEA)](/docs/network/latency/pairs/iad-sea-rtt) | **62.3 ms** | Excellent | | [🇧🇷 São Paulo (GRU)](/docs/network/latency/pairs/iad-gru-rtt) | **101.4 ms** | Good | | [🇭🇰 Hong Kong (HKG)](/docs/network/latency/pairs/iad-hkg-rtt) | **187.2 ms** | Fair | | [🇦🇺 Melbourne (MEL)](/docs/network/latency/pairs/iad-mel-rtt) | **203.9 ms** | Fair | | [🇸🇬 Singapore (SIN)](/docs/network/latency/pairs/iad-sin-rtt) | **211.5 ms** | Fair | | [🇦🇺 Sydney (SYD)](/docs/network/latency/pairs/iad-syd-rtt) | **193.6 ms** | Fair | | [🇹🇼 Taipei (TPE)](/docs/network/latency/pairs/iad-tpe-rtt) | **172.3 ms** | Fair | | [🇯🇵 Tokyo (TYO)](/docs/network/latency/pairs/iad-tyo-rtt) | **141.9 ms** | Good | | [🇿🇦 Johannesburg (JNB)](/docs/network/latency/pairs/iad-jnb-rtt) | **228.3 ms** | Fair | **🇺🇸 Los Angeles (LAX): 19 outbound routes** | Destination | RTT | Tier | | ------------------------------------------------------------------ | -----------: | --------- | | [🇳🇱 Amsterdam (AMS)](/docs/network/latency/pairs/lax-ams-rtt) | **132.1 ms** | Good | | [🇩🇪 Berlin (BER)](/docs/network/latency/pairs/lax-ber-rtt) | **139.3 ms** | Good | | [🇩🇪 Frankfurt (FRA)](/docs/network/latency/pairs/lax-fra-rtt) | **142.3 ms** | Good | | [🇬🇧 London (LON)](/docs/network/latency/pairs/lax-lon-rtt) | **129.2 ms** | Good | | [🇫🇷 Marseille (MRS)](/docs/network/latency/pairs/lax-mrs-rtt) | **143.9 ms** | Good | | [🇷🇺 Moscow (MOW)](/docs/network/latency/pairs/lax-mow-rtt) | **177 ms** | Fair | | [🇫🇷 Paris (PAR)](/docs/network/latency/pairs/lax-par-rtt) | **135.9 ms** | Good | | [🇺🇸 Ashburn (IAD)](/docs/network/latency/pairs/lax-iad-rtt) | **60.4 ms** | Excellent | | [🇺🇸 Miami (MIA)](/docs/network/latency/pairs/lax-mia-rtt) | **57.3 ms** | Excellent | | [🇺🇸 New York (NYC)](/docs/network/latency/pairs/lax-nyc-rtt) | **57.9 ms** | Excellent | | [🇺🇸 Seattle (SEA)](/docs/network/latency/pairs/lax-sea-rtt) | **26.9 ms** | Ultra-Low | | [🇧🇷 São Paulo (GRU)](/docs/network/latency/pairs/lax-gru-rtt) | **131.3 ms** | Good | | [🇭🇰 Hong Kong (HKG)](/docs/network/latency/pairs/lax-hkg-rtt) | **146.5 ms** | Good | | [🇦🇺 Melbourne (MEL)](/docs/network/latency/pairs/lax-mel-rtt) | **148.2 ms** | Good | | [🇸🇬 Singapore (SIN)](/docs/network/latency/pairs/lax-sin-rtt) | **167.7 ms** | Fair | | [🇦🇺 Sydney (SYD)](/docs/network/latency/pairs/lax-syd-rtt) | **137.6 ms** | Good | | [🇹🇼 Taipei (TPE)](/docs/network/latency/pairs/lax-tpe-rtt) | **132.9 ms** | Good | | [🇯🇵 Tokyo (TYO)](/docs/network/latency/pairs/lax-tyo-rtt) | **101.2 ms** | Good | | [🇿🇦 Johannesburg (JNB)](/docs/network/latency/pairs/lax-jnb-rtt) | **287.2 ms** | High | **🇺🇸 Miami (MIA): 19 outbound routes** | Destination | RTT | Tier | | ------------------------------------------------------------------ | -----------: | --------- | | [🇳🇱 Amsterdam (AMS)](/docs/network/latency/pairs/mia-ams-rtt) | **106 ms** | Good | | [🇩🇪 Berlin (BER)](/docs/network/latency/pairs/mia-ber-rtt) | **117.3 ms** | Good | | [🇩🇪 Frankfurt (FRA)](/docs/network/latency/pairs/mia-fra-rtt) | **112.2 ms** | Good | | [🇬🇧 London (LON)](/docs/network/latency/pairs/mia-lon-rtt) | **101.8 ms** | Good | | [🇫🇷 Marseille (MRS)](/docs/network/latency/pairs/mia-mrs-rtt) | **113.7 ms** | Good | | [🇷🇺 Moscow (MOW)](/docs/network/latency/pairs/mia-mow-rtt) | **145.5 ms** | Good | | [🇫🇷 Paris (PAR)](/docs/network/latency/pairs/mia-par-rtt) | **106.2 ms** | Good | | [🇺🇸 Ashburn (IAD)](/docs/network/latency/pairs/mia-iad-rtt) | **27.3 ms** | Ultra-Low | | [🇺🇸 Los Angeles (LAX)](/docs/network/latency/pairs/mia-lax-rtt) | **56.8 ms** | Excellent | | [🇺🇸 New York (NYC)](/docs/network/latency/pairs/mia-nyc-rtt) | **33.3 ms** | Excellent | | [🇺🇸 Seattle (SEA)](/docs/network/latency/pairs/mia-sea-rtt) | **81.9 ms** | Good | | [🇧🇷 São Paulo (GRU)](/docs/network/latency/pairs/mia-gru-rtt) | **128.7 ms** | Good | | [🇭🇰 Hong Kong (HKG)](/docs/network/latency/pairs/mia-hkg-rtt) | **199.5 ms** | Fair | | [🇦🇺 Melbourne (MEL)](/docs/network/latency/pairs/mia-mel-rtt) | **200.7 ms** | Fair | | [🇸🇬 Singapore (SIN)](/docs/network/latency/pairs/mia-sin-rtt) | **223.1 ms** | Fair | | [🇦🇺 Sydney (SYD)](/docs/network/latency/pairs/mia-syd-rtt) | **191.5 ms** | Fair | | [🇹🇼 Taipei (TPE)](/docs/network/latency/pairs/mia-tpe-rtt) | **185.1 ms** | Fair | | [🇯🇵 Tokyo (TYO)](/docs/network/latency/pairs/mia-tyo-rtt) | **156.3 ms** | Fair | | [🇿🇦 Johannesburg (JNB)](/docs/network/latency/pairs/mia-jnb-rtt) | **259.8 ms** | High | **🇺🇸 New York (NYC): 19 outbound routes** | Destination | RTT | Tier | | ------------------------------------------------------------------ | -----------: | --------- | | [🇳🇱 Amsterdam (AMS)](/docs/network/latency/pairs/nyc-ams-rtt) | **68.7 ms** | Excellent | | [🇩🇪 Berlin (BER)](/docs/network/latency/pairs/nyc-ber-rtt) | **78.8 ms** | Excellent | | [🇩🇪 Frankfurt (FRA)](/docs/network/latency/pairs/nyc-fra-rtt) | **73.9 ms** | Excellent | | [🇬🇧 London (LON)](/docs/network/latency/pairs/nyc-lon-rtt) | **63.5 ms** | Excellent | | [🇫🇷 Marseille (MRS)](/docs/network/latency/pairs/nyc-mrs-rtt) | **80.2 ms** | Good | | [🇷🇺 Moscow (MOW)](/docs/network/latency/pairs/nyc-mow-rtt) | **107 ms** | Good | | [🇫🇷 Paris (PAR)](/docs/network/latency/pairs/nyc-par-rtt) | **68.9 ms** | Excellent | | [🇺🇸 Ashburn (IAD)](/docs/network/latency/pairs/nyc-iad-rtt) | **6 ms** | Ultra-Low | | [🇺🇸 Los Angeles (LAX)](/docs/network/latency/pairs/nyc-lax-rtt) | **58.9 ms** | Excellent | | [🇺🇸 Miami (MIA)](/docs/network/latency/pairs/nyc-mia-rtt) | **32.9 ms** | Excellent | | [🇺🇸 Seattle (SEA)](/docs/network/latency/pairs/nyc-sea-rtt) | **58.7 ms** | Excellent | | [🇧🇷 São Paulo (GRU)](/docs/network/latency/pairs/nyc-gru-rtt) | **106.9 ms** | Good | | [🇭🇰 Hong Kong (HKG)](/docs/network/latency/pairs/nyc-hkg-rtt) | **187.1 ms** | Fair | | [🇦🇺 Melbourne (MEL)](/docs/network/latency/pairs/nyc-mel-rtt) | **205.1 ms** | Fair | | [🇸🇬 Singapore (SIN)](/docs/network/latency/pairs/nyc-sin-rtt) | **208.7 ms** | Fair | | [🇦🇺 Sydney (SYD)](/docs/network/latency/pairs/nyc-syd-rtt) | **192.6 ms** | Fair | | [🇹🇼 Taipei (TPE)](/docs/network/latency/pairs/nyc-tpe-rtt) | **173.4 ms** | Fair | | [🇯🇵 Tokyo (TYO)](/docs/network/latency/pairs/nyc-tyo-rtt) | **143.1 ms** | Good | | [🇿🇦 Johannesburg (JNB)](/docs/network/latency/pairs/nyc-jnb-rtt) | **221.5 ms** | Fair | **🇺🇸 Seattle (SEA): 19 outbound routes** | Destination | RTT | Tier | | ------------------------------------------------------------------ | -----------: | --------- | | [🇳🇱 Amsterdam (AMS)](/docs/network/latency/pairs/sea-ams-rtt) | **128.6 ms** | Good | | [🇩🇪 Berlin (BER)](/docs/network/latency/pairs/sea-ber-rtt) | **140.9 ms** | Good | | [🇩🇪 Frankfurt (FRA)](/docs/network/latency/pairs/sea-fra-rtt) | **133.2 ms** | Good | | [🇬🇧 London (LON)](/docs/network/latency/pairs/sea-lon-rtt) | **124.1 ms** | Good | | [🇫🇷 Marseille (MRS)](/docs/network/latency/pairs/sea-mrs-rtt) | **141.2 ms** | Good | | [🇷🇺 Moscow (MOW)](/docs/network/latency/pairs/sea-mow-rtt) | **164.6 ms** | Fair | | [🇫🇷 Paris (PAR)](/docs/network/latency/pairs/sea-par-rtt) | **129.2 ms** | Good | | [🇺🇸 Ashburn (IAD)](/docs/network/latency/pairs/sea-iad-rtt) | **61.5 ms** | Excellent | | [🇺🇸 Los Angeles (LAX)](/docs/network/latency/pairs/sea-lax-rtt) | **25.9 ms** | Ultra-Low | | [🇺🇸 Miami (MIA)](/docs/network/latency/pairs/sea-mia-rtt) | **81.7 ms** | Good | | [🇺🇸 New York (NYC)](/docs/network/latency/pairs/sea-nyc-rtt) | **59.5 ms** | Excellent | | [🇧🇷 São Paulo (GRU)](/docs/network/latency/pairs/sea-gru-rtt) | **155.7 ms** | Fair | | [🇭🇰 Hong Kong (HKG)](/docs/network/latency/pairs/sea-hkg-rtt) | **130.4 ms** | Good | | [🇦🇺 Melbourne (MEL)](/docs/network/latency/pairs/sea-mel-rtt) | **170.1 ms** | Fair | | [🇸🇬 Singapore (SIN)](/docs/network/latency/pairs/sea-sin-rtt) | **151.1 ms** | Fair | | [🇦🇺 Sydney (SYD)](/docs/network/latency/pairs/sea-syd-rtt) | **161.3 ms** | Fair | | [🇹🇼 Taipei (TPE)](/docs/network/latency/pairs/sea-tpe-rtt) | **114.7 ms** | Good | | [🇯🇵 Tokyo (TYO)](/docs/network/latency/pairs/sea-tyo-rtt) | **85 ms** | Good | | [🇿🇦 Johannesburg (JNB)](/docs/network/latency/pairs/sea-jnb-rtt) | **282.1 ms** | High | ### Routes from South America (19) **🇧🇷 São Paulo (GRU): 19 outbound routes** | Destination | RTT | Tier | | ------------------------------------------------------------------ | -----------: | --------- | | [🇳🇱 Amsterdam (AMS)](/docs/network/latency/pairs/gru-ams-rtt) | **180 ms** | Fair | | [🇩🇪 Berlin (BER)](/docs/network/latency/pairs/gru-ber-rtt) | **191.3 ms** | Fair | | [🇩🇪 Frankfurt (FRA)](/docs/network/latency/pairs/gru-fra-rtt) | **186.2 ms** | Fair | | [🇬🇧 London (LON)](/docs/network/latency/pairs/gru-lon-rtt) | **175.8 ms** | Fair | | [🇫🇷 Marseille (MRS)](/docs/network/latency/pairs/gru-mrs-rtt) | **187.7 ms** | Fair | | [🇷🇺 Moscow (MOW)](/docs/network/latency/pairs/gru-mow-rtt) | **219.5 ms** | Fair | | [🇫🇷 Paris (PAR)](/docs/network/latency/pairs/gru-par-rtt) | **180.2 ms** | Fair | | [🇺🇸 Ashburn (IAD)](/docs/network/latency/pairs/gru-iad-rtt) | **128 ms** | Good | | [🇺🇸 Los Angeles (LAX)](/docs/network/latency/pairs/gru-lax-rtt) | **130.8 ms** | Good | | [🇺🇸 Miami (MIA)](/docs/network/latency/pairs/gru-mia-rtt) | **74 ms** | Excellent | | [🇺🇸 New York (NYC)](/docs/network/latency/pairs/gru-nyc-rtt) | **122 ms** | Good | | [🇺🇸 Seattle (SEA)](/docs/network/latency/pairs/gru-sea-rtt) | **155.9 ms** | Fair | | [🇭🇰 Hong Kong (HKG)](/docs/network/latency/pairs/gru-hkg-rtt) | **273.5 ms** | High | | [🇦🇺 Melbourne (MEL)](/docs/network/latency/pairs/gru-mel-rtt) | **274.7 ms** | High | | [🇸🇬 Singapore (SIN)](/docs/network/latency/pairs/gru-sin-rtt) | **297.1 ms** | High | | [🇦🇺 Sydney (SYD)](/docs/network/latency/pairs/gru-syd-rtt) | **265.5 ms** | High | | [🇹🇼 Taipei (TPE)](/docs/network/latency/pairs/gru-tpe-rtt) | **259.1 ms** | High | | [🇯🇵 Tokyo (TYO)](/docs/network/latency/pairs/gru-tyo-rtt) | **230.3 ms** | Fair | | [🇿🇦 Johannesburg (JNB)](/docs/network/latency/pairs/gru-jnb-rtt) | **333.8 ms** | High | ### Routes from Asia Pacific (114) **🇭🇰 Hong Kong (HKG): 19 outbound routes** | Destination | RTT | Tier | | ------------------------------------------------------------------ | -----------: | --------- | | [🇳🇱 Amsterdam (AMS)](/docs/network/latency/pairs/hkg-ams-rtt) | **154.4 ms** | Fair | | [🇩🇪 Berlin (BER)](/docs/network/latency/pairs/hkg-ber-rtt) | **145.1 ms** | Good | | [🇩🇪 Frankfurt (FRA)](/docs/network/latency/pairs/hkg-fra-rtt) | **150.7 ms** | Fair | | [🇬🇧 London (LON)](/docs/network/latency/pairs/hkg-lon-rtt) | **158.7 ms** | Fair | | [🇫🇷 Marseille (MRS)](/docs/network/latency/pairs/hkg-mrs-rtt) | **165.4 ms** | Fair | | [🇷🇺 Moscow (MOW)](/docs/network/latency/pairs/hkg-mow-rtt) | **118.2 ms** | Good | | [🇫🇷 Paris (PAR)](/docs/network/latency/pairs/hkg-par-rtt) | **159.4 ms** | Fair | | [🇺🇸 Ashburn (IAD)](/docs/network/latency/pairs/hkg-iad-rtt) | **186.2 ms** | Fair | | [🇺🇸 Los Angeles (LAX)](/docs/network/latency/pairs/hkg-lax-rtt) | **145 ms** | Good | | [🇺🇸 Miami (MIA)](/docs/network/latency/pairs/hkg-mia-rtt) | **201.1 ms** | Fair | | [🇺🇸 New York (NYC)](/docs/network/latency/pairs/hkg-nyc-rtt) | **184.7 ms** | Fair | | [🇺🇸 Seattle (SEA)](/docs/network/latency/pairs/hkg-sea-rtt) | **131.6 ms** | Good | | [🇧🇷 São Paulo (GRU)](/docs/network/latency/pairs/hkg-gru-rtt) | **287.6 ms** | High | | [🇦🇺 Melbourne (MEL)](/docs/network/latency/pairs/hkg-mel-rtt) | **138.4 ms** | Good | | [🇸🇬 Singapore (SIN)](/docs/network/latency/pairs/hkg-sin-rtt) | **31.5 ms** | Excellent | | [🇦🇺 Sydney (SYD)](/docs/network/latency/pairs/hkg-syd-rtt) | **137.2 ms** | Good | | [🇹🇼 Taipei (TPE)](/docs/network/latency/pairs/hkg-tpe-rtt) | **14.7 ms** | Ultra-Low | | [🇯🇵 Tokyo (TYO)](/docs/network/latency/pairs/hkg-tyo-rtt) | **44.9 ms** | Excellent | | [🇿🇦 Johannesburg (JNB)](/docs/network/latency/pairs/hkg-jnb-rtt) | **316.7 ms** | High | **🇦🇺 Melbourne (MEL): 19 outbound routes** | Destination | RTT | Tier | | ------------------------------------------------------------------ | -----------: | --------- | | [🇳🇱 Amsterdam (AMS)](/docs/network/latency/pairs/mel-ams-rtt) | **252.8 ms** | High | | [🇩🇪 Berlin (BER)](/docs/network/latency/pairs/mel-ber-rtt) | **254.7 ms** | High | | [🇩🇪 Frankfurt (FRA)](/docs/network/latency/pairs/mel-fra-rtt) | **248.6 ms** | Fair | | [🇬🇧 London (LON)](/docs/network/latency/pairs/mel-lon-rtt) | **252 ms** | High | | [🇫🇷 Marseille (MRS)](/docs/network/latency/pairs/mel-mrs-rtt) | **233.5 ms** | Fair | | [🇷🇺 Moscow (MOW)](/docs/network/latency/pairs/mel-mow-rtt) | **236.1 ms** | Fair | | [🇫🇷 Paris (PAR)](/docs/network/latency/pairs/mel-par-rtt) | **246.5 ms** | Fair | | [🇺🇸 Ashburn (IAD)](/docs/network/latency/pairs/mel-iad-rtt) | **203.9 ms** | Fair | | [🇺🇸 Los Angeles (LAX)](/docs/network/latency/pairs/mel-lax-rtt) | **148.4 ms** | Good | | [🇺🇸 Miami (MIA)](/docs/network/latency/pairs/mel-mia-rtt) | **201 ms** | Fair | | [🇺🇸 New York (NYC)](/docs/network/latency/pairs/mel-nyc-rtt) | **204.2 ms** | Fair | | [🇺🇸 Seattle (SEA)](/docs/network/latency/pairs/mel-sea-rtt) | **172.1 ms** | Fair | | [🇧🇷 São Paulo (GRU)](/docs/network/latency/pairs/mel-gru-rtt) | **305.3 ms** | High | | [🇭🇰 Hong Kong (HKG)](/docs/network/latency/pairs/mel-hkg-rtt) | **138.5 ms** | Good | | [🇸🇬 Singapore (SIN)](/docs/network/latency/pairs/mel-sin-rtt) | **88.4 ms** | Good | | [🇦🇺 Sydney (SYD)](/docs/network/latency/pairs/mel-syd-rtt) | **9.8 ms** | Ultra-Low | | [🇹🇼 Taipei (TPE)](/docs/network/latency/pairs/mel-tpe-rtt) | **143.2 ms** | Good | | [🇯🇵 Tokyo (TYO)](/docs/network/latency/pairs/mel-tyo-rtt) | **112 ms** | Good | | [🇿🇦 Johannesburg (JNB)](/docs/network/latency/pairs/mel-jnb-rtt) | **405.5 ms** | High | **🇸🇬 Singapore (SIN): 19 outbound routes** | Destination | RTT | Tier | | ------------------------------------------------------------------ | -----------: | --------- | | [🇳🇱 Amsterdam (AMS)](/docs/network/latency/pairs/sin-ams-rtt) | **155.8 ms** | Fair | | [🇩🇪 Berlin (BER)](/docs/network/latency/pairs/sin-ber-rtt) | **160.3 ms** | Fair | | [🇩🇪 Frankfurt (FRA)](/docs/network/latency/pairs/sin-fra-rtt) | **151.6 ms** | Fair | | [🇬🇧 London (LON)](/docs/network/latency/pairs/sin-lon-rtt) | **155.4 ms** | Fair | | [🇫🇷 Marseille (MRS)](/docs/network/latency/pairs/sin-mrs-rtt) | **140.6 ms** | Good | | [🇷🇺 Moscow (MOW)](/docs/network/latency/pairs/sin-mow-rtt) | **147.7 ms** | Good | | [🇫🇷 Paris (PAR)](/docs/network/latency/pairs/sin-par-rtt) | **151.4 ms** | Fair | | [🇺🇸 Ashburn (IAD)](/docs/network/latency/pairs/sin-iad-rtt) | **210 ms** | Fair | | [🇺🇸 Los Angeles (LAX)](/docs/network/latency/pairs/sin-lax-rtt) | **168.2 ms** | Fair | | [🇺🇸 Miami (MIA)](/docs/network/latency/pairs/sin-mia-rtt) | **221.7 ms** | Fair | | [🇺🇸 New York (NYC)](/docs/network/latency/pairs/sin-nyc-rtt) | **209.3 ms** | Fair | | [🇺🇸 Seattle (SEA)](/docs/network/latency/pairs/sin-sea-rtt) | **151.5 ms** | Fair | | [🇧🇷 São Paulo (GRU)](/docs/network/latency/pairs/sin-gru-rtt) | **311.4 ms** | High | | [🇭🇰 Hong Kong (HKG)](/docs/network/latency/pairs/sin-hkg-rtt) | **30.8 ms** | Excellent | | [🇦🇺 Melbourne (MEL)](/docs/network/latency/pairs/sin-mel-rtt) | **88.6 ms** | Good | | [🇦🇺 Sydney (SYD)](/docs/network/latency/pairs/sin-syd-rtt) | **94.5 ms** | Good | | [🇹🇼 Taipei (TPE)](/docs/network/latency/pairs/sin-tpe-rtt) | **45.8 ms** | Excellent | | [🇯🇵 Tokyo (TYO)](/docs/network/latency/pairs/sin-tyo-rtt) | **68.9 ms** | Excellent | | [🇿🇦 Johannesburg (JNB)](/docs/network/latency/pairs/sin-jnb-rtt) | **312.6 ms** | High | **🇦🇺 Sydney (SYD): 19 outbound routes** | Destination | RTT | Tier | | ------------------------------------------------------------------ | -----------: | --------- | | [🇳🇱 Amsterdam (AMS)](/docs/network/latency/pairs/syd-ams-rtt) | **258.9 ms** | High | | [🇩🇪 Berlin (BER)](/docs/network/latency/pairs/syd-ber-rtt) | **260.5 ms** | High | | [🇩🇪 Frankfurt (FRA)](/docs/network/latency/pairs/syd-fra-rtt) | **254.4 ms** | High | | [🇬🇧 London (LON)](/docs/network/latency/pairs/syd-lon-rtt) | **256.5 ms** | High | | [🇫🇷 Marseille (MRS)](/docs/network/latency/pairs/syd-mrs-rtt) | **240.5 ms** | Fair | | [🇷🇺 Moscow (MOW)](/docs/network/latency/pairs/syd-mow-rtt) | **242.2 ms** | Fair | | [🇫🇷 Paris (PAR)](/docs/network/latency/pairs/syd-par-rtt) | **250.3 ms** | High | | [🇺🇸 Ashburn (IAD)](/docs/network/latency/pairs/syd-iad-rtt) | **195.3 ms** | Fair | | [🇺🇸 Los Angeles (LAX)](/docs/network/latency/pairs/syd-lax-rtt) | **136.5 ms** | Good | | [🇺🇸 Miami (MIA)](/docs/network/latency/pairs/syd-mia-rtt) | **191.5 ms** | Fair | | [🇺🇸 New York (NYC)](/docs/network/latency/pairs/syd-nyc-rtt) | **191.9 ms** | Fair | | [🇺🇸 Seattle (SEA)](/docs/network/latency/pairs/syd-sea-rtt) | **161.9 ms** | Fair | | [🇧🇷 São Paulo (GRU)](/docs/network/latency/pairs/syd-gru-rtt) | **296.7 ms** | High | | [🇭🇰 Hong Kong (HKG)](/docs/network/latency/pairs/syd-hkg-rtt) | **135.7 ms** | Good | | [🇦🇺 Melbourne (MEL)](/docs/network/latency/pairs/syd-mel-rtt) | **9.8 ms** | Ultra-Low | | [🇸🇬 Singapore (SIN)](/docs/network/latency/pairs/syd-sin-rtt) | **94.5 ms** | Good | | [🇹🇼 Taipei (TPE)](/docs/network/latency/pairs/syd-tpe-rtt) | **131.7 ms** | Good | | [🇯🇵 Tokyo (TYO)](/docs/network/latency/pairs/syd-tyo-rtt) | **101.3 ms** | Good | | [🇿🇦 Johannesburg (JNB)](/docs/network/latency/pairs/syd-jnb-rtt) | **412.5 ms** | High | **🇹🇼 Taipei (TPE): 19 outbound routes** | Destination | RTT | Tier | | ------------------------------------------------------------------ | -----------: | --------- | | [🇳🇱 Amsterdam (AMS)](/docs/network/latency/pairs/tpe-ams-rtt) | **166.5 ms** | Fair | | [🇩🇪 Berlin (BER)](/docs/network/latency/pairs/tpe-ber-rtt) | **158.6 ms** | Fair | | [🇩🇪 Frankfurt (FRA)](/docs/network/latency/pairs/tpe-fra-rtt) | **166 ms** | Fair | | [🇬🇧 London (LON)](/docs/network/latency/pairs/tpe-lon-rtt) | **173 ms** | Fair | | [🇫🇷 Marseille (MRS)](/docs/network/latency/pairs/tpe-mrs-rtt) | **179.5 ms** | Fair | | [🇷🇺 Moscow (MOW)](/docs/network/latency/pairs/tpe-mow-rtt) | **132.7 ms** | Good | | [🇫🇷 Paris (PAR)](/docs/network/latency/pairs/tpe-par-rtt) | **173.8 ms** | Fair | | [🇺🇸 Ashburn (IAD)](/docs/network/latency/pairs/tpe-iad-rtt) | **171.6 ms** | Fair | | [🇺🇸 Los Angeles (LAX)](/docs/network/latency/pairs/tpe-lax-rtt) | **132 ms** | Good | | [🇺🇸 Miami (MIA)](/docs/network/latency/pairs/tpe-mia-rtt) | **185.7 ms** | Fair | | [🇺🇸 New York (NYC)](/docs/network/latency/pairs/tpe-nyc-rtt) | **174.2 ms** | Fair | | [🇺🇸 Seattle (SEA)](/docs/network/latency/pairs/tpe-sea-rtt) | **115.8 ms** | Good | | [🇧🇷 São Paulo (GRU)](/docs/network/latency/pairs/tpe-gru-rtt) | **273 ms** | High | | [🇭🇰 Hong Kong (HKG)](/docs/network/latency/pairs/tpe-hkg-rtt) | **15.5 ms** | Ultra-Low | | [🇦🇺 Melbourne (MEL)](/docs/network/latency/pairs/tpe-mel-rtt) | **142 ms** | Good | | [🇸🇬 Singapore (SIN)](/docs/network/latency/pairs/tpe-sin-rtt) | **44.3 ms** | Excellent | | [🇦🇺 Sydney (SYD)](/docs/network/latency/pairs/tpe-syd-rtt) | **133.2 ms** | Good | | [🇯🇵 Tokyo (TYO)](/docs/network/latency/pairs/tpe-tyo-rtt) | **31.9 ms** | Excellent | | [🇿🇦 Johannesburg (JNB)](/docs/network/latency/pairs/tpe-jnb-rtt) | **331 ms** | High | **🇯🇵 Tokyo (TYO): 19 outbound routes** | Destination | RTT | Tier | | ------------------------------------------------------------------ | -----------: | --------- | | [🇳🇱 Amsterdam (AMS)](/docs/network/latency/pairs/tyo-ams-rtt) | **197.4 ms** | Fair | | [🇩🇪 Berlin (BER)](/docs/network/latency/pairs/tyo-ber-rtt) | **186.1 ms** | Fair | | [🇩🇪 Frankfurt (FRA)](/docs/network/latency/pairs/tyo-fra-rtt) | **193.2 ms** | Fair | | [🇬🇧 London (LON)](/docs/network/latency/pairs/tyo-lon-rtt) | **202.7 ms** | Fair | | [🇫🇷 Marseille (MRS)](/docs/network/latency/pairs/tyo-mrs-rtt) | **205.6 ms** | Fair | | [🇷🇺 Moscow (MOW)](/docs/network/latency/pairs/tyo-mow-rtt) | **161.7 ms** | Fair | | [🇫🇷 Paris (PAR)](/docs/network/latency/pairs/tyo-par-rtt) | **204.8 ms** | Fair | | [🇺🇸 Ashburn (IAD)](/docs/network/latency/pairs/tyo-iad-rtt) | **142.9 ms** | Good | | [🇺🇸 Los Angeles (LAX)](/docs/network/latency/pairs/tyo-lax-rtt) | **101.2 ms** | Good | | [🇺🇸 Miami (MIA)](/docs/network/latency/pairs/tyo-mia-rtt) | **156.5 ms** | Fair | | [🇺🇸 New York (NYC)](/docs/network/latency/pairs/tyo-nyc-rtt) | **144.6 ms** | Good | | [🇺🇸 Seattle (SEA)](/docs/network/latency/pairs/tyo-sea-rtt) | **84.8 ms** | Good | | [🇧🇷 São Paulo (GRU)](/docs/network/latency/pairs/tyo-gru-rtt) | **230.5 ms** | Fair | | [🇭🇰 Hong Kong (HKG)](/docs/network/latency/pairs/tyo-hkg-rtt) | **44.8 ms** | Excellent | | [🇦🇺 Melbourne (MEL)](/docs/network/latency/pairs/tyo-mel-rtt) | **113.9 ms** | Good | | [🇸🇬 Singapore (SIN)](/docs/network/latency/pairs/tyo-sin-rtt) | **69.4 ms** | Excellent | | [🇦🇺 Sydney (SYD)](/docs/network/latency/pairs/tyo-syd-rtt) | **101.9 ms** | Good | | [🇹🇼 Taipei (TPE)](/docs/network/latency/pairs/tyo-tpe-rtt) | **32.3 ms** | Excellent | | [🇿🇦 Johannesburg (JNB)](/docs/network/latency/pairs/tyo-jnb-rtt) | **359.2 ms** | High | ### Routes from Africa (19) **🇿🇦 Johannesburg (JNB): 19 outbound routes** | Destination | RTT | Tier | | ----------------------------------------------------------------- | -----------: | ---- | | [🇳🇱 Amsterdam (AMS)](/docs/network/latency/pairs/jnb-ams-rtt) | **163.2 ms** | Fair | | [🇩🇪 Berlin (BER)](/docs/network/latency/pairs/jnb-ber-rtt) | **172.1 ms** | Fair | | [🇩🇪 Frankfurt (FRA)](/docs/network/latency/pairs/jnb-fra-rtt) | **166 ms** | Fair | | [🇬🇧 London (LON)](/docs/network/latency/pairs/jnb-lon-rtt) | **158 ms** | Fair | | [🇫🇷 Marseille (MRS)](/docs/network/latency/pairs/jnb-mrs-rtt) | **172 ms** | Fair | | [🇷🇺 Moscow (MOW)](/docs/network/latency/pairs/jnb-mow-rtt) | **200.7 ms** | Fair | | [🇫🇷 Paris (PAR)](/docs/network/latency/pairs/jnb-par-rtt) | **164.4 ms** | Fair | | [🇺🇸 Ashburn (IAD)](/docs/network/latency/pairs/jnb-iad-rtt) | **229 ms** | Fair | | [🇺🇸 Los Angeles (LAX)](/docs/network/latency/pairs/jnb-lax-rtt) | **285.5 ms** | High | | [🇺🇸 Miami (MIA)](/docs/network/latency/pairs/jnb-mia-rtt) | **259.9 ms** | High | | [🇺🇸 New York (NYC)](/docs/network/latency/pairs/jnb-nyc-rtt) | **221.8 ms** | Fair | | [🇺🇸 Seattle (SEA)](/docs/network/latency/pairs/jnb-sea-rtt) | **281.2 ms** | High | | [🇧🇷 São Paulo (GRU)](/docs/network/latency/pairs/jnb-gru-rtt) | **328.7 ms** | High | | [🇭🇰 Hong Kong (HKG)](/docs/network/latency/pairs/jnb-hkg-rtt) | **318 ms** | High | | [🇦🇺 Melbourne (MEL)](/docs/network/latency/pairs/jnb-mel-rtt) | **406.2 ms** | High | | [🇸🇬 Singapore (SIN)](/docs/network/latency/pairs/jnb-sin-rtt) | **311.6 ms** | High | | [🇦🇺 Sydney (SYD)](/docs/network/latency/pairs/jnb-syd-rtt) | **412.6 ms** | High | | [🇹🇼 Taipei (TPE)](/docs/network/latency/pairs/jnb-tpe-rtt) | **329.8 ms** | High | | [🇯🇵 Tokyo (TYO)](/docs/network/latency/pairs/jnb-tyo-rtt) | **360.5 ms** | High | *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse by [region & corridor](/docs/network/latency/regions).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Johannesburg, South Africa → Amsterdam, Netherlands RTT 🇿🇦 **Johannesburg, South Africa (JNB)** → 🇳🇱 **Amsterdam, Netherlands (AMS)** Measurement round: `2026-08-16T04:07:28Z`. Johannesburg to Amsterdam averaged 163.2 ms across 50 replies. The observed range was 158.68 to 177.43 ms; these figures describe one probe round, not a long-term service target. This is number 2 of 19 outgoing routes by average RTT. That ranking compares different destinations and is not adjusted for distance. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------- | | RTT | **163.2 ms** | | Jitter | **3.46 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **88.05 ms** | | Fiber Efficiency | **54%** | | Latency Tier | Fair | | Source Region | [Africa](/docs/network/latency/regions/africa) | | Destination Region | [Europe](/docs/network/latency/regions/europe) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Johannesburg](https://www.geonames.org/993800) (-26.20227, 28.04363) and [Amsterdam](https://www.geonames.org/2759794) (52.37403, 4.88969). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **8,991.8 km** | | Vacuum RTT floor | **59.99 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **88.05 ms** | | Low-latency fiber material floor | **87.7 ms** | | Engineering floor (5% path allowance) | **92.47 ms** | | Research 1.33× mapped-fiber reference | **117.11 ms** | | Estimated unamplified path loss | **1888.3 dB** | | Transparent optical spans / inline amplifiers | **119 / 118** | | Published RTT inflation over fiber floor | **1.85×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **158.68 ms** | | Average RTT | **163.2 ms** | | Maximum RTT | **177.43 ms** | | Standard deviation | **4.23 ms** | | Stdev / average | **2.6%** | ## Route Context * Departure PoP: [Johannesburg (JNB)](/docs/network/latency/jnb-johannesburg) * Destination PoP: [Amsterdam (AMS)](/docs/network/latency/ams-amsterdam) * Region overview: [Africa](/docs/network/latency/regions/africa) → [Europe](/docs/network/latency/regions/europe) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Amsterdam to Johannesburg latency and RTT](/docs/network/latency/pairs/ams-jnb-rtt): 171.9 ms **Fastest routes departing Johannesburg (JNB)** * [Johannesburg to London latency and RTT](/docs/network/latency/pairs/jnb-lon-rtt): 158 ms * [Johannesburg to Paris latency and RTT](/docs/network/latency/pairs/jnb-par-rtt): 164.4 ms * [Johannesburg to Frankfurt latency and RTT](/docs/network/latency/pairs/jnb-fra-rtt): 166 ms * [Johannesburg to Marseille latency and RTT](/docs/network/latency/pairs/jnb-mrs-rtt): 172 ms * [Johannesburg to Berlin latency and RTT](/docs/network/latency/pairs/jnb-ber-rtt): 172.1 ms **Fastest routes arriving at Amsterdam (AMS)** * [London to Amsterdam latency and RTT](/docs/network/latency/pairs/lon-ams-rtt): 5.2 ms * [Frankfurt to Amsterdam latency and RTT](/docs/network/latency/pairs/fra-ams-rtt): 5.9 ms * [Paris to Amsterdam latency and RTT](/docs/network/latency/pairs/par-ams-rtt): 7 ms * [Berlin to Amsterdam latency and RTT](/docs/network/latency/pairs/ber-ams-rtt): 7.2 ms * [Marseille to Amsterdam latency and RTT](/docs/network/latency/pairs/mrs-ams-rtt): 20.3 ms **Same corridor (Africa → Europe)** * [Johannesburg to Moscow latency and RTT](/docs/network/latency/pairs/jnb-mow-rtt): 200.7 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/jnb-ams.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Johannesburg → Amsterdam round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/jnb-ams-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Johannesburg, South Africa → Berlin, Germany RTT 🇿🇦 **Johannesburg, South Africa (JNB)** → 🇩🇪 **Berlin, Germany (BER)** Measurement round: `2026-08-16T04:07:28Z`. Johannesburg to Berlin averaged 172.1 ms across 50 replies. The observed range was 168.33 to 187.75 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 3.76 ms, or 2.2% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------- | | RTT | **172.1 ms** | | Jitter | **2.87 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **86.53 ms** | | Fiber Efficiency | **50.3%** | | Latency Tier | Fair | | Source Region | [Africa](/docs/network/latency/regions/africa) | | Destination Region | [Europe](/docs/network/latency/regions/europe) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Johannesburg](https://www.geonames.org/993800) (-26.20227, 28.04363) and [Berlin](https://www.geonames.org/2950159) (52.52437, 13.41053). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **8,836.6 km** | | Vacuum RTT floor | **58.95 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **86.53 ms** | | Low-latency fiber material floor | **86.19 ms** | | Engineering floor (5% path allowance) | **90.87 ms** | | Research 1.33× mapped-fiber reference | **115.09 ms** | | Estimated unamplified path loss | **1855.7 dB** | | Transparent optical spans / inline amplifiers | **116 / 115** | | Published RTT inflation over fiber floor | **1.99×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **168.33 ms** | | Average RTT | **172.1 ms** | | Maximum RTT | **187.75 ms** | | Standard deviation | **3.76 ms** | | Stdev / average | **2.2%** | ## Route Context * Departure PoP: [Johannesburg (JNB)](/docs/network/latency/jnb-johannesburg) * Destination PoP: [Berlin (BER)](/docs/network/latency/ber-berlin) * Region overview: [Africa](/docs/network/latency/regions/africa) → [Europe](/docs/network/latency/regions/europe) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Berlin to Johannesburg latency and RTT](/docs/network/latency/pairs/ber-jnb-rtt): 172.1 ms **Fastest routes departing Johannesburg (JNB)** * [Johannesburg to London latency and RTT](/docs/network/latency/pairs/jnb-lon-rtt): 158 ms * [Johannesburg to Amsterdam latency and RTT](/docs/network/latency/pairs/jnb-ams-rtt): 163.2 ms * [Johannesburg to Paris latency and RTT](/docs/network/latency/pairs/jnb-par-rtt): 164.4 ms * [Johannesburg to Frankfurt latency and RTT](/docs/network/latency/pairs/jnb-fra-rtt): 166 ms * [Johannesburg to Marseille latency and RTT](/docs/network/latency/pairs/jnb-mrs-rtt): 172 ms **Fastest routes arriving at Berlin (BER)** * [Frankfurt to Berlin latency and RTT](/docs/network/latency/pairs/fra-ber-rtt): 6.1 ms * [Amsterdam to Berlin latency and RTT](/docs/network/latency/pairs/ams-ber-rtt): 7.2 ms * [Paris to Berlin latency and RTT](/docs/network/latency/pairs/par-ber-rtt): 15.5 ms * [London to Berlin latency and RTT](/docs/network/latency/pairs/lon-ber-rtt): 16.2 ms * [Marseille to Berlin latency and RTT](/docs/network/latency/pairs/mrs-ber-rtt): 20.6 ms **Same corridor (Africa → Europe)** * [Johannesburg to Moscow latency and RTT](/docs/network/latency/pairs/jnb-mow-rtt): 200.7 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/jnb-ber.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Johannesburg → Berlin round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/jnb-ber-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Johannesburg, South Africa → Frankfurt, Germany RTT 🇿🇦 **Johannesburg, South Africa (JNB)** → 🇩🇪 **Frankfurt, Germany (FRA)** Measurement round: `2026-08-16T04:07:28Z`. Johannesburg to Frankfurt averaged 166 ms across 50 replies. The observed range was 156.26 to 186.26 ms; these figures describe one probe round, not a long-term service target. This is number 4 of 19 outgoing routes by average RTT. That ranking compares different destinations and is not adjusted for distance. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------- | | RTT | **166 ms** | | Jitter | **7.73 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **84.87 ms** | | Fiber Efficiency | **51.1%** | | Latency Tier | Fair | | Source Region | [Africa](/docs/network/latency/regions/africa) | | Destination Region | [Europe](/docs/network/latency/regions/europe) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Johannesburg](https://www.geonames.org/993800) (-26.20227, 28.04363) and [Frankfurt am Main](https://www.geonames.org/2925533) (50.11552, 8.68417). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **8,666.9 km** | | Vacuum RTT floor | **57.82 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **84.87 ms** | | Low-latency fiber material floor | **84.53 ms** | | Engineering floor (5% path allowance) | **89.13 ms** | | Research 1.33× mapped-fiber reference | **112.88 ms** | | Estimated unamplified path loss | **1820.1 dB** | | Transparent optical spans / inline amplifiers | **114 / 113** | | Published RTT inflation over fiber floor | **1.96×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **156.26 ms** | | Average RTT | **166 ms** | | Maximum RTT | **186.26 ms** | | Standard deviation | **7.97 ms** | | Stdev / average | **4.8%** | ## Route Context * Departure PoP: [Johannesburg (JNB)](/docs/network/latency/jnb-johannesburg) * Destination PoP: [Frankfurt (FRA)](/docs/network/latency/fra-frankfurt) * Region overview: [Africa](/docs/network/latency/regions/africa) → [Europe](/docs/network/latency/regions/europe) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Frankfurt to Johannesburg latency and RTT](/docs/network/latency/pairs/fra-jnb-rtt): 188.4 ms **Fastest routes departing Johannesburg (JNB)** * [Johannesburg to London latency and RTT](/docs/network/latency/pairs/jnb-lon-rtt): 158 ms * [Johannesburg to Amsterdam latency and RTT](/docs/network/latency/pairs/jnb-ams-rtt): 163.2 ms * [Johannesburg to Paris latency and RTT](/docs/network/latency/pairs/jnb-par-rtt): 164.4 ms * [Johannesburg to Marseille latency and RTT](/docs/network/latency/pairs/jnb-mrs-rtt): 172 ms * [Johannesburg to Berlin latency and RTT](/docs/network/latency/pairs/jnb-ber-rtt): 172.1 ms **Fastest routes arriving at Frankfurt (FRA)** * [Amsterdam to Frankfurt latency and RTT](/docs/network/latency/pairs/ams-fra-rtt): 5.9 ms * [Berlin to Frankfurt latency and RTT](/docs/network/latency/pairs/ber-fra-rtt): 6.1 ms * [Paris to Frankfurt latency and RTT](/docs/network/latency/pairs/par-fra-rtt): 7.6 ms * [London to Frankfurt latency and RTT](/docs/network/latency/pairs/lon-fra-rtt): 13.6 ms * [Marseille to Frankfurt latency and RTT](/docs/network/latency/pairs/mrs-fra-rtt): 16 ms **Same corridor (Africa → Europe)** * [Johannesburg to Moscow latency and RTT](/docs/network/latency/pairs/jnb-mow-rtt): 200.7 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/jnb-fra.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Johannesburg → Frankfurt round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/jnb-fra-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Johannesburg, South Africa → São Paulo, Brazil RTT 🇿🇦 **Johannesburg, South Africa (JNB)** → 🇧🇷 **São Paulo, Brazil (GRU)** Measurement round: `2026-08-16T04:07:28Z`. Johannesburg to São Paulo averaged 328.7 ms, compared with a 72.88 ms city-centre fiber reference. The difference cannot be assigned to a particular cable or carrier from RTT alone. Standard deviation was 9.87 ms, or 3.0% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **328.7 ms** | | Jitter | **7.53 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **72.88 ms** | | Fiber Efficiency | **22.2%** | | Latency Tier | High | | Source Region | [Africa](/docs/network/latency/regions/africa) | | Destination Region | [South America](/docs/network/latency/regions/south-america) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Johannesburg](https://www.geonames.org/993800) (-26.20227, 28.04363) and [São Paulo](https://www.geonames.org/3448439) (-23.54750, -46.63611). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **7,441.9 km** | | Vacuum RTT floor | **49.65 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **72.88 ms** | | Low-latency fiber material floor | **72.58 ms** | | Engineering floor (5% path allowance) | **76.53 ms** | | Research 1.33× mapped-fiber reference | **96.93 ms** | | Estimated unamplified path loss | **1562.8 dB** | | Transparent optical spans / inline amplifiers | **98 / 97** | | Published RTT inflation over fiber floor | **4.51×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **319.02 ms** | | Average RTT | **328.7 ms** | | Maximum RTT | **366.86 ms** | | Standard deviation | **9.87 ms** | | Stdev / average | **3.0%** | ## Route Context * Departure PoP: [Johannesburg (JNB)](/docs/network/latency/jnb-johannesburg) * Destination PoP: [São Paulo (GRU)](/docs/network/latency/gru-sao-paulo) * Region overview: [Africa](/docs/network/latency/regions/africa) → [South America](/docs/network/latency/regions/south-america) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [São Paulo to Johannesburg latency and RTT](/docs/network/latency/pairs/gru-jnb-rtt): 333.8 ms **Fastest routes departing Johannesburg (JNB)** * [Johannesburg to London latency and RTT](/docs/network/latency/pairs/jnb-lon-rtt): 158 ms * [Johannesburg to Amsterdam latency and RTT](/docs/network/latency/pairs/jnb-ams-rtt): 163.2 ms * [Johannesburg to Paris latency and RTT](/docs/network/latency/pairs/jnb-par-rtt): 164.4 ms * [Johannesburg to Frankfurt latency and RTT](/docs/network/latency/pairs/jnb-fra-rtt): 166 ms * [Johannesburg to Marseille latency and RTT](/docs/network/latency/pairs/jnb-mrs-rtt): 172 ms **Fastest routes arriving at São Paulo (GRU)** * [Ashburn to São Paulo latency and RTT](/docs/network/latency/pairs/iad-gru-rtt): 101.4 ms * [New York to São Paulo latency and RTT](/docs/network/latency/pairs/nyc-gru-rtt): 106.9 ms * [Miami to São Paulo latency and RTT](/docs/network/latency/pairs/mia-gru-rtt): 128.7 ms * [Los Angeles to São Paulo latency and RTT](/docs/network/latency/pairs/lax-gru-rtt): 131.3 ms * [Seattle to São Paulo latency and RTT](/docs/network/latency/pairs/sea-gru-rtt): 155.7 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/jnb-gru.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Johannesburg → São Paulo round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/jnb-gru-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Johannesburg, South Africa → Hong Kong RTT 🇿🇦 **Johannesburg, South Africa (JNB)** → 🇭🇰 **Hong Kong (HKG)** Measurement round: `2026-08-16T04:07:28Z`. Johannesburg to Hong Kong averaged 318 ms, compared with a 104.93 ms city-centre fiber reference. The difference cannot be assigned to a particular cable or carrier from RTT alone. Standard deviation was 12.41 ms, or 3.9% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------------------- | | RTT | **318 ms** | | Jitter | **9.12 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **104.93 ms** | | Fiber Efficiency | **33%** | | Latency Tier | High | | Source Region | [Africa](/docs/network/latency/regions/africa) | | Destination Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Johannesburg](https://www.geonames.org/993800) (-26.20227, 28.04363) and [Hong Kong](https://www.geonames.org/1819729) (22.27832, 114.17469). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | --------------- | | WGS-84 geodesic distance | **10,715.5 km** | | Vacuum RTT floor | **71.49 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **104.93 ms** | | Low-latency fiber material floor | **104.51 ms** | | Engineering floor (5% path allowance) | **110.2 ms** | | Research 1.33× mapped-fiber reference | **139.56 ms** | | Estimated unamplified path loss | **2250.3 dB** | | Transparent optical spans / inline amplifiers | **141 / 140** | | Published RTT inflation over fiber floor | **3.03×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **302.56 ms** | | Average RTT | **318 ms** | | Maximum RTT | **363.62 ms** | | Standard deviation | **12.41 ms** | | Stdev / average | **3.9%** | ## Route Context * Departure PoP: [Johannesburg (JNB)](/docs/network/latency/jnb-johannesburg) * Destination PoP: [Hong Kong (HKG)](/docs/network/latency/hkg-hong-kong) * Region overview: [Africa](/docs/network/latency/regions/africa) → [Asia Pacific](/docs/network/latency/regions/asia-pacific) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Hong Kong to Johannesburg latency and RTT](/docs/network/latency/pairs/hkg-jnb-rtt): 316.7 ms **Fastest routes departing Johannesburg (JNB)** * [Johannesburg to London latency and RTT](/docs/network/latency/pairs/jnb-lon-rtt): 158 ms * [Johannesburg to Amsterdam latency and RTT](/docs/network/latency/pairs/jnb-ams-rtt): 163.2 ms * [Johannesburg to Paris latency and RTT](/docs/network/latency/pairs/jnb-par-rtt): 164.4 ms * [Johannesburg to Frankfurt latency and RTT](/docs/network/latency/pairs/jnb-fra-rtt): 166 ms * [Johannesburg to Marseille latency and RTT](/docs/network/latency/pairs/jnb-mrs-rtt): 172 ms **Fastest routes arriving at Hong Kong (HKG)** * [Taipei to Hong Kong latency and RTT](/docs/network/latency/pairs/tpe-hkg-rtt): 15.5 ms * [Singapore to Hong Kong latency and RTT](/docs/network/latency/pairs/sin-hkg-rtt): 30.8 ms * [Tokyo to Hong Kong latency and RTT](/docs/network/latency/pairs/tyo-hkg-rtt): 44.8 ms * [Moscow to Hong Kong latency and RTT](/docs/network/latency/pairs/mow-hkg-rtt): 118.6 ms * [Seattle to Hong Kong latency and RTT](/docs/network/latency/pairs/sea-hkg-rtt): 130.4 ms **Same corridor (Africa → Asia Pacific)** * [Johannesburg to Singapore latency and RTT](/docs/network/latency/pairs/jnb-sin-rtt): 311.6 ms * [Johannesburg to Taipei latency and RTT](/docs/network/latency/pairs/jnb-tpe-rtt): 329.8 ms * [Johannesburg to Tokyo latency and RTT](/docs/network/latency/pairs/jnb-tyo-rtt): 360.5 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/jnb-hkg.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Johannesburg → Hong Kong round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/jnb-hkg-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Johannesburg, South Africa → Ashburn, USA RTT 🇿🇦 **Johannesburg, South Africa (JNB)** → 🇺🇸 **Ashburn, USA (IAD)** Measurement round: `2026-08-16T04:07:28Z`. Johannesburg to Ashburn averaged 229 ms across 50 replies. The observed range was 223.41 to 246.68 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 5.3 ms, or 2.3% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **229 ms** | | Jitter | **4.4 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **128.11 ms** | | Fiber Efficiency | **55.9%** | | Latency Tier | Fair | | Source Region | [Africa](/docs/network/latency/regions/africa) | | Destination Region | [North America](/docs/network/latency/regions/north-america) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Johannesburg](https://www.geonames.org/993800) (-26.20227, 28.04363) and [Ashburn](https://www.geonames.org/4744870) (39.04372, -77.48749). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | ------------- | | WGS-84 geodesic distance | **13,082 km** | | Vacuum RTT floor | **87.27 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **128.11 ms** | | Low-latency fiber material floor | **127.59 ms** | | Engineering floor (5% path allowance) | **134.53 ms** | | Research 1.33× mapped-fiber reference | **170.38 ms** | | Estimated unamplified path loss | **2747.2 dB** | | Transparent optical spans / inline amplifiers | **172 / 171** | | Published RTT inflation over fiber floor | **1.79×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **223.41 ms** | | Average RTT | **229 ms** | | Maximum RTT | **246.68 ms** | | Standard deviation | **5.3 ms** | | Stdev / average | **2.3%** | ## Route Context * Departure PoP: [Johannesburg (JNB)](/docs/network/latency/jnb-johannesburg) * Destination PoP: [Ashburn (IAD)](/docs/network/latency/iad-ashburn) * Region overview: [Africa](/docs/network/latency/regions/africa) → [North America](/docs/network/latency/regions/north-america) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Ashburn to Johannesburg latency and RTT](/docs/network/latency/pairs/iad-jnb-rtt): 228.3 ms **Fastest routes departing Johannesburg (JNB)** * [Johannesburg to London latency and RTT](/docs/network/latency/pairs/jnb-lon-rtt): 158 ms * [Johannesburg to Amsterdam latency and RTT](/docs/network/latency/pairs/jnb-ams-rtt): 163.2 ms * [Johannesburg to Paris latency and RTT](/docs/network/latency/pairs/jnb-par-rtt): 164.4 ms * [Johannesburg to Frankfurt latency and RTT](/docs/network/latency/pairs/jnb-fra-rtt): 166 ms * [Johannesburg to Marseille latency and RTT](/docs/network/latency/pairs/jnb-mrs-rtt): 172 ms **Fastest routes arriving at Ashburn (IAD)** * [New York to Ashburn latency and RTT](/docs/network/latency/pairs/nyc-iad-rtt): 6 ms * [Miami to Ashburn latency and RTT](/docs/network/latency/pairs/mia-iad-rtt): 27.3 ms * [Los Angeles to Ashburn latency and RTT](/docs/network/latency/pairs/lax-iad-rtt): 60.4 ms * [Seattle to Ashburn latency and RTT](/docs/network/latency/pairs/sea-iad-rtt): 61.5 ms * [London to Ashburn latency and RTT](/docs/network/latency/pairs/lon-iad-rtt): 71 ms **Same corridor (Africa → North America)** * [Johannesburg to New York latency and RTT](/docs/network/latency/pairs/jnb-nyc-rtt): 221.8 ms * [Johannesburg to Miami latency and RTT](/docs/network/latency/pairs/jnb-mia-rtt): 259.9 ms * [Johannesburg to Seattle latency and RTT](/docs/network/latency/pairs/jnb-sea-rtt): 281.2 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/jnb-iad.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Johannesburg → Ashburn round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/jnb-iad-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Johannesburg, South Africa → Los Angeles, USA RTT 🇿🇦 **Johannesburg, South Africa (JNB)** → 🇺🇸 **Los Angeles, USA (LAX)** Measurement round: `2026-08-16T04:07:28Z`. Johannesburg to Los Angeles averaged 285.5 ms across 50 replies. The observed range was 274.6 to 308.95 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 8.74 ms, or 3.1% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **285.5 ms** | | Jitter | **7.72 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **163.35 ms** | | Fiber Efficiency | **57.2%** | | Latency Tier | High | | Source Region | [Africa](/docs/network/latency/regions/africa) | | Destination Region | [North America](/docs/network/latency/regions/north-america) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Johannesburg](https://www.geonames.org/993800) (-26.20227, 28.04363) and [Los Angeles](https://www.geonames.org/5368361) (34.05223, -118.24368). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | --------------- | | WGS-84 geodesic distance | **16,680.3 km** | | Vacuum RTT floor | **111.28 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **163.35 ms** | | Low-latency fiber material floor | **162.69 ms** | | Engineering floor (5% path allowance) | **171.54 ms** | | Research 1.33× mapped-fiber reference | **217.25 ms** | | Estimated unamplified path loss | **3502.9 dB** | | Transparent optical spans / inline amplifiers | **219 / 218** | | Published RTT inflation over fiber floor | **1.75×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **274.6 ms** | | Average RTT | **285.5 ms** | | Maximum RTT | **308.95 ms** | | Standard deviation | **8.74 ms** | | Stdev / average | **3.1%** | ## Route Context * Departure PoP: [Johannesburg (JNB)](/docs/network/latency/jnb-johannesburg) * Destination PoP: [Los Angeles (LAX)](/docs/network/latency/lax-los-angeles) * Region overview: [Africa](/docs/network/latency/regions/africa) → [North America](/docs/network/latency/regions/north-america) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Los Angeles to Johannesburg latency and RTT](/docs/network/latency/pairs/lax-jnb-rtt): 287.2 ms **Fastest routes departing Johannesburg (JNB)** * [Johannesburg to London latency and RTT](/docs/network/latency/pairs/jnb-lon-rtt): 158 ms * [Johannesburg to Amsterdam latency and RTT](/docs/network/latency/pairs/jnb-ams-rtt): 163.2 ms * [Johannesburg to Paris latency and RTT](/docs/network/latency/pairs/jnb-par-rtt): 164.4 ms * [Johannesburg to Frankfurt latency and RTT](/docs/network/latency/pairs/jnb-fra-rtt): 166 ms * [Johannesburg to Marseille latency and RTT](/docs/network/latency/pairs/jnb-mrs-rtt): 172 ms **Fastest routes arriving at Los Angeles (LAX)** * [Seattle to Los Angeles latency and RTT](/docs/network/latency/pairs/sea-lax-rtt): 25.9 ms * [Miami to Los Angeles latency and RTT](/docs/network/latency/pairs/mia-lax-rtt): 56.8 ms * [New York to Los Angeles latency and RTT](/docs/network/latency/pairs/nyc-lax-rtt): 58.9 ms * [Ashburn to Los Angeles latency and RTT](/docs/network/latency/pairs/iad-lax-rtt): 60.7 ms * [Tokyo to Los Angeles latency and RTT](/docs/network/latency/pairs/tyo-lax-rtt): 101.2 ms **Same corridor (Africa → North America)** * [Johannesburg to New York latency and RTT](/docs/network/latency/pairs/jnb-nyc-rtt): 221.8 ms * [Johannesburg to Ashburn latency and RTT](/docs/network/latency/pairs/jnb-iad-rtt): 229 ms * [Johannesburg to Miami latency and RTT](/docs/network/latency/pairs/jnb-mia-rtt): 259.9 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/jnb-lax.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Johannesburg → Los Angeles round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/jnb-lax-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Johannesburg, South Africa → London, UK RTT 🇿🇦 **Johannesburg, South Africa (JNB)** → 🇬🇧 **London, UK (LON)** Measurement round: `2026-08-16T04:07:28Z`. Johannesburg to London averaged 158 ms across 50 replies. The observed range was 151.59 to 170.89 ms; these figures describe one probe round, not a long-term service target. This is number 1 of 19 outgoing routes by average RTT. That ranking compares different destinations and is not adjusted for distance. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------- | | RTT | **158 ms** | | Jitter | **3.47 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **88.52 ms** | | Fiber Efficiency | **56%** | | Latency Tier | Fair | | Source Region | [Africa](/docs/network/latency/regions/africa) | | Destination Region | [Europe](/docs/network/latency/regions/europe) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Johannesburg](https://www.geonames.org/993800) (-26.20227, 28.04363) and [London](https://www.geonames.org/2643743) (51.50853, -0.12574). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **9,039.3 km** | | Vacuum RTT floor | **60.3 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **88.52 ms** | | Low-latency fiber material floor | **88.16 ms** | | Engineering floor (5% path allowance) | **92.96 ms** | | Research 1.33× mapped-fiber reference | **117.73 ms** | | Estimated unamplified path loss | **1898.3 dB** | | Transparent optical spans / inline amplifiers | **119 / 118** | | Published RTT inflation over fiber floor | **1.78×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **151.59 ms** | | Average RTT | **158 ms** | | Maximum RTT | **170.89 ms** | | Standard deviation | **4.48 ms** | | Stdev / average | **2.8%** | ## Route Context * Departure PoP: [Johannesburg (JNB)](/docs/network/latency/jnb-johannesburg) * Destination PoP: [London (LON)](/docs/network/latency/lon-london) * Region overview: [Africa](/docs/network/latency/regions/africa) → [Europe](/docs/network/latency/regions/europe) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [London to Johannesburg latency and RTT](/docs/network/latency/pairs/lon-jnb-rtt): 179.6 ms **Fastest routes departing Johannesburg (JNB)** * [Johannesburg to Amsterdam latency and RTT](/docs/network/latency/pairs/jnb-ams-rtt): 163.2 ms * [Johannesburg to Paris latency and RTT](/docs/network/latency/pairs/jnb-par-rtt): 164.4 ms * [Johannesburg to Frankfurt latency and RTT](/docs/network/latency/pairs/jnb-fra-rtt): 166 ms * [Johannesburg to Marseille latency and RTT](/docs/network/latency/pairs/jnb-mrs-rtt): 172 ms * [Johannesburg to Berlin latency and RTT](/docs/network/latency/pairs/jnb-ber-rtt): 172.1 ms **Fastest routes arriving at London (LON)** * [Amsterdam to London latency and RTT](/docs/network/latency/pairs/ams-lon-rtt): 5.2 ms * [Paris to London latency and RTT](/docs/network/latency/pairs/par-lon-rtt): 6.4 ms * [Frankfurt to London latency and RTT](/docs/network/latency/pairs/fra-lon-rtt): 12.9 ms * [Berlin to London latency and RTT](/docs/network/latency/pairs/ber-lon-rtt): 15.2 ms * [Marseille to London latency and RTT](/docs/network/latency/pairs/mrs-lon-rtt): 18.9 ms **Same corridor (Africa → Europe)** * [Johannesburg to Moscow latency and RTT](/docs/network/latency/pairs/jnb-mow-rtt): 200.7 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/jnb-lon.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Johannesburg → London round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/jnb-lon-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Johannesburg, South Africa → Melbourne, Australia RTT 🇿🇦 **Johannesburg, South Africa (JNB)** → 🇦🇺 **Melbourne, Australia (MEL)** Measurement round: `2026-08-16T04:07:28Z`. Johannesburg to Melbourne averaged 406.2 ms, compared with a 101.33 ms city-centre fiber reference. The difference cannot be assigned to a particular cable or carrier from RTT alone. Standard deviation was 16.4 ms, or 4.0% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------------------- | | RTT | **406.2 ms** | | Jitter | **10.85 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **101.33 ms** | | Fiber Efficiency | **24.9%** | | Latency Tier | High | | Source Region | [Africa](/docs/network/latency/regions/africa) | | Destination Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Johannesburg](https://www.geonames.org/993800) (-26.20227, 28.04363) and [Melbourne](https://www.geonames.org/2158177) (-37.81400, 144.96332). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | ------------- | | WGS-84 geodesic distance | **10,347 km** | | Vacuum RTT floor | **69.03 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **101.33 ms** | | Low-latency fiber material floor | **100.92 ms** | | Engineering floor (5% path allowance) | **106.41 ms** | | Research 1.33× mapped-fiber reference | **134.76 ms** | | Estimated unamplified path loss | **2172.9 dB** | | Transparent optical spans / inline amplifiers | **136 / 135** | | Published RTT inflation over fiber floor | **4.01×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **391.17 ms** | | Average RTT | **406.2 ms** | | Maximum RTT | **483.64 ms** | | Standard deviation | **16.4 ms** | | Stdev / average | **4.0%** | ## Route Context * Departure PoP: [Johannesburg (JNB)](/docs/network/latency/jnb-johannesburg) * Destination PoP: [Melbourne (MEL)](/docs/network/latency/mel-melbourne) * Region overview: [Africa](/docs/network/latency/regions/africa) → [Asia Pacific](/docs/network/latency/regions/asia-pacific) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Melbourne to Johannesburg latency and RTT](/docs/network/latency/pairs/mel-jnb-rtt): 405.5 ms **Fastest routes departing Johannesburg (JNB)** * [Johannesburg to London latency and RTT](/docs/network/latency/pairs/jnb-lon-rtt): 158 ms * [Johannesburg to Amsterdam latency and RTT](/docs/network/latency/pairs/jnb-ams-rtt): 163.2 ms * [Johannesburg to Paris latency and RTT](/docs/network/latency/pairs/jnb-par-rtt): 164.4 ms * [Johannesburg to Frankfurt latency and RTT](/docs/network/latency/pairs/jnb-fra-rtt): 166 ms * [Johannesburg to Marseille latency and RTT](/docs/network/latency/pairs/jnb-mrs-rtt): 172 ms **Fastest routes arriving at Melbourne (MEL)** * [Sydney to Melbourne latency and RTT](/docs/network/latency/pairs/syd-mel-rtt): 9.8 ms * [Singapore to Melbourne latency and RTT](/docs/network/latency/pairs/sin-mel-rtt): 88.6 ms * [Tokyo to Melbourne latency and RTT](/docs/network/latency/pairs/tyo-mel-rtt): 113.9 ms * [Hong Kong to Melbourne latency and RTT](/docs/network/latency/pairs/hkg-mel-rtt): 138.4 ms * [Taipei to Melbourne latency and RTT](/docs/network/latency/pairs/tpe-mel-rtt): 142 ms **Same corridor (Africa → Asia Pacific)** * [Johannesburg to Singapore latency and RTT](/docs/network/latency/pairs/jnb-sin-rtt): 311.6 ms * [Johannesburg to Hong Kong latency and RTT](/docs/network/latency/pairs/jnb-hkg-rtt): 318 ms * [Johannesburg to Taipei latency and RTT](/docs/network/latency/pairs/jnb-tpe-rtt): 329.8 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/jnb-mel.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Johannesburg → Melbourne round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/jnb-mel-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Johannesburg, South Africa → Miami, USA RTT 🇿🇦 **Johannesburg, South Africa (JNB)** → 🇺🇸 **Miami, USA (MIA)** Measurement round: `2026-08-16T04:07:28Z`. Johannesburg to Miami averaged 259.9 ms, compared with a 126.78 ms city-centre fiber reference. The difference cannot be assigned to a particular cable or carrier from RTT alone. Standard deviation was 5.02 ms, or 1.9% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **259.9 ms** | | Jitter | **4.08 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **126.78 ms** | | Fiber Efficiency | **48.8%** | | Latency Tier | High | | Source Region | [Africa](/docs/network/latency/regions/africa) | | Destination Region | [North America](/docs/network/latency/regions/north-america) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Johannesburg](https://www.geonames.org/993800) (-26.20227, 28.04363) and [Miami](https://www.geonames.org/4164138) (25.77427, -80.19366). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | --------------- | | WGS-84 geodesic distance | **12,945.8 km** | | Vacuum RTT floor | **86.36 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **126.78 ms** | | Low-latency fiber material floor | **126.27 ms** | | Engineering floor (5% path allowance) | **133.13 ms** | | Research 1.33× mapped-fiber reference | **168.61 ms** | | Estimated unamplified path loss | **2718.6 dB** | | Transparent optical spans / inline amplifiers | **170 / 169** | | Published RTT inflation over fiber floor | **2.05×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **252.44 ms** | | Average RTT | **259.9 ms** | | Maximum RTT | **275.29 ms** | | Standard deviation | **5.02 ms** | | Stdev / average | **1.9%** | ## Route Context * Departure PoP: [Johannesburg (JNB)](/docs/network/latency/jnb-johannesburg) * Destination PoP: [Miami (MIA)](/docs/network/latency/mia-miami) * Region overview: [Africa](/docs/network/latency/regions/africa) → [North America](/docs/network/latency/regions/north-america) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Miami to Johannesburg latency and RTT](/docs/network/latency/pairs/mia-jnb-rtt): 259.8 ms **Fastest routes departing Johannesburg (JNB)** * [Johannesburg to London latency and RTT](/docs/network/latency/pairs/jnb-lon-rtt): 158 ms * [Johannesburg to Amsterdam latency and RTT](/docs/network/latency/pairs/jnb-ams-rtt): 163.2 ms * [Johannesburg to Paris latency and RTT](/docs/network/latency/pairs/jnb-par-rtt): 164.4 ms * [Johannesburg to Frankfurt latency and RTT](/docs/network/latency/pairs/jnb-fra-rtt): 166 ms * [Johannesburg to Marseille latency and RTT](/docs/network/latency/pairs/jnb-mrs-rtt): 172 ms **Fastest routes arriving at Miami (MIA)** * [Ashburn to Miami latency and RTT](/docs/network/latency/pairs/iad-mia-rtt): 27.4 ms * [New York to Miami latency and RTT](/docs/network/latency/pairs/nyc-mia-rtt): 32.9 ms * [Los Angeles to Miami latency and RTT](/docs/network/latency/pairs/lax-mia-rtt): 57.3 ms * [São Paulo to Miami latency and RTT](/docs/network/latency/pairs/gru-mia-rtt): 74 ms * [Seattle to Miami latency and RTT](/docs/network/latency/pairs/sea-mia-rtt): 81.7 ms **Same corridor (Africa → North America)** * [Johannesburg to New York latency and RTT](/docs/network/latency/pairs/jnb-nyc-rtt): 221.8 ms * [Johannesburg to Ashburn latency and RTT](/docs/network/latency/pairs/jnb-iad-rtt): 229 ms * [Johannesburg to Seattle latency and RTT](/docs/network/latency/pairs/jnb-sea-rtt): 281.2 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/jnb-mia.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Johannesburg → Miami round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/jnb-mia-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Johannesburg, South Africa → Moscow, Russia RTT 🇿🇦 **Johannesburg, South Africa (JNB)** → 🇷🇺 **Moscow, Russia (MOW)** Measurement round: `2026-08-16T04:07:28Z`. Johannesburg to Moscow averaged 200.7 ms, compared with a 89.37 ms city-centre fiber reference. The difference cannot be assigned to a particular cable or carrier from RTT alone. Standard deviation was 4.34 ms, or 2.2% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------- | | RTT | **200.7 ms** | | Jitter | **4.03 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **89.37 ms** | | Fiber Efficiency | **44.5%** | | Latency Tier | Fair | | Source Region | [Africa](/docs/network/latency/regions/africa) | | Destination Region | [Europe](/docs/network/latency/regions/europe) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Johannesburg](https://www.geonames.org/993800) (-26.20227, 28.04363) and [Moscow](https://www.geonames.org/524901) (55.75204, 37.61781). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **9,125.7 km** | | Vacuum RTT floor | **60.88 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **89.37 ms** | | Low-latency fiber material floor | **89.01 ms** | | Engineering floor (5% path allowance) | **93.85 ms** | | Research 1.33× mapped-fiber reference | **118.86 ms** | | Estimated unamplified path loss | **1916.4 dB** | | Transparent optical spans / inline amplifiers | **120 / 119** | | Published RTT inflation over fiber floor | **2.25×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **196.07 ms** | | Average RTT | **200.7 ms** | | Maximum RTT | **218.52 ms** | | Standard deviation | **4.34 ms** | | Stdev / average | **2.2%** | ## Route Context * Departure PoP: [Johannesburg (JNB)](/docs/network/latency/jnb-johannesburg) * Destination PoP: [Moscow (MOW)](/docs/network/latency/mow-moscow) * Region overview: [Africa](/docs/network/latency/regions/africa) → [Europe](/docs/network/latency/regions/europe) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Moscow to Johannesburg latency and RTT](/docs/network/latency/pairs/mow-jnb-rtt): 201.9 ms **Fastest routes departing Johannesburg (JNB)** * [Johannesburg to London latency and RTT](/docs/network/latency/pairs/jnb-lon-rtt): 158 ms * [Johannesburg to Amsterdam latency and RTT](/docs/network/latency/pairs/jnb-ams-rtt): 163.2 ms * [Johannesburg to Paris latency and RTT](/docs/network/latency/pairs/jnb-par-rtt): 164.4 ms * [Johannesburg to Frankfurt latency and RTT](/docs/network/latency/pairs/jnb-fra-rtt): 166 ms * [Johannesburg to Marseille latency and RTT](/docs/network/latency/pairs/jnb-mrs-rtt): 172 ms **Fastest routes arriving at Moscow (MOW)** * [Berlin to Moscow latency and RTT](/docs/network/latency/pairs/ber-mow-rtt): 28.6 ms * [Frankfurt to Moscow latency and RTT](/docs/network/latency/pairs/fra-mow-rtt): 34.7 ms * [Amsterdam to Moscow latency and RTT](/docs/network/latency/pairs/ams-mow-rtt): 38.1 ms * [London to Moscow latency and RTT](/docs/network/latency/pairs/lon-mow-rtt): 43.5 ms * [Paris to Moscow latency and RTT](/docs/network/latency/pairs/par-mow-rtt): 44.7 ms **Same corridor (Africa → Europe)** * [Johannesburg to Berlin latency and RTT](/docs/network/latency/pairs/jnb-ber-rtt): 172.1 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/jnb-mow.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Johannesburg → Moscow round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/jnb-mow-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Johannesburg, South Africa → Marseille, France RTT 🇿🇦 **Johannesburg, South Africa (JNB)** → 🇫🇷 **Marseille, France (MRS)** Measurement round: `2026-08-16T04:07:28Z`. Johannesburg to Marseille averaged 172 ms, compared with a 78.71 ms city-centre fiber reference. The difference cannot be assigned to a particular cable or carrier from RTT alone. This is number 5 of 19 outgoing routes by average RTT. That ranking compares different destinations and is not adjusted for distance. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------- | | RTT | **172 ms** | | Jitter | **2.46 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **78.71 ms** | | Fiber Efficiency | **45.8%** | | Latency Tier | Fair | | Source Region | [Africa](/docs/network/latency/regions/africa) | | Destination Region | [Europe](/docs/network/latency/regions/europe) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Johannesburg](https://www.geonames.org/993800) (-26.20227, 28.04363) and [Marseille](https://www.geonames.org/2995469) (43.29695, 5.38107). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **8,037.1 km** | | Vacuum RTT floor | **53.62 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **78.71 ms** | | Low-latency fiber material floor | **78.39 ms** | | Engineering floor (5% path allowance) | **82.65 ms** | | Research 1.33× mapped-fiber reference | **104.68 ms** | | Estimated unamplified path loss | **1687.8 dB** | | Transparent optical spans / inline amplifiers | **106 / 105** | | Published RTT inflation over fiber floor | **2.19×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **168.24 ms** | | Average RTT | **172 ms** | | Maximum RTT | **182.06 ms** | | Standard deviation | **3.29 ms** | | Stdev / average | **1.9%** | ## Route Context * Departure PoP: [Johannesburg (JNB)](/docs/network/latency/jnb-johannesburg) * Destination PoP: [Marseille (MRS)](/docs/network/latency/mrs-marseille) * Region overview: [Africa](/docs/network/latency/regions/africa) → [Europe](/docs/network/latency/regions/europe) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Marseille to Johannesburg latency and RTT](/docs/network/latency/pairs/mrs-jnb-rtt): 198.5 ms **Fastest routes departing Johannesburg (JNB)** * [Johannesburg to London latency and RTT](/docs/network/latency/pairs/jnb-lon-rtt): 158 ms * [Johannesburg to Amsterdam latency and RTT](/docs/network/latency/pairs/jnb-ams-rtt): 163.2 ms * [Johannesburg to Paris latency and RTT](/docs/network/latency/pairs/jnb-par-rtt): 164.4 ms * [Johannesburg to Frankfurt latency and RTT](/docs/network/latency/pairs/jnb-fra-rtt): 166 ms * [Johannesburg to Berlin latency and RTT](/docs/network/latency/pairs/jnb-ber-rtt): 172.1 ms **Fastest routes arriving at Marseille (MRS)** * [Paris to Marseille latency and RTT](/docs/network/latency/pairs/par-mrs-rtt): 8.9 ms * [Frankfurt to Marseille latency and RTT](/docs/network/latency/pairs/fra-mrs-rtt): 16.4 ms * [London to Marseille latency and RTT](/docs/network/latency/pairs/lon-mrs-rtt): 17.9 ms * [Amsterdam to Marseille latency and RTT](/docs/network/latency/pairs/ams-mrs-rtt): 19.6 ms * [Berlin to Marseille latency and RTT](/docs/network/latency/pairs/ber-mrs-rtt): 20.2 ms **Same corridor (Africa → Europe)** * [Johannesburg to Moscow latency and RTT](/docs/network/latency/pairs/jnb-mow-rtt): 200.7 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/jnb-mrs.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Johannesburg → Marseille round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/jnb-mrs-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Johannesburg, South Africa → New York, USA RTT 🇿🇦 **Johannesburg, South Africa (JNB)** → 🇺🇸 **New York, USA (NYC)** Measurement round: `2026-08-16T04:07:28Z`. Johannesburg to New York averaged 221.8 ms across 50 replies. The observed range was 213.14 to 261.72 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 8.43 ms, or 3.8% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **221.8 ms** | | Jitter | **6.44 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **125.67 ms** | | Fiber Efficiency | **56.7%** | | Latency Tier | Fair | | Source Region | [Africa](/docs/network/latency/regions/africa) | | Destination Region | [North America](/docs/network/latency/regions/north-america) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Johannesburg](https://www.geonames.org/993800) (-26.20227, 28.04363) and [New York City](https://www.geonames.org/5128581) (40.71427, -74.00597). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | --------------- | | WGS-84 geodesic distance | **12,832.9 km** | | Vacuum RTT floor | **85.61 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **125.67 ms** | | Low-latency fiber material floor | **125.16 ms** | | Engineering floor (5% path allowance) | **131.97 ms** | | Research 1.33× mapped-fiber reference | **167.14 ms** | | Estimated unamplified path loss | **2694.9 dB** | | Transparent optical spans / inline amplifiers | **169 / 168** | | Published RTT inflation over fiber floor | **1.76×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **213.14 ms** | | Average RTT | **221.8 ms** | | Maximum RTT | **261.72 ms** | | Standard deviation | **8.43 ms** | | Stdev / average | **3.8%** | ## Route Context * Departure PoP: [Johannesburg (JNB)](/docs/network/latency/jnb-johannesburg) * Destination PoP: [New York (NYC)](/docs/network/latency/nyc-new-york) * Region overview: [Africa](/docs/network/latency/regions/africa) → [North America](/docs/network/latency/regions/north-america) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [New York to Johannesburg latency and RTT](/docs/network/latency/pairs/nyc-jnb-rtt): 221.5 ms **Fastest routes departing Johannesburg (JNB)** * [Johannesburg to London latency and RTT](/docs/network/latency/pairs/jnb-lon-rtt): 158 ms * [Johannesburg to Amsterdam latency and RTT](/docs/network/latency/pairs/jnb-ams-rtt): 163.2 ms * [Johannesburg to Paris latency and RTT](/docs/network/latency/pairs/jnb-par-rtt): 164.4 ms * [Johannesburg to Frankfurt latency and RTT](/docs/network/latency/pairs/jnb-fra-rtt): 166 ms * [Johannesburg to Marseille latency and RTT](/docs/network/latency/pairs/jnb-mrs-rtt): 172 ms **Fastest routes arriving at New York (NYC)** * [Ashburn to New York latency and RTT](/docs/network/latency/pairs/iad-nyc-rtt): 6 ms * [Miami to New York latency and RTT](/docs/network/latency/pairs/mia-nyc-rtt): 33.3 ms * [Los Angeles to New York latency and RTT](/docs/network/latency/pairs/lax-nyc-rtt): 57.9 ms * [Seattle to New York latency and RTT](/docs/network/latency/pairs/sea-nyc-rtt): 59.5 ms * [London to New York latency and RTT](/docs/network/latency/pairs/lon-nyc-rtt): 63.8 ms **Same corridor (Africa → North America)** * [Johannesburg to Ashburn latency and RTT](/docs/network/latency/pairs/jnb-iad-rtt): 229 ms * [Johannesburg to Miami latency and RTT](/docs/network/latency/pairs/jnb-mia-rtt): 259.9 ms * [Johannesburg to Seattle latency and RTT](/docs/network/latency/pairs/jnb-sea-rtt): 281.2 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/jnb-nyc.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Johannesburg → New York round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/jnb-nyc-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Johannesburg, South Africa → Paris, France RTT 🇿🇦 **Johannesburg, South Africa (JNB)** → 🇫🇷 **Paris, France (PAR)** Measurement round: `2026-08-16T04:07:28Z`. Johannesburg to Paris averaged 164.4 ms across 50 replies. The observed range was 158.64 to 177.46 ms; these figures describe one probe round, not a long-term service target. This is number 3 of 19 outgoing routes by average RTT. That ranking compares different destinations and is not adjusted for distance. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------- | | RTT | **164.4 ms** | | Jitter | **4.24 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **85.17 ms** | | Fiber Efficiency | **51.8%** | | Latency Tier | Fair | | Source Region | [Africa](/docs/network/latency/regions/africa) | | Destination Region | [Europe](/docs/network/latency/regions/europe) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Johannesburg](https://www.geonames.org/993800) (-26.20227, 28.04363) and [Paris](https://www.geonames.org/2988507) (48.85341, 2.34880). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **8,697.3 km** | | Vacuum RTT floor | **58.02 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **85.17 ms** | | Low-latency fiber material floor | **84.83 ms** | | Engineering floor (5% path allowance) | **89.44 ms** | | Research 1.33× mapped-fiber reference | **113.28 ms** | | Estimated unamplified path loss | **1826.4 dB** | | Transparent optical spans / inline amplifiers | **115 / 114** | | Published RTT inflation over fiber floor | **1.93×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **158.64 ms** | | Average RTT | **164.4 ms** | | Maximum RTT | **177.46 ms** | | Standard deviation | **4.15 ms** | | Stdev / average | **2.5%** | ## Route Context * Departure PoP: [Johannesburg (JNB)](/docs/network/latency/jnb-johannesburg) * Destination PoP: [Paris (PAR)](/docs/network/latency/par-paris) * Region overview: [Africa](/docs/network/latency/regions/africa) → [Europe](/docs/network/latency/regions/europe) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Paris to Johannesburg latency and RTT](/docs/network/latency/pairs/par-jnb-rtt): 180.9 ms **Fastest routes departing Johannesburg (JNB)** * [Johannesburg to London latency and RTT](/docs/network/latency/pairs/jnb-lon-rtt): 158 ms * [Johannesburg to Amsterdam latency and RTT](/docs/network/latency/pairs/jnb-ams-rtt): 163.2 ms * [Johannesburg to Frankfurt latency and RTT](/docs/network/latency/pairs/jnb-fra-rtt): 166 ms * [Johannesburg to Marseille latency and RTT](/docs/network/latency/pairs/jnb-mrs-rtt): 172 ms * [Johannesburg to Berlin latency and RTT](/docs/network/latency/pairs/jnb-ber-rtt): 172.1 ms **Fastest routes arriving at Paris (PAR)** * [London to Paris latency and RTT](/docs/network/latency/pairs/lon-par-rtt): 6.4 ms * [Amsterdam to Paris latency and RTT](/docs/network/latency/pairs/ams-par-rtt): 7 ms * [Frankfurt to Paris latency and RTT](/docs/network/latency/pairs/fra-par-rtt): 7.6 ms * [Marseille to Paris latency and RTT](/docs/network/latency/pairs/mrs-par-rtt): 8.9 ms * [Berlin to Paris latency and RTT](/docs/network/latency/pairs/ber-par-rtt): 15.6 ms **Same corridor (Africa → Europe)** * [Johannesburg to Moscow latency and RTT](/docs/network/latency/pairs/jnb-mow-rtt): 200.7 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/jnb-par.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Johannesburg → Paris round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/jnb-par-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Johannesburg, South Africa → Seattle, USA RTT 🇿🇦 **Johannesburg, South Africa (JNB)** → 🇺🇸 **Seattle, USA (SEA)** Measurement round: `2026-08-16T04:07:28Z`. Johannesburg to Seattle averaged 281.2 ms across 50 replies. The observed range was 267.91 to 315.6 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 9.73 ms, or 3.5% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **281.2 ms** | | Jitter | **9.21 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **161.61 ms** | | Fiber Efficiency | **57.5%** | | Latency Tier | High | | Source Region | [Africa](/docs/network/latency/regions/africa) | | Destination Region | [North America](/docs/network/latency/regions/north-america) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Johannesburg](https://www.geonames.org/993800) (-26.20227, 28.04363) and [Seattle](https://www.geonames.org/5809844) (47.60621, -122.33207). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | --------------- | | WGS-84 geodesic distance | **16,502.7 km** | | Vacuum RTT floor | **110.09 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **161.61 ms** | | Low-latency fiber material floor | **160.96 ms** | | Engineering floor (5% path allowance) | **169.71 ms** | | Research 1.33× mapped-fiber reference | **214.94 ms** | | Estimated unamplified path loss | **3465.6 dB** | | Transparent optical spans / inline amplifiers | **217 / 216** | | Published RTT inflation over fiber floor | **1.74×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **267.91 ms** | | Average RTT | **281.2 ms** | | Maximum RTT | **315.6 ms** | | Standard deviation | **9.73 ms** | | Stdev / average | **3.5%** | ## Route Context * Departure PoP: [Johannesburg (JNB)](/docs/network/latency/jnb-johannesburg) * Destination PoP: [Seattle (SEA)](/docs/network/latency/sea-seattle) * Region overview: [Africa](/docs/network/latency/regions/africa) → [North America](/docs/network/latency/regions/north-america) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Seattle to Johannesburg latency and RTT](/docs/network/latency/pairs/sea-jnb-rtt): 282.1 ms **Fastest routes departing Johannesburg (JNB)** * [Johannesburg to London latency and RTT](/docs/network/latency/pairs/jnb-lon-rtt): 158 ms * [Johannesburg to Amsterdam latency and RTT](/docs/network/latency/pairs/jnb-ams-rtt): 163.2 ms * [Johannesburg to Paris latency and RTT](/docs/network/latency/pairs/jnb-par-rtt): 164.4 ms * [Johannesburg to Frankfurt latency and RTT](/docs/network/latency/pairs/jnb-fra-rtt): 166 ms * [Johannesburg to Marseille latency and RTT](/docs/network/latency/pairs/jnb-mrs-rtt): 172 ms **Fastest routes arriving at Seattle (SEA)** * [Los Angeles to Seattle latency and RTT](/docs/network/latency/pairs/lax-sea-rtt): 26.9 ms * [New York to Seattle latency and RTT](/docs/network/latency/pairs/nyc-sea-rtt): 58.7 ms * [Ashburn to Seattle latency and RTT](/docs/network/latency/pairs/iad-sea-rtt): 62.3 ms * [Miami to Seattle latency and RTT](/docs/network/latency/pairs/mia-sea-rtt): 81.9 ms * [Tokyo to Seattle latency and RTT](/docs/network/latency/pairs/tyo-sea-rtt): 84.8 ms **Same corridor (Africa → North America)** * [Johannesburg to New York latency and RTT](/docs/network/latency/pairs/jnb-nyc-rtt): 221.8 ms * [Johannesburg to Ashburn latency and RTT](/docs/network/latency/pairs/jnb-iad-rtt): 229 ms * [Johannesburg to Miami latency and RTT](/docs/network/latency/pairs/jnb-mia-rtt): 259.9 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/jnb-sea.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Johannesburg → Seattle round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/jnb-sea-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Johannesburg, South Africa → Singapore RTT 🇿🇦 **Johannesburg, South Africa (JNB)** → 🇸🇬 **Singapore (SIN)** Measurement round: `2026-08-16T04:07:28Z`. Johannesburg to Singapore averaged 311.6 ms, compared with a 84.85 ms city-centre fiber reference. The difference cannot be assigned to a particular cable or carrier from RTT alone. Standard deviation was 9.83 ms, or 3.2% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------------------- | | RTT | **311.6 ms** | | Jitter | **7.03 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **84.85 ms** | | Fiber Efficiency | **27.2%** | | Latency Tier | High | | Source Region | [Africa](/docs/network/latency/regions/africa) | | Destination Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Johannesburg](https://www.geonames.org/993800) (-26.20227, 28.04363) and [Singapore](https://www.geonames.org/1880252) (1.28967, 103.85007). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **8,664.2 km** | | Vacuum RTT floor | **57.8 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **84.85 ms** | | Low-latency fiber material floor | **84.51 ms** | | Engineering floor (5% path allowance) | **89.1 ms** | | Research 1.33× mapped-fiber reference | **112.85 ms** | | Estimated unamplified path loss | **1819.5 dB** | | Transparent optical spans / inline amplifiers | **114 / 113** | | Published RTT inflation over fiber floor | **3.67×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **301.09 ms** | | Average RTT | **311.6 ms** | | Maximum RTT | **347.05 ms** | | Standard deviation | **9.83 ms** | | Stdev / average | **3.2%** | ## Route Context * Departure PoP: [Johannesburg (JNB)](/docs/network/latency/jnb-johannesburg) * Destination PoP: [Singapore (SIN)](/docs/network/latency/sin-singapore) * Region overview: [Africa](/docs/network/latency/regions/africa) → [Asia Pacific](/docs/network/latency/regions/asia-pacific) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Singapore to Johannesburg latency and RTT](/docs/network/latency/pairs/sin-jnb-rtt): 312.6 ms **Fastest routes departing Johannesburg (JNB)** * [Johannesburg to London latency and RTT](/docs/network/latency/pairs/jnb-lon-rtt): 158 ms * [Johannesburg to Amsterdam latency and RTT](/docs/network/latency/pairs/jnb-ams-rtt): 163.2 ms * [Johannesburg to Paris latency and RTT](/docs/network/latency/pairs/jnb-par-rtt): 164.4 ms * [Johannesburg to Frankfurt latency and RTT](/docs/network/latency/pairs/jnb-fra-rtt): 166 ms * [Johannesburg to Marseille latency and RTT](/docs/network/latency/pairs/jnb-mrs-rtt): 172 ms **Fastest routes arriving at Singapore (SIN)** * [Hong Kong to Singapore latency and RTT](/docs/network/latency/pairs/hkg-sin-rtt): 31.5 ms * [Taipei to Singapore latency and RTT](/docs/network/latency/pairs/tpe-sin-rtt): 44.3 ms * [Tokyo to Singapore latency and RTT](/docs/network/latency/pairs/tyo-sin-rtt): 69.4 ms * [Melbourne to Singapore latency and RTT](/docs/network/latency/pairs/mel-sin-rtt): 88.4 ms * [Sydney to Singapore latency and RTT](/docs/network/latency/pairs/syd-sin-rtt): 94.5 ms **Same corridor (Africa → Asia Pacific)** * [Johannesburg to Hong Kong latency and RTT](/docs/network/latency/pairs/jnb-hkg-rtt): 318 ms * [Johannesburg to Taipei latency and RTT](/docs/network/latency/pairs/jnb-tpe-rtt): 329.8 ms * [Johannesburg to Tokyo latency and RTT](/docs/network/latency/pairs/jnb-tyo-rtt): 360.5 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/jnb-sin.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Johannesburg → Singapore round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/jnb-sin-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Johannesburg, South Africa → Sydney, Australia RTT 🇿🇦 **Johannesburg, South Africa (JNB)** → 🇦🇺 **Sydney, Australia (SYD)** Measurement round: `2026-08-16T04:07:28Z`. Johannesburg to Sydney averaged 412.6 ms, compared with a 108.32 ms city-centre fiber reference. The difference cannot be assigned to a particular cable or carrier from RTT alone. Standard deviation was 16.76 ms, or 4.1% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------------------- | | RTT | **412.6 ms** | | Jitter | **13.13 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **108.32 ms** | | Fiber Efficiency | **26.3%** | | Latency Tier | High | | Source Region | [Africa](/docs/network/latency/regions/africa) | | Destination Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Johannesburg](https://www.geonames.org/993800) (-26.20227, 28.04363) and [Sydney](https://www.geonames.org/2147714) (-33.86785, 151.20732). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | --------------- | | WGS-84 geodesic distance | **11,060.7 km** | | Vacuum RTT floor | **73.79 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **108.32 ms** | | Low-latency fiber material floor | **107.88 ms** | | Engineering floor (5% path allowance) | **113.75 ms** | | Research 1.33× mapped-fiber reference | **144.06 ms** | | Estimated unamplified path loss | **2322.8 dB** | | Transparent optical spans / inline amplifiers | **146 / 145** | | Published RTT inflation over fiber floor | **3.81×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **391.98 ms** | | Average RTT | **412.6 ms** | | Maximum RTT | **456.66 ms** | | Standard deviation | **16.76 ms** | | Stdev / average | **4.1%** | ## Route Context * Departure PoP: [Johannesburg (JNB)](/docs/network/latency/jnb-johannesburg) * Destination PoP: [Sydney (SYD)](/docs/network/latency/syd-sydney) * Region overview: [Africa](/docs/network/latency/regions/africa) → [Asia Pacific](/docs/network/latency/regions/asia-pacific) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Sydney to Johannesburg latency and RTT](/docs/network/latency/pairs/syd-jnb-rtt): 412.5 ms **Fastest routes departing Johannesburg (JNB)** * [Johannesburg to London latency and RTT](/docs/network/latency/pairs/jnb-lon-rtt): 158 ms * [Johannesburg to Amsterdam latency and RTT](/docs/network/latency/pairs/jnb-ams-rtt): 163.2 ms * [Johannesburg to Paris latency and RTT](/docs/network/latency/pairs/jnb-par-rtt): 164.4 ms * [Johannesburg to Frankfurt latency and RTT](/docs/network/latency/pairs/jnb-fra-rtt): 166 ms * [Johannesburg to Marseille latency and RTT](/docs/network/latency/pairs/jnb-mrs-rtt): 172 ms **Fastest routes arriving at Sydney (SYD)** * [Melbourne to Sydney latency and RTT](/docs/network/latency/pairs/mel-syd-rtt): 9.8 ms * [Singapore to Sydney latency and RTT](/docs/network/latency/pairs/sin-syd-rtt): 94.5 ms * [Tokyo to Sydney latency and RTT](/docs/network/latency/pairs/tyo-syd-rtt): 101.9 ms * [Taipei to Sydney latency and RTT](/docs/network/latency/pairs/tpe-syd-rtt): 133.2 ms * [Hong Kong to Sydney latency and RTT](/docs/network/latency/pairs/hkg-syd-rtt): 137.2 ms **Same corridor (Africa → Asia Pacific)** * [Johannesburg to Singapore latency and RTT](/docs/network/latency/pairs/jnb-sin-rtt): 311.6 ms * [Johannesburg to Hong Kong latency and RTT](/docs/network/latency/pairs/jnb-hkg-rtt): 318 ms * [Johannesburg to Taipei latency and RTT](/docs/network/latency/pairs/jnb-tpe-rtt): 329.8 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/jnb-syd.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Johannesburg → Sydney round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/jnb-syd-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Johannesburg, South Africa → Taipei, Taiwan RTT 🇿🇦 **Johannesburg, South Africa (JNB)** → 🇹🇼 **Taipei, Taiwan (TPE)** Measurement round: `2026-08-16T04:07:28Z`. Johannesburg to Taipei averaged 329.8 ms, compared with a 112.87 ms city-centre fiber reference. The difference cannot be assigned to a particular cable or carrier from RTT alone. Standard deviation was 9.77 ms, or 3.0% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------------------- | | RTT | **329.8 ms** | | Jitter | **9.42 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **112.87 ms** | | Fiber Efficiency | **34.2%** | | Latency Tier | High | | Source Region | [Africa](/docs/network/latency/regions/africa) | | Destination Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Johannesburg](https://www.geonames.org/993800) (-26.20227, 28.04363) and [Taipei](https://www.geonames.org/1668341) (25.05306, 121.52639). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | --------------- | | WGS-84 geodesic distance | **11,525.4 km** | | Vacuum RTT floor | **76.89 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **112.87 ms** | | Low-latency fiber material floor | **112.41 ms** | | Engineering floor (5% path allowance) | **118.52 ms** | | Research 1.33× mapped-fiber reference | **150.11 ms** | | Estimated unamplified path loss | **2420.3 dB** | | Transparent optical spans / inline amplifiers | **152 / 151** | | Published RTT inflation over fiber floor | **2.92×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **316.93 ms** | | Average RTT | **329.8 ms** | | Maximum RTT | **358.31 ms** | | Standard deviation | **9.77 ms** | | Stdev / average | **3.0%** | ## Route Context * Departure PoP: [Johannesburg (JNB)](/docs/network/latency/jnb-johannesburg) * Destination PoP: [Taipei (TPE)](/docs/network/latency/tpe-taipei) * Region overview: [Africa](/docs/network/latency/regions/africa) → [Asia Pacific](/docs/network/latency/regions/asia-pacific) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Taipei to Johannesburg latency and RTT](/docs/network/latency/pairs/tpe-jnb-rtt): 331 ms **Fastest routes departing Johannesburg (JNB)** * [Johannesburg to London latency and RTT](/docs/network/latency/pairs/jnb-lon-rtt): 158 ms * [Johannesburg to Amsterdam latency and RTT](/docs/network/latency/pairs/jnb-ams-rtt): 163.2 ms * [Johannesburg to Paris latency and RTT](/docs/network/latency/pairs/jnb-par-rtt): 164.4 ms * [Johannesburg to Frankfurt latency and RTT](/docs/network/latency/pairs/jnb-fra-rtt): 166 ms * [Johannesburg to Marseille latency and RTT](/docs/network/latency/pairs/jnb-mrs-rtt): 172 ms **Fastest routes arriving at Taipei (TPE)** * [Hong Kong to Taipei latency and RTT](/docs/network/latency/pairs/hkg-tpe-rtt): 14.7 ms * [Tokyo to Taipei latency and RTT](/docs/network/latency/pairs/tyo-tpe-rtt): 32.3 ms * [Singapore to Taipei latency and RTT](/docs/network/latency/pairs/sin-tpe-rtt): 45.8 ms * [Seattle to Taipei latency and RTT](/docs/network/latency/pairs/sea-tpe-rtt): 114.7 ms * [Sydney to Taipei latency and RTT](/docs/network/latency/pairs/syd-tpe-rtt): 131.7 ms **Same corridor (Africa → Asia Pacific)** * [Johannesburg to Singapore latency and RTT](/docs/network/latency/pairs/jnb-sin-rtt): 311.6 ms * [Johannesburg to Hong Kong latency and RTT](/docs/network/latency/pairs/jnb-hkg-rtt): 318 ms * [Johannesburg to Tokyo latency and RTT](/docs/network/latency/pairs/jnb-tyo-rtt): 360.5 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/jnb-tpe.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Johannesburg → Taipei round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/jnb-tpe-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Johannesburg, South Africa → Tokyo, Japan RTT 🇿🇦 **Johannesburg, South Africa (JNB)** → 🇯🇵 **Tokyo, Japan (TYO)** Measurement round: `2026-08-16T04:07:28Z`. Johannesburg to Tokyo averaged 360.5 ms, compared with a 132.57 ms city-centre fiber reference. The difference cannot be assigned to a particular cable or carrier from RTT alone. Standard deviation was 16.2 ms, or 4.5% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------------------- | | RTT | **360.5 ms** | | Jitter | **13.91 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **132.57 ms** | | Fiber Efficiency | **36.8%** | | Latency Tier | High | | Source Region | [Africa](/docs/network/latency/regions/africa) | | Destination Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Johannesburg](https://www.geonames.org/993800) (-26.20227, 28.04363) and [Tokyo](https://www.geonames.org/1850147) (35.68950, 139.69171). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | --------------- | | WGS-84 geodesic distance | **13,537.7 km** | | Vacuum RTT floor | **90.31 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **132.57 ms** | | Low-latency fiber material floor | **132.04 ms** | | Engineering floor (5% path allowance) | **139.22 ms** | | Research 1.33× mapped-fiber reference | **176.32 ms** | | Estimated unamplified path loss | **2842.9 dB** | | Transparent optical spans / inline amplifiers | **178 / 177** | | Published RTT inflation over fiber floor | **2.72×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **339.64 ms** | | Average RTT | **360.5 ms** | | Maximum RTT | **427.1 ms** | | Standard deviation | **16.2 ms** | | Stdev / average | **4.5%** | ## Route Context * Departure PoP: [Johannesburg (JNB)](/docs/network/latency/jnb-johannesburg) * Destination PoP: [Tokyo (TYO)](/docs/network/latency/tyo-tokyo) * Region overview: [Africa](/docs/network/latency/regions/africa) → [Asia Pacific](/docs/network/latency/regions/asia-pacific) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Tokyo to Johannesburg latency and RTT](/docs/network/latency/pairs/tyo-jnb-rtt): 359.2 ms **Fastest routes departing Johannesburg (JNB)** * [Johannesburg to London latency and RTT](/docs/network/latency/pairs/jnb-lon-rtt): 158 ms * [Johannesburg to Amsterdam latency and RTT](/docs/network/latency/pairs/jnb-ams-rtt): 163.2 ms * [Johannesburg to Paris latency and RTT](/docs/network/latency/pairs/jnb-par-rtt): 164.4 ms * [Johannesburg to Frankfurt latency and RTT](/docs/network/latency/pairs/jnb-fra-rtt): 166 ms * [Johannesburg to Marseille latency and RTT](/docs/network/latency/pairs/jnb-mrs-rtt): 172 ms **Fastest routes arriving at Tokyo (TYO)** * [Taipei to Tokyo latency and RTT](/docs/network/latency/pairs/tpe-tyo-rtt): 31.9 ms * [Hong Kong to Tokyo latency and RTT](/docs/network/latency/pairs/hkg-tyo-rtt): 44.9 ms * [Singapore to Tokyo latency and RTT](/docs/network/latency/pairs/sin-tyo-rtt): 68.9 ms * [Seattle to Tokyo latency and RTT](/docs/network/latency/pairs/sea-tyo-rtt): 85 ms * [Los Angeles to Tokyo latency and RTT](/docs/network/latency/pairs/lax-tyo-rtt): 101.2 ms **Same corridor (Africa → Asia Pacific)** * [Johannesburg to Singapore latency and RTT](/docs/network/latency/pairs/jnb-sin-rtt): 311.6 ms * [Johannesburg to Hong Kong latency and RTT](/docs/network/latency/pairs/jnb-hkg-rtt): 318 ms * [Johannesburg to Taipei latency and RTT](/docs/network/latency/pairs/jnb-tpe-rtt): 329.8 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/jnb-tyo.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Johannesburg → Tokyo round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/jnb-tyo-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Los Angeles, USA → Amsterdam, Netherlands RTT 🇺🇸 **Los Angeles, USA (LAX)** → 🇳🇱 **Amsterdam, Netherlands (AMS)** Measurement round: `2026-08-16T04:07:28Z`. Los Angeles to Amsterdam averaged 132.1 ms across 50 replies. The observed range was 127 to 143.89 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 4.46 ms, or 3.4% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **132.1 ms** | | Jitter | **3.59 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **87.74 ms** | | Fiber Efficiency | **66.4%** | | Latency Tier | Good | | Source Region | [North America](/docs/network/latency/regions/north-america) | | Destination Region | [Europe](/docs/network/latency/regions/europe) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Los Angeles](https://www.geonames.org/5368361) (34.05223, -118.24368) and [Amsterdam](https://www.geonames.org/2759794) (52.37403, 4.88969). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | ------------- | | WGS-84 geodesic distance | **8,960 km** | | Vacuum RTT floor | **59.77 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **87.74 ms** | | Low-latency fiber material floor | **87.39 ms** | | Engineering floor (5% path allowance) | **92.14 ms** | | Research 1.33× mapped-fiber reference | **116.7 ms** | | Estimated unamplified path loss | **1881.6 dB** | | Transparent optical spans / inline amplifiers | **118 / 117** | | Published RTT inflation over fiber floor | **1.51×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **127 ms** | | Average RTT | **132.1 ms** | | Maximum RTT | **143.89 ms** | | Standard deviation | **4.46 ms** | | Stdev / average | **3.4%** | ## Route Context * Departure PoP: [Los Angeles (LAX)](/docs/network/latency/lax-los-angeles) * Destination PoP: [Amsterdam (AMS)](/docs/network/latency/ams-amsterdam) * Region overview: [North America](/docs/network/latency/regions/north-america) → [Europe](/docs/network/latency/regions/europe) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Amsterdam to Los Angeles latency and RTT](/docs/network/latency/pairs/ams-lax-rtt): 133.7 ms **Fastest routes departing Los Angeles (LAX)** * [Los Angeles to Seattle latency and RTT](/docs/network/latency/pairs/lax-sea-rtt): 26.9 ms * [Los Angeles to Miami latency and RTT](/docs/network/latency/pairs/lax-mia-rtt): 57.3 ms * [Los Angeles to New York latency and RTT](/docs/network/latency/pairs/lax-nyc-rtt): 57.9 ms * [Los Angeles to Ashburn latency and RTT](/docs/network/latency/pairs/lax-iad-rtt): 60.4 ms * [Los Angeles to Tokyo latency and RTT](/docs/network/latency/pairs/lax-tyo-rtt): 101.2 ms **Fastest routes arriving at Amsterdam (AMS)** * [London to Amsterdam latency and RTT](/docs/network/latency/pairs/lon-ams-rtt): 5.2 ms * [Frankfurt to Amsterdam latency and RTT](/docs/network/latency/pairs/fra-ams-rtt): 5.9 ms * [Paris to Amsterdam latency and RTT](/docs/network/latency/pairs/par-ams-rtt): 7 ms * [Berlin to Amsterdam latency and RTT](/docs/network/latency/pairs/ber-ams-rtt): 7.2 ms * [Marseille to Amsterdam latency and RTT](/docs/network/latency/pairs/mrs-ams-rtt): 20.3 ms **Same corridor (North America → Europe)** * [New York to London latency and RTT](/docs/network/latency/pairs/nyc-lon-rtt): 63.5 ms * [New York to Amsterdam latency and RTT](/docs/network/latency/pairs/nyc-ams-rtt): 68.7 ms * [New York to Paris latency and RTT](/docs/network/latency/pairs/nyc-par-rtt): 68.9 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/lax-ams.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Los Angeles → Amsterdam round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/lax-ams-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Los Angeles, USA → Berlin, Germany RTT 🇺🇸 **Los Angeles, USA (LAX)** → 🇩🇪 **Berlin, Germany (BER)** Measurement round: `2026-08-16T04:07:28Z`. Los Angeles to Berlin averaged 139.3 ms across 50 replies. The observed range was 136.33 to 148.84 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 2.96 ms, or 2.1% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **139.3 ms** | | Jitter | **1.96 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **91.39 ms** | | Fiber Efficiency | **65.6%** | | Latency Tier | Good | | Source Region | [North America](/docs/network/latency/regions/north-america) | | Destination Region | [Europe](/docs/network/latency/regions/europe) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Los Angeles](https://www.geonames.org/5368361) (34.05223, -118.24368) and [Berlin](https://www.geonames.org/2950159) (52.52437, 13.41053). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **9,331.9 km** | | Vacuum RTT floor | **62.26 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **91.39 ms** | | Low-latency fiber material floor | **91.02 ms** | | Engineering floor (5% path allowance) | **95.97 ms** | | Research 1.33× mapped-fiber reference | **121.54 ms** | | Estimated unamplified path loss | **1959.7 dB** | | Transparent optical spans / inline amplifiers | **123 / 122** | | Published RTT inflation over fiber floor | **1.52×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **136.33 ms** | | Average RTT | **139.3 ms** | | Maximum RTT | **148.84 ms** | | Standard deviation | **2.96 ms** | | Stdev / average | **2.1%** | ## Route Context * Departure PoP: [Los Angeles (LAX)](/docs/network/latency/lax-los-angeles) * Destination PoP: [Berlin (BER)](/docs/network/latency/ber-berlin) * Region overview: [North America](/docs/network/latency/regions/north-america) → [Europe](/docs/network/latency/regions/europe) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Berlin to Los Angeles latency and RTT](/docs/network/latency/pairs/ber-lax-rtt): 140.9 ms **Fastest routes departing Los Angeles (LAX)** * [Los Angeles to Seattle latency and RTT](/docs/network/latency/pairs/lax-sea-rtt): 26.9 ms * [Los Angeles to Miami latency and RTT](/docs/network/latency/pairs/lax-mia-rtt): 57.3 ms * [Los Angeles to New York latency and RTT](/docs/network/latency/pairs/lax-nyc-rtt): 57.9 ms * [Los Angeles to Ashburn latency and RTT](/docs/network/latency/pairs/lax-iad-rtt): 60.4 ms * [Los Angeles to Tokyo latency and RTT](/docs/network/latency/pairs/lax-tyo-rtt): 101.2 ms **Fastest routes arriving at Berlin (BER)** * [Frankfurt to Berlin latency and RTT](/docs/network/latency/pairs/fra-ber-rtt): 6.1 ms * [Amsterdam to Berlin latency and RTT](/docs/network/latency/pairs/ams-ber-rtt): 7.2 ms * [Paris to Berlin latency and RTT](/docs/network/latency/pairs/par-ber-rtt): 15.5 ms * [London to Berlin latency and RTT](/docs/network/latency/pairs/lon-ber-rtt): 16.2 ms * [Marseille to Berlin latency and RTT](/docs/network/latency/pairs/mrs-ber-rtt): 20.6 ms **Same corridor (North America → Europe)** * [New York to London latency and RTT](/docs/network/latency/pairs/nyc-lon-rtt): 63.5 ms * [New York to Amsterdam latency and RTT](/docs/network/latency/pairs/nyc-ams-rtt): 68.7 ms * [New York to Paris latency and RTT](/docs/network/latency/pairs/nyc-par-rtt): 68.9 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/lax-ber.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Los Angeles → Berlin round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/lax-ber-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Los Angeles, USA → Frankfurt, Germany RTT 🇺🇸 **Los Angeles, USA (LAX)** → 🇩🇪 **Frankfurt, Germany (FRA)** Measurement round: `2026-08-16T04:07:28Z`. Los Angeles to Frankfurt averaged 142.3 ms across 50 replies. The observed range was 137.29 to 154.07 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 3.29 ms, or 2.3% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **142.3 ms** | | Jitter | **3.06 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **91.31 ms** | | Fiber Efficiency | **64.2%** | | Latency Tier | Good | | Source Region | [North America](/docs/network/latency/regions/north-america) | | Destination Region | [Europe](/docs/network/latency/regions/europe) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Los Angeles](https://www.geonames.org/5368361) (34.05223, -118.24368) and [Frankfurt am Main](https://www.geonames.org/2925533) (50.11552, 8.68417). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **9,324.5 km** | | Vacuum RTT floor | **62.21 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **91.31 ms** | | Low-latency fiber material floor | **90.95 ms** | | Engineering floor (5% path allowance) | **95.89 ms** | | Research 1.33× mapped-fiber reference | **121.45 ms** | | Estimated unamplified path loss | **1958.1 dB** | | Transparent optical spans / inline amplifiers | **123 / 122** | | Published RTT inflation over fiber floor | **1.56×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **137.29 ms** | | Average RTT | **142.3 ms** | | Maximum RTT | **154.07 ms** | | Standard deviation | **3.29 ms** | | Stdev / average | **2.3%** | ## Route Context * Departure PoP: [Los Angeles (LAX)](/docs/network/latency/lax-los-angeles) * Destination PoP: [Frankfurt (FRA)](/docs/network/latency/fra-frankfurt) * Region overview: [North America](/docs/network/latency/regions/north-america) → [Europe](/docs/network/latency/regions/europe) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Frankfurt to Los Angeles latency and RTT](/docs/network/latency/pairs/fra-lax-rtt): 141.8 ms **Fastest routes departing Los Angeles (LAX)** * [Los Angeles to Seattle latency and RTT](/docs/network/latency/pairs/lax-sea-rtt): 26.9 ms * [Los Angeles to Miami latency and RTT](/docs/network/latency/pairs/lax-mia-rtt): 57.3 ms * [Los Angeles to New York latency and RTT](/docs/network/latency/pairs/lax-nyc-rtt): 57.9 ms * [Los Angeles to Ashburn latency and RTT](/docs/network/latency/pairs/lax-iad-rtt): 60.4 ms * [Los Angeles to Tokyo latency and RTT](/docs/network/latency/pairs/lax-tyo-rtt): 101.2 ms **Fastest routes arriving at Frankfurt (FRA)** * [Amsterdam to Frankfurt latency and RTT](/docs/network/latency/pairs/ams-fra-rtt): 5.9 ms * [Berlin to Frankfurt latency and RTT](/docs/network/latency/pairs/ber-fra-rtt): 6.1 ms * [Paris to Frankfurt latency and RTT](/docs/network/latency/pairs/par-fra-rtt): 7.6 ms * [London to Frankfurt latency and RTT](/docs/network/latency/pairs/lon-fra-rtt): 13.6 ms * [Marseille to Frankfurt latency and RTT](/docs/network/latency/pairs/mrs-fra-rtt): 16 ms **Same corridor (North America → Europe)** * [New York to London latency and RTT](/docs/network/latency/pairs/nyc-lon-rtt): 63.5 ms * [New York to Amsterdam latency and RTT](/docs/network/latency/pairs/nyc-ams-rtt): 68.7 ms * [New York to Paris latency and RTT](/docs/network/latency/pairs/nyc-par-rtt): 68.9 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/lax-fra.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Los Angeles → Frankfurt round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/lax-fra-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Los Angeles, USA → São Paulo, Brazil RTT 🇺🇸 **Los Angeles, USA (LAX)** → 🇧🇷 **São Paulo, Brazil (GRU)** Measurement round: `2026-08-16T04:07:28Z`. Los Angeles to São Paulo averaged 131.3 ms across 50 replies. The observed range was 125.74 to 147.12 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 4.94 ms, or 3.8% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **131.3 ms** | | Jitter | **5.03 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **96.89 ms** | | Fiber Efficiency | **73.8%** | | Latency Tier | Good | | Source Region | [North America](/docs/network/latency/regions/north-america) | | Destination Region | [South America](/docs/network/latency/regions/south-america) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Los Angeles](https://www.geonames.org/5368361) (34.05223, -118.24368) and [São Paulo](https://www.geonames.org/3448439) (-23.54750, -46.63611). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **9,894.5 km** | | Vacuum RTT floor | **66.01 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **96.89 ms** | | Low-latency fiber material floor | **96.51 ms** | | Engineering floor (5% path allowance) | **101.75 ms** | | Research 1.33× mapped-fiber reference | **128.87 ms** | | Estimated unamplified path loss | **2077.8 dB** | | Transparent optical spans / inline amplifiers | **130 / 129** | | Published RTT inflation over fiber floor | **1.36×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **125.74 ms** | | Average RTT | **131.3 ms** | | Maximum RTT | **147.12 ms** | | Standard deviation | **4.94 ms** | | Stdev / average | **3.8%** | ## Route Context * Departure PoP: [Los Angeles (LAX)](/docs/network/latency/lax-los-angeles) * Destination PoP: [São Paulo (GRU)](/docs/network/latency/gru-sao-paulo) * Region overview: [North America](/docs/network/latency/regions/north-america) → [South America](/docs/network/latency/regions/south-america) * Corridor overview: [North America to South America](/docs/network/latency/regions/north-america-to-south-america) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [São Paulo to Los Angeles latency and RTT](/docs/network/latency/pairs/gru-lax-rtt): 130.8 ms **Fastest routes departing Los Angeles (LAX)** * [Los Angeles to Seattle latency and RTT](/docs/network/latency/pairs/lax-sea-rtt): 26.9 ms * [Los Angeles to Miami latency and RTT](/docs/network/latency/pairs/lax-mia-rtt): 57.3 ms * [Los Angeles to New York latency and RTT](/docs/network/latency/pairs/lax-nyc-rtt): 57.9 ms * [Los Angeles to Ashburn latency and RTT](/docs/network/latency/pairs/lax-iad-rtt): 60.4 ms * [Los Angeles to Tokyo latency and RTT](/docs/network/latency/pairs/lax-tyo-rtt): 101.2 ms **Fastest routes arriving at São Paulo (GRU)** * [Ashburn to São Paulo latency and RTT](/docs/network/latency/pairs/iad-gru-rtt): 101.4 ms * [New York to São Paulo latency and RTT](/docs/network/latency/pairs/nyc-gru-rtt): 106.9 ms * [Miami to São Paulo latency and RTT](/docs/network/latency/pairs/mia-gru-rtt): 128.7 ms * [Seattle to São Paulo latency and RTT](/docs/network/latency/pairs/sea-gru-rtt): 155.7 ms * [London to São Paulo latency and RTT](/docs/network/latency/pairs/lon-gru-rtt): 170.7 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/lax-gru.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Los Angeles → São Paulo round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/lax-gru-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Los Angeles, USA → Hong Kong RTT 🇺🇸 **Los Angeles, USA (LAX)** → 🇭🇰 **Hong Kong (HKG)** Measurement round: `2026-08-16T04:07:28Z`. Los Angeles to Hong Kong averaged 146.5 ms across 50 replies. The observed range was 143.01 to 158.89 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 3.29 ms, or 2.2% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **146.5 ms** | | Jitter | **2.27 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **114.3 ms** | | Fiber Efficiency | **78%** | | Latency Tier | Good | | Source Region | [North America](/docs/network/latency/regions/north-america) | | Destination Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Los Angeles](https://www.geonames.org/5368361) (34.05223, -118.24368) and [Hong Kong](https://www.geonames.org/1819729) (22.27832, 114.17469). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | --------------- | | WGS-84 geodesic distance | **11,671.5 km** | | Vacuum RTT floor | **77.86 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **114.3 ms** | | Low-latency fiber material floor | **113.84 ms** | | Engineering floor (5% path allowance) | **120.03 ms** | | Research 1.33× mapped-fiber reference | **152.01 ms** | | Estimated unamplified path loss | **2451 dB** | | Transparent optical spans / inline amplifiers | **154 / 153** | | Published RTT inflation over fiber floor | **1.28×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **143.01 ms** | | Average RTT | **146.5 ms** | | Maximum RTT | **158.89 ms** | | Standard deviation | **3.29 ms** | | Stdev / average | **2.2%** | ## Route Context * Departure PoP: [Los Angeles (LAX)](/docs/network/latency/lax-los-angeles) * Destination PoP: [Hong Kong (HKG)](/docs/network/latency/hkg-hong-kong) * Region overview: [North America](/docs/network/latency/regions/north-america) → [Asia Pacific](/docs/network/latency/regions/asia-pacific) * Corridor overview: [North America to Asia Pacific](/docs/network/latency/regions/north-america-to-asia-pacific) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Hong Kong to Los Angeles latency and RTT](/docs/network/latency/pairs/hkg-lax-rtt): 145 ms **Fastest routes departing Los Angeles (LAX)** * [Los Angeles to Seattle latency and RTT](/docs/network/latency/pairs/lax-sea-rtt): 26.9 ms * [Los Angeles to Miami latency and RTT](/docs/network/latency/pairs/lax-mia-rtt): 57.3 ms * [Los Angeles to New York latency and RTT](/docs/network/latency/pairs/lax-nyc-rtt): 57.9 ms * [Los Angeles to Ashburn latency and RTT](/docs/network/latency/pairs/lax-iad-rtt): 60.4 ms * [Los Angeles to Tokyo latency and RTT](/docs/network/latency/pairs/lax-tyo-rtt): 101.2 ms **Fastest routes arriving at Hong Kong (HKG)** * [Taipei to Hong Kong latency and RTT](/docs/network/latency/pairs/tpe-hkg-rtt): 15.5 ms * [Singapore to Hong Kong latency and RTT](/docs/network/latency/pairs/sin-hkg-rtt): 30.8 ms * [Tokyo to Hong Kong latency and RTT](/docs/network/latency/pairs/tyo-hkg-rtt): 44.8 ms * [Moscow to Hong Kong latency and RTT](/docs/network/latency/pairs/mow-hkg-rtt): 118.6 ms * [Seattle to Hong Kong latency and RTT](/docs/network/latency/pairs/sea-hkg-rtt): 130.4 ms **Same corridor (North America → Asia Pacific)** * [Seattle to Tokyo latency and RTT](/docs/network/latency/pairs/sea-tyo-rtt): 85 ms * [Seattle to Taipei latency and RTT](/docs/network/latency/pairs/sea-tpe-rtt): 114.7 ms * [Los Angeles to Taipei latency and RTT](/docs/network/latency/pairs/lax-tpe-rtt): 132.9 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/lax-hkg.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Los Angeles → Hong Kong round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/lax-hkg-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Los Angeles, USA → Ashburn, USA RTT 🇺🇸 **Los Angeles, USA (LAX)** → 🇺🇸 **Ashburn, USA (IAD)** Measurement round: `2026-08-16T04:07:28Z`. Los Angeles to Ashburn averaged 60.4 ms across 50 replies. The observed range was 57.19 to 70.55 ms; these figures describe one probe round, not a long-term service target. This is number 4 of 19 outgoing routes by average RTT. That ranking compares different destinations and is not adjusted for distance. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **60.4 ms** | | Jitter | **2.06 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **35.84 ms** | | Fiber Efficiency | **59.3%** | | Latency Tier | Excellent | | Source Region | [North America](/docs/network/latency/regions/north-america) | | Destination Region | [North America](/docs/network/latency/regions/north-america) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Los Angeles](https://www.geonames.org/5368361) (34.05223, -118.24368) and [Ashburn](https://www.geonames.org/4744870) (39.04372, -77.48749). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **3,660.1 km** | | Vacuum RTT floor | **24.42 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **35.84 ms** | | Low-latency fiber material floor | **35.7 ms** | | Engineering floor (5% path allowance) | **37.64 ms** | | Research 1.33× mapped-fiber reference | **47.67 ms** | | Estimated unamplified path loss | **768.6 dB** | | Transparent optical spans / inline amplifiers | **49 / 48** | | Published RTT inflation over fiber floor | **1.69×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------ | | Minimum RTT | **57.19 ms** | | Average RTT | **60.4 ms** | | Maximum RTT | **70.55 ms** | | Standard deviation | **2.61 ms** | | Stdev / average | **4.3%** | ## Route Context * Departure PoP: [Los Angeles (LAX)](/docs/network/latency/lax-los-angeles) * Destination PoP: [Ashburn (IAD)](/docs/network/latency/iad-ashburn) * Region overview: [North America](/docs/network/latency/regions/north-america) → [North America](/docs/network/latency/regions/north-america) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Ashburn to Los Angeles latency and RTT](/docs/network/latency/pairs/iad-lax-rtt): 60.7 ms **Fastest routes departing Los Angeles (LAX)** * [Los Angeles to Seattle latency and RTT](/docs/network/latency/pairs/lax-sea-rtt): 26.9 ms * [Los Angeles to Miami latency and RTT](/docs/network/latency/pairs/lax-mia-rtt): 57.3 ms * [Los Angeles to New York latency and RTT](/docs/network/latency/pairs/lax-nyc-rtt): 57.9 ms * [Los Angeles to Tokyo latency and RTT](/docs/network/latency/pairs/lax-tyo-rtt): 101.2 ms * [Los Angeles to London latency and RTT](/docs/network/latency/pairs/lax-lon-rtt): 129.2 ms **Fastest routes arriving at Ashburn (IAD)** * [New York to Ashburn latency and RTT](/docs/network/latency/pairs/nyc-iad-rtt): 6 ms * [Miami to Ashburn latency and RTT](/docs/network/latency/pairs/mia-iad-rtt): 27.3 ms * [Seattle to Ashburn latency and RTT](/docs/network/latency/pairs/sea-iad-rtt): 61.5 ms * [London to Ashburn latency and RTT](/docs/network/latency/pairs/lon-iad-rtt): 71 ms * [Paris to Ashburn latency and RTT](/docs/network/latency/pairs/par-iad-rtt): 75.5 ms **Same corridor (North America → North America)** * [Ashburn to New York latency and RTT](/docs/network/latency/pairs/iad-nyc-rtt): 6 ms * [Seattle to Los Angeles latency and RTT](/docs/network/latency/pairs/sea-lax-rtt): 25.9 ms * [Ashburn to Miami latency and RTT](/docs/network/latency/pairs/iad-mia-rtt): 27.4 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/lax-iad.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Los Angeles → Ashburn round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/lax-iad-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Los Angeles, USA → Johannesburg, South Africa RTT 🇺🇸 **Los Angeles, USA (LAX)** → 🇿🇦 **Johannesburg, South Africa (JNB)** Measurement round: `2026-08-16T04:07:28Z`. Los Angeles to Johannesburg averaged 287.2 ms across 50 replies. The observed range was 281.06 to 306.43 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 5.36 ms, or 1.9% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **287.2 ms** | | Jitter | **4.1 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **163.35 ms** | | Fiber Efficiency | **56.9%** | | Latency Tier | High | | Source Region | [North America](/docs/network/latency/regions/north-america) | | Destination Region | [Africa](/docs/network/latency/regions/africa) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Los Angeles](https://www.geonames.org/5368361) (34.05223, -118.24368) and [Johannesburg](https://www.geonames.org/993800) (-26.20227, 28.04363). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | --------------- | | WGS-84 geodesic distance | **16,680.3 km** | | Vacuum RTT floor | **111.28 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **163.35 ms** | | Low-latency fiber material floor | **162.69 ms** | | Engineering floor (5% path allowance) | **171.54 ms** | | Research 1.33× mapped-fiber reference | **217.25 ms** | | Estimated unamplified path loss | **3502.9 dB** | | Transparent optical spans / inline amplifiers | **219 / 218** | | Published RTT inflation over fiber floor | **1.76×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **281.06 ms** | | Average RTT | **287.2 ms** | | Maximum RTT | **306.43 ms** | | Standard deviation | **5.36 ms** | | Stdev / average | **1.9%** | ## Route Context * Departure PoP: [Los Angeles (LAX)](/docs/network/latency/lax-los-angeles) * Destination PoP: [Johannesburg (JNB)](/docs/network/latency/jnb-johannesburg) * Region overview: [North America](/docs/network/latency/regions/north-america) → [Africa](/docs/network/latency/regions/africa) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Johannesburg to Los Angeles latency and RTT](/docs/network/latency/pairs/jnb-lax-rtt): 285.5 ms **Fastest routes departing Los Angeles (LAX)** * [Los Angeles to Seattle latency and RTT](/docs/network/latency/pairs/lax-sea-rtt): 26.9 ms * [Los Angeles to Miami latency and RTT](/docs/network/latency/pairs/lax-mia-rtt): 57.3 ms * [Los Angeles to New York latency and RTT](/docs/network/latency/pairs/lax-nyc-rtt): 57.9 ms * [Los Angeles to Ashburn latency and RTT](/docs/network/latency/pairs/lax-iad-rtt): 60.4 ms * [Los Angeles to Tokyo latency and RTT](/docs/network/latency/pairs/lax-tyo-rtt): 101.2 ms **Fastest routes arriving at Johannesburg (JNB)** * [Amsterdam to Johannesburg latency and RTT](/docs/network/latency/pairs/ams-jnb-rtt): 171.9 ms * [Berlin to Johannesburg latency and RTT](/docs/network/latency/pairs/ber-jnb-rtt): 172.1 ms * [London to Johannesburg latency and RTT](/docs/network/latency/pairs/lon-jnb-rtt): 179.6 ms * [Paris to Johannesburg latency and RTT](/docs/network/latency/pairs/par-jnb-rtt): 180.9 ms * [Frankfurt to Johannesburg latency and RTT](/docs/network/latency/pairs/fra-jnb-rtt): 188.4 ms **Same corridor (North America → Africa)** * [New York to Johannesburg latency and RTT](/docs/network/latency/pairs/nyc-jnb-rtt): 221.5 ms * [Ashburn to Johannesburg latency and RTT](/docs/network/latency/pairs/iad-jnb-rtt): 228.3 ms * [Miami to Johannesburg latency and RTT](/docs/network/latency/pairs/mia-jnb-rtt): 259.8 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/lax-jnb.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Los Angeles → Johannesburg round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/lax-jnb-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Los Angeles, USA → London, UK RTT 🇺🇸 **Los Angeles, USA (LAX)** → 🇬🇧 **London, UK (LON)** Measurement round: `2026-08-16T04:07:28Z`. Los Angeles to London averaged 129.2 ms across 50 replies. The observed range was 123.19 to 146.01 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 4.9 ms, or 3.8% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **129.2 ms** | | Jitter | **3.99 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **85.95 ms** | | Fiber Efficiency | **66.5%** | | Latency Tier | Good | | Source Region | [North America](/docs/network/latency/regions/north-america) | | Destination Region | [Europe](/docs/network/latency/regions/europe) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Los Angeles](https://www.geonames.org/5368361) (34.05223, -118.24368) and [London](https://www.geonames.org/2643743) (51.50853, -0.12574). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **8,776.7 km** | | Vacuum RTT floor | **58.55 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **85.95 ms** | | Low-latency fiber material floor | **85.6 ms** | | Engineering floor (5% path allowance) | **90.26 ms** | | Research 1.33× mapped-fiber reference | **114.31 ms** | | Estimated unamplified path loss | **1843.1 dB** | | Transparent optical spans / inline amplifiers | **116 / 115** | | Published RTT inflation over fiber floor | **1.5×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **123.19 ms** | | Average RTT | **129.2 ms** | | Maximum RTT | **146.01 ms** | | Standard deviation | **4.9 ms** | | Stdev / average | **3.8%** | ## Route Context * Departure PoP: [Los Angeles (LAX)](/docs/network/latency/lax-los-angeles) * Destination PoP: [London (LON)](/docs/network/latency/lon-london) * Region overview: [North America](/docs/network/latency/regions/north-america) → [Europe](/docs/network/latency/regions/europe) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [London to Los Angeles latency and RTT](/docs/network/latency/pairs/lon-lax-rtt): 127.5 ms **Fastest routes departing Los Angeles (LAX)** * [Los Angeles to Seattle latency and RTT](/docs/network/latency/pairs/lax-sea-rtt): 26.9 ms * [Los Angeles to Miami latency and RTT](/docs/network/latency/pairs/lax-mia-rtt): 57.3 ms * [Los Angeles to New York latency and RTT](/docs/network/latency/pairs/lax-nyc-rtt): 57.9 ms * [Los Angeles to Ashburn latency and RTT](/docs/network/latency/pairs/lax-iad-rtt): 60.4 ms * [Los Angeles to Tokyo latency and RTT](/docs/network/latency/pairs/lax-tyo-rtt): 101.2 ms **Fastest routes arriving at London (LON)** * [Amsterdam to London latency and RTT](/docs/network/latency/pairs/ams-lon-rtt): 5.2 ms * [Paris to London latency and RTT](/docs/network/latency/pairs/par-lon-rtt): 6.4 ms * [Frankfurt to London latency and RTT](/docs/network/latency/pairs/fra-lon-rtt): 12.9 ms * [Berlin to London latency and RTT](/docs/network/latency/pairs/ber-lon-rtt): 15.2 ms * [Marseille to London latency and RTT](/docs/network/latency/pairs/mrs-lon-rtt): 18.9 ms **Same corridor (North America → Europe)** * [New York to London latency and RTT](/docs/network/latency/pairs/nyc-lon-rtt): 63.5 ms * [New York to Amsterdam latency and RTT](/docs/network/latency/pairs/nyc-ams-rtt): 68.7 ms * [New York to Paris latency and RTT](/docs/network/latency/pairs/nyc-par-rtt): 68.9 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/lax-lon.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Los Angeles → London round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/lax-lon-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Los Angeles, USA → Melbourne, Australia RTT 🇺🇸 **Los Angeles, USA (LAX)** → 🇦🇺 **Melbourne, Australia (MEL)** Measurement round: `2026-08-16T04:07:28Z`. Los Angeles to Melbourne averaged 148.2 ms across 50 replies. The observed range was 142.14 to 162.19 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 5.38 ms, or 3.6% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **148.2 ms** | | Jitter | **4.01 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **125 ms** | | Fiber Efficiency | **84.3%** | | Latency Tier | Good | | Source Region | [North America](/docs/network/latency/regions/north-america) | | Destination Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Los Angeles](https://www.geonames.org/5368361) (34.05223, -118.24368) and [Melbourne](https://www.geonames.org/2158177) (-37.81400, 144.96332). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | ------------- | | WGS-84 geodesic distance | **12,764 km** | | Vacuum RTT floor | **85.15 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **125 ms** | | Low-latency fiber material floor | **124.49 ms** | | Engineering floor (5% path allowance) | **131.26 ms** | | Research 1.33× mapped-fiber reference | **166.24 ms** | | Estimated unamplified path loss | **2680.4 dB** | | Transparent optical spans / inline amplifiers | **168 / 167** | | Published RTT inflation over fiber floor | **1.19×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **142.14 ms** | | Average RTT | **148.2 ms** | | Maximum RTT | **162.19 ms** | | Standard deviation | **5.38 ms** | | Stdev / average | **3.6%** | ## Route Context * Departure PoP: [Los Angeles (LAX)](/docs/network/latency/lax-los-angeles) * Destination PoP: [Melbourne (MEL)](/docs/network/latency/mel-melbourne) * Region overview: [North America](/docs/network/latency/regions/north-america) → [Asia Pacific](/docs/network/latency/regions/asia-pacific) * Corridor overview: [North America to Asia Pacific](/docs/network/latency/regions/north-america-to-asia-pacific) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Melbourne to Los Angeles latency and RTT](/docs/network/latency/pairs/mel-lax-rtt): 148.4 ms **Fastest routes departing Los Angeles (LAX)** * [Los Angeles to Seattle latency and RTT](/docs/network/latency/pairs/lax-sea-rtt): 26.9 ms * [Los Angeles to Miami latency and RTT](/docs/network/latency/pairs/lax-mia-rtt): 57.3 ms * [Los Angeles to New York latency and RTT](/docs/network/latency/pairs/lax-nyc-rtt): 57.9 ms * [Los Angeles to Ashburn latency and RTT](/docs/network/latency/pairs/lax-iad-rtt): 60.4 ms * [Los Angeles to Tokyo latency and RTT](/docs/network/latency/pairs/lax-tyo-rtt): 101.2 ms **Fastest routes arriving at Melbourne (MEL)** * [Sydney to Melbourne latency and RTT](/docs/network/latency/pairs/syd-mel-rtt): 9.8 ms * [Singapore to Melbourne latency and RTT](/docs/network/latency/pairs/sin-mel-rtt): 88.6 ms * [Tokyo to Melbourne latency and RTT](/docs/network/latency/pairs/tyo-mel-rtt): 113.9 ms * [Hong Kong to Melbourne latency and RTT](/docs/network/latency/pairs/hkg-mel-rtt): 138.4 ms * [Taipei to Melbourne latency and RTT](/docs/network/latency/pairs/tpe-mel-rtt): 142 ms **Same corridor (North America → Asia Pacific)** * [Seattle to Tokyo latency and RTT](/docs/network/latency/pairs/sea-tyo-rtt): 85 ms * [Seattle to Taipei latency and RTT](/docs/network/latency/pairs/sea-tpe-rtt): 114.7 ms * [Seattle to Hong Kong latency and RTT](/docs/network/latency/pairs/sea-hkg-rtt): 130.4 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/lax-mel.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Los Angeles → Melbourne round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/lax-mel-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Los Angeles, USA → Miami, USA RTT 🇺🇸 **Los Angeles, USA (LAX)** → 🇺🇸 **Miami, USA (MIA)** Measurement round: `2026-08-16T04:07:28Z`. Los Angeles to Miami averaged 57.3 ms across 50 replies. The observed range was 55.45 to 63.93 ms; these figures describe one probe round, not a long-term service target. This is number 2 of 19 outgoing routes by average RTT. That ranking compares different destinations and is not adjusted for distance. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **57.3 ms** | | Jitter | **1.26 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **36.86 ms** | | Fiber Efficiency | **64.3%** | | Latency Tier | Excellent | | Source Region | [North America](/docs/network/latency/regions/north-america) | | Destination Region | [North America](/docs/network/latency/regions/north-america) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Los Angeles](https://www.geonames.org/5368361) (34.05223, -118.24368) and [Miami](https://www.geonames.org/4164138) (25.77427, -80.19366). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **3,764.3 km** | | Vacuum RTT floor | **25.11 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **36.86 ms** | | Low-latency fiber material floor | **36.71 ms** | | Engineering floor (5% path allowance) | **38.71 ms** | | Research 1.33× mapped-fiber reference | **49.03 ms** | | Estimated unamplified path loss | **790.5 dB** | | Transparent optical spans / inline amplifiers | **50 / 49** | | Published RTT inflation over fiber floor | **1.55×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------ | | Minimum RTT | **55.45 ms** | | Average RTT | **57.3 ms** | | Maximum RTT | **63.93 ms** | | Standard deviation | **1.73 ms** | | Stdev / average | **3.0%** | ## Route Context * Departure PoP: [Los Angeles (LAX)](/docs/network/latency/lax-los-angeles) * Destination PoP: [Miami (MIA)](/docs/network/latency/mia-miami) * Region overview: [North America](/docs/network/latency/regions/north-america) → [North America](/docs/network/latency/regions/north-america) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Miami to Los Angeles latency and RTT](/docs/network/latency/pairs/mia-lax-rtt): 56.8 ms **Fastest routes departing Los Angeles (LAX)** * [Los Angeles to Seattle latency and RTT](/docs/network/latency/pairs/lax-sea-rtt): 26.9 ms * [Los Angeles to New York latency and RTT](/docs/network/latency/pairs/lax-nyc-rtt): 57.9 ms * [Los Angeles to Ashburn latency and RTT](/docs/network/latency/pairs/lax-iad-rtt): 60.4 ms * [Los Angeles to Tokyo latency and RTT](/docs/network/latency/pairs/lax-tyo-rtt): 101.2 ms * [Los Angeles to London latency and RTT](/docs/network/latency/pairs/lax-lon-rtt): 129.2 ms **Fastest routes arriving at Miami (MIA)** * [Ashburn to Miami latency and RTT](/docs/network/latency/pairs/iad-mia-rtt): 27.4 ms * [New York to Miami latency and RTT](/docs/network/latency/pairs/nyc-mia-rtt): 32.9 ms * [São Paulo to Miami latency and RTT](/docs/network/latency/pairs/gru-mia-rtt): 74 ms * [Seattle to Miami latency and RTT](/docs/network/latency/pairs/sea-mia-rtt): 81.7 ms * [London to Miami latency and RTT](/docs/network/latency/pairs/lon-mia-rtt): 101.9 ms **Same corridor (North America → North America)** * [Ashburn to New York latency and RTT](/docs/network/latency/pairs/iad-nyc-rtt): 6 ms * [New York to Ashburn latency and RTT](/docs/network/latency/pairs/nyc-iad-rtt): 6 ms * [Seattle to Los Angeles latency and RTT](/docs/network/latency/pairs/sea-lax-rtt): 25.9 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/lax-mia.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Los Angeles → Miami round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/lax-mia-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Los Angeles, USA → Moscow, Russia RTT 🇺🇸 **Los Angeles, USA (LAX)** → 🇷🇺 **Moscow, Russia (MOW)** Measurement round: `2026-08-16T04:07:28Z`. Los Angeles to Moscow averaged 177 ms across 50 replies. The observed range was 169.62 to 197.78 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 6.43 ms, or 3.6% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **177 ms** | | Jitter | **4.63 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **95.9 ms** | | Fiber Efficiency | **54.2%** | | Latency Tier | Fair | | Source Region | [North America](/docs/network/latency/regions/north-america) | | Destination Region | [Europe](/docs/network/latency/regions/europe) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Los Angeles](https://www.geonames.org/5368361) (34.05223, -118.24368) and [Moscow](https://www.geonames.org/524901) (55.75204, 37.61781). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **9,793.2 km** | | Vacuum RTT floor | **65.33 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **95.9 ms** | | Low-latency fiber material floor | **95.52 ms** | | Engineering floor (5% path allowance) | **100.71 ms** | | Research 1.33× mapped-fiber reference | **127.55 ms** | | Estimated unamplified path loss | **2056.6 dB** | | Transparent optical spans / inline amplifiers | **129 / 128** | | Published RTT inflation over fiber floor | **1.85×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **169.62 ms** | | Average RTT | **177 ms** | | Maximum RTT | **197.78 ms** | | Standard deviation | **6.43 ms** | | Stdev / average | **3.6%** | ## Route Context * Departure PoP: [Los Angeles (LAX)](/docs/network/latency/lax-los-angeles) * Destination PoP: [Moscow (MOW)](/docs/network/latency/mow-moscow) * Region overview: [North America](/docs/network/latency/regions/north-america) → [Europe](/docs/network/latency/regions/europe) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Moscow to Los Angeles latency and RTT](/docs/network/latency/pairs/mow-lax-rtt): 177.7 ms **Fastest routes departing Los Angeles (LAX)** * [Los Angeles to Seattle latency and RTT](/docs/network/latency/pairs/lax-sea-rtt): 26.9 ms * [Los Angeles to Miami latency and RTT](/docs/network/latency/pairs/lax-mia-rtt): 57.3 ms * [Los Angeles to New York latency and RTT](/docs/network/latency/pairs/lax-nyc-rtt): 57.9 ms * [Los Angeles to Ashburn latency and RTT](/docs/network/latency/pairs/lax-iad-rtt): 60.4 ms * [Los Angeles to Tokyo latency and RTT](/docs/network/latency/pairs/lax-tyo-rtt): 101.2 ms **Fastest routes arriving at Moscow (MOW)** * [Berlin to Moscow latency and RTT](/docs/network/latency/pairs/ber-mow-rtt): 28.6 ms * [Frankfurt to Moscow latency and RTT](/docs/network/latency/pairs/fra-mow-rtt): 34.7 ms * [Amsterdam to Moscow latency and RTT](/docs/network/latency/pairs/ams-mow-rtt): 38.1 ms * [London to Moscow latency and RTT](/docs/network/latency/pairs/lon-mow-rtt): 43.5 ms * [Paris to Moscow latency and RTT](/docs/network/latency/pairs/par-mow-rtt): 44.7 ms **Same corridor (North America → Europe)** * [New York to London latency and RTT](/docs/network/latency/pairs/nyc-lon-rtt): 63.5 ms * [New York to Amsterdam latency and RTT](/docs/network/latency/pairs/nyc-ams-rtt): 68.7 ms * [New York to Paris latency and RTT](/docs/network/latency/pairs/nyc-par-rtt): 68.9 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/lax-mow.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Los Angeles → Moscow round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/lax-mow-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Los Angeles, USA → Marseille, France RTT 🇺🇸 **Los Angeles, USA (LAX)** → 🇫🇷 **Marseille, France (MRS)** Measurement round: `2026-08-16T04:07:28Z`. Los Angeles to Marseille averaged 143.9 ms across 50 replies. The observed range was 139.48 to 166.33 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 4.83 ms, or 3.4% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **143.9 ms** | | Jitter | **3.12 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **95.09 ms** | | Fiber Efficiency | **66.1%** | | Latency Tier | Good | | Source Region | [North America](/docs/network/latency/regions/north-america) | | Destination Region | [Europe](/docs/network/latency/regions/europe) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Los Angeles](https://www.geonames.org/5368361) (34.05223, -118.24368) and [Marseille](https://www.geonames.org/2995469) (43.29695, 5.38107). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **9,710.4 km** | | Vacuum RTT floor | **64.78 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **95.09 ms** | | Low-latency fiber material floor | **94.71 ms** | | Engineering floor (5% path allowance) | **99.86 ms** | | Research 1.33× mapped-fiber reference | **126.47 ms** | | Estimated unamplified path loss | **2039.2 dB** | | Transparent optical spans / inline amplifiers | **128 / 127** | | Published RTT inflation over fiber floor | **1.51×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **139.48 ms** | | Average RTT | **143.9 ms** | | Maximum RTT | **166.33 ms** | | Standard deviation | **4.83 ms** | | Stdev / average | **3.4%** | ## Route Context * Departure PoP: [Los Angeles (LAX)](/docs/network/latency/lax-los-angeles) * Destination PoP: [Marseille (MRS)](/docs/network/latency/mrs-marseille) * Region overview: [North America](/docs/network/latency/regions/north-america) → [Europe](/docs/network/latency/regions/europe) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Marseille to Los Angeles latency and RTT](/docs/network/latency/pairs/mrs-lax-rtt): 144.5 ms **Fastest routes departing Los Angeles (LAX)** * [Los Angeles to Seattle latency and RTT](/docs/network/latency/pairs/lax-sea-rtt): 26.9 ms * [Los Angeles to Miami latency and RTT](/docs/network/latency/pairs/lax-mia-rtt): 57.3 ms * [Los Angeles to New York latency and RTT](/docs/network/latency/pairs/lax-nyc-rtt): 57.9 ms * [Los Angeles to Ashburn latency and RTT](/docs/network/latency/pairs/lax-iad-rtt): 60.4 ms * [Los Angeles to Tokyo latency and RTT](/docs/network/latency/pairs/lax-tyo-rtt): 101.2 ms **Fastest routes arriving at Marseille (MRS)** * [Paris to Marseille latency and RTT](/docs/network/latency/pairs/par-mrs-rtt): 8.9 ms * [Frankfurt to Marseille latency and RTT](/docs/network/latency/pairs/fra-mrs-rtt): 16.4 ms * [London to Marseille latency and RTT](/docs/network/latency/pairs/lon-mrs-rtt): 17.9 ms * [Amsterdam to Marseille latency and RTT](/docs/network/latency/pairs/ams-mrs-rtt): 19.6 ms * [Berlin to Marseille latency and RTT](/docs/network/latency/pairs/ber-mrs-rtt): 20.2 ms **Same corridor (North America → Europe)** * [New York to London latency and RTT](/docs/network/latency/pairs/nyc-lon-rtt): 63.5 ms * [New York to Amsterdam latency and RTT](/docs/network/latency/pairs/nyc-ams-rtt): 68.7 ms * [New York to Paris latency and RTT](/docs/network/latency/pairs/nyc-par-rtt): 68.9 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/lax-mrs.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Los Angeles → Marseille round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/lax-mrs-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Los Angeles, USA → New York, USA RTT 🇺🇸 **Los Angeles, USA (LAX)** → 🇺🇸 **New York, USA (NYC)** Measurement round: `2026-08-16T04:07:28Z`. Los Angeles to New York averaged 57.9 ms across 50 replies. The observed range was 55.78 to 62.74 ms; these figures describe one probe round, not a long-term service target. This is number 3 of 19 outgoing routes by average RTT. That ranking compares different destinations and is not adjusted for distance. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **57.9 ms** | | Jitter | **1.19 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **38.63 ms** | | Fiber Efficiency | **66.7%** | | Latency Tier | Excellent | | Source Region | [North America](/docs/network/latency/regions/north-america) | | Destination Region | [North America](/docs/network/latency/regions/north-america) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Los Angeles](https://www.geonames.org/5368361) (34.05223, -118.24368) and [New York City](https://www.geonames.org/5128581) (40.71427, -74.00597). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **3,944.4 km** | | Vacuum RTT floor | **26.31 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **38.63 ms** | | Low-latency fiber material floor | **38.47 ms** | | Engineering floor (5% path allowance) | **40.56 ms** | | Research 1.33× mapped-fiber reference | **51.37 ms** | | Estimated unamplified path loss | **828.3 dB** | | Transparent optical spans / inline amplifiers | **52 / 51** | | Published RTT inflation over fiber floor | **1.5×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------ | | Minimum RTT | **55.78 ms** | | Average RTT | **57.9 ms** | | Maximum RTT | **62.74 ms** | | Standard deviation | **1.33 ms** | | Stdev / average | **2.3%** | ## Route Context * Departure PoP: [Los Angeles (LAX)](/docs/network/latency/lax-los-angeles) * Destination PoP: [New York (NYC)](/docs/network/latency/nyc-new-york) * Region overview: [North America](/docs/network/latency/regions/north-america) → [North America](/docs/network/latency/regions/north-america) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [New York to Los Angeles latency and RTT](/docs/network/latency/pairs/nyc-lax-rtt): 58.9 ms **Fastest routes departing Los Angeles (LAX)** * [Los Angeles to Seattle latency and RTT](/docs/network/latency/pairs/lax-sea-rtt): 26.9 ms * [Los Angeles to Miami latency and RTT](/docs/network/latency/pairs/lax-mia-rtt): 57.3 ms * [Los Angeles to Ashburn latency and RTT](/docs/network/latency/pairs/lax-iad-rtt): 60.4 ms * [Los Angeles to Tokyo latency and RTT](/docs/network/latency/pairs/lax-tyo-rtt): 101.2 ms * [Los Angeles to London latency and RTT](/docs/network/latency/pairs/lax-lon-rtt): 129.2 ms **Fastest routes arriving at New York (NYC)** * [Ashburn to New York latency and RTT](/docs/network/latency/pairs/iad-nyc-rtt): 6 ms * [Miami to New York latency and RTT](/docs/network/latency/pairs/mia-nyc-rtt): 33.3 ms * [Seattle to New York latency and RTT](/docs/network/latency/pairs/sea-nyc-rtt): 59.5 ms * [London to New York latency and RTT](/docs/network/latency/pairs/lon-nyc-rtt): 63.8 ms * [Amsterdam to New York latency and RTT](/docs/network/latency/pairs/ams-nyc-rtt): 68.6 ms **Same corridor (North America → North America)** * [New York to Ashburn latency and RTT](/docs/network/latency/pairs/nyc-iad-rtt): 6 ms * [Seattle to Los Angeles latency and RTT](/docs/network/latency/pairs/sea-lax-rtt): 25.9 ms * [Miami to Ashburn latency and RTT](/docs/network/latency/pairs/mia-iad-rtt): 27.3 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/lax-nyc.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Los Angeles → New York round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/lax-nyc-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Los Angeles, USA → Paris, France RTT 🇺🇸 **Los Angeles, USA (LAX)** → 🇫🇷 **Paris, France (PAR)** Measurement round: `2026-08-16T04:07:28Z`. Los Angeles to Paris averaged 135.9 ms across 50 replies. The observed range was 128.01 to 155.68 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 6.27 ms, or 4.6% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **135.9 ms** | | Jitter | **6.02 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **89.18 ms** | | Fiber Efficiency | **65.6%** | | Latency Tier | Good | | Source Region | [North America](/docs/network/latency/regions/north-america) | | Destination Region | [Europe](/docs/network/latency/regions/europe) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Los Angeles](https://www.geonames.org/5368361) (34.05223, -118.24368) and [Paris](https://www.geonames.org/2988507) (48.85341, 2.34880). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **9,107.2 km** | | Vacuum RTT floor | **60.76 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **89.18 ms** | | Low-latency fiber material floor | **88.83 ms** | | Engineering floor (5% path allowance) | **93.66 ms** | | Research 1.33× mapped-fiber reference | **118.62 ms** | | Estimated unamplified path loss | **1912.5 dB** | | Transparent optical spans / inline amplifiers | **120 / 119** | | Published RTT inflation over fiber floor | **1.52×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **128.01 ms** | | Average RTT | **135.9 ms** | | Maximum RTT | **155.68 ms** | | Standard deviation | **6.27 ms** | | Stdev / average | **4.6%** | ## Route Context * Departure PoP: [Los Angeles (LAX)](/docs/network/latency/lax-los-angeles) * Destination PoP: [Paris (PAR)](/docs/network/latency/par-paris) * Region overview: [North America](/docs/network/latency/regions/north-america) → [Europe](/docs/network/latency/regions/europe) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Paris to Los Angeles latency and RTT](/docs/network/latency/pairs/par-lax-rtt): 136.5 ms **Fastest routes departing Los Angeles (LAX)** * [Los Angeles to Seattle latency and RTT](/docs/network/latency/pairs/lax-sea-rtt): 26.9 ms * [Los Angeles to Miami latency and RTT](/docs/network/latency/pairs/lax-mia-rtt): 57.3 ms * [Los Angeles to New York latency and RTT](/docs/network/latency/pairs/lax-nyc-rtt): 57.9 ms * [Los Angeles to Ashburn latency and RTT](/docs/network/latency/pairs/lax-iad-rtt): 60.4 ms * [Los Angeles to Tokyo latency and RTT](/docs/network/latency/pairs/lax-tyo-rtt): 101.2 ms **Fastest routes arriving at Paris (PAR)** * [London to Paris latency and RTT](/docs/network/latency/pairs/lon-par-rtt): 6.4 ms * [Amsterdam to Paris latency and RTT](/docs/network/latency/pairs/ams-par-rtt): 7 ms * [Frankfurt to Paris latency and RTT](/docs/network/latency/pairs/fra-par-rtt): 7.6 ms * [Marseille to Paris latency and RTT](/docs/network/latency/pairs/mrs-par-rtt): 8.9 ms * [Berlin to Paris latency and RTT](/docs/network/latency/pairs/ber-par-rtt): 15.6 ms **Same corridor (North America → Europe)** * [New York to London latency and RTT](/docs/network/latency/pairs/nyc-lon-rtt): 63.5 ms * [New York to Amsterdam latency and RTT](/docs/network/latency/pairs/nyc-ams-rtt): 68.7 ms * [New York to Paris latency and RTT](/docs/network/latency/pairs/nyc-par-rtt): 68.9 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/lax-par.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Los Angeles → Paris round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/lax-par-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Los Angeles, USA → Seattle, USA RTT 🇺🇸 **Los Angeles, USA (LAX)** → 🇺🇸 **Seattle, USA (SEA)** Measurement round: `2026-08-16T04:07:28Z`. Los Angeles to Seattle averaged 26.9 ms across 50 replies. The observed range was 25.5 to 29.88 ms; these figures describe one probe round, not a long-term service target. This is number 1 of 19 outgoing routes by average RTT. That ranking compares different destinations and is not adjusted for distance. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **26.9 ms** | | Jitter | **0.89 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **15.12 ms** | | Fiber Efficiency | **56.2%** | | Latency Tier | Ultra-Low | | Source Region | [North America](/docs/network/latency/regions/north-america) | | Destination Region | [North America](/docs/network/latency/regions/north-america) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Los Angeles](https://www.geonames.org/5368361) (34.05223, -118.24368) and [Seattle](https://www.geonames.org/5809844) (47.60621, -122.33207). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **1,543.6 km** | | Vacuum RTT floor | **10.3 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **15.12 ms** | | Low-latency fiber material floor | **15.06 ms** | | Engineering floor (5% path allowance) | **15.87 ms** | | Research 1.33× mapped-fiber reference | **20.1 ms** | | Estimated unamplified path loss | **324.2 dB** | | Transparent optical spans / inline amplifiers | **21 / 20** | | Published RTT inflation over fiber floor | **1.78×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------ | | Minimum RTT | **25.5 ms** | | Average RTT | **26.9 ms** | | Maximum RTT | **29.88 ms** | | Standard deviation | **1.13 ms** | | Stdev / average | **4.2%** | ## Route Context * Departure PoP: [Los Angeles (LAX)](/docs/network/latency/lax-los-angeles) * Destination PoP: [Seattle (SEA)](/docs/network/latency/sea-seattle) * Region overview: [North America](/docs/network/latency/regions/north-america) → [North America](/docs/network/latency/regions/north-america) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Seattle to Los Angeles latency and RTT](/docs/network/latency/pairs/sea-lax-rtt): 25.9 ms **Fastest routes departing Los Angeles (LAX)** * [Los Angeles to Miami latency and RTT](/docs/network/latency/pairs/lax-mia-rtt): 57.3 ms * [Los Angeles to New York latency and RTT](/docs/network/latency/pairs/lax-nyc-rtt): 57.9 ms * [Los Angeles to Ashburn latency and RTT](/docs/network/latency/pairs/lax-iad-rtt): 60.4 ms * [Los Angeles to Tokyo latency and RTT](/docs/network/latency/pairs/lax-tyo-rtt): 101.2 ms * [Los Angeles to London latency and RTT](/docs/network/latency/pairs/lax-lon-rtt): 129.2 ms **Fastest routes arriving at Seattle (SEA)** * [New York to Seattle latency and RTT](/docs/network/latency/pairs/nyc-sea-rtt): 58.7 ms * [Ashburn to Seattle latency and RTT](/docs/network/latency/pairs/iad-sea-rtt): 62.3 ms * [Miami to Seattle latency and RTT](/docs/network/latency/pairs/mia-sea-rtt): 81.9 ms * [Tokyo to Seattle latency and RTT](/docs/network/latency/pairs/tyo-sea-rtt): 84.8 ms * [Taipei to Seattle latency and RTT](/docs/network/latency/pairs/tpe-sea-rtt): 115.8 ms **Same corridor (North America → North America)** * [Ashburn to New York latency and RTT](/docs/network/latency/pairs/iad-nyc-rtt): 6 ms * [New York to Ashburn latency and RTT](/docs/network/latency/pairs/nyc-iad-rtt): 6 ms * [Miami to Ashburn latency and RTT](/docs/network/latency/pairs/mia-iad-rtt): 27.3 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/lax-sea.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Los Angeles → Seattle round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/lax-sea-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Los Angeles, USA → Singapore RTT 🇺🇸 **Los Angeles, USA (LAX)** → 🇸🇬 **Singapore (SIN)** Measurement round: `2026-08-16T04:07:28Z`. Los Angeles to Singapore averaged 167.7 ms across 50 replies. The observed range was 163.93 to 178.4 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 3.29 ms, or 2.0% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **167.7 ms** | | Jitter | **2.59 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **138.43 ms** | | Fiber Efficiency | **82.5%** | | Latency Tier | Fair | | Source Region | [North America](/docs/network/latency/regions/north-america) | | Destination Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Los Angeles](https://www.geonames.org/5368361) (34.05223, -118.24368) and [Singapore](https://www.geonames.org/1880252) (1.28967, 103.85007). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | --------------- | | WGS-84 geodesic distance | **14,135.8 km** | | Vacuum RTT floor | **94.3 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **138.43 ms** | | Low-latency fiber material floor | **137.87 ms** | | Engineering floor (5% path allowance) | **145.37 ms** | | Research 1.33× mapped-fiber reference | **184.11 ms** | | Estimated unamplified path loss | **2968.5 dB** | | Transparent optical spans / inline amplifiers | **186 / 185** | | Published RTT inflation over fiber floor | **1.21×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **163.93 ms** | | Average RTT | **167.7 ms** | | Maximum RTT | **178.4 ms** | | Standard deviation | **3.29 ms** | | Stdev / average | **2.0%** | ## Route Context * Departure PoP: [Los Angeles (LAX)](/docs/network/latency/lax-los-angeles) * Destination PoP: [Singapore (SIN)](/docs/network/latency/sin-singapore) * Region overview: [North America](/docs/network/latency/regions/north-america) → [Asia Pacific](/docs/network/latency/regions/asia-pacific) * Corridor overview: [North America to Asia Pacific](/docs/network/latency/regions/north-america-to-asia-pacific) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Singapore to Los Angeles latency and RTT](/docs/network/latency/pairs/sin-lax-rtt): 168.2 ms **Fastest routes departing Los Angeles (LAX)** * [Los Angeles to Seattle latency and RTT](/docs/network/latency/pairs/lax-sea-rtt): 26.9 ms * [Los Angeles to Miami latency and RTT](/docs/network/latency/pairs/lax-mia-rtt): 57.3 ms * [Los Angeles to New York latency and RTT](/docs/network/latency/pairs/lax-nyc-rtt): 57.9 ms * [Los Angeles to Ashburn latency and RTT](/docs/network/latency/pairs/lax-iad-rtt): 60.4 ms * [Los Angeles to Tokyo latency and RTT](/docs/network/latency/pairs/lax-tyo-rtt): 101.2 ms **Fastest routes arriving at Singapore (SIN)** * [Hong Kong to Singapore latency and RTT](/docs/network/latency/pairs/hkg-sin-rtt): 31.5 ms * [Taipei to Singapore latency and RTT](/docs/network/latency/pairs/tpe-sin-rtt): 44.3 ms * [Tokyo to Singapore latency and RTT](/docs/network/latency/pairs/tyo-sin-rtt): 69.4 ms * [Melbourne to Singapore latency and RTT](/docs/network/latency/pairs/mel-sin-rtt): 88.4 ms * [Sydney to Singapore latency and RTT](/docs/network/latency/pairs/syd-sin-rtt): 94.5 ms **Same corridor (North America → Asia Pacific)** * [Seattle to Tokyo latency and RTT](/docs/network/latency/pairs/sea-tyo-rtt): 85 ms * [Seattle to Taipei latency and RTT](/docs/network/latency/pairs/sea-tpe-rtt): 114.7 ms * [Seattle to Hong Kong latency and RTT](/docs/network/latency/pairs/sea-hkg-rtt): 130.4 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/lax-sin.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Los Angeles → Singapore round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/lax-sin-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Los Angeles, USA → Sydney, Australia RTT 🇺🇸 **Los Angeles, USA (LAX)** → 🇦🇺 **Sydney, Australia (SYD)** Measurement round: `2026-08-16T04:07:28Z`. Los Angeles to Sydney averaged 137.6 ms across 50 replies. The observed range was 134.53 to 146.99 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 2.45 ms, or 1.8% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **137.6 ms** | | Jitter | **1.96 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **118.13 ms** | | Fiber Efficiency | **85.9%** | | Latency Tier | Good | | Source Region | [North America](/docs/network/latency/regions/north-america) | | Destination Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Los Angeles](https://www.geonames.org/5368361) (34.05223, -118.24368) and [Sydney](https://www.geonames.org/2147714) (-33.86785, 151.20732). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | --------------- | | WGS-84 geodesic distance | **12,063.2 km** | | Vacuum RTT floor | **80.48 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **118.13 ms** | | Low-latency fiber material floor | **117.66 ms** | | Engineering floor (5% path allowance) | **124.05 ms** | | Research 1.33× mapped-fiber reference | **157.12 ms** | | Estimated unamplified path loss | **2533.3 dB** | | Transparent optical spans / inline amplifiers | **159 / 158** | | Published RTT inflation over fiber floor | **1.16×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **134.53 ms** | | Average RTT | **137.6 ms** | | Maximum RTT | **146.99 ms** | | Standard deviation | **2.45 ms** | | Stdev / average | **1.8%** | ## Route Context * Departure PoP: [Los Angeles (LAX)](/docs/network/latency/lax-los-angeles) * Destination PoP: [Sydney (SYD)](/docs/network/latency/syd-sydney) * Region overview: [North America](/docs/network/latency/regions/north-america) → [Asia Pacific](/docs/network/latency/regions/asia-pacific) * Corridor overview: [North America to Asia Pacific](/docs/network/latency/regions/north-america-to-asia-pacific) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Sydney to Los Angeles latency and RTT](/docs/network/latency/pairs/syd-lax-rtt): 136.5 ms **Fastest routes departing Los Angeles (LAX)** * [Los Angeles to Seattle latency and RTT](/docs/network/latency/pairs/lax-sea-rtt): 26.9 ms * [Los Angeles to Miami latency and RTT](/docs/network/latency/pairs/lax-mia-rtt): 57.3 ms * [Los Angeles to New York latency and RTT](/docs/network/latency/pairs/lax-nyc-rtt): 57.9 ms * [Los Angeles to Ashburn latency and RTT](/docs/network/latency/pairs/lax-iad-rtt): 60.4 ms * [Los Angeles to Tokyo latency and RTT](/docs/network/latency/pairs/lax-tyo-rtt): 101.2 ms **Fastest routes arriving at Sydney (SYD)** * [Melbourne to Sydney latency and RTT](/docs/network/latency/pairs/mel-syd-rtt): 9.8 ms * [Singapore to Sydney latency and RTT](/docs/network/latency/pairs/sin-syd-rtt): 94.5 ms * [Tokyo to Sydney latency and RTT](/docs/network/latency/pairs/tyo-syd-rtt): 101.9 ms * [Taipei to Sydney latency and RTT](/docs/network/latency/pairs/tpe-syd-rtt): 133.2 ms * [Hong Kong to Sydney latency and RTT](/docs/network/latency/pairs/hkg-syd-rtt): 137.2 ms **Same corridor (North America → Asia Pacific)** * [Seattle to Tokyo latency and RTT](/docs/network/latency/pairs/sea-tyo-rtt): 85 ms * [Seattle to Taipei latency and RTT](/docs/network/latency/pairs/sea-tpe-rtt): 114.7 ms * [Seattle to Hong Kong latency and RTT](/docs/network/latency/pairs/sea-hkg-rtt): 130.4 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/lax-syd.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Los Angeles → Sydney round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/lax-syd-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Los Angeles, USA → Taipei, Taiwan RTT 🇺🇸 **Los Angeles, USA (LAX)** → 🇹🇼 **Taipei, Taiwan (TPE)** Measurement round: `2026-08-16T04:07:28Z`. Los Angeles to Taipei averaged 132.9 ms across 50 replies. The observed range was 128.75 to 148.78 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 4.24 ms, or 3.2% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **132.9 ms** | | Jitter | **3.59 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **107 ms** | | Fiber Efficiency | **80.5%** | | Latency Tier | Good | | Source Region | [North America](/docs/network/latency/regions/north-america) | | Destination Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Los Angeles](https://www.geonames.org/5368361) (34.05223, -118.24368) and [Taipei](https://www.geonames.org/1668341) (25.05306, 121.52639). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | --------------- | | WGS-84 geodesic distance | **10,926.8 km** | | Vacuum RTT floor | **72.9 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **107 ms** | | Low-latency fiber material floor | **106.57 ms** | | Engineering floor (5% path allowance) | **112.37 ms** | | Research 1.33× mapped-fiber reference | **142.32 ms** | | Estimated unamplified path loss | **2294.6 dB** | | Transparent optical spans / inline amplifiers | **144 / 143** | | Published RTT inflation over fiber floor | **1.24×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **128.75 ms** | | Average RTT | **132.9 ms** | | Maximum RTT | **148.78 ms** | | Standard deviation | **4.24 ms** | | Stdev / average | **3.2%** | ## Route Context * Departure PoP: [Los Angeles (LAX)](/docs/network/latency/lax-los-angeles) * Destination PoP: [Taipei (TPE)](/docs/network/latency/tpe-taipei) * Region overview: [North America](/docs/network/latency/regions/north-america) → [Asia Pacific](/docs/network/latency/regions/asia-pacific) * Corridor overview: [North America to Asia Pacific](/docs/network/latency/regions/north-america-to-asia-pacific) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Taipei to Los Angeles latency and RTT](/docs/network/latency/pairs/tpe-lax-rtt): 132 ms **Fastest routes departing Los Angeles (LAX)** * [Los Angeles to Seattle latency and RTT](/docs/network/latency/pairs/lax-sea-rtt): 26.9 ms * [Los Angeles to Miami latency and RTT](/docs/network/latency/pairs/lax-mia-rtt): 57.3 ms * [Los Angeles to New York latency and RTT](/docs/network/latency/pairs/lax-nyc-rtt): 57.9 ms * [Los Angeles to Ashburn latency and RTT](/docs/network/latency/pairs/lax-iad-rtt): 60.4 ms * [Los Angeles to Tokyo latency and RTT](/docs/network/latency/pairs/lax-tyo-rtt): 101.2 ms **Fastest routes arriving at Taipei (TPE)** * [Hong Kong to Taipei latency and RTT](/docs/network/latency/pairs/hkg-tpe-rtt): 14.7 ms * [Tokyo to Taipei latency and RTT](/docs/network/latency/pairs/tyo-tpe-rtt): 32.3 ms * [Singapore to Taipei latency and RTT](/docs/network/latency/pairs/sin-tpe-rtt): 45.8 ms * [Seattle to Taipei latency and RTT](/docs/network/latency/pairs/sea-tpe-rtt): 114.7 ms * [Sydney to Taipei latency and RTT](/docs/network/latency/pairs/syd-tpe-rtt): 131.7 ms **Same corridor (North America → Asia Pacific)** * [Seattle to Tokyo latency and RTT](/docs/network/latency/pairs/sea-tyo-rtt): 85 ms * [Seattle to Hong Kong latency and RTT](/docs/network/latency/pairs/sea-hkg-rtt): 130.4 ms * [Los Angeles to Sydney latency and RTT](/docs/network/latency/pairs/lax-syd-rtt): 137.6 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/lax-tpe.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Los Angeles → Taipei round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/lax-tpe-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Los Angeles, USA → Tokyo, Japan RTT 🇺🇸 **Los Angeles, USA (LAX)** → 🇯🇵 **Tokyo, Japan (TYO)** Measurement round: `2026-08-16T04:07:28Z`. Los Angeles to Tokyo averaged 101.2 ms across 50 replies. The observed range was 97.1 to 112.57 ms; these figures describe one probe round, not a long-term service target. This is number 5 of 19 outgoing routes by average RTT. That ranking compares different destinations and is not adjusted for distance. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **101.2 ms** | | Jitter | **3.3 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **86.51 ms** | | Fiber Efficiency | **85.5%** | | Latency Tier | Good | | Source Region | [North America](/docs/network/latency/regions/north-america) | | Destination Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Los Angeles](https://www.geonames.org/5368361) (34.05223, -118.24368) and [Tokyo](https://www.geonames.org/1850147) (35.68950, 139.69171). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **8,834.5 km** | | Vacuum RTT floor | **58.94 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **86.51 ms** | | Low-latency fiber material floor | **86.17 ms** | | Engineering floor (5% path allowance) | **90.85 ms** | | Research 1.33× mapped-fiber reference | **115.06 ms** | | Estimated unamplified path loss | **1855.3 dB** | | Transparent optical spans / inline amplifiers | **116 / 115** | | Published RTT inflation over fiber floor | **1.17×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **97.1 ms** | | Average RTT | **101.2 ms** | | Maximum RTT | **112.57 ms** | | Standard deviation | **3.72 ms** | | Stdev / average | **3.7%** | ## Route Context * Departure PoP: [Los Angeles (LAX)](/docs/network/latency/lax-los-angeles) * Destination PoP: [Tokyo (TYO)](/docs/network/latency/tyo-tokyo) * Region overview: [North America](/docs/network/latency/regions/north-america) → [Asia Pacific](/docs/network/latency/regions/asia-pacific) * Corridor overview: [North America to Asia Pacific](/docs/network/latency/regions/north-america-to-asia-pacific) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Tokyo to Los Angeles latency and RTT](/docs/network/latency/pairs/tyo-lax-rtt): 101.2 ms **Fastest routes departing Los Angeles (LAX)** * [Los Angeles to Seattle latency and RTT](/docs/network/latency/pairs/lax-sea-rtt): 26.9 ms * [Los Angeles to Miami latency and RTT](/docs/network/latency/pairs/lax-mia-rtt): 57.3 ms * [Los Angeles to New York latency and RTT](/docs/network/latency/pairs/lax-nyc-rtt): 57.9 ms * [Los Angeles to Ashburn latency and RTT](/docs/network/latency/pairs/lax-iad-rtt): 60.4 ms * [Los Angeles to London latency and RTT](/docs/network/latency/pairs/lax-lon-rtt): 129.2 ms **Fastest routes arriving at Tokyo (TYO)** * [Taipei to Tokyo latency and RTT](/docs/network/latency/pairs/tpe-tyo-rtt): 31.9 ms * [Hong Kong to Tokyo latency and RTT](/docs/network/latency/pairs/hkg-tyo-rtt): 44.9 ms * [Singapore to Tokyo latency and RTT](/docs/network/latency/pairs/sin-tyo-rtt): 68.9 ms * [Seattle to Tokyo latency and RTT](/docs/network/latency/pairs/sea-tyo-rtt): 85 ms * [Sydney to Tokyo latency and RTT](/docs/network/latency/pairs/syd-tyo-rtt): 101.3 ms **Same corridor (North America → Asia Pacific)** * [Seattle to Taipei latency and RTT](/docs/network/latency/pairs/sea-tpe-rtt): 114.7 ms * [Seattle to Hong Kong latency and RTT](/docs/network/latency/pairs/sea-hkg-rtt): 130.4 ms * [Los Angeles to Taipei latency and RTT](/docs/network/latency/pairs/lax-tpe-rtt): 132.9 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/lax-tyo.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Los Angeles → Tokyo round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/lax-tyo-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # London, UK → Amsterdam, Netherlands RTT 🇬🇧 **London, UK (LON)** → 🇳🇱 **Amsterdam, Netherlands (AMS)** Measurement round: `2026-08-16T04:07:28Z`. London to Amsterdam averaged 5.2 ms across 50 replies. The observed range was 5.03 to 5.75 ms; these figures describe one probe round, not a long-term service target. This is number 1 of 19 outgoing routes by average RTT. That ranking compares different destinations and is not adjusted for distance. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------- | | RTT | **5.2 ms** | | Jitter | **0.13 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **3.51 ms** | | Fiber Efficiency | **67.4%** | | Latency Tier | Ultra-Low | | Source Region | [Europe](/docs/network/latency/regions/europe) | | Destination Region | [Europe](/docs/network/latency/regions/europe) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [London](https://www.geonames.org/2643743) (51.50853, -0.12574) and [Amsterdam](https://www.geonames.org/2759794) (52.37403, 4.88969). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | ----------- | | WGS-84 geodesic distance | **358 km** | | Vacuum RTT floor | **2.39 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **3.51 ms** | | Low-latency fiber material floor | **3.49 ms** | | Engineering floor (5% path allowance) | **3.68 ms** | | Research 1.33× mapped-fiber reference | **4.66 ms** | | Estimated unamplified path loss | **75.2 dB** | | Transparent optical spans / inline amplifiers | **5 / 4** | | Published RTT inflation over fiber floor | **1.48×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ----------- | | Minimum RTT | **5.03 ms** | | Average RTT | **5.2 ms** | | Maximum RTT | **5.75 ms** | | Standard deviation | **0.16 ms** | | Stdev / average | **3.1%** | ## Route Context * Departure PoP: [London (LON)](/docs/network/latency/lon-london) * Destination PoP: [Amsterdam (AMS)](/docs/network/latency/ams-amsterdam) * Region overview: [Europe](/docs/network/latency/regions/europe) → [Europe](/docs/network/latency/regions/europe) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Amsterdam to London latency and RTT](/docs/network/latency/pairs/ams-lon-rtt): 5.2 ms **Fastest routes departing London (LON)** * [London to Paris latency and RTT](/docs/network/latency/pairs/lon-par-rtt): 6.4 ms * [London to Frankfurt latency and RTT](/docs/network/latency/pairs/lon-fra-rtt): 13.6 ms * [London to Berlin latency and RTT](/docs/network/latency/pairs/lon-ber-rtt): 16.2 ms * [London to Marseille latency and RTT](/docs/network/latency/pairs/lon-mrs-rtt): 17.9 ms * [London to Moscow latency and RTT](/docs/network/latency/pairs/lon-mow-rtt): 43.5 ms **Fastest routes arriving at Amsterdam (AMS)** * [Frankfurt to Amsterdam latency and RTT](/docs/network/latency/pairs/fra-ams-rtt): 5.9 ms * [Paris to Amsterdam latency and RTT](/docs/network/latency/pairs/par-ams-rtt): 7 ms * [Berlin to Amsterdam latency and RTT](/docs/network/latency/pairs/ber-ams-rtt): 7.2 ms * [Marseille to Amsterdam latency and RTT](/docs/network/latency/pairs/mrs-ams-rtt): 20.3 ms * [Moscow to Amsterdam latency and RTT](/docs/network/latency/pairs/mow-ams-rtt): 37.6 ms **Same corridor (Europe → Europe)** * [Amsterdam to Frankfurt latency and RTT](/docs/network/latency/pairs/ams-fra-rtt): 5.9 ms * [Berlin to Frankfurt latency and RTT](/docs/network/latency/pairs/ber-fra-rtt): 6.1 ms * [Frankfurt to Berlin latency and RTT](/docs/network/latency/pairs/fra-ber-rtt): 6.1 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/lon-ams.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, London → Amsterdam round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/lon-ams-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # London, UK → Berlin, Germany RTT 🇬🇧 **London, UK (LON)** → 🇩🇪 **Berlin, Germany (BER)** Measurement round: `2026-08-16T04:07:28Z`. London to Berlin averaged 16.2 ms across 50 replies. The observed range was 15.5 to 19.35 ms; these figures describe one probe round, not a long-term service target. This is number 4 of 19 outgoing routes by average RTT. That ranking compares different destinations and is not adjusted for distance. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------- | | RTT | **16.2 ms** | | Jitter | **0.64 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **9.15 ms** | | Fiber Efficiency | **56.5%** | | Latency Tier | Ultra-Low | | Source Region | [Europe](/docs/network/latency/regions/europe) | | Destination Region | [Europe](/docs/network/latency/regions/europe) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [London](https://www.geonames.org/2643743) (51.50853, -0.12574) and [Berlin](https://www.geonames.org/2950159) (52.52437, 13.41053). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | ------------ | | WGS-84 geodesic distance | **934.7 km** | | Vacuum RTT floor | **6.24 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **9.15 ms** | | Low-latency fiber material floor | **9.12 ms** | | Engineering floor (5% path allowance) | **9.61 ms** | | Research 1.33× mapped-fiber reference | **12.17 ms** | | Estimated unamplified path loss | **196.3 dB** | | Transparent optical spans / inline amplifiers | **13 / 12** | | Published RTT inflation over fiber floor | **1.77×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------ | | Minimum RTT | **15.5 ms** | | Average RTT | **16.2 ms** | | Maximum RTT | **19.35 ms** | | Standard deviation | **0.73 ms** | | Stdev / average | **4.5%** | ## Route Context * Departure PoP: [London (LON)](/docs/network/latency/lon-london) * Destination PoP: [Berlin (BER)](/docs/network/latency/ber-berlin) * Region overview: [Europe](/docs/network/latency/regions/europe) → [Europe](/docs/network/latency/regions/europe) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Berlin to London latency and RTT](/docs/network/latency/pairs/ber-lon-rtt): 15.2 ms **Fastest routes departing London (LON)** * [London to Amsterdam latency and RTT](/docs/network/latency/pairs/lon-ams-rtt): 5.2 ms * [London to Paris latency and RTT](/docs/network/latency/pairs/lon-par-rtt): 6.4 ms * [London to Frankfurt latency and RTT](/docs/network/latency/pairs/lon-fra-rtt): 13.6 ms * [London to Marseille latency and RTT](/docs/network/latency/pairs/lon-mrs-rtt): 17.9 ms * [London to Moscow latency and RTT](/docs/network/latency/pairs/lon-mow-rtt): 43.5 ms **Fastest routes arriving at Berlin (BER)** * [Frankfurt to Berlin latency and RTT](/docs/network/latency/pairs/fra-ber-rtt): 6.1 ms * [Amsterdam to Berlin latency and RTT](/docs/network/latency/pairs/ams-ber-rtt): 7.2 ms * [Paris to Berlin latency and RTT](/docs/network/latency/pairs/par-ber-rtt): 15.5 ms * [Marseille to Berlin latency and RTT](/docs/network/latency/pairs/mrs-ber-rtt): 20.6 ms * [Moscow to Berlin latency and RTT](/docs/network/latency/pairs/mow-ber-rtt): 27.5 ms **Same corridor (Europe → Europe)** * [Amsterdam to London latency and RTT](/docs/network/latency/pairs/ams-lon-rtt): 5.2 ms * [Amsterdam to Frankfurt latency and RTT](/docs/network/latency/pairs/ams-fra-rtt): 5.9 ms * [Frankfurt to Amsterdam latency and RTT](/docs/network/latency/pairs/fra-ams-rtt): 5.9 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/lon-ber.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, London → Berlin round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/lon-ber-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # London, UK → Frankfurt, Germany RTT 🇬🇧 **London, UK (LON)** → 🇩🇪 **Frankfurt, Germany (FRA)** Measurement round: `2026-08-16T04:07:28Z`. London to Frankfurt averaged 13.6 ms, compared with a 6.26 ms city-centre fiber reference. The difference cannot be assigned to a particular cable or carrier from RTT alone. This is number 3 of 19 outgoing routes by average RTT. That ranking compares different destinations and is not adjusted for distance. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------- | | RTT | **13.6 ms** | | Jitter | **0.29 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **6.26 ms** | | Fiber Efficiency | **46.1%** | | Latency Tier | Ultra-Low | | Source Region | [Europe](/docs/network/latency/regions/europe) | | Destination Region | [Europe](/docs/network/latency/regions/europe) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [London](https://www.geonames.org/2643743) (51.50853, -0.12574) and [Frankfurt am Main](https://www.geonames.org/2925533) (50.11552, 8.68417). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | ------------ | | WGS-84 geodesic distance | **639.5 km** | | Vacuum RTT floor | **4.27 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **6.26 ms** | | Low-latency fiber material floor | **6.24 ms** | | Engineering floor (5% path allowance) | **6.58 ms** | | Research 1.33× mapped-fiber reference | **8.33 ms** | | Estimated unamplified path loss | **134.3 dB** | | Transparent optical spans / inline amplifiers | **9 / 8** | | Published RTT inflation over fiber floor | **2.17×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------ | | Minimum RTT | **13.18 ms** | | Average RTT | **13.6 ms** | | Maximum RTT | **15.43 ms** | | Standard deviation | **0.42 ms** | | Stdev / average | **3.1%** | ## Route Context * Departure PoP: [London (LON)](/docs/network/latency/lon-london) * Destination PoP: [Frankfurt (FRA)](/docs/network/latency/fra-frankfurt) * Region overview: [Europe](/docs/network/latency/regions/europe) → [Europe](/docs/network/latency/regions/europe) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Frankfurt to London latency and RTT](/docs/network/latency/pairs/fra-lon-rtt): 12.9 ms **Fastest routes departing London (LON)** * [London to Amsterdam latency and RTT](/docs/network/latency/pairs/lon-ams-rtt): 5.2 ms * [London to Paris latency and RTT](/docs/network/latency/pairs/lon-par-rtt): 6.4 ms * [London to Berlin latency and RTT](/docs/network/latency/pairs/lon-ber-rtt): 16.2 ms * [London to Marseille latency and RTT](/docs/network/latency/pairs/lon-mrs-rtt): 17.9 ms * [London to Moscow latency and RTT](/docs/network/latency/pairs/lon-mow-rtt): 43.5 ms **Fastest routes arriving at Frankfurt (FRA)** * [Amsterdam to Frankfurt latency and RTT](/docs/network/latency/pairs/ams-fra-rtt): 5.9 ms * [Berlin to Frankfurt latency and RTT](/docs/network/latency/pairs/ber-fra-rtt): 6.1 ms * [Paris to Frankfurt latency and RTT](/docs/network/latency/pairs/par-fra-rtt): 7.6 ms * [Marseille to Frankfurt latency and RTT](/docs/network/latency/pairs/mrs-fra-rtt): 16 ms * [Moscow to Frankfurt latency and RTT](/docs/network/latency/pairs/mow-fra-rtt): 35.9 ms **Same corridor (Europe → Europe)** * [Amsterdam to London latency and RTT](/docs/network/latency/pairs/ams-lon-rtt): 5.2 ms * [Frankfurt to Amsterdam latency and RTT](/docs/network/latency/pairs/fra-ams-rtt): 5.9 ms * [Frankfurt to Berlin latency and RTT](/docs/network/latency/pairs/fra-ber-rtt): 6.1 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/lon-fra.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, London → Frankfurt round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/lon-fra-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # London, UK → São Paulo, Brazil RTT 🇬🇧 **London, UK (LON)** → 🇧🇷 **São Paulo, Brazil (GRU)** Measurement round: `2026-08-16T04:07:28Z`. London to São Paulo averaged 170.7 ms across 50 replies. The observed range was 165.28 to 185.16 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 4.35 ms, or 2.5% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **170.7 ms** | | Jitter | **4.17 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **92.76 ms** | | Fiber Efficiency | **54.3%** | | Latency Tier | Fair | | Source Region | [Europe](/docs/network/latency/regions/europe) | | Destination Region | [South America](/docs/network/latency/regions/south-america) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [London](https://www.geonames.org/2643743) (51.50853, -0.12574) and [São Paulo](https://www.geonames.org/3448439) (-23.54750, -46.63611). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **9,472.8 km** | | Vacuum RTT floor | **63.2 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **92.76 ms** | | Low-latency fiber material floor | **92.39 ms** | | Engineering floor (5% path allowance) | **97.42 ms** | | Research 1.33× mapped-fiber reference | **123.38 ms** | | Estimated unamplified path loss | **1989.3 dB** | | Transparent optical spans / inline amplifiers | **125 / 124** | | Published RTT inflation over fiber floor | **1.84×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **165.28 ms** | | Average RTT | **170.7 ms** | | Maximum RTT | **185.16 ms** | | Standard deviation | **4.35 ms** | | Stdev / average | **2.5%** | ## Route Context * Departure PoP: [London (LON)](/docs/network/latency/lon-london) * Destination PoP: [São Paulo (GRU)](/docs/network/latency/gru-sao-paulo) * Region overview: [Europe](/docs/network/latency/regions/europe) → [South America](/docs/network/latency/regions/south-america) * Corridor overview: [Europe to South America](/docs/network/latency/regions/europe-to-south-america) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [São Paulo to London latency and RTT](/docs/network/latency/pairs/gru-lon-rtt): 175.8 ms **Fastest routes departing London (LON)** * [London to Amsterdam latency and RTT](/docs/network/latency/pairs/lon-ams-rtt): 5.2 ms * [London to Paris latency and RTT](/docs/network/latency/pairs/lon-par-rtt): 6.4 ms * [London to Frankfurt latency and RTT](/docs/network/latency/pairs/lon-fra-rtt): 13.6 ms * [London to Berlin latency and RTT](/docs/network/latency/pairs/lon-ber-rtt): 16.2 ms * [London to Marseille latency and RTT](/docs/network/latency/pairs/lon-mrs-rtt): 17.9 ms **Fastest routes arriving at São Paulo (GRU)** * [Ashburn to São Paulo latency and RTT](/docs/network/latency/pairs/iad-gru-rtt): 101.4 ms * [New York to São Paulo latency and RTT](/docs/network/latency/pairs/nyc-gru-rtt): 106.9 ms * [Miami to São Paulo latency and RTT](/docs/network/latency/pairs/mia-gru-rtt): 128.7 ms * [Los Angeles to São Paulo latency and RTT](/docs/network/latency/pairs/lax-gru-rtt): 131.3 ms * [Seattle to São Paulo latency and RTT](/docs/network/latency/pairs/sea-gru-rtt): 155.7 ms **Same corridor (Europe → South America)** * [Amsterdam to São Paulo latency and RTT](/docs/network/latency/pairs/ams-gru-rtt): 175.5 ms * [Paris to São Paulo latency and RTT](/docs/network/latency/pairs/par-gru-rtt): 175.8 ms * [Frankfurt to São Paulo latency and RTT](/docs/network/latency/pairs/fra-gru-rtt): 181.3 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/lon-gru.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, London → São Paulo round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/lon-gru-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # London, UK → Hong Kong RTT 🇬🇧 **London, UK (LON)** → 🇭🇰 **Hong Kong (HKG)** Measurement round: `2026-08-16T04:07:28Z`. London to Hong Kong averaged 160.6 ms across 50 replies. The observed range was 151.57 to 189.08 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 7.9 ms, or 4.9% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------------------- | | RTT | **160.6 ms** | | Jitter | **6.03 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **94.44 ms** | | Fiber Efficiency | **58.8%** | | Latency Tier | Fair | | Source Region | [Europe](/docs/network/latency/regions/europe) | | Destination Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [London](https://www.geonames.org/2643743) (51.50853, -0.12574) and [Hong Kong](https://www.geonames.org/1819729) (22.27832, 114.17469). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **9,643.6 km** | | Vacuum RTT floor | **64.34 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **94.44 ms** | | Low-latency fiber material floor | **94.06 ms** | | Engineering floor (5% path allowance) | **99.17 ms** | | Research 1.33× mapped-fiber reference | **125.6 ms** | | Estimated unamplified path loss | **2025.2 dB** | | Transparent optical spans / inline amplifiers | **127 / 126** | | Published RTT inflation over fiber floor | **1.7×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **151.57 ms** | | Average RTT | **160.6 ms** | | Maximum RTT | **189.08 ms** | | Standard deviation | **7.9 ms** | | Stdev / average | **4.9%** | ## Route Context * Departure PoP: [London (LON)](/docs/network/latency/lon-london) * Destination PoP: [Hong Kong (HKG)](/docs/network/latency/hkg-hong-kong) * Region overview: [Europe](/docs/network/latency/regions/europe) → [Asia Pacific](/docs/network/latency/regions/asia-pacific) * Corridor overview: [Europe to Asia Pacific](/docs/network/latency/regions/europe-to-asia-pacific) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Hong Kong to London latency and RTT](/docs/network/latency/pairs/hkg-lon-rtt): 158.7 ms **Fastest routes departing London (LON)** * [London to Amsterdam latency and RTT](/docs/network/latency/pairs/lon-ams-rtt): 5.2 ms * [London to Paris latency and RTT](/docs/network/latency/pairs/lon-par-rtt): 6.4 ms * [London to Frankfurt latency and RTT](/docs/network/latency/pairs/lon-fra-rtt): 13.6 ms * [London to Berlin latency and RTT](/docs/network/latency/pairs/lon-ber-rtt): 16.2 ms * [London to Marseille latency and RTT](/docs/network/latency/pairs/lon-mrs-rtt): 17.9 ms **Fastest routes arriving at Hong Kong (HKG)** * [Taipei to Hong Kong latency and RTT](/docs/network/latency/pairs/tpe-hkg-rtt): 15.5 ms * [Singapore to Hong Kong latency and RTT](/docs/network/latency/pairs/sin-hkg-rtt): 30.8 ms * [Tokyo to Hong Kong latency and RTT](/docs/network/latency/pairs/tyo-hkg-rtt): 44.8 ms * [Moscow to Hong Kong latency and RTT](/docs/network/latency/pairs/mow-hkg-rtt): 118.6 ms * [Seattle to Hong Kong latency and RTT](/docs/network/latency/pairs/sea-hkg-rtt): 130.4 ms **Same corridor (Europe → Asia Pacific)** * [Moscow to Taipei latency and RTT](/docs/network/latency/pairs/mow-tpe-rtt): 132.1 ms * [Marseille to Singapore latency and RTT](/docs/network/latency/pairs/mrs-sin-rtt): 139.6 ms * [Berlin to Hong Kong latency and RTT](/docs/network/latency/pairs/ber-hkg-rtt): 144.9 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/lon-hkg.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, London → Hong Kong round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/lon-hkg-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # London, UK → Ashburn, USA RTT 🇬🇧 **London, UK (LON)** → 🇺🇸 **Ashburn, USA (IAD)** Measurement round: `2026-08-16T04:07:28Z`. London to Ashburn averaged 71 ms across 50 replies. The observed range was 68.31 to 80.67 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 2.3 ms, or 3.2% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **71 ms** | | Jitter | **1.76 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **58.1 ms** | | Fiber Efficiency | **81.8%** | | Latency Tier | Excellent | | Source Region | [Europe](/docs/network/latency/regions/europe) | | Destination Region | [North America](/docs/network/latency/regions/north-america) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [London](https://www.geonames.org/2643743) (51.50853, -0.12574) and [Ashburn](https://www.geonames.org/4744870) (39.04372, -77.48749). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **5,932.8 km** | | Vacuum RTT floor | **39.58 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **58.1 ms** | | Low-latency fiber material floor | **57.87 ms** | | Engineering floor (5% path allowance) | **61.01 ms** | | Research 1.33× mapped-fiber reference | **77.27 ms** | | Estimated unamplified path loss | **1245.9 dB** | | Transparent optical spans / inline amplifiers | **78 / 77** | | Published RTT inflation over fiber floor | **1.22×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------ | | Minimum RTT | **68.31 ms** | | Average RTT | **71 ms** | | Maximum RTT | **80.67 ms** | | Standard deviation | **2.3 ms** | | Stdev / average | **3.2%** | ## Route Context * Departure PoP: [London (LON)](/docs/network/latency/lon-london) * Destination PoP: [Ashburn (IAD)](/docs/network/latency/iad-ashburn) * Region overview: [Europe](/docs/network/latency/regions/europe) → [North America](/docs/network/latency/regions/north-america) * Corridor overview: [Europe to North America](/docs/network/latency/regions/europe-to-north-america) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Ashburn to London latency and RTT](/docs/network/latency/pairs/iad-lon-rtt): 70.3 ms **Fastest routes departing London (LON)** * [London to Amsterdam latency and RTT](/docs/network/latency/pairs/lon-ams-rtt): 5.2 ms * [London to Paris latency and RTT](/docs/network/latency/pairs/lon-par-rtt): 6.4 ms * [London to Frankfurt latency and RTT](/docs/network/latency/pairs/lon-fra-rtt): 13.6 ms * [London to Berlin latency and RTT](/docs/network/latency/pairs/lon-ber-rtt): 16.2 ms * [London to Marseille latency and RTT](/docs/network/latency/pairs/lon-mrs-rtt): 17.9 ms **Fastest routes arriving at Ashburn (IAD)** * [New York to Ashburn latency and RTT](/docs/network/latency/pairs/nyc-iad-rtt): 6 ms * [Miami to Ashburn latency and RTT](/docs/network/latency/pairs/mia-iad-rtt): 27.3 ms * [Los Angeles to Ashburn latency and RTT](/docs/network/latency/pairs/lax-iad-rtt): 60.4 ms * [Seattle to Ashburn latency and RTT](/docs/network/latency/pairs/sea-iad-rtt): 61.5 ms * [Paris to Ashburn latency and RTT](/docs/network/latency/pairs/par-iad-rtt): 75.5 ms **Same corridor (Europe → North America)** * [London to New York latency and RTT](/docs/network/latency/pairs/lon-nyc-rtt): 63.8 ms * [Amsterdam to New York latency and RTT](/docs/network/latency/pairs/ams-nyc-rtt): 68.6 ms * [Paris to New York latency and RTT](/docs/network/latency/pairs/par-nyc-rtt): 68.9 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/lon-iad.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, London → Ashburn round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/lon-iad-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # London, UK → Johannesburg, South Africa RTT 🇬🇧 **London, UK (LON)** → 🇿🇦 **Johannesburg, South Africa (JNB)** Measurement round: `2026-08-16T04:07:28Z`. London to Johannesburg averaged 179.6 ms, compared with a 88.52 ms city-centre fiber reference. The difference cannot be assigned to a particular cable or carrier from RTT alone. Standard deviation was 5.96 ms, or 3.3% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------- | | RTT | **179.6 ms** | | Jitter | **4.3 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **88.52 ms** | | Fiber Efficiency | **49.3%** | | Latency Tier | Fair | | Source Region | [Europe](/docs/network/latency/regions/europe) | | Destination Region | [Africa](/docs/network/latency/regions/africa) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [London](https://www.geonames.org/2643743) (51.50853, -0.12574) and [Johannesburg](https://www.geonames.org/993800) (-26.20227, 28.04363). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **9,039.3 km** | | Vacuum RTT floor | **60.3 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **88.52 ms** | | Low-latency fiber material floor | **88.16 ms** | | Engineering floor (5% path allowance) | **92.96 ms** | | Research 1.33× mapped-fiber reference | **117.73 ms** | | Estimated unamplified path loss | **1898.3 dB** | | Transparent optical spans / inline amplifiers | **119 / 118** | | Published RTT inflation over fiber floor | **2.03×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **171.6 ms** | | Average RTT | **179.6 ms** | | Maximum RTT | **203.66 ms** | | Standard deviation | **5.96 ms** | | Stdev / average | **3.3%** | ## Route Context * Departure PoP: [London (LON)](/docs/network/latency/lon-london) * Destination PoP: [Johannesburg (JNB)](/docs/network/latency/jnb-johannesburg) * Region overview: [Europe](/docs/network/latency/regions/europe) → [Africa](/docs/network/latency/regions/africa) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Johannesburg to London latency and RTT](/docs/network/latency/pairs/jnb-lon-rtt): 158 ms **Fastest routes departing London (LON)** * [London to Amsterdam latency and RTT](/docs/network/latency/pairs/lon-ams-rtt): 5.2 ms * [London to Paris latency and RTT](/docs/network/latency/pairs/lon-par-rtt): 6.4 ms * [London to Frankfurt latency and RTT](/docs/network/latency/pairs/lon-fra-rtt): 13.6 ms * [London to Berlin latency and RTT](/docs/network/latency/pairs/lon-ber-rtt): 16.2 ms * [London to Marseille latency and RTT](/docs/network/latency/pairs/lon-mrs-rtt): 17.9 ms **Fastest routes arriving at Johannesburg (JNB)** * [Amsterdam to Johannesburg latency and RTT](/docs/network/latency/pairs/ams-jnb-rtt): 171.9 ms * [Berlin to Johannesburg latency and RTT](/docs/network/latency/pairs/ber-jnb-rtt): 172.1 ms * [Paris to Johannesburg latency and RTT](/docs/network/latency/pairs/par-jnb-rtt): 180.9 ms * [Frankfurt to Johannesburg latency and RTT](/docs/network/latency/pairs/fra-jnb-rtt): 188.4 ms * [Marseille to Johannesburg latency and RTT](/docs/network/latency/pairs/mrs-jnb-rtt): 198.5 ms **Same corridor (Europe → Africa)** * [Moscow to Johannesburg latency and RTT](/docs/network/latency/pairs/mow-jnb-rtt): 201.9 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/lon-jnb.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, London → Johannesburg round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/lon-jnb-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # London, UK → Los Angeles, USA RTT 🇬🇧 **London, UK (LON)** → 🇺🇸 **Los Angeles, USA (LAX)** Measurement round: `2026-08-16T04:07:28Z`. London to Los Angeles averaged 127.5 ms across 50 replies. The observed range was 123.84 to 138.75 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 3.03 ms, or 2.4% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **127.5 ms** | | Jitter | **2.52 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **85.95 ms** | | Fiber Efficiency | **67.4%** | | Latency Tier | Good | | Source Region | [Europe](/docs/network/latency/regions/europe) | | Destination Region | [North America](/docs/network/latency/regions/north-america) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [London](https://www.geonames.org/2643743) (51.50853, -0.12574) and [Los Angeles](https://www.geonames.org/5368361) (34.05223, -118.24368). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **8,776.7 km** | | Vacuum RTT floor | **58.55 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **85.95 ms** | | Low-latency fiber material floor | **85.6 ms** | | Engineering floor (5% path allowance) | **90.26 ms** | | Research 1.33× mapped-fiber reference | **114.31 ms** | | Estimated unamplified path loss | **1843.1 dB** | | Transparent optical spans / inline amplifiers | **116 / 115** | | Published RTT inflation over fiber floor | **1.48×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **123.84 ms** | | Average RTT | **127.5 ms** | | Maximum RTT | **138.75 ms** | | Standard deviation | **3.03 ms** | | Stdev / average | **2.4%** | ## Route Context * Departure PoP: [London (LON)](/docs/network/latency/lon-london) * Destination PoP: [Los Angeles (LAX)](/docs/network/latency/lax-los-angeles) * Region overview: [Europe](/docs/network/latency/regions/europe) → [North America](/docs/network/latency/regions/north-america) * Corridor overview: [Europe to North America](/docs/network/latency/regions/europe-to-north-america) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Los Angeles to London latency and RTT](/docs/network/latency/pairs/lax-lon-rtt): 129.2 ms **Fastest routes departing London (LON)** * [London to Amsterdam latency and RTT](/docs/network/latency/pairs/lon-ams-rtt): 5.2 ms * [London to Paris latency and RTT](/docs/network/latency/pairs/lon-par-rtt): 6.4 ms * [London to Frankfurt latency and RTT](/docs/network/latency/pairs/lon-fra-rtt): 13.6 ms * [London to Berlin latency and RTT](/docs/network/latency/pairs/lon-ber-rtt): 16.2 ms * [London to Marseille latency and RTT](/docs/network/latency/pairs/lon-mrs-rtt): 17.9 ms **Fastest routes arriving at Los Angeles (LAX)** * [Seattle to Los Angeles latency and RTT](/docs/network/latency/pairs/sea-lax-rtt): 25.9 ms * [Miami to Los Angeles latency and RTT](/docs/network/latency/pairs/mia-lax-rtt): 56.8 ms * [New York to Los Angeles latency and RTT](/docs/network/latency/pairs/nyc-lax-rtt): 58.9 ms * [Ashburn to Los Angeles latency and RTT](/docs/network/latency/pairs/iad-lax-rtt): 60.7 ms * [Tokyo to Los Angeles latency and RTT](/docs/network/latency/pairs/tyo-lax-rtt): 101.2 ms **Same corridor (Europe → North America)** * [London to New York latency and RTT](/docs/network/latency/pairs/lon-nyc-rtt): 63.8 ms * [Amsterdam to New York latency and RTT](/docs/network/latency/pairs/ams-nyc-rtt): 68.6 ms * [Paris to New York latency and RTT](/docs/network/latency/pairs/par-nyc-rtt): 68.9 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/lon-lax.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, London → Los Angeles round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/lon-lax-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # London, UK → Melbourne, Australia RTT 🇬🇧 **London, UK (LON)** → 🇦🇺 **Melbourne, Australia (MEL)** Measurement round: `2026-08-16T04:07:28Z`. London to Melbourne averaged 249.6 ms across 50 replies. The observed range was 242.04 to 269.23 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 5.89 ms, or 2.4% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------------------- | | RTT | **249.6 ms** | | Jitter | **4.83 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **165.48 ms** | | Fiber Efficiency | **66.3%** | | Latency Tier | Fair | | Source Region | [Europe](/docs/network/latency/regions/europe) | | Destination Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [London](https://www.geonames.org/2643743) (51.50853, -0.12574) and [Melbourne](https://www.geonames.org/2158177) (-37.81400, 144.96332). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | --------------- | | WGS-84 geodesic distance | **16,898.4 km** | | Vacuum RTT floor | **112.73 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **165.48 ms** | | Low-latency fiber material floor | **164.82 ms** | | Engineering floor (5% path allowance) | **173.78 ms** | | Research 1.33× mapped-fiber reference | **220.09 ms** | | Estimated unamplified path loss | **3548.7 dB** | | Transparent optical spans / inline amplifiers | **222 / 221** | | Published RTT inflation over fiber floor | **1.51×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **242.04 ms** | | Average RTT | **249.6 ms** | | Maximum RTT | **269.23 ms** | | Standard deviation | **5.89 ms** | | Stdev / average | **2.4%** | ## Route Context * Departure PoP: [London (LON)](/docs/network/latency/lon-london) * Destination PoP: [Melbourne (MEL)](/docs/network/latency/mel-melbourne) * Region overview: [Europe](/docs/network/latency/regions/europe) → [Asia Pacific](/docs/network/latency/regions/asia-pacific) * Corridor overview: [Europe to Asia Pacific](/docs/network/latency/regions/europe-to-asia-pacific) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Melbourne to London latency and RTT](/docs/network/latency/pairs/mel-lon-rtt): 252 ms **Fastest routes departing London (LON)** * [London to Amsterdam latency and RTT](/docs/network/latency/pairs/lon-ams-rtt): 5.2 ms * [London to Paris latency and RTT](/docs/network/latency/pairs/lon-par-rtt): 6.4 ms * [London to Frankfurt latency and RTT](/docs/network/latency/pairs/lon-fra-rtt): 13.6 ms * [London to Berlin latency and RTT](/docs/network/latency/pairs/lon-ber-rtt): 16.2 ms * [London to Marseille latency and RTT](/docs/network/latency/pairs/lon-mrs-rtt): 17.9 ms **Fastest routes arriving at Melbourne (MEL)** * [Sydney to Melbourne latency and RTT](/docs/network/latency/pairs/syd-mel-rtt): 9.8 ms * [Singapore to Melbourne latency and RTT](/docs/network/latency/pairs/sin-mel-rtt): 88.6 ms * [Tokyo to Melbourne latency and RTT](/docs/network/latency/pairs/tyo-mel-rtt): 113.9 ms * [Hong Kong to Melbourne latency and RTT](/docs/network/latency/pairs/hkg-mel-rtt): 138.4 ms * [Taipei to Melbourne latency and RTT](/docs/network/latency/pairs/tpe-mel-rtt): 142 ms **Same corridor (Europe → Asia Pacific)** * [Moscow to Hong Kong latency and RTT](/docs/network/latency/pairs/mow-hkg-rtt): 118.6 ms * [Moscow to Taipei latency and RTT](/docs/network/latency/pairs/mow-tpe-rtt): 132.1 ms * [Marseille to Singapore latency and RTT](/docs/network/latency/pairs/mrs-sin-rtt): 139.6 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/lon-mel.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, London → Melbourne round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/lon-mel-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # London, UK → Miami, USA RTT 🇬🇧 **London, UK (LON)** → 🇺🇸 **Miami, USA (MIA)** Measurement round: `2026-08-16T04:07:28Z`. London to Miami averaged 101.9 ms across 50 replies. The observed range was 97.83 to 115.14 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 3.85 ms, or 3.8% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **101.9 ms** | | Jitter | **3.08 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **69.9 ms** | | Fiber Efficiency | **68.6%** | | Latency Tier | Good | | Source Region | [Europe](/docs/network/latency/regions/europe) | | Destination Region | [North America](/docs/network/latency/regions/north-america) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [London](https://www.geonames.org/2643743) (51.50853, -0.12574) and [Miami](https://www.geonames.org/4164138) (25.77427, -80.19366). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **7,138.4 km** | | Vacuum RTT floor | **47.62 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **69.9 ms** | | Low-latency fiber material floor | **69.62 ms** | | Engineering floor (5% path allowance) | **73.41 ms** | | Research 1.33× mapped-fiber reference | **92.97 ms** | | Estimated unamplified path loss | **1499.1 dB** | | Transparent optical spans / inline amplifiers | **94 / 93** | | Published RTT inflation over fiber floor | **1.46×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **97.83 ms** | | Average RTT | **101.9 ms** | | Maximum RTT | **115.14 ms** | | Standard deviation | **3.85 ms** | | Stdev / average | **3.8%** | ## Route Context * Departure PoP: [London (LON)](/docs/network/latency/lon-london) * Destination PoP: [Miami (MIA)](/docs/network/latency/mia-miami) * Region overview: [Europe](/docs/network/latency/regions/europe) → [North America](/docs/network/latency/regions/north-america) * Corridor overview: [Europe to North America](/docs/network/latency/regions/europe-to-north-america) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Miami to London latency and RTT](/docs/network/latency/pairs/mia-lon-rtt): 101.8 ms **Fastest routes departing London (LON)** * [London to Amsterdam latency and RTT](/docs/network/latency/pairs/lon-ams-rtt): 5.2 ms * [London to Paris latency and RTT](/docs/network/latency/pairs/lon-par-rtt): 6.4 ms * [London to Frankfurt latency and RTT](/docs/network/latency/pairs/lon-fra-rtt): 13.6 ms * [London to Berlin latency and RTT](/docs/network/latency/pairs/lon-ber-rtt): 16.2 ms * [London to Marseille latency and RTT](/docs/network/latency/pairs/lon-mrs-rtt): 17.9 ms **Fastest routes arriving at Miami (MIA)** * [Ashburn to Miami latency and RTT](/docs/network/latency/pairs/iad-mia-rtt): 27.4 ms * [New York to Miami latency and RTT](/docs/network/latency/pairs/nyc-mia-rtt): 32.9 ms * [Los Angeles to Miami latency and RTT](/docs/network/latency/pairs/lax-mia-rtt): 57.3 ms * [São Paulo to Miami latency and RTT](/docs/network/latency/pairs/gru-mia-rtt): 74 ms * [Seattle to Miami latency and RTT](/docs/network/latency/pairs/sea-mia-rtt): 81.7 ms **Same corridor (Europe → North America)** * [London to New York latency and RTT](/docs/network/latency/pairs/lon-nyc-rtt): 63.8 ms * [Amsterdam to New York latency and RTT](/docs/network/latency/pairs/ams-nyc-rtt): 68.6 ms * [Paris to New York latency and RTT](/docs/network/latency/pairs/par-nyc-rtt): 68.9 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/lon-mia.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, London → Miami round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/lon-mia-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # London, UK → Moscow, Russia RTT 🇬🇧 **London, UK (LON)** → 🇷🇺 **Moscow, Russia (MOW)** Measurement round: `2026-08-16T04:07:28Z`. London to Moscow averaged 43.5 ms across 50 replies. The observed range was 41.81 to 48.52 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 1.38 ms, or 3.2% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------- | | RTT | **43.5 ms** | | Jitter | **1.21 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **24.56 ms** | | Fiber Efficiency | **56.5%** | | Latency Tier | Excellent | | Source Region | [Europe](/docs/network/latency/regions/europe) | | Destination Region | [Europe](/docs/network/latency/regions/europe) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [London](https://www.geonames.org/2643743) (51.50853, -0.12574) and [Moscow](https://www.geonames.org/524901) (55.75204, 37.61781). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **2,508.5 km** | | Vacuum RTT floor | **16.73 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **24.56 ms** | | Low-latency fiber material floor | **24.47 ms** | | Engineering floor (5% path allowance) | **25.8 ms** | | Research 1.33× mapped-fiber reference | **32.67 ms** | | Estimated unamplified path loss | **526.8 dB** | | Transparent optical spans / inline amplifiers | **33 / 32** | | Published RTT inflation over fiber floor | **1.77×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------ | | Minimum RTT | **41.81 ms** | | Average RTT | **43.5 ms** | | Maximum RTT | **48.52 ms** | | Standard deviation | **1.38 ms** | | Stdev / average | **3.2%** | ## Route Context * Departure PoP: [London (LON)](/docs/network/latency/lon-london) * Destination PoP: [Moscow (MOW)](/docs/network/latency/mow-moscow) * Region overview: [Europe](/docs/network/latency/regions/europe) → [Europe](/docs/network/latency/regions/europe) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Moscow to London latency and RTT](/docs/network/latency/pairs/mow-lon-rtt): 42.4 ms **Fastest routes departing London (LON)** * [London to Amsterdam latency and RTT](/docs/network/latency/pairs/lon-ams-rtt): 5.2 ms * [London to Paris latency and RTT](/docs/network/latency/pairs/lon-par-rtt): 6.4 ms * [London to Frankfurt latency and RTT](/docs/network/latency/pairs/lon-fra-rtt): 13.6 ms * [London to Berlin latency and RTT](/docs/network/latency/pairs/lon-ber-rtt): 16.2 ms * [London to Marseille latency and RTT](/docs/network/latency/pairs/lon-mrs-rtt): 17.9 ms **Fastest routes arriving at Moscow (MOW)** * [Berlin to Moscow latency and RTT](/docs/network/latency/pairs/ber-mow-rtt): 28.6 ms * [Frankfurt to Moscow latency and RTT](/docs/network/latency/pairs/fra-mow-rtt): 34.7 ms * [Amsterdam to Moscow latency and RTT](/docs/network/latency/pairs/ams-mow-rtt): 38.1 ms * [Paris to Moscow latency and RTT](/docs/network/latency/pairs/par-mow-rtt): 44.7 ms * [Marseille to Moscow latency and RTT](/docs/network/latency/pairs/mrs-mow-rtt): 49.6 ms **Same corridor (Europe → Europe)** * [Amsterdam to London latency and RTT](/docs/network/latency/pairs/ams-lon-rtt): 5.2 ms * [Amsterdam to Frankfurt latency and RTT](/docs/network/latency/pairs/ams-fra-rtt): 5.9 ms * [Frankfurt to Amsterdam latency and RTT](/docs/network/latency/pairs/fra-ams-rtt): 5.9 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/lon-mow.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, London → Moscow round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/lon-mow-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # London, UK → Marseille, France RTT 🇬🇧 **London, UK (LON)** → 🇫🇷 **Marseille, France (MRS)** Measurement round: `2026-08-16T04:07:28Z`. London to Marseille averaged 17.9 ms across 50 replies. The observed range was 17.02 to 21.69 ms; these figures describe one probe round, not a long-term service target. This is number 5 of 19 outgoing routes by average RTT. That ranking compares different destinations and is not adjusted for distance. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------- | | RTT | **17.9 ms** | | Jitter | **0.62 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **9.82 ms** | | Fiber Efficiency | **54.8%** | | Latency Tier | Ultra-Low | | Source Region | [Europe](/docs/network/latency/regions/europe) | | Destination Region | [Europe](/docs/network/latency/regions/europe) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [London](https://www.geonames.org/2643743) (51.50853, -0.12574) and [Marseille](https://www.geonames.org/2995469) (43.29695, 5.38107). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **1,002.4 km** | | Vacuum RTT floor | **6.69 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **9.82 ms** | | Low-latency fiber material floor | **9.78 ms** | | Engineering floor (5% path allowance) | **10.31 ms** | | Research 1.33× mapped-fiber reference | **13.06 ms** | | Estimated unamplified path loss | **210.5 dB** | | Transparent optical spans / inline amplifiers | **14 / 13** | | Published RTT inflation over fiber floor | **1.82×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------ | | Minimum RTT | **17.02 ms** | | Average RTT | **17.9 ms** | | Maximum RTT | **21.69 ms** | | Standard deviation | **0.84 ms** | | Stdev / average | **4.7%** | ## Route Context * Departure PoP: [London (LON)](/docs/network/latency/lon-london) * Destination PoP: [Marseille (MRS)](/docs/network/latency/mrs-marseille) * Region overview: [Europe](/docs/network/latency/regions/europe) → [Europe](/docs/network/latency/regions/europe) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Marseille to London latency and RTT](/docs/network/latency/pairs/mrs-lon-rtt): 18.9 ms **Fastest routes departing London (LON)** * [London to Amsterdam latency and RTT](/docs/network/latency/pairs/lon-ams-rtt): 5.2 ms * [London to Paris latency and RTT](/docs/network/latency/pairs/lon-par-rtt): 6.4 ms * [London to Frankfurt latency and RTT](/docs/network/latency/pairs/lon-fra-rtt): 13.6 ms * [London to Berlin latency and RTT](/docs/network/latency/pairs/lon-ber-rtt): 16.2 ms * [London to Moscow latency and RTT](/docs/network/latency/pairs/lon-mow-rtt): 43.5 ms **Fastest routes arriving at Marseille (MRS)** * [Paris to Marseille latency and RTT](/docs/network/latency/pairs/par-mrs-rtt): 8.9 ms * [Frankfurt to Marseille latency and RTT](/docs/network/latency/pairs/fra-mrs-rtt): 16.4 ms * [Amsterdam to Marseille latency and RTT](/docs/network/latency/pairs/ams-mrs-rtt): 19.6 ms * [Berlin to Marseille latency and RTT](/docs/network/latency/pairs/ber-mrs-rtt): 20.2 ms * [Moscow to Marseille latency and RTT](/docs/network/latency/pairs/mow-mrs-rtt): 49.9 ms **Same corridor (Europe → Europe)** * [Amsterdam to London latency and RTT](/docs/network/latency/pairs/ams-lon-rtt): 5.2 ms * [Amsterdam to Frankfurt latency and RTT](/docs/network/latency/pairs/ams-fra-rtt): 5.9 ms * [Frankfurt to Amsterdam latency and RTT](/docs/network/latency/pairs/fra-ams-rtt): 5.9 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/lon-mrs.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, London → Marseille round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/lon-mrs-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # London, UK → New York, USA RTT 🇬🇧 **London, UK (LON)** → 🇺🇸 **New York, USA (NYC)** Measurement round: `2026-08-16T04:07:28Z`. London to New York averaged 63.8 ms across 50 replies. The observed range was 60.82 to 74.75 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 2.78 ms, or 4.4% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **63.8 ms** | | Jitter | **2.55 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **54.69 ms** | | Fiber Efficiency | **85.7%** | | Latency Tier | Excellent | | Source Region | [Europe](/docs/network/latency/regions/europe) | | Destination Region | [North America](/docs/network/latency/regions/north-america) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [London](https://www.geonames.org/2643743) (51.50853, -0.12574) and [New York City](https://www.geonames.org/5128581) (40.71427, -74.00597). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **5,585.2 km** | | Vacuum RTT floor | **37.26 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **54.69 ms** | | Low-latency fiber material floor | **54.48 ms** | | Engineering floor (5% path allowance) | **57.44 ms** | | Research 1.33× mapped-fiber reference | **72.74 ms** | | Estimated unamplified path loss | **1172.9 dB** | | Transparent optical spans / inline amplifiers | **74 / 73** | | Published RTT inflation over fiber floor | **1.17×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------ | | Minimum RTT | **60.82 ms** | | Average RTT | **63.8 ms** | | Maximum RTT | **74.75 ms** | | Standard deviation | **2.78 ms** | | Stdev / average | **4.4%** | ## Route Context * Departure PoP: [London (LON)](/docs/network/latency/lon-london) * Destination PoP: [New York (NYC)](/docs/network/latency/nyc-new-york) * Region overview: [Europe](/docs/network/latency/regions/europe) → [North America](/docs/network/latency/regions/north-america) * Corridor overview: [Europe to North America](/docs/network/latency/regions/europe-to-north-america) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [New York to London latency and RTT](/docs/network/latency/pairs/nyc-lon-rtt): 63.5 ms **Fastest routes departing London (LON)** * [London to Amsterdam latency and RTT](/docs/network/latency/pairs/lon-ams-rtt): 5.2 ms * [London to Paris latency and RTT](/docs/network/latency/pairs/lon-par-rtt): 6.4 ms * [London to Frankfurt latency and RTT](/docs/network/latency/pairs/lon-fra-rtt): 13.6 ms * [London to Berlin latency and RTT](/docs/network/latency/pairs/lon-ber-rtt): 16.2 ms * [London to Marseille latency and RTT](/docs/network/latency/pairs/lon-mrs-rtt): 17.9 ms **Fastest routes arriving at New York (NYC)** * [Ashburn to New York latency and RTT](/docs/network/latency/pairs/iad-nyc-rtt): 6 ms * [Miami to New York latency and RTT](/docs/network/latency/pairs/mia-nyc-rtt): 33.3 ms * [Los Angeles to New York latency and RTT](/docs/network/latency/pairs/lax-nyc-rtt): 57.9 ms * [Seattle to New York latency and RTT](/docs/network/latency/pairs/sea-nyc-rtt): 59.5 ms * [Amsterdam to New York latency and RTT](/docs/network/latency/pairs/ams-nyc-rtt): 68.6 ms **Same corridor (Europe → North America)** * [Paris to New York latency and RTT](/docs/network/latency/pairs/par-nyc-rtt): 68.9 ms * [London to Ashburn latency and RTT](/docs/network/latency/pairs/lon-iad-rtt): 71 ms * [Frankfurt to New York latency and RTT](/docs/network/latency/pairs/fra-nyc-rtt): 74.4 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/lon-nyc.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, London → New York round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/lon-nyc-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # London, UK → Paris, France RTT 🇬🇧 **London, UK (LON)** → 🇫🇷 **Paris, France (PAR)** Measurement round: `2026-08-16T04:07:28Z`. London to Paris averaged 6.4 ms across 50 replies. The observed range was 6.23 to 6.9 ms; these figures describe one probe round, not a long-term service target. This is number 2 of 19 outgoing routes by average RTT. That ranking compares different destinations and is not adjusted for distance. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------- | | RTT | **6.4 ms** | | Jitter | **0.11 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **3.37 ms** | | Fiber Efficiency | **52.7%** | | Latency Tier | Ultra-Low | | Source Region | [Europe](/docs/network/latency/regions/europe) | | Destination Region | [Europe](/docs/network/latency/regions/europe) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [London](https://www.geonames.org/2643743) (51.50853, -0.12574) and [Paris](https://www.geonames.org/2988507) (48.85341, 2.34880). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | ------------ | | WGS-84 geodesic distance | **344.1 km** | | Vacuum RTT floor | **2.3 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **3.37 ms** | | Low-latency fiber material floor | **3.36 ms** | | Engineering floor (5% path allowance) | **3.54 ms** | | Research 1.33× mapped-fiber reference | **4.48 ms** | | Estimated unamplified path loss | **72.3 dB** | | Transparent optical spans / inline amplifiers | **5 / 4** | | Published RTT inflation over fiber floor | **1.9×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ----------- | | Minimum RTT | **6.23 ms** | | Average RTT | **6.4 ms** | | Maximum RTT | **6.9 ms** | | Standard deviation | **0.15 ms** | | Stdev / average | **2.3%** | ## Route Context * Departure PoP: [London (LON)](/docs/network/latency/lon-london) * Destination PoP: [Paris (PAR)](/docs/network/latency/par-paris) * Region overview: [Europe](/docs/network/latency/regions/europe) → [Europe](/docs/network/latency/regions/europe) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Paris to London latency and RTT](/docs/network/latency/pairs/par-lon-rtt): 6.4 ms **Fastest routes departing London (LON)** * [London to Amsterdam latency and RTT](/docs/network/latency/pairs/lon-ams-rtt): 5.2 ms * [London to Frankfurt latency and RTT](/docs/network/latency/pairs/lon-fra-rtt): 13.6 ms * [London to Berlin latency and RTT](/docs/network/latency/pairs/lon-ber-rtt): 16.2 ms * [London to Marseille latency and RTT](/docs/network/latency/pairs/lon-mrs-rtt): 17.9 ms * [London to Moscow latency and RTT](/docs/network/latency/pairs/lon-mow-rtt): 43.5 ms **Fastest routes arriving at Paris (PAR)** * [Amsterdam to Paris latency and RTT](/docs/network/latency/pairs/ams-par-rtt): 7 ms * [Frankfurt to Paris latency and RTT](/docs/network/latency/pairs/fra-par-rtt): 7.6 ms * [Marseille to Paris latency and RTT](/docs/network/latency/pairs/mrs-par-rtt): 8.9 ms * [Berlin to Paris latency and RTT](/docs/network/latency/pairs/ber-par-rtt): 15.6 ms * [Moscow to Paris latency and RTT](/docs/network/latency/pairs/mow-par-rtt): 44 ms **Same corridor (Europe → Europe)** * [Amsterdam to London latency and RTT](/docs/network/latency/pairs/ams-lon-rtt): 5.2 ms * [Amsterdam to Frankfurt latency and RTT](/docs/network/latency/pairs/ams-fra-rtt): 5.9 ms * [Frankfurt to Amsterdam latency and RTT](/docs/network/latency/pairs/fra-ams-rtt): 5.9 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/lon-par.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, London → Paris round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/lon-par-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # London, UK → Seattle, USA RTT 🇬🇧 **London, UK (LON)** → 🇺🇸 **Seattle, USA (SEA)** Measurement round: `2026-08-16T04:07:28Z`. London to Seattle averaged 123.2 ms across 50 replies. The observed range was 121.09 to 129.68 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 1.95 ms, or 1.6% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **123.2 ms** | | Jitter | **1.63 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **75.63 ms** | | Fiber Efficiency | **61.4%** | | Latency Tier | Good | | Source Region | [Europe](/docs/network/latency/regions/europe) | | Destination Region | [North America](/docs/network/latency/regions/north-america) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [London](https://www.geonames.org/2643743) (51.50853, -0.12574) and [Seattle](https://www.geonames.org/5809844) (47.60621, -122.33207). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **7,722.7 km** | | Vacuum RTT floor | **51.52 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **75.63 ms** | | Low-latency fiber material floor | **75.32 ms** | | Engineering floor (5% path allowance) | **79.42 ms** | | Research 1.33× mapped-fiber reference | **100.58 ms** | | Estimated unamplified path loss | **1621.8 dB** | | Transparent optical spans / inline amplifiers | **102 / 101** | | Published RTT inflation over fiber floor | **1.63×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **121.09 ms** | | Average RTT | **123.2 ms** | | Maximum RTT | **129.68 ms** | | Standard deviation | **1.95 ms** | | Stdev / average | **1.6%** | ## Route Context * Departure PoP: [London (LON)](/docs/network/latency/lon-london) * Destination PoP: [Seattle (SEA)](/docs/network/latency/sea-seattle) * Region overview: [Europe](/docs/network/latency/regions/europe) → [North America](/docs/network/latency/regions/north-america) * Corridor overview: [Europe to North America](/docs/network/latency/regions/europe-to-north-america) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Seattle to London latency and RTT](/docs/network/latency/pairs/sea-lon-rtt): 124.1 ms **Fastest routes departing London (LON)** * [London to Amsterdam latency and RTT](/docs/network/latency/pairs/lon-ams-rtt): 5.2 ms * [London to Paris latency and RTT](/docs/network/latency/pairs/lon-par-rtt): 6.4 ms * [London to Frankfurt latency and RTT](/docs/network/latency/pairs/lon-fra-rtt): 13.6 ms * [London to Berlin latency and RTT](/docs/network/latency/pairs/lon-ber-rtt): 16.2 ms * [London to Marseille latency and RTT](/docs/network/latency/pairs/lon-mrs-rtt): 17.9 ms **Fastest routes arriving at Seattle (SEA)** * [Los Angeles to Seattle latency and RTT](/docs/network/latency/pairs/lax-sea-rtt): 26.9 ms * [New York to Seattle latency and RTT](/docs/network/latency/pairs/nyc-sea-rtt): 58.7 ms * [Ashburn to Seattle latency and RTT](/docs/network/latency/pairs/iad-sea-rtt): 62.3 ms * [Miami to Seattle latency and RTT](/docs/network/latency/pairs/mia-sea-rtt): 81.9 ms * [Tokyo to Seattle latency and RTT](/docs/network/latency/pairs/tyo-sea-rtt): 84.8 ms **Same corridor (Europe → North America)** * [London to New York latency and RTT](/docs/network/latency/pairs/lon-nyc-rtt): 63.8 ms * [Amsterdam to New York latency and RTT](/docs/network/latency/pairs/ams-nyc-rtt): 68.6 ms * [Paris to New York latency and RTT](/docs/network/latency/pairs/par-nyc-rtt): 68.9 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/lon-sea.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, London → Seattle round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/lon-sea-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # London, UK → Singapore RTT 🇬🇧 **London, UK (LON)** → 🇸🇬 **Singapore (SIN)** Measurement round: `2026-08-16T04:07:28Z`. London to Singapore averaged 156.5 ms across 50 replies. The observed range was 151.92 to 167.3 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 3.54 ms, or 2.3% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------------------- | | RTT | **156.5 ms** | | Jitter | **3.13 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **106.35 ms** | | Fiber Efficiency | **68%** | | Latency Tier | Fair | | Source Region | [Europe](/docs/network/latency/regions/europe) | | Destination Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [London](https://www.geonames.org/2643743) (51.50853, -0.12574) and [Singapore](https://www.geonames.org/1880252) (1.28967, 103.85007). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | --------------- | | WGS-84 geodesic distance | **10,859.7 km** | | Vacuum RTT floor | **72.45 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **106.35 ms** | | Low-latency fiber material floor | **105.92 ms** | | Engineering floor (5% path allowance) | **111.68 ms** | | Research 1.33× mapped-fiber reference | **141.44 ms** | | Estimated unamplified path loss | **2280.5 dB** | | Transparent optical spans / inline amplifiers | **143 / 142** | | Published RTT inflation over fiber floor | **1.47×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **151.92 ms** | | Average RTT | **156.5 ms** | | Maximum RTT | **167.3 ms** | | Standard deviation | **3.54 ms** | | Stdev / average | **2.3%** | ## Route Context * Departure PoP: [London (LON)](/docs/network/latency/lon-london) * Destination PoP: [Singapore (SIN)](/docs/network/latency/sin-singapore) * Region overview: [Europe](/docs/network/latency/regions/europe) → [Asia Pacific](/docs/network/latency/regions/asia-pacific) * Corridor overview: [Europe to Asia Pacific](/docs/network/latency/regions/europe-to-asia-pacific) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Singapore to London latency and RTT](/docs/network/latency/pairs/sin-lon-rtt): 155.4 ms **Fastest routes departing London (LON)** * [London to Amsterdam latency and RTT](/docs/network/latency/pairs/lon-ams-rtt): 5.2 ms * [London to Paris latency and RTT](/docs/network/latency/pairs/lon-par-rtt): 6.4 ms * [London to Frankfurt latency and RTT](/docs/network/latency/pairs/lon-fra-rtt): 13.6 ms * [London to Berlin latency and RTT](/docs/network/latency/pairs/lon-ber-rtt): 16.2 ms * [London to Marseille latency and RTT](/docs/network/latency/pairs/lon-mrs-rtt): 17.9 ms **Fastest routes arriving at Singapore (SIN)** * [Hong Kong to Singapore latency and RTT](/docs/network/latency/pairs/hkg-sin-rtt): 31.5 ms * [Taipei to Singapore latency and RTT](/docs/network/latency/pairs/tpe-sin-rtt): 44.3 ms * [Tokyo to Singapore latency and RTT](/docs/network/latency/pairs/tyo-sin-rtt): 69.4 ms * [Melbourne to Singapore latency and RTT](/docs/network/latency/pairs/mel-sin-rtt): 88.4 ms * [Sydney to Singapore latency and RTT](/docs/network/latency/pairs/syd-sin-rtt): 94.5 ms **Same corridor (Europe → Asia Pacific)** * [Moscow to Hong Kong latency and RTT](/docs/network/latency/pairs/mow-hkg-rtt): 118.6 ms * [Moscow to Taipei latency and RTT](/docs/network/latency/pairs/mow-tpe-rtt): 132.1 ms * [Marseille to Singapore latency and RTT](/docs/network/latency/pairs/mrs-sin-rtt): 139.6 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/lon-sin.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, London → Singapore round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/lon-sin-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # London, UK → Sydney, Australia RTT 🇬🇧 **London, UK (LON)** → 🇦🇺 **Sydney, Australia (SYD)** Measurement round: `2026-08-16T04:07:28Z`. London to Sydney averaged 258.1 ms across 50 replies. The observed range was 249.91 to 280.95 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 6.66 ms, or 2.6% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------------------- | | RTT | **258.1 ms** | | Jitter | **5.27 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **166.37 ms** | | Fiber Efficiency | **64.5%** | | Latency Tier | High | | Source Region | [Europe](/docs/network/latency/regions/europe) | | Destination Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [London](https://www.geonames.org/2643743) (51.50853, -0.12574) and [Sydney](https://www.geonames.org/2147714) (-33.86785, 151.20732). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | --------------- | | WGS-84 geodesic distance | **16,988.9 km** | | Vacuum RTT floor | **113.34 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **166.37 ms** | | Low-latency fiber material floor | **165.7 ms** | | Engineering floor (5% path allowance) | **174.71 ms** | | Research 1.33× mapped-fiber reference | **221.27 ms** | | Estimated unamplified path loss | **3567.7 dB** | | Transparent optical spans / inline amplifiers | **223 / 222** | | Published RTT inflation over fiber floor | **1.55×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **249.91 ms** | | Average RTT | **258.1 ms** | | Maximum RTT | **280.95 ms** | | Standard deviation | **6.66 ms** | | Stdev / average | **2.6%** | ## Route Context * Departure PoP: [London (LON)](/docs/network/latency/lon-london) * Destination PoP: [Sydney (SYD)](/docs/network/latency/syd-sydney) * Region overview: [Europe](/docs/network/latency/regions/europe) → [Asia Pacific](/docs/network/latency/regions/asia-pacific) * Corridor overview: [Europe to Asia Pacific](/docs/network/latency/regions/europe-to-asia-pacific) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Sydney to London latency and RTT](/docs/network/latency/pairs/syd-lon-rtt): 256.5 ms **Fastest routes departing London (LON)** * [London to Amsterdam latency and RTT](/docs/network/latency/pairs/lon-ams-rtt): 5.2 ms * [London to Paris latency and RTT](/docs/network/latency/pairs/lon-par-rtt): 6.4 ms * [London to Frankfurt latency and RTT](/docs/network/latency/pairs/lon-fra-rtt): 13.6 ms * [London to Berlin latency and RTT](/docs/network/latency/pairs/lon-ber-rtt): 16.2 ms * [London to Marseille latency and RTT](/docs/network/latency/pairs/lon-mrs-rtt): 17.9 ms **Fastest routes arriving at Sydney (SYD)** * [Melbourne to Sydney latency and RTT](/docs/network/latency/pairs/mel-syd-rtt): 9.8 ms * [Singapore to Sydney latency and RTT](/docs/network/latency/pairs/sin-syd-rtt): 94.5 ms * [Tokyo to Sydney latency and RTT](/docs/network/latency/pairs/tyo-syd-rtt): 101.9 ms * [Taipei to Sydney latency and RTT](/docs/network/latency/pairs/tpe-syd-rtt): 133.2 ms * [Hong Kong to Sydney latency and RTT](/docs/network/latency/pairs/hkg-syd-rtt): 137.2 ms **Same corridor (Europe → Asia Pacific)** * [Moscow to Hong Kong latency and RTT](/docs/network/latency/pairs/mow-hkg-rtt): 118.6 ms * [Moscow to Taipei latency and RTT](/docs/network/latency/pairs/mow-tpe-rtt): 132.1 ms * [Marseille to Singapore latency and RTT](/docs/network/latency/pairs/mrs-sin-rtt): 139.6 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/lon-syd.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, London → Sydney round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/lon-syd-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # London, UK → Taipei, Taiwan RTT 🇬🇧 **London, UK (LON)** → 🇹🇼 **Taipei, Taiwan (TPE)** Measurement round: `2026-08-16T04:07:28Z`. London to Taipei averaged 173.8 ms across 50 replies. The observed range was 166.08 to 201.98 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 7.4 ms, or 4.3% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------------------- | | RTT | **173.8 ms** | | Jitter | **5.49 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **95.97 ms** | | Fiber Efficiency | **55.2%** | | Latency Tier | Fair | | Source Region | [Europe](/docs/network/latency/regions/europe) | | Destination Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [London](https://www.geonames.org/2643743) (51.50853, -0.12574) and [Taipei](https://www.geonames.org/1668341) (25.05306, 121.52639). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **9,799.7 km** | | Vacuum RTT floor | **65.38 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **95.97 ms** | | Low-latency fiber material floor | **95.58 ms** | | Engineering floor (5% path allowance) | **100.78 ms** | | Research 1.33× mapped-fiber reference | **127.63 ms** | | Estimated unamplified path loss | **2057.9 dB** | | Transparent optical spans / inline amplifiers | **129 / 128** | | Published RTT inflation over fiber floor | **1.81×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **166.08 ms** | | Average RTT | **173.8 ms** | | Maximum RTT | **201.98 ms** | | Standard deviation | **7.4 ms** | | Stdev / average | **4.3%** | ## Route Context * Departure PoP: [London (LON)](/docs/network/latency/lon-london) * Destination PoP: [Taipei (TPE)](/docs/network/latency/tpe-taipei) * Region overview: [Europe](/docs/network/latency/regions/europe) → [Asia Pacific](/docs/network/latency/regions/asia-pacific) * Corridor overview: [Europe to Asia Pacific](/docs/network/latency/regions/europe-to-asia-pacific) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Taipei to London latency and RTT](/docs/network/latency/pairs/tpe-lon-rtt): 173 ms **Fastest routes departing London (LON)** * [London to Amsterdam latency and RTT](/docs/network/latency/pairs/lon-ams-rtt): 5.2 ms * [London to Paris latency and RTT](/docs/network/latency/pairs/lon-par-rtt): 6.4 ms * [London to Frankfurt latency and RTT](/docs/network/latency/pairs/lon-fra-rtt): 13.6 ms * [London to Berlin latency and RTT](/docs/network/latency/pairs/lon-ber-rtt): 16.2 ms * [London to Marseille latency and RTT](/docs/network/latency/pairs/lon-mrs-rtt): 17.9 ms **Fastest routes arriving at Taipei (TPE)** * [Hong Kong to Taipei latency and RTT](/docs/network/latency/pairs/hkg-tpe-rtt): 14.7 ms * [Tokyo to Taipei latency and RTT](/docs/network/latency/pairs/tyo-tpe-rtt): 32.3 ms * [Singapore to Taipei latency and RTT](/docs/network/latency/pairs/sin-tpe-rtt): 45.8 ms * [Seattle to Taipei latency and RTT](/docs/network/latency/pairs/sea-tpe-rtt): 114.7 ms * [Sydney to Taipei latency and RTT](/docs/network/latency/pairs/syd-tpe-rtt): 131.7 ms **Same corridor (Europe → Asia Pacific)** * [Moscow to Hong Kong latency and RTT](/docs/network/latency/pairs/mow-hkg-rtt): 118.6 ms * [Moscow to Taipei latency and RTT](/docs/network/latency/pairs/mow-tpe-rtt): 132.1 ms * [Marseille to Singapore latency and RTT](/docs/network/latency/pairs/mrs-sin-rtt): 139.6 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/lon-tpe.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, London → Taipei round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/lon-tpe-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # London, UK → Tokyo, Japan RTT 🇬🇧 **London, UK (LON)** → 🇯🇵 **Tokyo, Japan (TYO)** Measurement round: `2026-08-16T04:07:28Z`. London to Tokyo averaged 202.9 ms, compared with a 93.84 ms city-centre fiber reference. The difference cannot be assigned to a particular cable or carrier from RTT alone. Standard deviation was 4.43 ms, or 2.2% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------------------- | | RTT | **202.9 ms** | | Jitter | **3.56 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **93.84 ms** | | Fiber Efficiency | **46.2%** | | Latency Tier | Fair | | Source Region | [Europe](/docs/network/latency/regions/europe) | | Destination Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [London](https://www.geonames.org/2643743) (51.50853, -0.12574) and [Tokyo](https://www.geonames.org/1850147) (35.68950, 139.69171). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **9,582.1 km** | | Vacuum RTT floor | **63.93 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **93.84 ms** | | Low-latency fiber material floor | **93.46 ms** | | Engineering floor (5% path allowance) | **98.54 ms** | | Research 1.33× mapped-fiber reference | **124.8 ms** | | Estimated unamplified path loss | **2012.2 dB** | | Transparent optical spans / inline amplifiers | **126 / 125** | | Published RTT inflation over fiber floor | **2.16×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **197.73 ms** | | Average RTT | **202.9 ms** | | Maximum RTT | **219 ms** | | Standard deviation | **4.43 ms** | | Stdev / average | **2.2%** | ## Route Context * Departure PoP: [London (LON)](/docs/network/latency/lon-london) * Destination PoP: [Tokyo (TYO)](/docs/network/latency/tyo-tokyo) * Region overview: [Europe](/docs/network/latency/regions/europe) → [Asia Pacific](/docs/network/latency/regions/asia-pacific) * Corridor overview: [Europe to Asia Pacific](/docs/network/latency/regions/europe-to-asia-pacific) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Tokyo to London latency and RTT](/docs/network/latency/pairs/tyo-lon-rtt): 202.7 ms **Fastest routes departing London (LON)** * [London to Amsterdam latency and RTT](/docs/network/latency/pairs/lon-ams-rtt): 5.2 ms * [London to Paris latency and RTT](/docs/network/latency/pairs/lon-par-rtt): 6.4 ms * [London to Frankfurt latency and RTT](/docs/network/latency/pairs/lon-fra-rtt): 13.6 ms * [London to Berlin latency and RTT](/docs/network/latency/pairs/lon-ber-rtt): 16.2 ms * [London to Marseille latency and RTT](/docs/network/latency/pairs/lon-mrs-rtt): 17.9 ms **Fastest routes arriving at Tokyo (TYO)** * [Taipei to Tokyo latency and RTT](/docs/network/latency/pairs/tpe-tyo-rtt): 31.9 ms * [Hong Kong to Tokyo latency and RTT](/docs/network/latency/pairs/hkg-tyo-rtt): 44.9 ms * [Singapore to Tokyo latency and RTT](/docs/network/latency/pairs/sin-tyo-rtt): 68.9 ms * [Seattle to Tokyo latency and RTT](/docs/network/latency/pairs/sea-tyo-rtt): 85 ms * [Los Angeles to Tokyo latency and RTT](/docs/network/latency/pairs/lax-tyo-rtt): 101.2 ms **Same corridor (Europe → Asia Pacific)** * [Moscow to Hong Kong latency and RTT](/docs/network/latency/pairs/mow-hkg-rtt): 118.6 ms * [Moscow to Taipei latency and RTT](/docs/network/latency/pairs/mow-tpe-rtt): 132.1 ms * [Marseille to Singapore latency and RTT](/docs/network/latency/pairs/mrs-sin-rtt): 139.6 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/lon-tyo.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, London → Tokyo round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/lon-tyo-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Melbourne, Australia → Amsterdam, Netherlands RTT 🇦🇺 **Melbourne, Australia (MEL)** → 🇳🇱 **Amsterdam, Netherlands (AMS)** Measurement round: `2026-08-16T04:07:28Z`. Melbourne to Amsterdam averaged 252.8 ms across 50 replies. The observed range was 241.73 to 276.97 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 7.18 ms, or 2.8% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------------------- | | RTT | **252.8 ms** | | Jitter | **6.4 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **161.98 ms** | | Fiber Efficiency | **64.1%** | | Latency Tier | High | | Source Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | | Destination Region | [Europe](/docs/network/latency/regions/europe) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Melbourne](https://www.geonames.org/2158177) (-37.81400, 144.96332) and [Amsterdam](https://www.geonames.org/2759794) (52.37403, 4.88969). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | --------------- | | WGS-84 geodesic distance | **16,540.5 km** | | Vacuum RTT floor | **110.35 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **161.98 ms** | | Low-latency fiber material floor | **161.33 ms** | | Engineering floor (5% path allowance) | **170.1 ms** | | Research 1.33× mapped-fiber reference | **215.43 ms** | | Estimated unamplified path loss | **3473.5 dB** | | Transparent optical spans / inline amplifiers | **218 / 217** | | Published RTT inflation over fiber floor | **1.56×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **241.73 ms** | | Average RTT | **252.8 ms** | | Maximum RTT | **276.97 ms** | | Standard deviation | **7.18 ms** | | Stdev / average | **2.8%** | ## Route Context * Departure PoP: [Melbourne (MEL)](/docs/network/latency/mel-melbourne) * Destination PoP: [Amsterdam (AMS)](/docs/network/latency/ams-amsterdam) * Region overview: [Asia Pacific](/docs/network/latency/regions/asia-pacific) → [Europe](/docs/network/latency/regions/europe) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Amsterdam to Melbourne latency and RTT](/docs/network/latency/pairs/ams-mel-rtt): 252.5 ms **Fastest routes departing Melbourne (MEL)** * [Melbourne to Sydney latency and RTT](/docs/network/latency/pairs/mel-syd-rtt): 9.8 ms * [Melbourne to Singapore latency and RTT](/docs/network/latency/pairs/mel-sin-rtt): 88.4 ms * [Melbourne to Tokyo latency and RTT](/docs/network/latency/pairs/mel-tyo-rtt): 112 ms * [Melbourne to Hong Kong latency and RTT](/docs/network/latency/pairs/mel-hkg-rtt): 138.5 ms * [Melbourne to Taipei latency and RTT](/docs/network/latency/pairs/mel-tpe-rtt): 143.2 ms **Fastest routes arriving at Amsterdam (AMS)** * [London to Amsterdam latency and RTT](/docs/network/latency/pairs/lon-ams-rtt): 5.2 ms * [Frankfurt to Amsterdam latency and RTT](/docs/network/latency/pairs/fra-ams-rtt): 5.9 ms * [Paris to Amsterdam latency and RTT](/docs/network/latency/pairs/par-ams-rtt): 7 ms * [Berlin to Amsterdam latency and RTT](/docs/network/latency/pairs/ber-ams-rtt): 7.2 ms * [Marseille to Amsterdam latency and RTT](/docs/network/latency/pairs/mrs-ams-rtt): 20.3 ms **Same corridor (Asia Pacific → Europe)** * [Hong Kong to Moscow latency and RTT](/docs/network/latency/pairs/hkg-mow-rtt): 118.2 ms * [Taipei to Moscow latency and RTT](/docs/network/latency/pairs/tpe-mow-rtt): 132.7 ms * [Singapore to Marseille latency and RTT](/docs/network/latency/pairs/sin-mrs-rtt): 140.6 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/mel-ams.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Melbourne → Amsterdam round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/mel-ams-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Melbourne, Australia → Berlin, Germany RTT 🇦🇺 **Melbourne, Australia (MEL)** → 🇩🇪 **Berlin, Germany (BER)** Measurement round: `2026-08-16T04:07:28Z`. Melbourne to Berlin averaged 254.7 ms across 50 replies. The observed range was 239.97 to 301.16 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 11.1 ms, or 4.4% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------------------- | | RTT | **254.7 ms** | | Jitter | **10.6 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **156.34 ms** | | Fiber Efficiency | **61.4%** | | Latency Tier | High | | Source Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | | Destination Region | [Europe](/docs/network/latency/regions/europe) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Melbourne](https://www.geonames.org/2158177) (-37.81400, 144.96332) and [Berlin](https://www.geonames.org/2950159) (52.52437, 13.41053). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | --------------- | | WGS-84 geodesic distance | **15,965.1 km** | | Vacuum RTT floor | **106.51 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **156.34 ms** | | Low-latency fiber material floor | **155.71 ms** | | Engineering floor (5% path allowance) | **164.18 ms** | | Research 1.33× mapped-fiber reference | **207.94 ms** | | Estimated unamplified path loss | **3352.7 dB** | | Transparent optical spans / inline amplifiers | **210 / 209** | | Published RTT inflation over fiber floor | **1.63×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **239.97 ms** | | Average RTT | **254.7 ms** | | Maximum RTT | **301.16 ms** | | Standard deviation | **11.1 ms** | | Stdev / average | **4.4%** | ## Route Context * Departure PoP: [Melbourne (MEL)](/docs/network/latency/mel-melbourne) * Destination PoP: [Berlin (BER)](/docs/network/latency/ber-berlin) * Region overview: [Asia Pacific](/docs/network/latency/regions/asia-pacific) → [Europe](/docs/network/latency/regions/europe) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Berlin to Melbourne latency and RTT](/docs/network/latency/pairs/ber-mel-rtt): 256.2 ms **Fastest routes departing Melbourne (MEL)** * [Melbourne to Sydney latency and RTT](/docs/network/latency/pairs/mel-syd-rtt): 9.8 ms * [Melbourne to Singapore latency and RTT](/docs/network/latency/pairs/mel-sin-rtt): 88.4 ms * [Melbourne to Tokyo latency and RTT](/docs/network/latency/pairs/mel-tyo-rtt): 112 ms * [Melbourne to Hong Kong latency and RTT](/docs/network/latency/pairs/mel-hkg-rtt): 138.5 ms * [Melbourne to Taipei latency and RTT](/docs/network/latency/pairs/mel-tpe-rtt): 143.2 ms **Fastest routes arriving at Berlin (BER)** * [Frankfurt to Berlin latency and RTT](/docs/network/latency/pairs/fra-ber-rtt): 6.1 ms * [Amsterdam to Berlin latency and RTT](/docs/network/latency/pairs/ams-ber-rtt): 7.2 ms * [Paris to Berlin latency and RTT](/docs/network/latency/pairs/par-ber-rtt): 15.5 ms * [London to Berlin latency and RTT](/docs/network/latency/pairs/lon-ber-rtt): 16.2 ms * [Marseille to Berlin latency and RTT](/docs/network/latency/pairs/mrs-ber-rtt): 20.6 ms **Same corridor (Asia Pacific → Europe)** * [Hong Kong to Moscow latency and RTT](/docs/network/latency/pairs/hkg-mow-rtt): 118.2 ms * [Taipei to Moscow latency and RTT](/docs/network/latency/pairs/tpe-mow-rtt): 132.7 ms * [Singapore to Marseille latency and RTT](/docs/network/latency/pairs/sin-mrs-rtt): 140.6 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/mel-ber.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Melbourne → Berlin round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/mel-ber-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Melbourne, Australia → Frankfurt, Germany RTT 🇦🇺 **Melbourne, Australia (MEL)** → 🇩🇪 **Frankfurt, Germany (FRA)** Measurement round: `2026-08-16T04:07:28Z`. Melbourne to Frankfurt averaged 248.6 ms across 50 replies. The observed range was 238.22 to 288.5 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 10.01 ms, or 4.0% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------------------- | | RTT | **248.6 ms** | | Jitter | **8.5 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **159.77 ms** | | Fiber Efficiency | **64.3%** | | Latency Tier | Fair | | Source Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | | Destination Region | [Europe](/docs/network/latency/regions/europe) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Melbourne](https://www.geonames.org/2158177) (-37.81400, 144.96332) and [Frankfurt am Main](https://www.geonames.org/2925533) (50.11552, 8.68417). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | --------------- | | WGS-84 geodesic distance | **16,315.5 km** | | Vacuum RTT floor | **108.85 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **159.77 ms** | | Low-latency fiber material floor | **159.13 ms** | | Engineering floor (5% path allowance) | **167.78 ms** | | Research 1.33× mapped-fiber reference | **212.5 ms** | | Estimated unamplified path loss | **3426.2 dB** | | Transparent optical spans / inline amplifiers | **215 / 214** | | Published RTT inflation over fiber floor | **1.56×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **238.22 ms** | | Average RTT | **248.6 ms** | | Maximum RTT | **288.5 ms** | | Standard deviation | **10.01 ms** | | Stdev / average | **4.0%** | ## Route Context * Departure PoP: [Melbourne (MEL)](/docs/network/latency/mel-melbourne) * Destination PoP: [Frankfurt (FRA)](/docs/network/latency/fra-frankfurt) * Region overview: [Asia Pacific](/docs/network/latency/regions/asia-pacific) → [Europe](/docs/network/latency/regions/europe) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Frankfurt to Melbourne latency and RTT](/docs/network/latency/pairs/fra-mel-rtt): 250.1 ms **Fastest routes departing Melbourne (MEL)** * [Melbourne to Sydney latency and RTT](/docs/network/latency/pairs/mel-syd-rtt): 9.8 ms * [Melbourne to Singapore latency and RTT](/docs/network/latency/pairs/mel-sin-rtt): 88.4 ms * [Melbourne to Tokyo latency and RTT](/docs/network/latency/pairs/mel-tyo-rtt): 112 ms * [Melbourne to Hong Kong latency and RTT](/docs/network/latency/pairs/mel-hkg-rtt): 138.5 ms * [Melbourne to Taipei latency and RTT](/docs/network/latency/pairs/mel-tpe-rtt): 143.2 ms **Fastest routes arriving at Frankfurt (FRA)** * [Amsterdam to Frankfurt latency and RTT](/docs/network/latency/pairs/ams-fra-rtt): 5.9 ms * [Berlin to Frankfurt latency and RTT](/docs/network/latency/pairs/ber-fra-rtt): 6.1 ms * [Paris to Frankfurt latency and RTT](/docs/network/latency/pairs/par-fra-rtt): 7.6 ms * [London to Frankfurt latency and RTT](/docs/network/latency/pairs/lon-fra-rtt): 13.6 ms * [Marseille to Frankfurt latency and RTT](/docs/network/latency/pairs/mrs-fra-rtt): 16 ms **Same corridor (Asia Pacific → Europe)** * [Hong Kong to Moscow latency and RTT](/docs/network/latency/pairs/hkg-mow-rtt): 118.2 ms * [Taipei to Moscow latency and RTT](/docs/network/latency/pairs/tpe-mow-rtt): 132.7 ms * [Singapore to Marseille latency and RTT](/docs/network/latency/pairs/sin-mrs-rtt): 140.6 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/mel-fra.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Melbourne → Frankfurt round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/mel-fra-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Melbourne, Australia → São Paulo, Brazil RTT 🇦🇺 **Melbourne, Australia (MEL)** → 🇧🇷 **São Paulo, Brazil (GRU)** Measurement round: `2026-08-16T04:07:28Z`. Melbourne to São Paulo averaged 305.3 ms, compared with a 128.34 ms city-centre fiber reference. The difference cannot be assigned to a particular cable or carrier from RTT alone. Standard deviation was 8.1 ms, or 2.7% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **305.3 ms** | | Jitter | **6 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **128.34 ms** | | Fiber Efficiency | **42%** | | Latency Tier | High | | Source Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | | Destination Region | [South America](/docs/network/latency/regions/south-america) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Melbourne](https://www.geonames.org/2158177) (-37.81400, 144.96332) and [São Paulo](https://www.geonames.org/3448439) (-23.54750, -46.63611). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | --------------- | | WGS-84 geodesic distance | **13,105.1 km** | | Vacuum RTT floor | **87.43 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **128.34 ms** | | Low-latency fiber material floor | **127.82 ms** | | Engineering floor (5% path allowance) | **134.77 ms** | | Research 1.33× mapped-fiber reference | **170.69 ms** | | Estimated unamplified path loss | **2752.1 dB** | | Transparent optical spans / inline amplifiers | **173 / 172** | | Published RTT inflation over fiber floor | **2.38×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **295.61 ms** | | Average RTT | **305.3 ms** | | Maximum RTT | **333.89 ms** | | Standard deviation | **8.1 ms** | | Stdev / average | **2.7%** | ## Route Context * Departure PoP: [Melbourne (MEL)](/docs/network/latency/mel-melbourne) * Destination PoP: [São Paulo (GRU)](/docs/network/latency/gru-sao-paulo) * Region overview: [Asia Pacific](/docs/network/latency/regions/asia-pacific) → [South America](/docs/network/latency/regions/south-america) * Corridor overview: [Asia Pacific to South America](/docs/network/latency/regions/asia-pacific-to-south-america) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [São Paulo to Melbourne latency and RTT](/docs/network/latency/pairs/gru-mel-rtt): 274.7 ms **Fastest routes departing Melbourne (MEL)** * [Melbourne to Sydney latency and RTT](/docs/network/latency/pairs/mel-syd-rtt): 9.8 ms * [Melbourne to Singapore latency and RTT](/docs/network/latency/pairs/mel-sin-rtt): 88.4 ms * [Melbourne to Tokyo latency and RTT](/docs/network/latency/pairs/mel-tyo-rtt): 112 ms * [Melbourne to Hong Kong latency and RTT](/docs/network/latency/pairs/mel-hkg-rtt): 138.5 ms * [Melbourne to Taipei latency and RTT](/docs/network/latency/pairs/mel-tpe-rtt): 143.2 ms **Fastest routes arriving at São Paulo (GRU)** * [Ashburn to São Paulo latency and RTT](/docs/network/latency/pairs/iad-gru-rtt): 101.4 ms * [New York to São Paulo latency and RTT](/docs/network/latency/pairs/nyc-gru-rtt): 106.9 ms * [Miami to São Paulo latency and RTT](/docs/network/latency/pairs/mia-gru-rtt): 128.7 ms * [Los Angeles to São Paulo latency and RTT](/docs/network/latency/pairs/lax-gru-rtt): 131.3 ms * [Seattle to São Paulo latency and RTT](/docs/network/latency/pairs/sea-gru-rtt): 155.7 ms **Same corridor (Asia Pacific → South America)** * [Tokyo to São Paulo latency and RTT](/docs/network/latency/pairs/tyo-gru-rtt): 230.5 ms * [Taipei to São Paulo latency and RTT](/docs/network/latency/pairs/tpe-gru-rtt): 273 ms * [Hong Kong to São Paulo latency and RTT](/docs/network/latency/pairs/hkg-gru-rtt): 287.6 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/mel-gru.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Melbourne → São Paulo round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/mel-gru-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Melbourne, Australia → Hong Kong RTT 🇦🇺 **Melbourne, Australia (MEL)** → 🇭🇰 **Hong Kong (HKG)** Measurement round: `2026-08-16T04:07:28Z`. Melbourne to Hong Kong averaged 138.5 ms across 50 replies. The observed range was 133.73 to 150.99 ms; these figures describe one probe round, not a long-term service target. This is number 4 of 19 outgoing routes by average RTT. That ranking compares different destinations and is not adjusted for distance. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------------------- | | RTT | **138.5 ms** | | Jitter | **3.66 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **72.38 ms** | | Fiber Efficiency | **52.3%** | | Latency Tier | Good | | Source Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | | Destination Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Melbourne](https://www.geonames.org/2158177) (-37.81400, 144.96332) and [Hong Kong](https://www.geonames.org/1819729) (22.27832, 114.17469). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **7,391.1 km** | | Vacuum RTT floor | **49.31 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **72.38 ms** | | Low-latency fiber material floor | **72.09 ms** | | Engineering floor (5% path allowance) | **76.01 ms** | | Research 1.33× mapped-fiber reference | **96.26 ms** | | Estimated unamplified path loss | **1552.1 dB** | | Transparent optical spans / inline amplifiers | **98 / 97** | | Published RTT inflation over fiber floor | **1.91×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **133.73 ms** | | Average RTT | **138.5 ms** | | Maximum RTT | **150.99 ms** | | Standard deviation | **4.36 ms** | | Stdev / average | **3.1%** | ## Route Context * Departure PoP: [Melbourne (MEL)](/docs/network/latency/mel-melbourne) * Destination PoP: [Hong Kong (HKG)](/docs/network/latency/hkg-hong-kong) * Region overview: [Asia Pacific](/docs/network/latency/regions/asia-pacific) → [Asia Pacific](/docs/network/latency/regions/asia-pacific) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Hong Kong to Melbourne latency and RTT](/docs/network/latency/pairs/hkg-mel-rtt): 138.4 ms **Fastest routes departing Melbourne (MEL)** * [Melbourne to Sydney latency and RTT](/docs/network/latency/pairs/mel-syd-rtt): 9.8 ms * [Melbourne to Singapore latency and RTT](/docs/network/latency/pairs/mel-sin-rtt): 88.4 ms * [Melbourne to Tokyo latency and RTT](/docs/network/latency/pairs/mel-tyo-rtt): 112 ms * [Melbourne to Taipei latency and RTT](/docs/network/latency/pairs/mel-tpe-rtt): 143.2 ms * [Melbourne to Los Angeles latency and RTT](/docs/network/latency/pairs/mel-lax-rtt): 148.4 ms **Fastest routes arriving at Hong Kong (HKG)** * [Taipei to Hong Kong latency and RTT](/docs/network/latency/pairs/tpe-hkg-rtt): 15.5 ms * [Singapore to Hong Kong latency and RTT](/docs/network/latency/pairs/sin-hkg-rtt): 30.8 ms * [Tokyo to Hong Kong latency and RTT](/docs/network/latency/pairs/tyo-hkg-rtt): 44.8 ms * [Moscow to Hong Kong latency and RTT](/docs/network/latency/pairs/mow-hkg-rtt): 118.6 ms * [Seattle to Hong Kong latency and RTT](/docs/network/latency/pairs/sea-hkg-rtt): 130.4 ms **Same corridor (Asia Pacific → Asia Pacific)** * [Sydney to Melbourne latency and RTT](/docs/network/latency/pairs/syd-mel-rtt): 9.8 ms * [Hong Kong to Taipei latency and RTT](/docs/network/latency/pairs/hkg-tpe-rtt): 14.7 ms * [Hong Kong to Singapore latency and RTT](/docs/network/latency/pairs/hkg-sin-rtt): 31.5 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/mel-hkg.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Melbourne → Hong Kong round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/mel-hkg-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Melbourne, Australia → Ashburn, USA RTT 🇦🇺 **Melbourne, Australia (MEL)** → 🇺🇸 **Ashburn, USA (IAD)** Measurement round: `2026-08-16T04:07:28Z`. Melbourne to Ashburn averaged 203.9 ms across 50 replies. The observed range was 194.63 to 221.82 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 7.18 ms, or 3.5% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **203.9 ms** | | Jitter | **6.81 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **160.09 ms** | | Fiber Efficiency | **78.5%** | | Latency Tier | Fair | | Source Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | | Destination Region | [North America](/docs/network/latency/regions/north-america) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Melbourne](https://www.geonames.org/2158177) (-37.81400, 144.96332) and [Ashburn](https://www.geonames.org/4744870) (39.04372, -77.48749). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | --------------- | | WGS-84 geodesic distance | **16,348.2 km** | | Vacuum RTT floor | **109.06 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **160.09 ms** | | Low-latency fiber material floor | **159.45 ms** | | Engineering floor (5% path allowance) | **168.12 ms** | | Research 1.33× mapped-fiber reference | **212.92 ms** | | Estimated unamplified path loss | **3433.1 dB** | | Transparent optical spans / inline amplifiers | **215 / 214** | | Published RTT inflation over fiber floor | **1.27×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **194.63 ms** | | Average RTT | **203.9 ms** | | Maximum RTT | **221.82 ms** | | Standard deviation | **7.18 ms** | | Stdev / average | **3.5%** | ## Route Context * Departure PoP: [Melbourne (MEL)](/docs/network/latency/mel-melbourne) * Destination PoP: [Ashburn (IAD)](/docs/network/latency/iad-ashburn) * Region overview: [Asia Pacific](/docs/network/latency/regions/asia-pacific) → [North America](/docs/network/latency/regions/north-america) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Ashburn to Melbourne latency and RTT](/docs/network/latency/pairs/iad-mel-rtt): 203.9 ms **Fastest routes departing Melbourne (MEL)** * [Melbourne to Sydney latency and RTT](/docs/network/latency/pairs/mel-syd-rtt): 9.8 ms * [Melbourne to Singapore latency and RTT](/docs/network/latency/pairs/mel-sin-rtt): 88.4 ms * [Melbourne to Tokyo latency and RTT](/docs/network/latency/pairs/mel-tyo-rtt): 112 ms * [Melbourne to Hong Kong latency and RTT](/docs/network/latency/pairs/mel-hkg-rtt): 138.5 ms * [Melbourne to Taipei latency and RTT](/docs/network/latency/pairs/mel-tpe-rtt): 143.2 ms **Fastest routes arriving at Ashburn (IAD)** * [New York to Ashburn latency and RTT](/docs/network/latency/pairs/nyc-iad-rtt): 6 ms * [Miami to Ashburn latency and RTT](/docs/network/latency/pairs/mia-iad-rtt): 27.3 ms * [Los Angeles to Ashburn latency and RTT](/docs/network/latency/pairs/lax-iad-rtt): 60.4 ms * [Seattle to Ashburn latency and RTT](/docs/network/latency/pairs/sea-iad-rtt): 61.5 ms * [London to Ashburn latency and RTT](/docs/network/latency/pairs/lon-iad-rtt): 71 ms **Same corridor (Asia Pacific → North America)** * [Tokyo to Seattle latency and RTT](/docs/network/latency/pairs/tyo-sea-rtt): 84.8 ms * [Tokyo to Los Angeles latency and RTT](/docs/network/latency/pairs/tyo-lax-rtt): 101.2 ms * [Taipei to Seattle latency and RTT](/docs/network/latency/pairs/tpe-sea-rtt): 115.8 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/mel-iad.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Melbourne → Ashburn round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/mel-iad-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Melbourne, Australia → Johannesburg, South Africa RTT 🇦🇺 **Melbourne, Australia (MEL)** → 🇿🇦 **Johannesburg, South Africa (JNB)** Measurement round: `2026-08-16T04:07:28Z`. Melbourne to Johannesburg averaged 405.5 ms, compared with a 101.33 ms city-centre fiber reference. The difference cannot be assigned to a particular cable or carrier from RTT alone. Standard deviation was 16.64 ms, or 4.1% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------------------- | | RTT | **405.5 ms** | | Jitter | **12.98 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **101.33 ms** | | Fiber Efficiency | **25%** | | Latency Tier | High | | Source Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | | Destination Region | [Africa](/docs/network/latency/regions/africa) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Melbourne](https://www.geonames.org/2158177) (-37.81400, 144.96332) and [Johannesburg](https://www.geonames.org/993800) (-26.20227, 28.04363). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | ------------- | | WGS-84 geodesic distance | **10,347 km** | | Vacuum RTT floor | **69.03 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **101.33 ms** | | Low-latency fiber material floor | **100.92 ms** | | Engineering floor (5% path allowance) | **106.41 ms** | | Research 1.33× mapped-fiber reference | **134.76 ms** | | Estimated unamplified path loss | **2172.9 dB** | | Transparent optical spans / inline amplifiers | **136 / 135** | | Published RTT inflation over fiber floor | **4×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **387.75 ms** | | Average RTT | **405.5 ms** | | Maximum RTT | **464.78 ms** | | Standard deviation | **16.64 ms** | | Stdev / average | **4.1%** | ## Route Context * Departure PoP: [Melbourne (MEL)](/docs/network/latency/mel-melbourne) * Destination PoP: [Johannesburg (JNB)](/docs/network/latency/jnb-johannesburg) * Region overview: [Asia Pacific](/docs/network/latency/regions/asia-pacific) → [Africa](/docs/network/latency/regions/africa) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Johannesburg to Melbourne latency and RTT](/docs/network/latency/pairs/jnb-mel-rtt): 406.2 ms **Fastest routes departing Melbourne (MEL)** * [Melbourne to Sydney latency and RTT](/docs/network/latency/pairs/mel-syd-rtt): 9.8 ms * [Melbourne to Singapore latency and RTT](/docs/network/latency/pairs/mel-sin-rtt): 88.4 ms * [Melbourne to Tokyo latency and RTT](/docs/network/latency/pairs/mel-tyo-rtt): 112 ms * [Melbourne to Hong Kong latency and RTT](/docs/network/latency/pairs/mel-hkg-rtt): 138.5 ms * [Melbourne to Taipei latency and RTT](/docs/network/latency/pairs/mel-tpe-rtt): 143.2 ms **Fastest routes arriving at Johannesburg (JNB)** * [Amsterdam to Johannesburg latency and RTT](/docs/network/latency/pairs/ams-jnb-rtt): 171.9 ms * [Berlin to Johannesburg latency and RTT](/docs/network/latency/pairs/ber-jnb-rtt): 172.1 ms * [London to Johannesburg latency and RTT](/docs/network/latency/pairs/lon-jnb-rtt): 179.6 ms * [Paris to Johannesburg latency and RTT](/docs/network/latency/pairs/par-jnb-rtt): 180.9 ms * [Frankfurt to Johannesburg latency and RTT](/docs/network/latency/pairs/fra-jnb-rtt): 188.4 ms **Same corridor (Asia Pacific → Africa)** * [Singapore to Johannesburg latency and RTT](/docs/network/latency/pairs/sin-jnb-rtt): 312.6 ms * [Hong Kong to Johannesburg latency and RTT](/docs/network/latency/pairs/hkg-jnb-rtt): 316.7 ms * [Taipei to Johannesburg latency and RTT](/docs/network/latency/pairs/tpe-jnb-rtt): 331 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/mel-jnb.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Melbourne → Johannesburg round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/mel-jnb-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Melbourne, Australia → Los Angeles, USA RTT 🇦🇺 **Melbourne, Australia (MEL)** → 🇺🇸 **Los Angeles, USA (LAX)** Measurement round: `2026-08-16T04:07:28Z`. Melbourne to Los Angeles averaged 148.4 ms across 50 replies. The observed range was 145.47 to 157.53 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 2.96 ms, or 2.0% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **148.4 ms** | | Jitter | **2.8 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **125 ms** | | Fiber Efficiency | **84.2%** | | Latency Tier | Good | | Source Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | | Destination Region | [North America](/docs/network/latency/regions/north-america) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Melbourne](https://www.geonames.org/2158177) (-37.81400, 144.96332) and [Los Angeles](https://www.geonames.org/5368361) (34.05223, -118.24368). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | ------------- | | WGS-84 geodesic distance | **12,764 km** | | Vacuum RTT floor | **85.15 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **125 ms** | | Low-latency fiber material floor | **124.49 ms** | | Engineering floor (5% path allowance) | **131.26 ms** | | Research 1.33× mapped-fiber reference | **166.24 ms** | | Estimated unamplified path loss | **2680.4 dB** | | Transparent optical spans / inline amplifiers | **168 / 167** | | Published RTT inflation over fiber floor | **1.19×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **145.47 ms** | | Average RTT | **148.4 ms** | | Maximum RTT | **157.53 ms** | | Standard deviation | **2.96 ms** | | Stdev / average | **2.0%** | ## Route Context * Departure PoP: [Melbourne (MEL)](/docs/network/latency/mel-melbourne) * Destination PoP: [Los Angeles (LAX)](/docs/network/latency/lax-los-angeles) * Region overview: [Asia Pacific](/docs/network/latency/regions/asia-pacific) → [North America](/docs/network/latency/regions/north-america) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Los Angeles to Melbourne latency and RTT](/docs/network/latency/pairs/lax-mel-rtt): 148.2 ms **Fastest routes departing Melbourne (MEL)** * [Melbourne to Sydney latency and RTT](/docs/network/latency/pairs/mel-syd-rtt): 9.8 ms * [Melbourne to Singapore latency and RTT](/docs/network/latency/pairs/mel-sin-rtt): 88.4 ms * [Melbourne to Tokyo latency and RTT](/docs/network/latency/pairs/mel-tyo-rtt): 112 ms * [Melbourne to Hong Kong latency and RTT](/docs/network/latency/pairs/mel-hkg-rtt): 138.5 ms * [Melbourne to Taipei latency and RTT](/docs/network/latency/pairs/mel-tpe-rtt): 143.2 ms **Fastest routes arriving at Los Angeles (LAX)** * [Seattle to Los Angeles latency and RTT](/docs/network/latency/pairs/sea-lax-rtt): 25.9 ms * [Miami to Los Angeles latency and RTT](/docs/network/latency/pairs/mia-lax-rtt): 56.8 ms * [New York to Los Angeles latency and RTT](/docs/network/latency/pairs/nyc-lax-rtt): 58.9 ms * [Ashburn to Los Angeles latency and RTT](/docs/network/latency/pairs/iad-lax-rtt): 60.7 ms * [Tokyo to Los Angeles latency and RTT](/docs/network/latency/pairs/tyo-lax-rtt): 101.2 ms **Same corridor (Asia Pacific → North America)** * [Tokyo to Seattle latency and RTT](/docs/network/latency/pairs/tyo-sea-rtt): 84.8 ms * [Taipei to Seattle latency and RTT](/docs/network/latency/pairs/tpe-sea-rtt): 115.8 ms * [Hong Kong to Seattle latency and RTT](/docs/network/latency/pairs/hkg-sea-rtt): 131.6 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/mel-lax.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Melbourne → Los Angeles round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/mel-lax-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Melbourne, Australia → London, UK RTT 🇦🇺 **Melbourne, Australia (MEL)** → 🇬🇧 **London, UK (LON)** Measurement round: `2026-08-16T04:07:28Z`. Melbourne to London averaged 252 ms across 50 replies. The observed range was 239.89 to 289.22 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 10.82 ms, or 4.3% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------------------- | | RTT | **252 ms** | | Jitter | **8.6 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **165.48 ms** | | Fiber Efficiency | **65.7%** | | Latency Tier | High | | Source Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | | Destination Region | [Europe](/docs/network/latency/regions/europe) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Melbourne](https://www.geonames.org/2158177) (-37.81400, 144.96332) and [London](https://www.geonames.org/2643743) (51.50853, -0.12574). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | --------------- | | WGS-84 geodesic distance | **16,898.4 km** | | Vacuum RTT floor | **112.73 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **165.48 ms** | | Low-latency fiber material floor | **164.82 ms** | | Engineering floor (5% path allowance) | **173.78 ms** | | Research 1.33× mapped-fiber reference | **220.09 ms** | | Estimated unamplified path loss | **3548.7 dB** | | Transparent optical spans / inline amplifiers | **222 / 221** | | Published RTT inflation over fiber floor | **1.52×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **239.89 ms** | | Average RTT | **252 ms** | | Maximum RTT | **289.22 ms** | | Standard deviation | **10.82 ms** | | Stdev / average | **4.3%** | ## Route Context * Departure PoP: [Melbourne (MEL)](/docs/network/latency/mel-melbourne) * Destination PoP: [London (LON)](/docs/network/latency/lon-london) * Region overview: [Asia Pacific](/docs/network/latency/regions/asia-pacific) → [Europe](/docs/network/latency/regions/europe) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [London to Melbourne latency and RTT](/docs/network/latency/pairs/lon-mel-rtt): 249.6 ms **Fastest routes departing Melbourne (MEL)** * [Melbourne to Sydney latency and RTT](/docs/network/latency/pairs/mel-syd-rtt): 9.8 ms * [Melbourne to Singapore latency and RTT](/docs/network/latency/pairs/mel-sin-rtt): 88.4 ms * [Melbourne to Tokyo latency and RTT](/docs/network/latency/pairs/mel-tyo-rtt): 112 ms * [Melbourne to Hong Kong latency and RTT](/docs/network/latency/pairs/mel-hkg-rtt): 138.5 ms * [Melbourne to Taipei latency and RTT](/docs/network/latency/pairs/mel-tpe-rtt): 143.2 ms **Fastest routes arriving at London (LON)** * [Amsterdam to London latency and RTT](/docs/network/latency/pairs/ams-lon-rtt): 5.2 ms * [Paris to London latency and RTT](/docs/network/latency/pairs/par-lon-rtt): 6.4 ms * [Frankfurt to London latency and RTT](/docs/network/latency/pairs/fra-lon-rtt): 12.9 ms * [Berlin to London latency and RTT](/docs/network/latency/pairs/ber-lon-rtt): 15.2 ms * [Marseille to London latency and RTT](/docs/network/latency/pairs/mrs-lon-rtt): 18.9 ms **Same corridor (Asia Pacific → Europe)** * [Hong Kong to Moscow latency and RTT](/docs/network/latency/pairs/hkg-mow-rtt): 118.2 ms * [Taipei to Moscow latency and RTT](/docs/network/latency/pairs/tpe-mow-rtt): 132.7 ms * [Singapore to Marseille latency and RTT](/docs/network/latency/pairs/sin-mrs-rtt): 140.6 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/mel-lon.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Melbourne → London round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/mel-lon-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Melbourne, Australia → Miami, USA RTT 🇦🇺 **Melbourne, Australia (MEL)** → 🇺🇸 **Miami, USA (MIA)** Measurement round: `2026-08-16T04:07:28Z`. Melbourne to Miami averaged 201 ms across 50 replies. The observed range was 191.59 to 222.86 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 7.91 ms, or 3.9% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **201 ms** | | Jitter | **7.48 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **152.69 ms** | | Fiber Efficiency | **76%** | | Latency Tier | Fair | | Source Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | | Destination Region | [North America](/docs/network/latency/regions/north-america) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Melbourne](https://www.geonames.org/2158177) (-37.81400, 144.96332) and [Miami](https://www.geonames.org/4164138) (25.77427, -80.19366). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | --------------- | | WGS-84 geodesic distance | **15,591.9 km** | | Vacuum RTT floor | **104.02 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **152.69 ms** | | Low-latency fiber material floor | **152.07 ms** | | Engineering floor (5% path allowance) | **160.34 ms** | | Research 1.33× mapped-fiber reference | **203.07 ms** | | Estimated unamplified path loss | **3274.3 dB** | | Transparent optical spans / inline amplifiers | **205 / 204** | | Published RTT inflation over fiber floor | **1.32×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **191.59 ms** | | Average RTT | **201 ms** | | Maximum RTT | **222.86 ms** | | Standard deviation | **7.91 ms** | | Stdev / average | **3.9%** | ## Route Context * Departure PoP: [Melbourne (MEL)](/docs/network/latency/mel-melbourne) * Destination PoP: [Miami (MIA)](/docs/network/latency/mia-miami) * Region overview: [Asia Pacific](/docs/network/latency/regions/asia-pacific) → [North America](/docs/network/latency/regions/north-america) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Miami to Melbourne latency and RTT](/docs/network/latency/pairs/mia-mel-rtt): 200.7 ms **Fastest routes departing Melbourne (MEL)** * [Melbourne to Sydney latency and RTT](/docs/network/latency/pairs/mel-syd-rtt): 9.8 ms * [Melbourne to Singapore latency and RTT](/docs/network/latency/pairs/mel-sin-rtt): 88.4 ms * [Melbourne to Tokyo latency and RTT](/docs/network/latency/pairs/mel-tyo-rtt): 112 ms * [Melbourne to Hong Kong latency and RTT](/docs/network/latency/pairs/mel-hkg-rtt): 138.5 ms * [Melbourne to Taipei latency and RTT](/docs/network/latency/pairs/mel-tpe-rtt): 143.2 ms **Fastest routes arriving at Miami (MIA)** * [Ashburn to Miami latency and RTT](/docs/network/latency/pairs/iad-mia-rtt): 27.4 ms * [New York to Miami latency and RTT](/docs/network/latency/pairs/nyc-mia-rtt): 32.9 ms * [Los Angeles to Miami latency and RTT](/docs/network/latency/pairs/lax-mia-rtt): 57.3 ms * [São Paulo to Miami latency and RTT](/docs/network/latency/pairs/gru-mia-rtt): 74 ms * [Seattle to Miami latency and RTT](/docs/network/latency/pairs/sea-mia-rtt): 81.7 ms **Same corridor (Asia Pacific → North America)** * [Tokyo to Seattle latency and RTT](/docs/network/latency/pairs/tyo-sea-rtt): 84.8 ms * [Tokyo to Los Angeles latency and RTT](/docs/network/latency/pairs/tyo-lax-rtt): 101.2 ms * [Taipei to Seattle latency and RTT](/docs/network/latency/pairs/tpe-sea-rtt): 115.8 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/mel-mia.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Melbourne → Miami round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/mel-mia-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Melbourne, Australia → Moscow, Russia RTT 🇦🇺 **Melbourne, Australia (MEL)** → 🇷🇺 **Moscow, Russia (MOW)** Measurement round: `2026-08-16T04:07:28Z`. Melbourne to Moscow averaged 236.1 ms across 50 replies. The observed range was 226.33 to 266.71 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 8.22 ms, or 3.5% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------------------- | | RTT | **236.1 ms** | | Jitter | **7.04 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **141.13 ms** | | Fiber Efficiency | **59.8%** | | Latency Tier | Fair | | Source Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | | Destination Region | [Europe](/docs/network/latency/regions/europe) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Melbourne](https://www.geonames.org/2158177) (-37.81400, 144.96332) and [Moscow](https://www.geonames.org/524901) (55.75204, 37.61781). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | --------------- | | WGS-84 geodesic distance | **14,412.1 km** | | Vacuum RTT floor | **96.15 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **141.13 ms** | | Low-latency fiber material floor | **140.57 ms** | | Engineering floor (5% path allowance) | **148.21 ms** | | Research 1.33× mapped-fiber reference | **187.71 ms** | | Estimated unamplified path loss | **3026.5 dB** | | Transparent optical spans / inline amplifiers | **190 / 189** | | Published RTT inflation over fiber floor | **1.67×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **226.33 ms** | | Average RTT | **236.1 ms** | | Maximum RTT | **266.71 ms** | | Standard deviation | **8.22 ms** | | Stdev / average | **3.5%** | ## Route Context * Departure PoP: [Melbourne (MEL)](/docs/network/latency/mel-melbourne) * Destination PoP: [Moscow (MOW)](/docs/network/latency/mow-moscow) * Region overview: [Asia Pacific](/docs/network/latency/regions/asia-pacific) → [Europe](/docs/network/latency/regions/europe) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Moscow to Melbourne latency and RTT](/docs/network/latency/pairs/mow-mel-rtt): 235.9 ms **Fastest routes departing Melbourne (MEL)** * [Melbourne to Sydney latency and RTT](/docs/network/latency/pairs/mel-syd-rtt): 9.8 ms * [Melbourne to Singapore latency and RTT](/docs/network/latency/pairs/mel-sin-rtt): 88.4 ms * [Melbourne to Tokyo latency and RTT](/docs/network/latency/pairs/mel-tyo-rtt): 112 ms * [Melbourne to Hong Kong latency and RTT](/docs/network/latency/pairs/mel-hkg-rtt): 138.5 ms * [Melbourne to Taipei latency and RTT](/docs/network/latency/pairs/mel-tpe-rtt): 143.2 ms **Fastest routes arriving at Moscow (MOW)** * [Berlin to Moscow latency and RTT](/docs/network/latency/pairs/ber-mow-rtt): 28.6 ms * [Frankfurt to Moscow latency and RTT](/docs/network/latency/pairs/fra-mow-rtt): 34.7 ms * [Amsterdam to Moscow latency and RTT](/docs/network/latency/pairs/ams-mow-rtt): 38.1 ms * [London to Moscow latency and RTT](/docs/network/latency/pairs/lon-mow-rtt): 43.5 ms * [Paris to Moscow latency and RTT](/docs/network/latency/pairs/par-mow-rtt): 44.7 ms **Same corridor (Asia Pacific → Europe)** * [Hong Kong to Moscow latency and RTT](/docs/network/latency/pairs/hkg-mow-rtt): 118.2 ms * [Taipei to Moscow latency and RTT](/docs/network/latency/pairs/tpe-mow-rtt): 132.7 ms * [Singapore to Marseille latency and RTT](/docs/network/latency/pairs/sin-mrs-rtt): 140.6 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/mel-mow.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Melbourne → Moscow round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/mel-mow-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Melbourne, Australia → Marseille, France RTT 🇦🇺 **Melbourne, Australia (MEL)** → 🇫🇷 **Marseille, France (MRS)** Measurement round: `2026-08-16T04:07:28Z`. Melbourne to Marseille averaged 233.5 ms across 50 replies. The observed range was 221.74 to 271.93 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 10.3 ms, or 4.4% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------------------- | | RTT | **233.5 ms** | | Jitter | **8.57 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **162.36 ms** | | Fiber Efficiency | **69.5%** | | Latency Tier | Fair | | Source Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | | Destination Region | [Europe](/docs/network/latency/regions/europe) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Melbourne](https://www.geonames.org/2158177) (-37.81400, 144.96332) and [Marseille](https://www.geonames.org/2995469) (43.29695, 5.38107). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | --------------- | | WGS-84 geodesic distance | **16,579.2 km** | | Vacuum RTT floor | **110.6 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **162.36 ms** | | Low-latency fiber material floor | **161.7 ms** | | Engineering floor (5% path allowance) | **170.5 ms** | | Research 1.33× mapped-fiber reference | **215.93 ms** | | Estimated unamplified path loss | **3481.6 dB** | | Transparent optical spans / inline amplifiers | **218 / 217** | | Published RTT inflation over fiber floor | **1.44×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **221.74 ms** | | Average RTT | **233.5 ms** | | Maximum RTT | **271.93 ms** | | Standard deviation | **10.3 ms** | | Stdev / average | **4.4%** | ## Route Context * Departure PoP: [Melbourne (MEL)](/docs/network/latency/mel-melbourne) * Destination PoP: [Marseille (MRS)](/docs/network/latency/mrs-marseille) * Region overview: [Asia Pacific](/docs/network/latency/regions/asia-pacific) → [Europe](/docs/network/latency/regions/europe) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Marseille to Melbourne latency and RTT](/docs/network/latency/pairs/mrs-mel-rtt): 234.2 ms **Fastest routes departing Melbourne (MEL)** * [Melbourne to Sydney latency and RTT](/docs/network/latency/pairs/mel-syd-rtt): 9.8 ms * [Melbourne to Singapore latency and RTT](/docs/network/latency/pairs/mel-sin-rtt): 88.4 ms * [Melbourne to Tokyo latency and RTT](/docs/network/latency/pairs/mel-tyo-rtt): 112 ms * [Melbourne to Hong Kong latency and RTT](/docs/network/latency/pairs/mel-hkg-rtt): 138.5 ms * [Melbourne to Taipei latency and RTT](/docs/network/latency/pairs/mel-tpe-rtt): 143.2 ms **Fastest routes arriving at Marseille (MRS)** * [Paris to Marseille latency and RTT](/docs/network/latency/pairs/par-mrs-rtt): 8.9 ms * [Frankfurt to Marseille latency and RTT](/docs/network/latency/pairs/fra-mrs-rtt): 16.4 ms * [London to Marseille latency and RTT](/docs/network/latency/pairs/lon-mrs-rtt): 17.9 ms * [Amsterdam to Marseille latency and RTT](/docs/network/latency/pairs/ams-mrs-rtt): 19.6 ms * [Berlin to Marseille latency and RTT](/docs/network/latency/pairs/ber-mrs-rtt): 20.2 ms **Same corridor (Asia Pacific → Europe)** * [Hong Kong to Moscow latency and RTT](/docs/network/latency/pairs/hkg-mow-rtt): 118.2 ms * [Taipei to Moscow latency and RTT](/docs/network/latency/pairs/tpe-mow-rtt): 132.7 ms * [Singapore to Marseille latency and RTT](/docs/network/latency/pairs/sin-mrs-rtt): 140.6 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/mel-mrs.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Melbourne → Marseille round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/mel-mrs-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Melbourne, Australia → New York, USA RTT 🇦🇺 **Melbourne, Australia (MEL)** → 🇺🇸 **New York, USA (NYC)** Measurement round: `2026-08-16T04:07:28Z`. Melbourne to New York averaged 204.2 ms across 50 replies. The observed range was 191.39 to 227.39 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 8.13 ms, or 4.0% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **204.2 ms** | | Jitter | **7.67 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **163.27 ms** | | Fiber Efficiency | **80%** | | Latency Tier | Fair | | Source Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | | Destination Region | [North America](/docs/network/latency/regions/north-america) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Melbourne](https://www.geonames.org/2158177) (-37.81400, 144.96332) and [New York City](https://www.geonames.org/5128581) (40.71427, -74.00597). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | ------------- | | WGS-84 geodesic distance | **16,672 km** | | Vacuum RTT floor | **111.22 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **163.27 ms** | | Low-latency fiber material floor | **162.61 ms** | | Engineering floor (5% path allowance) | **171.45 ms** | | Research 1.33× mapped-fiber reference | **217.14 ms** | | Estimated unamplified path loss | **3501.1 dB** | | Transparent optical spans / inline amplifiers | **219 / 218** | | Published RTT inflation over fiber floor | **1.25×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **191.39 ms** | | Average RTT | **204.2 ms** | | Maximum RTT | **227.39 ms** | | Standard deviation | **8.13 ms** | | Stdev / average | **4.0%** | ## Route Context * Departure PoP: [Melbourne (MEL)](/docs/network/latency/mel-melbourne) * Destination PoP: [New York (NYC)](/docs/network/latency/nyc-new-york) * Region overview: [Asia Pacific](/docs/network/latency/regions/asia-pacific) → [North America](/docs/network/latency/regions/north-america) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [New York to Melbourne latency and RTT](/docs/network/latency/pairs/nyc-mel-rtt): 205.1 ms **Fastest routes departing Melbourne (MEL)** * [Melbourne to Sydney latency and RTT](/docs/network/latency/pairs/mel-syd-rtt): 9.8 ms * [Melbourne to Singapore latency and RTT](/docs/network/latency/pairs/mel-sin-rtt): 88.4 ms * [Melbourne to Tokyo latency and RTT](/docs/network/latency/pairs/mel-tyo-rtt): 112 ms * [Melbourne to Hong Kong latency and RTT](/docs/network/latency/pairs/mel-hkg-rtt): 138.5 ms * [Melbourne to Taipei latency and RTT](/docs/network/latency/pairs/mel-tpe-rtt): 143.2 ms **Fastest routes arriving at New York (NYC)** * [Ashburn to New York latency and RTT](/docs/network/latency/pairs/iad-nyc-rtt): 6 ms * [Miami to New York latency and RTT](/docs/network/latency/pairs/mia-nyc-rtt): 33.3 ms * [Los Angeles to New York latency and RTT](/docs/network/latency/pairs/lax-nyc-rtt): 57.9 ms * [Seattle to New York latency and RTT](/docs/network/latency/pairs/sea-nyc-rtt): 59.5 ms * [London to New York latency and RTT](/docs/network/latency/pairs/lon-nyc-rtt): 63.8 ms **Same corridor (Asia Pacific → North America)** * [Tokyo to Seattle latency and RTT](/docs/network/latency/pairs/tyo-sea-rtt): 84.8 ms * [Tokyo to Los Angeles latency and RTT](/docs/network/latency/pairs/tyo-lax-rtt): 101.2 ms * [Taipei to Seattle latency and RTT](/docs/network/latency/pairs/tpe-sea-rtt): 115.8 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/mel-nyc.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Melbourne → New York round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/mel-nyc-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Melbourne, Australia → Paris, France RTT 🇦🇺 **Melbourne, Australia (MEL)** → 🇫🇷 **Paris, France (PAR)** Measurement round: `2026-08-16T04:07:28Z`. Melbourne to Paris averaged 246.5 ms across 50 replies. The observed range was 234.28 to 275.63 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 8.75 ms, or 3.5% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------------------- | | RTT | **246.5 ms** | | Jitter | **6.84 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **164.4 ms** | | Fiber Efficiency | **66.7%** | | Latency Tier | Fair | | Source Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | | Destination Region | [Europe](/docs/network/latency/regions/europe) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Melbourne](https://www.geonames.org/2158177) (-37.81400, 144.96332) and [Paris](https://www.geonames.org/2988507) (48.85341, 2.34880). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | --------------- | | WGS-84 geodesic distance | **16,787.7 km** | | Vacuum RTT floor | **112 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **164.4 ms** | | Low-latency fiber material floor | **163.74 ms** | | Engineering floor (5% path allowance) | **172.64 ms** | | Research 1.33× mapped-fiber reference | **218.65 ms** | | Estimated unamplified path loss | **3525.4 dB** | | Transparent optical spans / inline amplifiers | **221 / 220** | | Published RTT inflation over fiber floor | **1.5×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **234.28 ms** | | Average RTT | **246.5 ms** | | Maximum RTT | **275.63 ms** | | Standard deviation | **8.75 ms** | | Stdev / average | **3.5%** | ## Route Context * Departure PoP: [Melbourne (MEL)](/docs/network/latency/mel-melbourne) * Destination PoP: [Paris (PAR)](/docs/network/latency/par-paris) * Region overview: [Asia Pacific](/docs/network/latency/regions/asia-pacific) → [Europe](/docs/network/latency/regions/europe) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Paris to Melbourne latency and RTT](/docs/network/latency/pairs/par-mel-rtt): 246.4 ms **Fastest routes departing Melbourne (MEL)** * [Melbourne to Sydney latency and RTT](/docs/network/latency/pairs/mel-syd-rtt): 9.8 ms * [Melbourne to Singapore latency and RTT](/docs/network/latency/pairs/mel-sin-rtt): 88.4 ms * [Melbourne to Tokyo latency and RTT](/docs/network/latency/pairs/mel-tyo-rtt): 112 ms * [Melbourne to Hong Kong latency and RTT](/docs/network/latency/pairs/mel-hkg-rtt): 138.5 ms * [Melbourne to Taipei latency and RTT](/docs/network/latency/pairs/mel-tpe-rtt): 143.2 ms **Fastest routes arriving at Paris (PAR)** * [London to Paris latency and RTT](/docs/network/latency/pairs/lon-par-rtt): 6.4 ms * [Amsterdam to Paris latency and RTT](/docs/network/latency/pairs/ams-par-rtt): 7 ms * [Frankfurt to Paris latency and RTT](/docs/network/latency/pairs/fra-par-rtt): 7.6 ms * [Marseille to Paris latency and RTT](/docs/network/latency/pairs/mrs-par-rtt): 8.9 ms * [Berlin to Paris latency and RTT](/docs/network/latency/pairs/ber-par-rtt): 15.6 ms **Same corridor (Asia Pacific → Europe)** * [Hong Kong to Moscow latency and RTT](/docs/network/latency/pairs/hkg-mow-rtt): 118.2 ms * [Taipei to Moscow latency and RTT](/docs/network/latency/pairs/tpe-mow-rtt): 132.7 ms * [Singapore to Marseille latency and RTT](/docs/network/latency/pairs/sin-mrs-rtt): 140.6 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/mel-par.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Melbourne → Paris round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/mel-par-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Melbourne, Australia → Seattle, USA RTT 🇦🇺 **Melbourne, Australia (MEL)** → 🇺🇸 **Seattle, USA (SEA)** Measurement round: `2026-08-16T04:07:28Z`. Melbourne to Seattle averaged 172.1 ms across 50 replies. The observed range was 169 to 187.7 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 3.51 ms, or 2.0% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **172.1 ms** | | Jitter | **2.55 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **128.93 ms** | | Fiber Efficiency | **74.9%** | | Latency Tier | Fair | | Source Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | | Destination Region | [North America](/docs/network/latency/regions/north-america) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Melbourne](https://www.geonames.org/2158177) (-37.81400, 144.96332) and [Seattle](https://www.geonames.org/5809844) (47.60621, -122.33207). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | ------------- | | WGS-84 geodesic distance | **13,166 km** | | Vacuum RTT floor | **87.83 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **128.93 ms** | | Low-latency fiber material floor | **128.41 ms** | | Engineering floor (5% path allowance) | **135.4 ms** | | Research 1.33× mapped-fiber reference | **171.48 ms** | | Estimated unamplified path loss | **2764.9 dB** | | Transparent optical spans / inline amplifiers | **173 / 172** | | Published RTT inflation over fiber floor | **1.33×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------ | | Minimum RTT | **169 ms** | | Average RTT | **172.1 ms** | | Maximum RTT | **187.7 ms** | | Standard deviation | **3.51 ms** | | Stdev / average | **2.0%** | ## Route Context * Departure PoP: [Melbourne (MEL)](/docs/network/latency/mel-melbourne) * Destination PoP: [Seattle (SEA)](/docs/network/latency/sea-seattle) * Region overview: [Asia Pacific](/docs/network/latency/regions/asia-pacific) → [North America](/docs/network/latency/regions/north-america) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Seattle to Melbourne latency and RTT](/docs/network/latency/pairs/sea-mel-rtt): 170.1 ms **Fastest routes departing Melbourne (MEL)** * [Melbourne to Sydney latency and RTT](/docs/network/latency/pairs/mel-syd-rtt): 9.8 ms * [Melbourne to Singapore latency and RTT](/docs/network/latency/pairs/mel-sin-rtt): 88.4 ms * [Melbourne to Tokyo latency and RTT](/docs/network/latency/pairs/mel-tyo-rtt): 112 ms * [Melbourne to Hong Kong latency and RTT](/docs/network/latency/pairs/mel-hkg-rtt): 138.5 ms * [Melbourne to Taipei latency and RTT](/docs/network/latency/pairs/mel-tpe-rtt): 143.2 ms **Fastest routes arriving at Seattle (SEA)** * [Los Angeles to Seattle latency and RTT](/docs/network/latency/pairs/lax-sea-rtt): 26.9 ms * [New York to Seattle latency and RTT](/docs/network/latency/pairs/nyc-sea-rtt): 58.7 ms * [Ashburn to Seattle latency and RTT](/docs/network/latency/pairs/iad-sea-rtt): 62.3 ms * [Miami to Seattle latency and RTT](/docs/network/latency/pairs/mia-sea-rtt): 81.9 ms * [Tokyo to Seattle latency and RTT](/docs/network/latency/pairs/tyo-sea-rtt): 84.8 ms **Same corridor (Asia Pacific → North America)** * [Tokyo to Los Angeles latency and RTT](/docs/network/latency/pairs/tyo-lax-rtt): 101.2 ms * [Taipei to Seattle latency and RTT](/docs/network/latency/pairs/tpe-sea-rtt): 115.8 ms * [Hong Kong to Seattle latency and RTT](/docs/network/latency/pairs/hkg-sea-rtt): 131.6 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/mel-sea.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Melbourne → Seattle round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/mel-sea-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Melbourne, Australia → Singapore RTT 🇦🇺 **Melbourne, Australia (MEL)** → 🇸🇬 **Singapore (SIN)** Measurement round: `2026-08-16T04:07:28Z`. Melbourne to Singapore averaged 88.4 ms across 50 replies. The observed range was 83.08 to 108.06 ms; these figures describe one probe round, not a long-term service target. This is number 2 of 19 outgoing routes by average RTT. That ranking compares different destinations and is not adjusted for distance. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------------------- | | RTT | **88.4 ms** | | Jitter | **3.46 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **59.22 ms** | | Fiber Efficiency | **67%** | | Latency Tier | Good | | Source Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | | Destination Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Melbourne](https://www.geonames.org/2158177) (-37.81400, 144.96332) and [Singapore](https://www.geonames.org/1880252) (1.28967, 103.85007). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **6,046.9 km** | | Vacuum RTT floor | **40.34 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **59.22 ms** | | Low-latency fiber material floor | **58.98 ms** | | Engineering floor (5% path allowance) | **62.18 ms** | | Research 1.33× mapped-fiber reference | **78.76 ms** | | Estimated unamplified path loss | **1269.9 dB** | | Transparent optical spans / inline amplifiers | **80 / 79** | | Published RTT inflation over fiber floor | **1.49×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **83.08 ms** | | Average RTT | **88.4 ms** | | Maximum RTT | **108.06 ms** | | Standard deviation | **4.37 ms** | | Stdev / average | **4.9%** | ## Route Context * Departure PoP: [Melbourne (MEL)](/docs/network/latency/mel-melbourne) * Destination PoP: [Singapore (SIN)](/docs/network/latency/sin-singapore) * Region overview: [Asia Pacific](/docs/network/latency/regions/asia-pacific) → [Asia Pacific](/docs/network/latency/regions/asia-pacific) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Singapore to Melbourne latency and RTT](/docs/network/latency/pairs/sin-mel-rtt): 88.6 ms **Fastest routes departing Melbourne (MEL)** * [Melbourne to Sydney latency and RTT](/docs/network/latency/pairs/mel-syd-rtt): 9.8 ms * [Melbourne to Tokyo latency and RTT](/docs/network/latency/pairs/mel-tyo-rtt): 112 ms * [Melbourne to Hong Kong latency and RTT](/docs/network/latency/pairs/mel-hkg-rtt): 138.5 ms * [Melbourne to Taipei latency and RTT](/docs/network/latency/pairs/mel-tpe-rtt): 143.2 ms * [Melbourne to Los Angeles latency and RTT](/docs/network/latency/pairs/mel-lax-rtt): 148.4 ms **Fastest routes arriving at Singapore (SIN)** * [Hong Kong to Singapore latency and RTT](/docs/network/latency/pairs/hkg-sin-rtt): 31.5 ms * [Taipei to Singapore latency and RTT](/docs/network/latency/pairs/tpe-sin-rtt): 44.3 ms * [Tokyo to Singapore latency and RTT](/docs/network/latency/pairs/tyo-sin-rtt): 69.4 ms * [Sydney to Singapore latency and RTT](/docs/network/latency/pairs/syd-sin-rtt): 94.5 ms * [Marseille to Singapore latency and RTT](/docs/network/latency/pairs/mrs-sin-rtt): 139.6 ms **Same corridor (Asia Pacific → Asia Pacific)** * [Sydney to Melbourne latency and RTT](/docs/network/latency/pairs/syd-mel-rtt): 9.8 ms * [Hong Kong to Taipei latency and RTT](/docs/network/latency/pairs/hkg-tpe-rtt): 14.7 ms * [Taipei to Hong Kong latency and RTT](/docs/network/latency/pairs/tpe-hkg-rtt): 15.5 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/mel-sin.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Melbourne → Singapore round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/mel-sin-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Melbourne, Australia → Sydney, Australia RTT 🇦🇺 **Melbourne, Australia (MEL)** → 🇦🇺 **Sydney, Australia (SYD)** Measurement round: `2026-08-16T04:07:28Z`. Melbourne to Sydney averaged 9.8 ms across 50 replies. The observed range was 9.36 to 11.08 ms; these figures describe one probe round, not a long-term service target. This is number 1 of 19 outgoing routes by average RTT. That ranking compares different destinations and is not adjusted for distance. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------------------- | | RTT | **9.8 ms** | | Jitter | **0.31 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **6.99 ms** | | Fiber Efficiency | **71.3%** | | Latency Tier | Ultra-Low | | Source Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | | Destination Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Melbourne](https://www.geonames.org/2158177) (-37.81400, 144.96332) and [Sydney](https://www.geonames.org/2147714) (-33.86785, 151.20732). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | ------------ | | WGS-84 geodesic distance | **713.8 km** | | Vacuum RTT floor | **4.76 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **6.99 ms** | | Low-latency fiber material floor | **6.96 ms** | | Engineering floor (5% path allowance) | **7.34 ms** | | Research 1.33× mapped-fiber reference | **9.3 ms** | | Estimated unamplified path loss | **149.9 dB** | | Transparent optical spans / inline amplifiers | **10 / 9** | | Published RTT inflation over fiber floor | **1.4×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------ | | Minimum RTT | **9.36 ms** | | Average RTT | **9.8 ms** | | Maximum RTT | **11.08 ms** | | Standard deviation | **0.4 ms** | | Stdev / average | **4.1%** | ## Route Context * Departure PoP: [Melbourne (MEL)](/docs/network/latency/mel-melbourne) * Destination PoP: [Sydney (SYD)](/docs/network/latency/syd-sydney) * Region overview: [Asia Pacific](/docs/network/latency/regions/asia-pacific) → [Asia Pacific](/docs/network/latency/regions/asia-pacific) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Sydney to Melbourne latency and RTT](/docs/network/latency/pairs/syd-mel-rtt): 9.8 ms **Fastest routes departing Melbourne (MEL)** * [Melbourne to Singapore latency and RTT](/docs/network/latency/pairs/mel-sin-rtt): 88.4 ms * [Melbourne to Tokyo latency and RTT](/docs/network/latency/pairs/mel-tyo-rtt): 112 ms * [Melbourne to Hong Kong latency and RTT](/docs/network/latency/pairs/mel-hkg-rtt): 138.5 ms * [Melbourne to Taipei latency and RTT](/docs/network/latency/pairs/mel-tpe-rtt): 143.2 ms * [Melbourne to Los Angeles latency and RTT](/docs/network/latency/pairs/mel-lax-rtt): 148.4 ms **Fastest routes arriving at Sydney (SYD)** * [Singapore to Sydney latency and RTT](/docs/network/latency/pairs/sin-syd-rtt): 94.5 ms * [Tokyo to Sydney latency and RTT](/docs/network/latency/pairs/tyo-syd-rtt): 101.9 ms * [Taipei to Sydney latency and RTT](/docs/network/latency/pairs/tpe-syd-rtt): 133.2 ms * [Hong Kong to Sydney latency and RTT](/docs/network/latency/pairs/hkg-syd-rtt): 137.2 ms * [Los Angeles to Sydney latency and RTT](/docs/network/latency/pairs/lax-syd-rtt): 137.6 ms **Same corridor (Asia Pacific → Asia Pacific)** * [Hong Kong to Taipei latency and RTT](/docs/network/latency/pairs/hkg-tpe-rtt): 14.7 ms * [Taipei to Hong Kong latency and RTT](/docs/network/latency/pairs/tpe-hkg-rtt): 15.5 ms * [Singapore to Hong Kong latency and RTT](/docs/network/latency/pairs/sin-hkg-rtt): 30.8 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/mel-syd.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Melbourne → Sydney round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/mel-syd-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Melbourne, Australia → Taipei, Taiwan RTT 🇦🇺 **Melbourne, Australia (MEL)** → 🇹🇼 **Taipei, Taiwan (TPE)** Measurement round: `2026-08-16T04:07:28Z`. Melbourne to Taipei averaged 143.2 ms across 50 replies. The observed range was 135 to 166.26 ms; these figures describe one probe round, not a long-term service target. This is number 5 of 19 outgoing routes by average RTT. That ranking compares different destinations and is not adjusted for distance. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------------------- | | RTT | **143.2 ms** | | Jitter | **6.42 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **72.25 ms** | | Fiber Efficiency | **50.5%** | | Latency Tier | Good | | Source Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | | Destination Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Melbourne](https://www.geonames.org/2158177) (-37.81400, 144.96332) and [Taipei](https://www.geonames.org/1668341) (25.05306, 121.52639). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **7,377.8 km** | | Vacuum RTT floor | **49.22 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **72.25 ms** | | Low-latency fiber material floor | **71.96 ms** | | Engineering floor (5% path allowance) | **75.87 ms** | | Research 1.33× mapped-fiber reference | **96.09 ms** | | Estimated unamplified path loss | **1549.3 dB** | | Transparent optical spans / inline amplifiers | **97 / 96** | | Published RTT inflation over fiber floor | **1.98×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **135 ms** | | Average RTT | **143.2 ms** | | Maximum RTT | **166.26 ms** | | Standard deviation | **7.11 ms** | | Stdev / average | **5.0%** | ## Route Context * Departure PoP: [Melbourne (MEL)](/docs/network/latency/mel-melbourne) * Destination PoP: [Taipei (TPE)](/docs/network/latency/tpe-taipei) * Region overview: [Asia Pacific](/docs/network/latency/regions/asia-pacific) → [Asia Pacific](/docs/network/latency/regions/asia-pacific) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Taipei to Melbourne latency and RTT](/docs/network/latency/pairs/tpe-mel-rtt): 142 ms **Fastest routes departing Melbourne (MEL)** * [Melbourne to Sydney latency and RTT](/docs/network/latency/pairs/mel-syd-rtt): 9.8 ms * [Melbourne to Singapore latency and RTT](/docs/network/latency/pairs/mel-sin-rtt): 88.4 ms * [Melbourne to Tokyo latency and RTT](/docs/network/latency/pairs/mel-tyo-rtt): 112 ms * [Melbourne to Hong Kong latency and RTT](/docs/network/latency/pairs/mel-hkg-rtt): 138.5 ms * [Melbourne to Los Angeles latency and RTT](/docs/network/latency/pairs/mel-lax-rtt): 148.4 ms **Fastest routes arriving at Taipei (TPE)** * [Hong Kong to Taipei latency and RTT](/docs/network/latency/pairs/hkg-tpe-rtt): 14.7 ms * [Tokyo to Taipei latency and RTT](/docs/network/latency/pairs/tyo-tpe-rtt): 32.3 ms * [Singapore to Taipei latency and RTT](/docs/network/latency/pairs/sin-tpe-rtt): 45.8 ms * [Seattle to Taipei latency and RTT](/docs/network/latency/pairs/sea-tpe-rtt): 114.7 ms * [Sydney to Taipei latency and RTT](/docs/network/latency/pairs/syd-tpe-rtt): 131.7 ms **Same corridor (Asia Pacific → Asia Pacific)** * [Sydney to Melbourne latency and RTT](/docs/network/latency/pairs/syd-mel-rtt): 9.8 ms * [Taipei to Hong Kong latency and RTT](/docs/network/latency/pairs/tpe-hkg-rtt): 15.5 ms * [Singapore to Hong Kong latency and RTT](/docs/network/latency/pairs/sin-hkg-rtt): 30.8 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/mel-tpe.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Melbourne → Taipei round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/mel-tpe-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Melbourne, Australia → Tokyo, Japan RTT 🇦🇺 **Melbourne, Australia (MEL)** → 🇯🇵 **Tokyo, Japan (TYO)** Measurement round: `2026-08-16T04:07:28Z`. Melbourne to Tokyo averaged 112 ms across 50 replies. The observed range was 105.34 to 140.34 ms; these figures describe one probe round, not a long-term service target. This is number 3 of 19 outgoing routes by average RTT. That ranking compares different destinations and is not adjusted for distance. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------------------- | | RTT | **112 ms** | | Jitter | **4.68 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **79.87 ms** | | Fiber Efficiency | **71.3%** | | Latency Tier | Good | | Source Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | | Destination Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Melbourne](https://www.geonames.org/2158177) (-37.81400, 144.96332) and [Tokyo](https://www.geonames.org/1850147) (35.68950, 139.69171). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **8,156.1 km** | | Vacuum RTT floor | **54.41 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **79.87 ms** | | Low-latency fiber material floor | **79.55 ms** | | Engineering floor (5% path allowance) | **83.88 ms** | | Research 1.33× mapped-fiber reference | **106.23 ms** | | Estimated unamplified path loss | **1712.8 dB** | | Transparent optical spans / inline amplifiers | **108 / 107** | | Published RTT inflation over fiber floor | **1.4×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **105.34 ms** | | Average RTT | **112 ms** | | Maximum RTT | **140.34 ms** | | Standard deviation | **5.58 ms** | | Stdev / average | **5.0%** | ## Route Context * Departure PoP: [Melbourne (MEL)](/docs/network/latency/mel-melbourne) * Destination PoP: [Tokyo (TYO)](/docs/network/latency/tyo-tokyo) * Region overview: [Asia Pacific](/docs/network/latency/regions/asia-pacific) → [Asia Pacific](/docs/network/latency/regions/asia-pacific) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Tokyo to Melbourne latency and RTT](/docs/network/latency/pairs/tyo-mel-rtt): 113.9 ms **Fastest routes departing Melbourne (MEL)** * [Melbourne to Sydney latency and RTT](/docs/network/latency/pairs/mel-syd-rtt): 9.8 ms * [Melbourne to Singapore latency and RTT](/docs/network/latency/pairs/mel-sin-rtt): 88.4 ms * [Melbourne to Hong Kong latency and RTT](/docs/network/latency/pairs/mel-hkg-rtt): 138.5 ms * [Melbourne to Taipei latency and RTT](/docs/network/latency/pairs/mel-tpe-rtt): 143.2 ms * [Melbourne to Los Angeles latency and RTT](/docs/network/latency/pairs/mel-lax-rtt): 148.4 ms **Fastest routes arriving at Tokyo (TYO)** * [Taipei to Tokyo latency and RTT](/docs/network/latency/pairs/tpe-tyo-rtt): 31.9 ms * [Hong Kong to Tokyo latency and RTT](/docs/network/latency/pairs/hkg-tyo-rtt): 44.9 ms * [Singapore to Tokyo latency and RTT](/docs/network/latency/pairs/sin-tyo-rtt): 68.9 ms * [Seattle to Tokyo latency and RTT](/docs/network/latency/pairs/sea-tyo-rtt): 85 ms * [Los Angeles to Tokyo latency and RTT](/docs/network/latency/pairs/lax-tyo-rtt): 101.2 ms **Same corridor (Asia Pacific → Asia Pacific)** * [Sydney to Melbourne latency and RTT](/docs/network/latency/pairs/syd-mel-rtt): 9.8 ms * [Hong Kong to Taipei latency and RTT](/docs/network/latency/pairs/hkg-tpe-rtt): 14.7 ms * [Taipei to Hong Kong latency and RTT](/docs/network/latency/pairs/tpe-hkg-rtt): 15.5 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/mel-tyo.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Melbourne → Tokyo round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/mel-tyo-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Miami, USA → Amsterdam, Netherlands RTT 🇺🇸 **Miami, USA (MIA)** → 🇳🇱 **Amsterdam, Netherlands (AMS)** Measurement round: `2026-08-16T04:07:28Z`. Miami to Amsterdam averaged 106 ms across 50 replies. The observed range was 103.48 to 113.5 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 2.2 ms, or 2.1% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **106 ms** | | Jitter | **1.97 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **73.03 ms** | | Fiber Efficiency | **68.9%** | | Latency Tier | Good | | Source Region | [North America](/docs/network/latency/regions/north-america) | | Destination Region | [Europe](/docs/network/latency/regions/europe) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Miami](https://www.geonames.org/4164138) (25.77427, -80.19366) and [Amsterdam](https://www.geonames.org/2759794) (52.37403, 4.88969). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **7,457.8 km** | | Vacuum RTT floor | **49.75 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **73.03 ms** | | Low-latency fiber material floor | **72.74 ms** | | Engineering floor (5% path allowance) | **76.69 ms** | | Research 1.33× mapped-fiber reference | **97.13 ms** | | Estimated unamplified path loss | **1566.1 dB** | | Transparent optical spans / inline amplifiers | **98 / 97** | | Published RTT inflation over fiber floor | **1.45×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **103.48 ms** | | Average RTT | **106 ms** | | Maximum RTT | **113.5 ms** | | Standard deviation | **2.2 ms** | | Stdev / average | **2.1%** | ## Route Context * Departure PoP: [Miami (MIA)](/docs/network/latency/mia-miami) * Destination PoP: [Amsterdam (AMS)](/docs/network/latency/ams-amsterdam) * Region overview: [North America](/docs/network/latency/regions/north-america) → [Europe](/docs/network/latency/regions/europe) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Amsterdam to Miami latency and RTT](/docs/network/latency/pairs/ams-mia-rtt): 106 ms **Fastest routes departing Miami (MIA)** * [Miami to Ashburn latency and RTT](/docs/network/latency/pairs/mia-iad-rtt): 27.3 ms * [Miami to New York latency and RTT](/docs/network/latency/pairs/mia-nyc-rtt): 33.3 ms * [Miami to Los Angeles latency and RTT](/docs/network/latency/pairs/mia-lax-rtt): 56.8 ms * [Miami to Seattle latency and RTT](/docs/network/latency/pairs/mia-sea-rtt): 81.9 ms * [Miami to London latency and RTT](/docs/network/latency/pairs/mia-lon-rtt): 101.8 ms **Fastest routes arriving at Amsterdam (AMS)** * [London to Amsterdam latency and RTT](/docs/network/latency/pairs/lon-ams-rtt): 5.2 ms * [Frankfurt to Amsterdam latency and RTT](/docs/network/latency/pairs/fra-ams-rtt): 5.9 ms * [Paris to Amsterdam latency and RTT](/docs/network/latency/pairs/par-ams-rtt): 7 ms * [Berlin to Amsterdam latency and RTT](/docs/network/latency/pairs/ber-ams-rtt): 7.2 ms * [Marseille to Amsterdam latency and RTT](/docs/network/latency/pairs/mrs-ams-rtt): 20.3 ms **Same corridor (North America → Europe)** * [New York to London latency and RTT](/docs/network/latency/pairs/nyc-lon-rtt): 63.5 ms * [New York to Amsterdam latency and RTT](/docs/network/latency/pairs/nyc-ams-rtt): 68.7 ms * [New York to Paris latency and RTT](/docs/network/latency/pairs/nyc-par-rtt): 68.9 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/mia-ams.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Miami → Amsterdam round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/mia-ams-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Miami, USA → Berlin, Germany RTT 🇺🇸 **Miami, USA (MIA)** → 🇩🇪 **Berlin, Germany (BER)** Measurement round: `2026-08-16T04:07:28Z`. Miami to Berlin averaged 117.3 ms across 50 replies. The observed range was 113.83 to 127.02 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 2.73 ms, or 2.3% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **117.3 ms** | | Jitter | **2.12 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **78.44 ms** | | Fiber Efficiency | **66.9%** | | Latency Tier | Good | | Source Region | [North America](/docs/network/latency/regions/north-america) | | Destination Region | [Europe](/docs/network/latency/regions/europe) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Miami](https://www.geonames.org/4164138) (25.77427, -80.19366) and [Berlin](https://www.geonames.org/2950159) (52.52437, 13.41053). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **8,009.8 km** | | Vacuum RTT floor | **53.44 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **78.44 ms** | | Low-latency fiber material floor | **78.12 ms** | | Engineering floor (5% path allowance) | **82.37 ms** | | Research 1.33× mapped-fiber reference | **104.32 ms** | | Estimated unamplified path loss | **1682.1 dB** | | Transparent optical spans / inline amplifiers | **106 / 105** | | Published RTT inflation over fiber floor | **1.5×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **113.83 ms** | | Average RTT | **117.3 ms** | | Maximum RTT | **127.02 ms** | | Standard deviation | **2.73 ms** | | Stdev / average | **2.3%** | ## Route Context * Departure PoP: [Miami (MIA)](/docs/network/latency/mia-miami) * Destination PoP: [Berlin (BER)](/docs/network/latency/ber-berlin) * Region overview: [North America](/docs/network/latency/regions/north-america) → [Europe](/docs/network/latency/regions/europe) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Berlin to Miami latency and RTT](/docs/network/latency/pairs/ber-mia-rtt): 118.3 ms **Fastest routes departing Miami (MIA)** * [Miami to Ashburn latency and RTT](/docs/network/latency/pairs/mia-iad-rtt): 27.3 ms * [Miami to New York latency and RTT](/docs/network/latency/pairs/mia-nyc-rtt): 33.3 ms * [Miami to Los Angeles latency and RTT](/docs/network/latency/pairs/mia-lax-rtt): 56.8 ms * [Miami to Seattle latency and RTT](/docs/network/latency/pairs/mia-sea-rtt): 81.9 ms * [Miami to London latency and RTT](/docs/network/latency/pairs/mia-lon-rtt): 101.8 ms **Fastest routes arriving at Berlin (BER)** * [Frankfurt to Berlin latency and RTT](/docs/network/latency/pairs/fra-ber-rtt): 6.1 ms * [Amsterdam to Berlin latency and RTT](/docs/network/latency/pairs/ams-ber-rtt): 7.2 ms * [Paris to Berlin latency and RTT](/docs/network/latency/pairs/par-ber-rtt): 15.5 ms * [London to Berlin latency and RTT](/docs/network/latency/pairs/lon-ber-rtt): 16.2 ms * [Marseille to Berlin latency and RTT](/docs/network/latency/pairs/mrs-ber-rtt): 20.6 ms **Same corridor (North America → Europe)** * [New York to London latency and RTT](/docs/network/latency/pairs/nyc-lon-rtt): 63.5 ms * [New York to Amsterdam latency and RTT](/docs/network/latency/pairs/nyc-ams-rtt): 68.7 ms * [New York to Paris latency and RTT](/docs/network/latency/pairs/nyc-par-rtt): 68.9 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/mia-ber.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Miami → Berlin round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/mia-ber-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Miami, USA → Frankfurt, Germany RTT 🇺🇸 **Miami, USA (MIA)** → 🇩🇪 **Frankfurt, Germany (FRA)** Measurement round: `2026-08-16T04:07:28Z`. Miami to Frankfurt averaged 112.2 ms across 50 replies. The observed range was 108.19 to 121.5 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 3.61 ms, or 3.2% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **112.2 ms** | | Jitter | **2.73 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **76.17 ms** | | Fiber Efficiency | **67.9%** | | Latency Tier | Good | | Source Region | [North America](/docs/network/latency/regions/north-america) | | Destination Region | [Europe](/docs/network/latency/regions/europe) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Miami](https://www.geonames.org/4164138) (25.77427, -80.19366) and [Frankfurt am Main](https://www.geonames.org/2925533) (50.11552, 8.68417). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **7,777.9 km** | | Vacuum RTT floor | **51.89 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **76.17 ms** | | Low-latency fiber material floor | **75.86 ms** | | Engineering floor (5% path allowance) | **79.99 ms** | | Research 1.33× mapped-fiber reference | **101.3 ms** | | Estimated unamplified path loss | **1633.4 dB** | | Transparent optical spans / inline amplifiers | **103 / 102** | | Published RTT inflation over fiber floor | **1.47×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **108.19 ms** | | Average RTT | **112.2 ms** | | Maximum RTT | **121.5 ms** | | Standard deviation | **3.61 ms** | | Stdev / average | **3.2%** | ## Route Context * Departure PoP: [Miami (MIA)](/docs/network/latency/mia-miami) * Destination PoP: [Frankfurt (FRA)](/docs/network/latency/fra-frankfurt) * Region overview: [North America](/docs/network/latency/regions/north-america) → [Europe](/docs/network/latency/regions/europe) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Frankfurt to Miami latency and RTT](/docs/network/latency/pairs/fra-mia-rtt): 112 ms **Fastest routes departing Miami (MIA)** * [Miami to Ashburn latency and RTT](/docs/network/latency/pairs/mia-iad-rtt): 27.3 ms * [Miami to New York latency and RTT](/docs/network/latency/pairs/mia-nyc-rtt): 33.3 ms * [Miami to Los Angeles latency and RTT](/docs/network/latency/pairs/mia-lax-rtt): 56.8 ms * [Miami to Seattle latency and RTT](/docs/network/latency/pairs/mia-sea-rtt): 81.9 ms * [Miami to London latency and RTT](/docs/network/latency/pairs/mia-lon-rtt): 101.8 ms **Fastest routes arriving at Frankfurt (FRA)** * [Amsterdam to Frankfurt latency and RTT](/docs/network/latency/pairs/ams-fra-rtt): 5.9 ms * [Berlin to Frankfurt latency and RTT](/docs/network/latency/pairs/ber-fra-rtt): 6.1 ms * [Paris to Frankfurt latency and RTT](/docs/network/latency/pairs/par-fra-rtt): 7.6 ms * [London to Frankfurt latency and RTT](/docs/network/latency/pairs/lon-fra-rtt): 13.6 ms * [Marseille to Frankfurt latency and RTT](/docs/network/latency/pairs/mrs-fra-rtt): 16 ms **Same corridor (North America → Europe)** * [New York to London latency and RTT](/docs/network/latency/pairs/nyc-lon-rtt): 63.5 ms * [New York to Amsterdam latency and RTT](/docs/network/latency/pairs/nyc-ams-rtt): 68.7 ms * [New York to Paris latency and RTT](/docs/network/latency/pairs/nyc-par-rtt): 68.9 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/mia-fra.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Miami → Frankfurt round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/mia-fra-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Miami, USA → São Paulo, Brazil RTT 🇺🇸 **Miami, USA (MIA)** → 🇧🇷 **São Paulo, Brazil (GRU)** Measurement round: `2026-08-16T04:07:28Z`. Miami to São Paulo averaged 128.7 ms, compared with a 64.11 ms city-centre fiber reference. The difference cannot be assigned to a particular cable or carrier from RTT alone. Standard deviation was 1.96 ms, or 1.5% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **128.7 ms** | | Jitter | **1.47 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **64.11 ms** | | Fiber Efficiency | **49.8%** | | Latency Tier | Good | | Source Region | [North America](/docs/network/latency/regions/north-america) | | Destination Region | [South America](/docs/network/latency/regions/south-america) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Miami](https://www.geonames.org/4164138) (25.77427, -80.19366) and [São Paulo](https://www.geonames.org/3448439) (-23.54750, -46.63611). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **6,546.4 km** | | Vacuum RTT floor | **43.67 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **64.11 ms** | | Low-latency fiber material floor | **63.85 ms** | | Engineering floor (5% path allowance) | **67.32 ms** | | Research 1.33× mapped-fiber reference | **85.26 ms** | | Estimated unamplified path loss | **1374.7 dB** | | Transparent optical spans / inline amplifiers | **86 / 85** | | Published RTT inflation over fiber floor | **2.01×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **126.78 ms** | | Average RTT | **128.7 ms** | | Maximum RTT | **136.62 ms** | | Standard deviation | **1.96 ms** | | Stdev / average | **1.5%** | ## Route Context * Departure PoP: [Miami (MIA)](/docs/network/latency/mia-miami) * Destination PoP: [São Paulo (GRU)](/docs/network/latency/gru-sao-paulo) * Region overview: [North America](/docs/network/latency/regions/north-america) → [South America](/docs/network/latency/regions/south-america) * Corridor overview: [North America to South America](/docs/network/latency/regions/north-america-to-south-america) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [São Paulo to Miami latency and RTT](/docs/network/latency/pairs/gru-mia-rtt): 74 ms **Fastest routes departing Miami (MIA)** * [Miami to Ashburn latency and RTT](/docs/network/latency/pairs/mia-iad-rtt): 27.3 ms * [Miami to New York latency and RTT](/docs/network/latency/pairs/mia-nyc-rtt): 33.3 ms * [Miami to Los Angeles latency and RTT](/docs/network/latency/pairs/mia-lax-rtt): 56.8 ms * [Miami to Seattle latency and RTT](/docs/network/latency/pairs/mia-sea-rtt): 81.9 ms * [Miami to London latency and RTT](/docs/network/latency/pairs/mia-lon-rtt): 101.8 ms **Fastest routes arriving at São Paulo (GRU)** * [Ashburn to São Paulo latency and RTT](/docs/network/latency/pairs/iad-gru-rtt): 101.4 ms * [New York to São Paulo latency and RTT](/docs/network/latency/pairs/nyc-gru-rtt): 106.9 ms * [Los Angeles to São Paulo latency and RTT](/docs/network/latency/pairs/lax-gru-rtt): 131.3 ms * [Seattle to São Paulo latency and RTT](/docs/network/latency/pairs/sea-gru-rtt): 155.7 ms * [London to São Paulo latency and RTT](/docs/network/latency/pairs/lon-gru-rtt): 170.7 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/mia-gru.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Miami → São Paulo round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/mia-gru-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Miami, USA → Hong Kong RTT 🇺🇸 **Miami, USA (MIA)** → 🇭🇰 **Hong Kong (HKG)** Measurement round: `2026-08-16T04:07:28Z`. Miami to Hong Kong averaged 199.5 ms across 50 replies. The observed range was 188.29 to 231.23 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 8.84 ms, or 4.4% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **199.5 ms** | | Jitter | **8.19 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **141.69 ms** | | Fiber Efficiency | **71%** | | Latency Tier | Fair | | Source Region | [North America](/docs/network/latency/regions/north-america) | | Destination Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Miami](https://www.geonames.org/4164138) (25.77427, -80.19366) and [Hong Kong](https://www.geonames.org/1819729) (22.27832, 114.17469). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | --------------- | | WGS-84 geodesic distance | **14,468.4 km** | | Vacuum RTT floor | **96.52 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **141.69 ms** | | Low-latency fiber material floor | **141.12 ms** | | Engineering floor (5% path allowance) | **148.79 ms** | | Research 1.33× mapped-fiber reference | **188.44 ms** | | Estimated unamplified path loss | **3038.4 dB** | | Transparent optical spans / inline amplifiers | **190 / 189** | | Published RTT inflation over fiber floor | **1.41×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **188.29 ms** | | Average RTT | **199.5 ms** | | Maximum RTT | **231.23 ms** | | Standard deviation | **8.84 ms** | | Stdev / average | **4.4%** | ## Route Context * Departure PoP: [Miami (MIA)](/docs/network/latency/mia-miami) * Destination PoP: [Hong Kong (HKG)](/docs/network/latency/hkg-hong-kong) * Region overview: [North America](/docs/network/latency/regions/north-america) → [Asia Pacific](/docs/network/latency/regions/asia-pacific) * Corridor overview: [North America to Asia Pacific](/docs/network/latency/regions/north-america-to-asia-pacific) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Hong Kong to Miami latency and RTT](/docs/network/latency/pairs/hkg-mia-rtt): 201.1 ms **Fastest routes departing Miami (MIA)** * [Miami to Ashburn latency and RTT](/docs/network/latency/pairs/mia-iad-rtt): 27.3 ms * [Miami to New York latency and RTT](/docs/network/latency/pairs/mia-nyc-rtt): 33.3 ms * [Miami to Los Angeles latency and RTT](/docs/network/latency/pairs/mia-lax-rtt): 56.8 ms * [Miami to Seattle latency and RTT](/docs/network/latency/pairs/mia-sea-rtt): 81.9 ms * [Miami to London latency and RTT](/docs/network/latency/pairs/mia-lon-rtt): 101.8 ms **Fastest routes arriving at Hong Kong (HKG)** * [Taipei to Hong Kong latency and RTT](/docs/network/latency/pairs/tpe-hkg-rtt): 15.5 ms * [Singapore to Hong Kong latency and RTT](/docs/network/latency/pairs/sin-hkg-rtt): 30.8 ms * [Tokyo to Hong Kong latency and RTT](/docs/network/latency/pairs/tyo-hkg-rtt): 44.8 ms * [Moscow to Hong Kong latency and RTT](/docs/network/latency/pairs/mow-hkg-rtt): 118.6 ms * [Seattle to Hong Kong latency and RTT](/docs/network/latency/pairs/sea-hkg-rtt): 130.4 ms **Same corridor (North America → Asia Pacific)** * [Seattle to Tokyo latency and RTT](/docs/network/latency/pairs/sea-tyo-rtt): 85 ms * [Los Angeles to Tokyo latency and RTT](/docs/network/latency/pairs/lax-tyo-rtt): 101.2 ms * [Seattle to Taipei latency and RTT](/docs/network/latency/pairs/sea-tpe-rtt): 114.7 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/mia-hkg.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Miami → Hong Kong round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/mia-hkg-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Miami, USA → Ashburn, USA RTT 🇺🇸 **Miami, USA (MIA)** → 🇺🇸 **Ashburn, USA (IAD)** Measurement round: `2026-08-16T04:07:28Z`. Miami to Ashburn averaged 27.3 ms across 50 replies. The observed range was 26.26 to 30.45 ms; these figures describe one probe round, not a long-term service target. This is number 1 of 19 outgoing routes by average RTT. That ranking compares different destinations and is not adjusted for distance. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **27.3 ms** | | Jitter | **0.94 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **14.62 ms** | | Fiber Efficiency | **53.6%** | | Latency Tier | Ultra-Low | | Source Region | [North America](/docs/network/latency/regions/north-america) | | Destination Region | [North America](/docs/network/latency/regions/north-america) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Miami](https://www.geonames.org/4164138) (25.77427, -80.19366) and [Ashburn](https://www.geonames.org/4744870) (39.04372, -77.48749). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **1,493.2 km** | | Vacuum RTT floor | **9.96 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **14.62 ms** | | Low-latency fiber material floor | **14.56 ms** | | Engineering floor (5% path allowance) | **15.36 ms** | | Research 1.33× mapped-fiber reference | **19.45 ms** | | Estimated unamplified path loss | **313.6 dB** | | Transparent optical spans / inline amplifiers | **20 / 19** | | Published RTT inflation over fiber floor | **1.87×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------ | | Minimum RTT | **26.26 ms** | | Average RTT | **27.3 ms** | | Maximum RTT | **30.45 ms** | | Standard deviation | **1.02 ms** | | Stdev / average | **3.7%** | ## Route Context * Departure PoP: [Miami (MIA)](/docs/network/latency/mia-miami) * Destination PoP: [Ashburn (IAD)](/docs/network/latency/iad-ashburn) * Region overview: [North America](/docs/network/latency/regions/north-america) → [North America](/docs/network/latency/regions/north-america) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Ashburn to Miami latency and RTT](/docs/network/latency/pairs/iad-mia-rtt): 27.4 ms **Fastest routes departing Miami (MIA)** * [Miami to New York latency and RTT](/docs/network/latency/pairs/mia-nyc-rtt): 33.3 ms * [Miami to Los Angeles latency and RTT](/docs/network/latency/pairs/mia-lax-rtt): 56.8 ms * [Miami to Seattle latency and RTT](/docs/network/latency/pairs/mia-sea-rtt): 81.9 ms * [Miami to London latency and RTT](/docs/network/latency/pairs/mia-lon-rtt): 101.8 ms * [Miami to Amsterdam latency and RTT](/docs/network/latency/pairs/mia-ams-rtt): 106 ms **Fastest routes arriving at Ashburn (IAD)** * [New York to Ashburn latency and RTT](/docs/network/latency/pairs/nyc-iad-rtt): 6 ms * [Los Angeles to Ashburn latency and RTT](/docs/network/latency/pairs/lax-iad-rtt): 60.4 ms * [Seattle to Ashburn latency and RTT](/docs/network/latency/pairs/sea-iad-rtt): 61.5 ms * [London to Ashburn latency and RTT](/docs/network/latency/pairs/lon-iad-rtt): 71 ms * [Paris to Ashburn latency and RTT](/docs/network/latency/pairs/par-iad-rtt): 75.5 ms **Same corridor (North America → North America)** * [Ashburn to New York latency and RTT](/docs/network/latency/pairs/iad-nyc-rtt): 6 ms * [Seattle to Los Angeles latency and RTT](/docs/network/latency/pairs/sea-lax-rtt): 25.9 ms * [Los Angeles to Seattle latency and RTT](/docs/network/latency/pairs/lax-sea-rtt): 26.9 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/mia-iad.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Miami → Ashburn round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/mia-iad-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Miami, USA → Johannesburg, South Africa RTT 🇺🇸 **Miami, USA (MIA)** → 🇿🇦 **Johannesburg, South Africa (JNB)** Measurement round: `2026-08-16T04:07:28Z`. Miami to Johannesburg averaged 259.8 ms, compared with a 126.78 ms city-centre fiber reference. The difference cannot be assigned to a particular cable or carrier from RTT alone. Standard deviation was 6.11 ms, or 2.4% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **259.8 ms** | | Jitter | **5.21 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **126.78 ms** | | Fiber Efficiency | **48.8%** | | Latency Tier | High | | Source Region | [North America](/docs/network/latency/regions/north-america) | | Destination Region | [Africa](/docs/network/latency/regions/africa) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Miami](https://www.geonames.org/4164138) (25.77427, -80.19366) and [Johannesburg](https://www.geonames.org/993800) (-26.20227, 28.04363). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | --------------- | | WGS-84 geodesic distance | **12,945.8 km** | | Vacuum RTT floor | **86.36 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **126.78 ms** | | Low-latency fiber material floor | **126.27 ms** | | Engineering floor (5% path allowance) | **133.13 ms** | | Research 1.33× mapped-fiber reference | **168.61 ms** | | Estimated unamplified path loss | **2718.6 dB** | | Transparent optical spans / inline amplifiers | **170 / 169** | | Published RTT inflation over fiber floor | **2.05×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **252.44 ms** | | Average RTT | **259.8 ms** | | Maximum RTT | **277.9 ms** | | Standard deviation | **6.11 ms** | | Stdev / average | **2.4%** | ## Route Context * Departure PoP: [Miami (MIA)](/docs/network/latency/mia-miami) * Destination PoP: [Johannesburg (JNB)](/docs/network/latency/jnb-johannesburg) * Region overview: [North America](/docs/network/latency/regions/north-america) → [Africa](/docs/network/latency/regions/africa) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Johannesburg to Miami latency and RTT](/docs/network/latency/pairs/jnb-mia-rtt): 259.9 ms **Fastest routes departing Miami (MIA)** * [Miami to Ashburn latency and RTT](/docs/network/latency/pairs/mia-iad-rtt): 27.3 ms * [Miami to New York latency and RTT](/docs/network/latency/pairs/mia-nyc-rtt): 33.3 ms * [Miami to Los Angeles latency and RTT](/docs/network/latency/pairs/mia-lax-rtt): 56.8 ms * [Miami to Seattle latency and RTT](/docs/network/latency/pairs/mia-sea-rtt): 81.9 ms * [Miami to London latency and RTT](/docs/network/latency/pairs/mia-lon-rtt): 101.8 ms **Fastest routes arriving at Johannesburg (JNB)** * [Amsterdam to Johannesburg latency and RTT](/docs/network/latency/pairs/ams-jnb-rtt): 171.9 ms * [Berlin to Johannesburg latency and RTT](/docs/network/latency/pairs/ber-jnb-rtt): 172.1 ms * [London to Johannesburg latency and RTT](/docs/network/latency/pairs/lon-jnb-rtt): 179.6 ms * [Paris to Johannesburg latency and RTT](/docs/network/latency/pairs/par-jnb-rtt): 180.9 ms * [Frankfurt to Johannesburg latency and RTT](/docs/network/latency/pairs/fra-jnb-rtt): 188.4 ms **Same corridor (North America → Africa)** * [New York to Johannesburg latency and RTT](/docs/network/latency/pairs/nyc-jnb-rtt): 221.5 ms * [Ashburn to Johannesburg latency and RTT](/docs/network/latency/pairs/iad-jnb-rtt): 228.3 ms * [Seattle to Johannesburg latency and RTT](/docs/network/latency/pairs/sea-jnb-rtt): 282.1 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/mia-jnb.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Miami → Johannesburg round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/mia-jnb-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Miami, USA → Los Angeles, USA RTT 🇺🇸 **Miami, USA (MIA)** → 🇺🇸 **Los Angeles, USA (LAX)** Measurement round: `2026-08-16T04:07:28Z`. Miami to Los Angeles averaged 56.8 ms across 50 replies. The observed range was 54.44 to 64.31 ms; these figures describe one probe round, not a long-term service target. This is number 3 of 19 outgoing routes by average RTT. That ranking compares different destinations and is not adjusted for distance. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **56.8 ms** | | Jitter | **1.82 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **36.86 ms** | | Fiber Efficiency | **64.9%** | | Latency Tier | Excellent | | Source Region | [North America](/docs/network/latency/regions/north-america) | | Destination Region | [North America](/docs/network/latency/regions/north-america) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Miami](https://www.geonames.org/4164138) (25.77427, -80.19366) and [Los Angeles](https://www.geonames.org/5368361) (34.05223, -118.24368). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **3,764.3 km** | | Vacuum RTT floor | **25.11 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **36.86 ms** | | Low-latency fiber material floor | **36.71 ms** | | Engineering floor (5% path allowance) | **38.71 ms** | | Research 1.33× mapped-fiber reference | **49.03 ms** | | Estimated unamplified path loss | **790.5 dB** | | Transparent optical spans / inline amplifiers | **50 / 49** | | Published RTT inflation over fiber floor | **1.54×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------ | | Minimum RTT | **54.44 ms** | | Average RTT | **56.8 ms** | | Maximum RTT | **64.31 ms** | | Standard deviation | **2.11 ms** | | Stdev / average | **3.7%** | ## Route Context * Departure PoP: [Miami (MIA)](/docs/network/latency/mia-miami) * Destination PoP: [Los Angeles (LAX)](/docs/network/latency/lax-los-angeles) * Region overview: [North America](/docs/network/latency/regions/north-america) → [North America](/docs/network/latency/regions/north-america) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Los Angeles to Miami latency and RTT](/docs/network/latency/pairs/lax-mia-rtt): 57.3 ms **Fastest routes departing Miami (MIA)** * [Miami to Ashburn latency and RTT](/docs/network/latency/pairs/mia-iad-rtt): 27.3 ms * [Miami to New York latency and RTT](/docs/network/latency/pairs/mia-nyc-rtt): 33.3 ms * [Miami to Seattle latency and RTT](/docs/network/latency/pairs/mia-sea-rtt): 81.9 ms * [Miami to London latency and RTT](/docs/network/latency/pairs/mia-lon-rtt): 101.8 ms * [Miami to Amsterdam latency and RTT](/docs/network/latency/pairs/mia-ams-rtt): 106 ms **Fastest routes arriving at Los Angeles (LAX)** * [Seattle to Los Angeles latency and RTT](/docs/network/latency/pairs/sea-lax-rtt): 25.9 ms * [New York to Los Angeles latency and RTT](/docs/network/latency/pairs/nyc-lax-rtt): 58.9 ms * [Ashburn to Los Angeles latency and RTT](/docs/network/latency/pairs/iad-lax-rtt): 60.7 ms * [Tokyo to Los Angeles latency and RTT](/docs/network/latency/pairs/tyo-lax-rtt): 101.2 ms * [London to Los Angeles latency and RTT](/docs/network/latency/pairs/lon-lax-rtt): 127.5 ms **Same corridor (North America → North America)** * [Ashburn to New York latency and RTT](/docs/network/latency/pairs/iad-nyc-rtt): 6 ms * [New York to Ashburn latency and RTT](/docs/network/latency/pairs/nyc-iad-rtt): 6 ms * [Los Angeles to Seattle latency and RTT](/docs/network/latency/pairs/lax-sea-rtt): 26.9 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/mia-lax.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Miami → Los Angeles round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/mia-lax-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Miami, USA → London, UK RTT 🇺🇸 **Miami, USA (MIA)** → 🇬🇧 **London, UK (LON)** Measurement round: `2026-08-16T04:07:28Z`. Miami to London averaged 101.8 ms across 50 replies. The observed range was 96.22 to 114.36 ms; these figures describe one probe round, not a long-term service target. This is number 5 of 19 outgoing routes by average RTT. That ranking compares different destinations and is not adjusted for distance. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **101.8 ms** | | Jitter | **2.93 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **69.9 ms** | | Fiber Efficiency | **68.7%** | | Latency Tier | Good | | Source Region | [North America](/docs/network/latency/regions/north-america) | | Destination Region | [Europe](/docs/network/latency/regions/europe) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Miami](https://www.geonames.org/4164138) (25.77427, -80.19366) and [London](https://www.geonames.org/2643743) (51.50853, -0.12574). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **7,138.4 km** | | Vacuum RTT floor | **47.62 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **69.9 ms** | | Low-latency fiber material floor | **69.62 ms** | | Engineering floor (5% path allowance) | **73.41 ms** | | Research 1.33× mapped-fiber reference | **92.97 ms** | | Estimated unamplified path loss | **1499.1 dB** | | Transparent optical spans / inline amplifiers | **94 / 93** | | Published RTT inflation over fiber floor | **1.46×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **96.22 ms** | | Average RTT | **101.8 ms** | | Maximum RTT | **114.36 ms** | | Standard deviation | **3.81 ms** | | Stdev / average | **3.7%** | ## Route Context * Departure PoP: [Miami (MIA)](/docs/network/latency/mia-miami) * Destination PoP: [London (LON)](/docs/network/latency/lon-london) * Region overview: [North America](/docs/network/latency/regions/north-america) → [Europe](/docs/network/latency/regions/europe) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [London to Miami latency and RTT](/docs/network/latency/pairs/lon-mia-rtt): 101.9 ms **Fastest routes departing Miami (MIA)** * [Miami to Ashburn latency and RTT](/docs/network/latency/pairs/mia-iad-rtt): 27.3 ms * [Miami to New York latency and RTT](/docs/network/latency/pairs/mia-nyc-rtt): 33.3 ms * [Miami to Los Angeles latency and RTT](/docs/network/latency/pairs/mia-lax-rtt): 56.8 ms * [Miami to Seattle latency and RTT](/docs/network/latency/pairs/mia-sea-rtt): 81.9 ms * [Miami to Amsterdam latency and RTT](/docs/network/latency/pairs/mia-ams-rtt): 106 ms **Fastest routes arriving at London (LON)** * [Amsterdam to London latency and RTT](/docs/network/latency/pairs/ams-lon-rtt): 5.2 ms * [Paris to London latency and RTT](/docs/network/latency/pairs/par-lon-rtt): 6.4 ms * [Frankfurt to London latency and RTT](/docs/network/latency/pairs/fra-lon-rtt): 12.9 ms * [Berlin to London latency and RTT](/docs/network/latency/pairs/ber-lon-rtt): 15.2 ms * [Marseille to London latency and RTT](/docs/network/latency/pairs/mrs-lon-rtt): 18.9 ms **Same corridor (North America → Europe)** * [New York to London latency and RTT](/docs/network/latency/pairs/nyc-lon-rtt): 63.5 ms * [New York to Amsterdam latency and RTT](/docs/network/latency/pairs/nyc-ams-rtt): 68.7 ms * [New York to Paris latency and RTT](/docs/network/latency/pairs/nyc-par-rtt): 68.9 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/mia-lon.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Miami → London round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/mia-lon-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Miami, USA → Melbourne, Australia RTT 🇺🇸 **Miami, USA (MIA)** → 🇦🇺 **Melbourne, Australia (MEL)** Measurement round: `2026-08-16T04:07:28Z`. Miami to Melbourne averaged 200.7 ms across 50 replies. The observed range was 190.44 to 231.45 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 9.1 ms, or 4.5% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **200.7 ms** | | Jitter | **6.99 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **152.69 ms** | | Fiber Efficiency | **76.1%** | | Latency Tier | Fair | | Source Region | [North America](/docs/network/latency/regions/north-america) | | Destination Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Miami](https://www.geonames.org/4164138) (25.77427, -80.19366) and [Melbourne](https://www.geonames.org/2158177) (-37.81400, 144.96332). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | --------------- | | WGS-84 geodesic distance | **15,591.9 km** | | Vacuum RTT floor | **104.02 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **152.69 ms** | | Low-latency fiber material floor | **152.07 ms** | | Engineering floor (5% path allowance) | **160.34 ms** | | Research 1.33× mapped-fiber reference | **203.07 ms** | | Estimated unamplified path loss | **3274.3 dB** | | Transparent optical spans / inline amplifiers | **205 / 204** | | Published RTT inflation over fiber floor | **1.31×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **190.44 ms** | | Average RTT | **200.7 ms** | | Maximum RTT | **231.45 ms** | | Standard deviation | **9.1 ms** | | Stdev / average | **4.5%** | ## Route Context * Departure PoP: [Miami (MIA)](/docs/network/latency/mia-miami) * Destination PoP: [Melbourne (MEL)](/docs/network/latency/mel-melbourne) * Region overview: [North America](/docs/network/latency/regions/north-america) → [Asia Pacific](/docs/network/latency/regions/asia-pacific) * Corridor overview: [North America to Asia Pacific](/docs/network/latency/regions/north-america-to-asia-pacific) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Melbourne to Miami latency and RTT](/docs/network/latency/pairs/mel-mia-rtt): 201 ms **Fastest routes departing Miami (MIA)** * [Miami to Ashburn latency and RTT](/docs/network/latency/pairs/mia-iad-rtt): 27.3 ms * [Miami to New York latency and RTT](/docs/network/latency/pairs/mia-nyc-rtt): 33.3 ms * [Miami to Los Angeles latency and RTT](/docs/network/latency/pairs/mia-lax-rtt): 56.8 ms * [Miami to Seattle latency and RTT](/docs/network/latency/pairs/mia-sea-rtt): 81.9 ms * [Miami to London latency and RTT](/docs/network/latency/pairs/mia-lon-rtt): 101.8 ms **Fastest routes arriving at Melbourne (MEL)** * [Sydney to Melbourne latency and RTT](/docs/network/latency/pairs/syd-mel-rtt): 9.8 ms * [Singapore to Melbourne latency and RTT](/docs/network/latency/pairs/sin-mel-rtt): 88.6 ms * [Tokyo to Melbourne latency and RTT](/docs/network/latency/pairs/tyo-mel-rtt): 113.9 ms * [Hong Kong to Melbourne latency and RTT](/docs/network/latency/pairs/hkg-mel-rtt): 138.4 ms * [Taipei to Melbourne latency and RTT](/docs/network/latency/pairs/tpe-mel-rtt): 142 ms **Same corridor (North America → Asia Pacific)** * [Seattle to Tokyo latency and RTT](/docs/network/latency/pairs/sea-tyo-rtt): 85 ms * [Los Angeles to Tokyo latency and RTT](/docs/network/latency/pairs/lax-tyo-rtt): 101.2 ms * [Seattle to Taipei latency and RTT](/docs/network/latency/pairs/sea-tpe-rtt): 114.7 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/mia-mel.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Miami → Melbourne round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/mia-mel-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Miami, USA → Moscow, Russia RTT 🇺🇸 **Miami, USA (MIA)** → 🇷🇺 **Moscow, Russia (MOW)** Measurement round: `2026-08-16T04:07:28Z`. Miami to Moscow averaged 145.5 ms across 50 replies. The observed range was 142.62 to 152.94 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 2.55 ms, or 1.8% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **145.5 ms** | | Jitter | **1.94 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **90.48 ms** | | Fiber Efficiency | **62.2%** | | Latency Tier | Good | | Source Region | [North America](/docs/network/latency/regions/north-america) | | Destination Region | [Europe](/docs/network/latency/regions/europe) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Miami](https://www.geonames.org/4164138) (25.77427, -80.19366) and [Moscow](https://www.geonames.org/524901) (55.75204, 37.61781). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **9,239.8 km** | | Vacuum RTT floor | **61.64 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **90.48 ms** | | Low-latency fiber material floor | **90.12 ms** | | Engineering floor (5% path allowance) | **95.02 ms** | | Research 1.33× mapped-fiber reference | **120.34 ms** | | Estimated unamplified path loss | **1940.4 dB** | | Transparent optical spans / inline amplifiers | **122 / 121** | | Published RTT inflation over fiber floor | **1.61×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **142.62 ms** | | Average RTT | **145.5 ms** | | Maximum RTT | **152.94 ms** | | Standard deviation | **2.55 ms** | | Stdev / average | **1.8%** | ## Route Context * Departure PoP: [Miami (MIA)](/docs/network/latency/mia-miami) * Destination PoP: [Moscow (MOW)](/docs/network/latency/mow-moscow) * Region overview: [North America](/docs/network/latency/regions/north-america) → [Europe](/docs/network/latency/regions/europe) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Moscow to Miami latency and RTT](/docs/network/latency/pairs/mow-mia-rtt): 144.5 ms **Fastest routes departing Miami (MIA)** * [Miami to Ashburn latency and RTT](/docs/network/latency/pairs/mia-iad-rtt): 27.3 ms * [Miami to New York latency and RTT](/docs/network/latency/pairs/mia-nyc-rtt): 33.3 ms * [Miami to Los Angeles latency and RTT](/docs/network/latency/pairs/mia-lax-rtt): 56.8 ms * [Miami to Seattle latency and RTT](/docs/network/latency/pairs/mia-sea-rtt): 81.9 ms * [Miami to London latency and RTT](/docs/network/latency/pairs/mia-lon-rtt): 101.8 ms **Fastest routes arriving at Moscow (MOW)** * [Berlin to Moscow latency and RTT](/docs/network/latency/pairs/ber-mow-rtt): 28.6 ms * [Frankfurt to Moscow latency and RTT](/docs/network/latency/pairs/fra-mow-rtt): 34.7 ms * [Amsterdam to Moscow latency and RTT](/docs/network/latency/pairs/ams-mow-rtt): 38.1 ms * [London to Moscow latency and RTT](/docs/network/latency/pairs/lon-mow-rtt): 43.5 ms * [Paris to Moscow latency and RTT](/docs/network/latency/pairs/par-mow-rtt): 44.7 ms **Same corridor (North America → Europe)** * [New York to London latency and RTT](/docs/network/latency/pairs/nyc-lon-rtt): 63.5 ms * [New York to Amsterdam latency and RTT](/docs/network/latency/pairs/nyc-ams-rtt): 68.7 ms * [New York to Paris latency and RTT](/docs/network/latency/pairs/nyc-par-rtt): 68.9 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/mia-mow.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Miami → Moscow round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/mia-mow-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Miami, USA → Marseille, France RTT 🇺🇸 **Miami, USA (MIA)** → 🇫🇷 **Marseille, France (MRS)** Measurement round: `2026-08-16T04:07:28Z`. Miami to Marseille averaged 113.7 ms across 50 replies. The observed range was 110.36 to 126.2 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 3.72 ms, or 3.3% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **113.7 ms** | | Jitter | **2.44 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **75.92 ms** | | Fiber Efficiency | **66.8%** | | Latency Tier | Good | | Source Region | [North America](/docs/network/latency/regions/north-america) | | Destination Region | [Europe](/docs/network/latency/regions/europe) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Miami](https://www.geonames.org/4164138) (25.77427, -80.19366) and [Marseille](https://www.geonames.org/2995469) (43.29695, 5.38107). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **7,752.4 km** | | Vacuum RTT floor | **51.72 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **75.92 ms** | | Low-latency fiber material floor | **75.61 ms** | | Engineering floor (5% path allowance) | **79.72 ms** | | Research 1.33× mapped-fiber reference | **100.97 ms** | | Estimated unamplified path loss | **1628 dB** | | Transparent optical spans / inline amplifiers | **102 / 101** | | Published RTT inflation over fiber floor | **1.5×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **110.36 ms** | | Average RTT | **113.7 ms** | | Maximum RTT | **126.2 ms** | | Standard deviation | **3.72 ms** | | Stdev / average | **3.3%** | ## Route Context * Departure PoP: [Miami (MIA)](/docs/network/latency/mia-miami) * Destination PoP: [Marseille (MRS)](/docs/network/latency/mrs-marseille) * Region overview: [North America](/docs/network/latency/regions/north-america) → [Europe](/docs/network/latency/regions/europe) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Marseille to Miami latency and RTT](/docs/network/latency/pairs/mrs-mia-rtt): 112.4 ms **Fastest routes departing Miami (MIA)** * [Miami to Ashburn latency and RTT](/docs/network/latency/pairs/mia-iad-rtt): 27.3 ms * [Miami to New York latency and RTT](/docs/network/latency/pairs/mia-nyc-rtt): 33.3 ms * [Miami to Los Angeles latency and RTT](/docs/network/latency/pairs/mia-lax-rtt): 56.8 ms * [Miami to Seattle latency and RTT](/docs/network/latency/pairs/mia-sea-rtt): 81.9 ms * [Miami to London latency and RTT](/docs/network/latency/pairs/mia-lon-rtt): 101.8 ms **Fastest routes arriving at Marseille (MRS)** * [Paris to Marseille latency and RTT](/docs/network/latency/pairs/par-mrs-rtt): 8.9 ms * [Frankfurt to Marseille latency and RTT](/docs/network/latency/pairs/fra-mrs-rtt): 16.4 ms * [London to Marseille latency and RTT](/docs/network/latency/pairs/lon-mrs-rtt): 17.9 ms * [Amsterdam to Marseille latency and RTT](/docs/network/latency/pairs/ams-mrs-rtt): 19.6 ms * [Berlin to Marseille latency and RTT](/docs/network/latency/pairs/ber-mrs-rtt): 20.2 ms **Same corridor (North America → Europe)** * [New York to London latency and RTT](/docs/network/latency/pairs/nyc-lon-rtt): 63.5 ms * [New York to Amsterdam latency and RTT](/docs/network/latency/pairs/nyc-ams-rtt): 68.7 ms * [New York to Paris latency and RTT](/docs/network/latency/pairs/nyc-par-rtt): 68.9 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/mia-mrs.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Miami → Marseille round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/mia-mrs-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Miami, USA → New York, USA RTT 🇺🇸 **Miami, USA (MIA)** → 🇺🇸 **New York, USA (NYC)** Measurement round: `2026-08-16T04:07:28Z`. Miami to New York averaged 33.3 ms across 50 replies. The observed range was 31.68 to 39.04 ms; these figures describe one probe round, not a long-term service target. This is number 2 of 19 outgoing routes by average RTT. That ranking compares different destinations and is not adjusted for distance. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **33.3 ms** | | Jitter | **1.24 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **17.17 ms** | | Fiber Efficiency | **51.6%** | | Latency Tier | Excellent | | Source Region | [North America](/docs/network/latency/regions/north-america) | | Destination Region | [North America](/docs/network/latency/regions/north-america) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Miami](https://www.geonames.org/4164138) (25.77427, -80.19366) and [New York City](https://www.geonames.org/5128581) (40.71427, -74.00597). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **1,753.2 km** | | Vacuum RTT floor | **11.7 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **17.17 ms** | | Low-latency fiber material floor | **17.1 ms** | | Engineering floor (5% path allowance) | **18.03 ms** | | Research 1.33× mapped-fiber reference | **22.83 ms** | | Estimated unamplified path loss | **368.2 dB** | | Transparent optical spans / inline amplifiers | **24 / 23** | | Published RTT inflation over fiber floor | **1.94×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------ | | Minimum RTT | **31.68 ms** | | Average RTT | **33.3 ms** | | Maximum RTT | **39.04 ms** | | Standard deviation | **1.51 ms** | | Stdev / average | **4.5%** | ## Route Context * Departure PoP: [Miami (MIA)](/docs/network/latency/mia-miami) * Destination PoP: [New York (NYC)](/docs/network/latency/nyc-new-york) * Region overview: [North America](/docs/network/latency/regions/north-america) → [North America](/docs/network/latency/regions/north-america) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [New York to Miami latency and RTT](/docs/network/latency/pairs/nyc-mia-rtt): 32.9 ms **Fastest routes departing Miami (MIA)** * [Miami to Ashburn latency and RTT](/docs/network/latency/pairs/mia-iad-rtt): 27.3 ms * [Miami to Los Angeles latency and RTT](/docs/network/latency/pairs/mia-lax-rtt): 56.8 ms * [Miami to Seattle latency and RTT](/docs/network/latency/pairs/mia-sea-rtt): 81.9 ms * [Miami to London latency and RTT](/docs/network/latency/pairs/mia-lon-rtt): 101.8 ms * [Miami to Amsterdam latency and RTT](/docs/network/latency/pairs/mia-ams-rtt): 106 ms **Fastest routes arriving at New York (NYC)** * [Ashburn to New York latency and RTT](/docs/network/latency/pairs/iad-nyc-rtt): 6 ms * [Los Angeles to New York latency and RTT](/docs/network/latency/pairs/lax-nyc-rtt): 57.9 ms * [Seattle to New York latency and RTT](/docs/network/latency/pairs/sea-nyc-rtt): 59.5 ms * [London to New York latency and RTT](/docs/network/latency/pairs/lon-nyc-rtt): 63.8 ms * [Amsterdam to New York latency and RTT](/docs/network/latency/pairs/ams-nyc-rtt): 68.6 ms **Same corridor (North America → North America)** * [New York to Ashburn latency and RTT](/docs/network/latency/pairs/nyc-iad-rtt): 6 ms * [Seattle to Los Angeles latency and RTT](/docs/network/latency/pairs/sea-lax-rtt): 25.9 ms * [Los Angeles to Seattle latency and RTT](/docs/network/latency/pairs/lax-sea-rtt): 26.9 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/mia-nyc.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Miami → New York round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/mia-nyc-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Miami, USA → Paris, France RTT 🇺🇸 **Miami, USA (MIA)** → 🇫🇷 **Paris, France (PAR)** Measurement round: `2026-08-16T04:07:28Z`. Miami to Paris averaged 106.2 ms across 50 replies. The observed range was 104.35 to 111.43 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 1.69 ms, or 1.6% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **106.2 ms** | | Jitter | **1.31 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **72.16 ms** | | Fiber Efficiency | **67.9%** | | Latency Tier | Good | | Source Region | [North America](/docs/network/latency/regions/north-america) | | Destination Region | [Europe](/docs/network/latency/regions/europe) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Miami](https://www.geonames.org/4164138) (25.77427, -80.19366) and [Paris](https://www.geonames.org/2988507) (48.85341, 2.34880). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **7,368.8 km** | | Vacuum RTT floor | **49.16 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **72.16 ms** | | Low-latency fiber material floor | **71.87 ms** | | Engineering floor (5% path allowance) | **75.78 ms** | | Research 1.33× mapped-fiber reference | **95.97 ms** | | Estimated unamplified path loss | **1547.4 dB** | | Transparent optical spans / inline amplifiers | **97 / 96** | | Published RTT inflation over fiber floor | **1.47×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **104.35 ms** | | Average RTT | **106.2 ms** | | Maximum RTT | **111.43 ms** | | Standard deviation | **1.69 ms** | | Stdev / average | **1.6%** | ## Route Context * Departure PoP: [Miami (MIA)](/docs/network/latency/mia-miami) * Destination PoP: [Paris (PAR)](/docs/network/latency/par-paris) * Region overview: [North America](/docs/network/latency/regions/north-america) → [Europe](/docs/network/latency/regions/europe) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Paris to Miami latency and RTT](/docs/network/latency/pairs/par-mia-rtt): 106.8 ms **Fastest routes departing Miami (MIA)** * [Miami to Ashburn latency and RTT](/docs/network/latency/pairs/mia-iad-rtt): 27.3 ms * [Miami to New York latency and RTT](/docs/network/latency/pairs/mia-nyc-rtt): 33.3 ms * [Miami to Los Angeles latency and RTT](/docs/network/latency/pairs/mia-lax-rtt): 56.8 ms * [Miami to Seattle latency and RTT](/docs/network/latency/pairs/mia-sea-rtt): 81.9 ms * [Miami to London latency and RTT](/docs/network/latency/pairs/mia-lon-rtt): 101.8 ms **Fastest routes arriving at Paris (PAR)** * [London to Paris latency and RTT](/docs/network/latency/pairs/lon-par-rtt): 6.4 ms * [Amsterdam to Paris latency and RTT](/docs/network/latency/pairs/ams-par-rtt): 7 ms * [Frankfurt to Paris latency and RTT](/docs/network/latency/pairs/fra-par-rtt): 7.6 ms * [Marseille to Paris latency and RTT](/docs/network/latency/pairs/mrs-par-rtt): 8.9 ms * [Berlin to Paris latency and RTT](/docs/network/latency/pairs/ber-par-rtt): 15.6 ms **Same corridor (North America → Europe)** * [New York to London latency and RTT](/docs/network/latency/pairs/nyc-lon-rtt): 63.5 ms * [New York to Amsterdam latency and RTT](/docs/network/latency/pairs/nyc-ams-rtt): 68.7 ms * [New York to Paris latency and RTT](/docs/network/latency/pairs/nyc-par-rtt): 68.9 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/mia-par.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Miami → Paris round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/mia-par-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Miami, USA → Seattle, USA RTT 🇺🇸 **Miami, USA (MIA)** → 🇺🇸 **Seattle, USA (SEA)** Measurement round: `2026-08-16T04:07:28Z`. Miami to Seattle averaged 81.9 ms across 50 replies. The observed range was 78.54 to 93.5 ms; these figures describe one probe round, not a long-term service target. This is number 4 of 19 outgoing routes by average RTT. That ranking compares different destinations and is not adjusted for distance. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **81.9 ms** | | Jitter | **2.63 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **43.08 ms** | | Fiber Efficiency | **52.6%** | | Latency Tier | Good | | Source Region | [North America](/docs/network/latency/regions/north-america) | | Destination Region | [North America](/docs/network/latency/regions/north-america) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Miami](https://www.geonames.org/4164138) (25.77427, -80.19366) and [Seattle](https://www.geonames.org/5809844) (47.60621, -122.33207). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **4,399.4 km** | | Vacuum RTT floor | **29.35 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **43.08 ms** | | Low-latency fiber material floor | **42.91 ms** | | Engineering floor (5% path allowance) | **45.24 ms** | | Research 1.33× mapped-fiber reference | **57.3 ms** | | Estimated unamplified path loss | **923.9 dB** | | Transparent optical spans / inline amplifiers | **58 / 57** | | Published RTT inflation over fiber floor | **1.9×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------ | | Minimum RTT | **78.54 ms** | | Average RTT | **81.9 ms** | | Maximum RTT | **93.5 ms** | | Standard deviation | **3.48 ms** | | Stdev / average | **4.2%** | ## Route Context * Departure PoP: [Miami (MIA)](/docs/network/latency/mia-miami) * Destination PoP: [Seattle (SEA)](/docs/network/latency/sea-seattle) * Region overview: [North America](/docs/network/latency/regions/north-america) → [North America](/docs/network/latency/regions/north-america) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Seattle to Miami latency and RTT](/docs/network/latency/pairs/sea-mia-rtt): 81.7 ms **Fastest routes departing Miami (MIA)** * [Miami to Ashburn latency and RTT](/docs/network/latency/pairs/mia-iad-rtt): 27.3 ms * [Miami to New York latency and RTT](/docs/network/latency/pairs/mia-nyc-rtt): 33.3 ms * [Miami to Los Angeles latency and RTT](/docs/network/latency/pairs/mia-lax-rtt): 56.8 ms * [Miami to London latency and RTT](/docs/network/latency/pairs/mia-lon-rtt): 101.8 ms * [Miami to Amsterdam latency and RTT](/docs/network/latency/pairs/mia-ams-rtt): 106 ms **Fastest routes arriving at Seattle (SEA)** * [Los Angeles to Seattle latency and RTT](/docs/network/latency/pairs/lax-sea-rtt): 26.9 ms * [New York to Seattle latency and RTT](/docs/network/latency/pairs/nyc-sea-rtt): 58.7 ms * [Ashburn to Seattle latency and RTT](/docs/network/latency/pairs/iad-sea-rtt): 62.3 ms * [Tokyo to Seattle latency and RTT](/docs/network/latency/pairs/tyo-sea-rtt): 84.8 ms * [Taipei to Seattle latency and RTT](/docs/network/latency/pairs/tpe-sea-rtt): 115.8 ms **Same corridor (North America → North America)** * [Ashburn to New York latency and RTT](/docs/network/latency/pairs/iad-nyc-rtt): 6 ms * [New York to Ashburn latency and RTT](/docs/network/latency/pairs/nyc-iad-rtt): 6 ms * [Seattle to Los Angeles latency and RTT](/docs/network/latency/pairs/sea-lax-rtt): 25.9 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/mia-sea.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Miami → Seattle round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/mia-sea-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Miami, USA → Singapore RTT 🇺🇸 **Miami, USA (MIA)** → 🇸🇬 **Singapore (SIN)** Measurement round: `2026-08-16T04:07:28Z`. Miami to Singapore averaged 223.1 ms across 50 replies. The observed range was 218.84 to 234.03 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 3.52 ms, or 1.6% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **223.1 ms** | | Jitter | **2.98 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **166.27 ms** | | Fiber Efficiency | **74.5%** | | Latency Tier | Fair | | Source Region | [North America](/docs/network/latency/regions/north-america) | | Destination Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Miami](https://www.geonames.org/4164138) (25.77427, -80.19366) and [Singapore](https://www.geonames.org/1880252) (1.28967, 103.85007). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | --------------- | | WGS-84 geodesic distance | **16,978.7 km** | | Vacuum RTT floor | **113.27 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **166.27 ms** | | Low-latency fiber material floor | **165.6 ms** | | Engineering floor (5% path allowance) | **174.6 ms** | | Research 1.33× mapped-fiber reference | **221.14 ms** | | Estimated unamplified path loss | **3565.5 dB** | | Transparent optical spans / inline amplifiers | **223 / 222** | | Published RTT inflation over fiber floor | **1.34×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **218.84 ms** | | Average RTT | **223.1 ms** | | Maximum RTT | **234.03 ms** | | Standard deviation | **3.52 ms** | | Stdev / average | **1.6%** | ## Route Context * Departure PoP: [Miami (MIA)](/docs/network/latency/mia-miami) * Destination PoP: [Singapore (SIN)](/docs/network/latency/sin-singapore) * Region overview: [North America](/docs/network/latency/regions/north-america) → [Asia Pacific](/docs/network/latency/regions/asia-pacific) * Corridor overview: [North America to Asia Pacific](/docs/network/latency/regions/north-america-to-asia-pacific) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Singapore to Miami latency and RTT](/docs/network/latency/pairs/sin-mia-rtt): 221.7 ms **Fastest routes departing Miami (MIA)** * [Miami to Ashburn latency and RTT](/docs/network/latency/pairs/mia-iad-rtt): 27.3 ms * [Miami to New York latency and RTT](/docs/network/latency/pairs/mia-nyc-rtt): 33.3 ms * [Miami to Los Angeles latency and RTT](/docs/network/latency/pairs/mia-lax-rtt): 56.8 ms * [Miami to Seattle latency and RTT](/docs/network/latency/pairs/mia-sea-rtt): 81.9 ms * [Miami to London latency and RTT](/docs/network/latency/pairs/mia-lon-rtt): 101.8 ms **Fastest routes arriving at Singapore (SIN)** * [Hong Kong to Singapore latency and RTT](/docs/network/latency/pairs/hkg-sin-rtt): 31.5 ms * [Taipei to Singapore latency and RTT](/docs/network/latency/pairs/tpe-sin-rtt): 44.3 ms * [Tokyo to Singapore latency and RTT](/docs/network/latency/pairs/tyo-sin-rtt): 69.4 ms * [Melbourne to Singapore latency and RTT](/docs/network/latency/pairs/mel-sin-rtt): 88.4 ms * [Sydney to Singapore latency and RTT](/docs/network/latency/pairs/syd-sin-rtt): 94.5 ms **Same corridor (North America → Asia Pacific)** * [Seattle to Tokyo latency and RTT](/docs/network/latency/pairs/sea-tyo-rtt): 85 ms * [Los Angeles to Tokyo latency and RTT](/docs/network/latency/pairs/lax-tyo-rtt): 101.2 ms * [Seattle to Taipei latency and RTT](/docs/network/latency/pairs/sea-tpe-rtt): 114.7 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/mia-sin.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Miami → Singapore round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/mia-sin-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Miami, USA → Sydney, Australia RTT 🇺🇸 **Miami, USA (MIA)** → 🇦🇺 **Sydney, Australia (SYD)** Measurement round: `2026-08-16T04:07:28Z`. Miami to Sydney averaged 191.5 ms across 50 replies. The observed range was 181.65 to 213.36 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 7.74 ms, or 4.0% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **191.5 ms** | | Jitter | **6.14 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **147.18 ms** | | Fiber Efficiency | **76.9%** | | Latency Tier | Fair | | Source Region | [North America](/docs/network/latency/regions/north-america) | | Destination Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Miami](https://www.geonames.org/4164138) (25.77427, -80.19366) and [Sydney](https://www.geonames.org/2147714) (-33.86785, 151.20732). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | --------------- | | WGS-84 geodesic distance | **15,029.6 km** | | Vacuum RTT floor | **100.27 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **147.18 ms** | | Low-latency fiber material floor | **146.59 ms** | | Engineering floor (5% path allowance) | **154.56 ms** | | Research 1.33× mapped-fiber reference | **195.75 ms** | | Estimated unamplified path loss | **3156.2 dB** | | Transparent optical spans / inline amplifiers | **198 / 197** | | Published RTT inflation over fiber floor | **1.3×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **181.65 ms** | | Average RTT | **191.5 ms** | | Maximum RTT | **213.36 ms** | | Standard deviation | **7.74 ms** | | Stdev / average | **4.0%** | ## Route Context * Departure PoP: [Miami (MIA)](/docs/network/latency/mia-miami) * Destination PoP: [Sydney (SYD)](/docs/network/latency/syd-sydney) * Region overview: [North America](/docs/network/latency/regions/north-america) → [Asia Pacific](/docs/network/latency/regions/asia-pacific) * Corridor overview: [North America to Asia Pacific](/docs/network/latency/regions/north-america-to-asia-pacific) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Sydney to Miami latency and RTT](/docs/network/latency/pairs/syd-mia-rtt): 191.5 ms **Fastest routes departing Miami (MIA)** * [Miami to Ashburn latency and RTT](/docs/network/latency/pairs/mia-iad-rtt): 27.3 ms * [Miami to New York latency and RTT](/docs/network/latency/pairs/mia-nyc-rtt): 33.3 ms * [Miami to Los Angeles latency and RTT](/docs/network/latency/pairs/mia-lax-rtt): 56.8 ms * [Miami to Seattle latency and RTT](/docs/network/latency/pairs/mia-sea-rtt): 81.9 ms * [Miami to London latency and RTT](/docs/network/latency/pairs/mia-lon-rtt): 101.8 ms **Fastest routes arriving at Sydney (SYD)** * [Melbourne to Sydney latency and RTT](/docs/network/latency/pairs/mel-syd-rtt): 9.8 ms * [Singapore to Sydney latency and RTT](/docs/network/latency/pairs/sin-syd-rtt): 94.5 ms * [Tokyo to Sydney latency and RTT](/docs/network/latency/pairs/tyo-syd-rtt): 101.9 ms * [Taipei to Sydney latency and RTT](/docs/network/latency/pairs/tpe-syd-rtt): 133.2 ms * [Hong Kong to Sydney latency and RTT](/docs/network/latency/pairs/hkg-syd-rtt): 137.2 ms **Same corridor (North America → Asia Pacific)** * [Seattle to Tokyo latency and RTT](/docs/network/latency/pairs/sea-tyo-rtt): 85 ms * [Los Angeles to Tokyo latency and RTT](/docs/network/latency/pairs/lax-tyo-rtt): 101.2 ms * [Seattle to Taipei latency and RTT](/docs/network/latency/pairs/sea-tpe-rtt): 114.7 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/mia-syd.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Miami → Sydney round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/mia-syd-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Miami, USA → Taipei, Taiwan RTT 🇺🇸 **Miami, USA (MIA)** → 🇹🇼 **Taipei, Taiwan (TPE)** Measurement round: `2026-08-16T04:07:28Z`. Miami to Taipei averaged 185.1 ms across 50 replies. The observed range was 178.34 to 197.2 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 5.13 ms, or 2.8% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **185.1 ms** | | Jitter | **4.64 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **136.3 ms** | | Fiber Efficiency | **73.6%** | | Latency Tier | Fair | | Source Region | [North America](/docs/network/latency/regions/north-america) | | Destination Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Miami](https://www.geonames.org/4164138) (25.77427, -80.19366) and [Taipei](https://www.geonames.org/1668341) (25.05306, 121.52639). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | --------------- | | WGS-84 geodesic distance | **13,918.7 km** | | Vacuum RTT floor | **92.86 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **136.3 ms** | | Low-latency fiber material floor | **135.76 ms** | | Engineering floor (5% path allowance) | **143.14 ms** | | Research 1.33× mapped-fiber reference | **181.28 ms** | | Estimated unamplified path loss | **2922.9 dB** | | Transparent optical spans / inline amplifiers | **183 / 182** | | Published RTT inflation over fiber floor | **1.36×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **178.34 ms** | | Average RTT | **185.1 ms** | | Maximum RTT | **197.2 ms** | | Standard deviation | **5.13 ms** | | Stdev / average | **2.8%** | ## Route Context * Departure PoP: [Miami (MIA)](/docs/network/latency/mia-miami) * Destination PoP: [Taipei (TPE)](/docs/network/latency/tpe-taipei) * Region overview: [North America](/docs/network/latency/regions/north-america) → [Asia Pacific](/docs/network/latency/regions/asia-pacific) * Corridor overview: [North America to Asia Pacific](/docs/network/latency/regions/north-america-to-asia-pacific) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Taipei to Miami latency and RTT](/docs/network/latency/pairs/tpe-mia-rtt): 185.7 ms **Fastest routes departing Miami (MIA)** * [Miami to Ashburn latency and RTT](/docs/network/latency/pairs/mia-iad-rtt): 27.3 ms * [Miami to New York latency and RTT](/docs/network/latency/pairs/mia-nyc-rtt): 33.3 ms * [Miami to Los Angeles latency and RTT](/docs/network/latency/pairs/mia-lax-rtt): 56.8 ms * [Miami to Seattle latency and RTT](/docs/network/latency/pairs/mia-sea-rtt): 81.9 ms * [Miami to London latency and RTT](/docs/network/latency/pairs/mia-lon-rtt): 101.8 ms **Fastest routes arriving at Taipei (TPE)** * [Hong Kong to Taipei latency and RTT](/docs/network/latency/pairs/hkg-tpe-rtt): 14.7 ms * [Tokyo to Taipei latency and RTT](/docs/network/latency/pairs/tyo-tpe-rtt): 32.3 ms * [Singapore to Taipei latency and RTT](/docs/network/latency/pairs/sin-tpe-rtt): 45.8 ms * [Seattle to Taipei latency and RTT](/docs/network/latency/pairs/sea-tpe-rtt): 114.7 ms * [Sydney to Taipei latency and RTT](/docs/network/latency/pairs/syd-tpe-rtt): 131.7 ms **Same corridor (North America → Asia Pacific)** * [Seattle to Tokyo latency and RTT](/docs/network/latency/pairs/sea-tyo-rtt): 85 ms * [Los Angeles to Tokyo latency and RTT](/docs/network/latency/pairs/lax-tyo-rtt): 101.2 ms * [Seattle to Hong Kong latency and RTT](/docs/network/latency/pairs/sea-hkg-rtt): 130.4 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/mia-tpe.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Miami → Taipei round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/mia-tpe-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Miami, USA → Tokyo, Japan RTT 🇺🇸 **Miami, USA (MIA)** → 🇯🇵 **Tokyo, Japan (TYO)** Measurement round: `2026-08-16T04:07:28Z`. Miami to Tokyo averaged 156.3 ms across 50 replies. The observed range was 149.17 to 178.72 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 6.19 ms, or 4.0% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **156.3 ms** | | Jitter | **4.99 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **117.71 ms** | | Fiber Efficiency | **75.3%** | | Latency Tier | Fair | | Source Region | [North America](/docs/network/latency/regions/north-america) | | Destination Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Miami](https://www.geonames.org/4164138) (25.77427, -80.19366) and [Tokyo](https://www.geonames.org/1850147) (35.68950, 139.69171). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | --------------- | | WGS-84 geodesic distance | **12,019.9 km** | | Vacuum RTT floor | **80.19 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **117.71 ms** | | Low-latency fiber material floor | **117.24 ms** | | Engineering floor (5% path allowance) | **123.61 ms** | | Research 1.33× mapped-fiber reference | **156.55 ms** | | Estimated unamplified path loss | **2524.2 dB** | | Transparent optical spans / inline amplifiers | **158 / 157** | | Published RTT inflation over fiber floor | **1.33×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **149.17 ms** | | Average RTT | **156.3 ms** | | Maximum RTT | **178.72 ms** | | Standard deviation | **6.19 ms** | | Stdev / average | **4.0%** | ## Route Context * Departure PoP: [Miami (MIA)](/docs/network/latency/mia-miami) * Destination PoP: [Tokyo (TYO)](/docs/network/latency/tyo-tokyo) * Region overview: [North America](/docs/network/latency/regions/north-america) → [Asia Pacific](/docs/network/latency/regions/asia-pacific) * Corridor overview: [North America to Asia Pacific](/docs/network/latency/regions/north-america-to-asia-pacific) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Tokyo to Miami latency and RTT](/docs/network/latency/pairs/tyo-mia-rtt): 156.5 ms **Fastest routes departing Miami (MIA)** * [Miami to Ashburn latency and RTT](/docs/network/latency/pairs/mia-iad-rtt): 27.3 ms * [Miami to New York latency and RTT](/docs/network/latency/pairs/mia-nyc-rtt): 33.3 ms * [Miami to Los Angeles latency and RTT](/docs/network/latency/pairs/mia-lax-rtt): 56.8 ms * [Miami to Seattle latency and RTT](/docs/network/latency/pairs/mia-sea-rtt): 81.9 ms * [Miami to London latency and RTT](/docs/network/latency/pairs/mia-lon-rtt): 101.8 ms **Fastest routes arriving at Tokyo (TYO)** * [Taipei to Tokyo latency and RTT](/docs/network/latency/pairs/tpe-tyo-rtt): 31.9 ms * [Hong Kong to Tokyo latency and RTT](/docs/network/latency/pairs/hkg-tyo-rtt): 44.9 ms * [Singapore to Tokyo latency and RTT](/docs/network/latency/pairs/sin-tyo-rtt): 68.9 ms * [Seattle to Tokyo latency and RTT](/docs/network/latency/pairs/sea-tyo-rtt): 85 ms * [Los Angeles to Tokyo latency and RTT](/docs/network/latency/pairs/lax-tyo-rtt): 101.2 ms **Same corridor (North America → Asia Pacific)** * [Seattle to Taipei latency and RTT](/docs/network/latency/pairs/sea-tpe-rtt): 114.7 ms * [Seattle to Hong Kong latency and RTT](/docs/network/latency/pairs/sea-hkg-rtt): 130.4 ms * [Los Angeles to Taipei latency and RTT](/docs/network/latency/pairs/lax-tpe-rtt): 132.9 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/mia-tyo.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Miami → Tokyo round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/mia-tyo-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Moscow, Russia → Amsterdam, Netherlands RTT 🇷🇺 **Moscow, Russia (MOW)** → 🇳🇱 **Amsterdam, Netherlands (AMS)** Measurement round: `2026-08-16T04:07:28Z`. Moscow to Amsterdam averaged 37.6 ms across 50 replies. The observed range was 37.09 to 39.9 ms; these figures describe one probe round, not a long-term service target. This is number 3 of 19 outgoing routes by average RTT. That ranking compares different destinations and is not adjusted for distance. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------- | | RTT | **37.6 ms** | | Jitter | **0.36 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **21.1 ms** | | Fiber Efficiency | **56.1%** | | Latency Tier | Excellent | | Source Region | [Europe](/docs/network/latency/regions/europe) | | Destination Region | [Europe](/docs/network/latency/regions/europe) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Moscow](https://www.geonames.org/524901) (55.75204, 37.61781) and [Amsterdam](https://www.geonames.org/2759794) (52.37403, 4.88969). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **2,154.6 km** | | Vacuum RTT floor | **14.37 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **21.1 ms** | | Low-latency fiber material floor | **21.02 ms** | | Engineering floor (5% path allowance) | **22.16 ms** | | Research 1.33× mapped-fiber reference | **28.06 ms** | | Estimated unamplified path loss | **452.5 dB** | | Transparent optical spans / inline amplifiers | **29 / 28** | | Published RTT inflation over fiber floor | **1.78×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------ | | Minimum RTT | **37.09 ms** | | Average RTT | **37.6 ms** | | Maximum RTT | **39.9 ms** | | Standard deviation | **0.57 ms** | | Stdev / average | **1.5%** | ## Route Context * Departure PoP: [Moscow (MOW)](/docs/network/latency/mow-moscow) * Destination PoP: [Amsterdam (AMS)](/docs/network/latency/ams-amsterdam) * Region overview: [Europe](/docs/network/latency/regions/europe) → [Europe](/docs/network/latency/regions/europe) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Amsterdam to Moscow latency and RTT](/docs/network/latency/pairs/ams-mow-rtt): 38.1 ms **Fastest routes departing Moscow (MOW)** * [Moscow to Berlin latency and RTT](/docs/network/latency/pairs/mow-ber-rtt): 27.5 ms * [Moscow to Frankfurt latency and RTT](/docs/network/latency/pairs/mow-fra-rtt): 35.9 ms * [Moscow to London latency and RTT](/docs/network/latency/pairs/mow-lon-rtt): 42.4 ms * [Moscow to Paris latency and RTT](/docs/network/latency/pairs/mow-par-rtt): 44 ms * [Moscow to Marseille latency and RTT](/docs/network/latency/pairs/mow-mrs-rtt): 49.9 ms **Fastest routes arriving at Amsterdam (AMS)** * [London to Amsterdam latency and RTT](/docs/network/latency/pairs/lon-ams-rtt): 5.2 ms * [Frankfurt to Amsterdam latency and RTT](/docs/network/latency/pairs/fra-ams-rtt): 5.9 ms * [Paris to Amsterdam latency and RTT](/docs/network/latency/pairs/par-ams-rtt): 7 ms * [Berlin to Amsterdam latency and RTT](/docs/network/latency/pairs/ber-ams-rtt): 7.2 ms * [Marseille to Amsterdam latency and RTT](/docs/network/latency/pairs/mrs-ams-rtt): 20.3 ms **Same corridor (Europe → Europe)** * [Amsterdam to London latency and RTT](/docs/network/latency/pairs/ams-lon-rtt): 5.2 ms * [Amsterdam to Frankfurt latency and RTT](/docs/network/latency/pairs/ams-fra-rtt): 5.9 ms * [Berlin to Frankfurt latency and RTT](/docs/network/latency/pairs/ber-fra-rtt): 6.1 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/mow-ams.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Moscow → Amsterdam round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/mow-ams-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Moscow, Russia → Berlin, Germany RTT 🇷🇺 **Moscow, Russia (MOW)** → 🇩🇪 **Berlin, Germany (BER)** Measurement round: `2026-08-16T04:07:28Z`. Moscow to Berlin averaged 27.5 ms across 50 replies. The observed range was 26.72 to 29.96 ms; these figures describe one probe round, not a long-term service target. This is number 1 of 19 outgoing routes by average RTT. That ranking compares different destinations and is not adjusted for distance. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------- | | RTT | **27.5 ms** | | Jitter | **0.54 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **15.8 ms** | | Fiber Efficiency | **57.5%** | | Latency Tier | Ultra-Low | | Source Region | [Europe](/docs/network/latency/regions/europe) | | Destination Region | [Europe](/docs/network/latency/regions/europe) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Moscow](https://www.geonames.org/524901) (55.75204, 37.61781) and [Berlin](https://www.geonames.org/2950159) (52.52437, 13.41053). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **1,613.5 km** | | Vacuum RTT floor | **10.76 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **15.8 ms** | | Low-latency fiber material floor | **15.74 ms** | | Engineering floor (5% path allowance) | **16.59 ms** | | Research 1.33× mapped-fiber reference | **21.01 ms** | | Estimated unamplified path loss | **338.8 dB** | | Transparent optical spans / inline amplifiers | **22 / 21** | | Published RTT inflation over fiber floor | **1.74×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------ | | Minimum RTT | **26.72 ms** | | Average RTT | **27.5 ms** | | Maximum RTT | **29.96 ms** | | Standard deviation | **0.74 ms** | | Stdev / average | **2.7%** | ## Route Context * Departure PoP: [Moscow (MOW)](/docs/network/latency/mow-moscow) * Destination PoP: [Berlin (BER)](/docs/network/latency/ber-berlin) * Region overview: [Europe](/docs/network/latency/regions/europe) → [Europe](/docs/network/latency/regions/europe) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Berlin to Moscow latency and RTT](/docs/network/latency/pairs/ber-mow-rtt): 28.6 ms **Fastest routes departing Moscow (MOW)** * [Moscow to Frankfurt latency and RTT](/docs/network/latency/pairs/mow-fra-rtt): 35.9 ms * [Moscow to Amsterdam latency and RTT](/docs/network/latency/pairs/mow-ams-rtt): 37.6 ms * [Moscow to London latency and RTT](/docs/network/latency/pairs/mow-lon-rtt): 42.4 ms * [Moscow to Paris latency and RTT](/docs/network/latency/pairs/mow-par-rtt): 44 ms * [Moscow to Marseille latency and RTT](/docs/network/latency/pairs/mow-mrs-rtt): 49.9 ms **Fastest routes arriving at Berlin (BER)** * [Frankfurt to Berlin latency and RTT](/docs/network/latency/pairs/fra-ber-rtt): 6.1 ms * [Amsterdam to Berlin latency and RTT](/docs/network/latency/pairs/ams-ber-rtt): 7.2 ms * [Paris to Berlin latency and RTT](/docs/network/latency/pairs/par-ber-rtt): 15.5 ms * [London to Berlin latency and RTT](/docs/network/latency/pairs/lon-ber-rtt): 16.2 ms * [Marseille to Berlin latency and RTT](/docs/network/latency/pairs/mrs-ber-rtt): 20.6 ms **Same corridor (Europe → Europe)** * [Amsterdam to London latency and RTT](/docs/network/latency/pairs/ams-lon-rtt): 5.2 ms * [London to Amsterdam latency and RTT](/docs/network/latency/pairs/lon-ams-rtt): 5.2 ms * [Amsterdam to Frankfurt latency and RTT](/docs/network/latency/pairs/ams-fra-rtt): 5.9 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/mow-ber.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Moscow → Berlin round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/mow-ber-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Moscow, Russia → Frankfurt, Germany RTT 🇷🇺 **Moscow, Russia (MOW)** → 🇩🇪 **Frankfurt, Germany (FRA)** Measurement round: `2026-08-16T04:07:28Z`. Moscow to Frankfurt averaged 35.9 ms across 50 replies. The observed range was 33.56 to 42.93 ms; these figures describe one probe round, not a long-term service target. This is number 2 of 19 outgoing routes by average RTT. That ranking compares different destinations and is not adjusted for distance. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------- | | RTT | **35.9 ms** | | Jitter | **1.52 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **19.85 ms** | | Fiber Efficiency | **55.3%** | | Latency Tier | Excellent | | Source Region | [Europe](/docs/network/latency/regions/europe) | | Destination Region | [Europe](/docs/network/latency/regions/europe) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Moscow](https://www.geonames.org/524901) (55.75204, 37.61781) and [Frankfurt am Main](https://www.geonames.org/2925533) (50.11552, 8.68417). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **2,026.8 km** | | Vacuum RTT floor | **13.52 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **19.85 ms** | | Low-latency fiber material floor | **19.77 ms** | | Engineering floor (5% path allowance) | **20.84 ms** | | Research 1.33× mapped-fiber reference | **26.4 ms** | | Estimated unamplified path loss | **425.6 dB** | | Transparent optical spans / inline amplifiers | **27 / 26** | | Published RTT inflation over fiber floor | **1.81×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------ | | Minimum RTT | **33.56 ms** | | Average RTT | **35.9 ms** | | Maximum RTT | **42.93 ms** | | Standard deviation | **1.78 ms** | | Stdev / average | **5.0%** | ## Route Context * Departure PoP: [Moscow (MOW)](/docs/network/latency/mow-moscow) * Destination PoP: [Frankfurt (FRA)](/docs/network/latency/fra-frankfurt) * Region overview: [Europe](/docs/network/latency/regions/europe) → [Europe](/docs/network/latency/regions/europe) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Frankfurt to Moscow latency and RTT](/docs/network/latency/pairs/fra-mow-rtt): 34.7 ms **Fastest routes departing Moscow (MOW)** * [Moscow to Berlin latency and RTT](/docs/network/latency/pairs/mow-ber-rtt): 27.5 ms * [Moscow to Amsterdam latency and RTT](/docs/network/latency/pairs/mow-ams-rtt): 37.6 ms * [Moscow to London latency and RTT](/docs/network/latency/pairs/mow-lon-rtt): 42.4 ms * [Moscow to Paris latency and RTT](/docs/network/latency/pairs/mow-par-rtt): 44 ms * [Moscow to Marseille latency and RTT](/docs/network/latency/pairs/mow-mrs-rtt): 49.9 ms **Fastest routes arriving at Frankfurt (FRA)** * [Amsterdam to Frankfurt latency and RTT](/docs/network/latency/pairs/ams-fra-rtt): 5.9 ms * [Berlin to Frankfurt latency and RTT](/docs/network/latency/pairs/ber-fra-rtt): 6.1 ms * [Paris to Frankfurt latency and RTT](/docs/network/latency/pairs/par-fra-rtt): 7.6 ms * [London to Frankfurt latency and RTT](/docs/network/latency/pairs/lon-fra-rtt): 13.6 ms * [Marseille to Frankfurt latency and RTT](/docs/network/latency/pairs/mrs-fra-rtt): 16 ms **Same corridor (Europe → Europe)** * [Amsterdam to London latency and RTT](/docs/network/latency/pairs/ams-lon-rtt): 5.2 ms * [London to Amsterdam latency and RTT](/docs/network/latency/pairs/lon-ams-rtt): 5.2 ms * [Frankfurt to Amsterdam latency and RTT](/docs/network/latency/pairs/fra-ams-rtt): 5.9 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/mow-fra.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Moscow → Frankfurt round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/mow-fra-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Moscow, Russia → São Paulo, Brazil RTT 🇷🇺 **Moscow, Russia (MOW)** → 🇧🇷 **São Paulo, Brazil (GRU)** Measurement round: `2026-08-16T04:07:28Z`. Moscow to São Paulo averaged 214.4 ms across 50 replies. The observed range was 203.61 to 257.23 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 10.55 ms, or 4.9% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **214.4 ms** | | Jitter | **8.94 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **115.47 ms** | | Fiber Efficiency | **53.9%** | | Latency Tier | Fair | | Source Region | [Europe](/docs/network/latency/regions/europe) | | Destination Region | [South America](/docs/network/latency/regions/south-america) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Moscow](https://www.geonames.org/524901) (55.75204, 37.61781) and [São Paulo](https://www.geonames.org/3448439) (-23.54750, -46.63611). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | --------------- | | WGS-84 geodesic distance | **11,791.1 km** | | Vacuum RTT floor | **78.66 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **115.47 ms** | | Low-latency fiber material floor | **115 ms** | | Engineering floor (5% path allowance) | **121.26 ms** | | Research 1.33× mapped-fiber reference | **153.57 ms** | | Estimated unamplified path loss | **2476.1 dB** | | Transparent optical spans / inline amplifiers | **155 / 154** | | Published RTT inflation over fiber floor | **1.86×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **203.61 ms** | | Average RTT | **214.4 ms** | | Maximum RTT | **257.23 ms** | | Standard deviation | **10.55 ms** | | Stdev / average | **4.9%** | ## Route Context * Departure PoP: [Moscow (MOW)](/docs/network/latency/mow-moscow) * Destination PoP: [São Paulo (GRU)](/docs/network/latency/gru-sao-paulo) * Region overview: [Europe](/docs/network/latency/regions/europe) → [South America](/docs/network/latency/regions/south-america) * Corridor overview: [Europe to South America](/docs/network/latency/regions/europe-to-south-america) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [São Paulo to Moscow latency and RTT](/docs/network/latency/pairs/gru-mow-rtt): 219.5 ms **Fastest routes departing Moscow (MOW)** * [Moscow to Berlin latency and RTT](/docs/network/latency/pairs/mow-ber-rtt): 27.5 ms * [Moscow to Frankfurt latency and RTT](/docs/network/latency/pairs/mow-fra-rtt): 35.9 ms * [Moscow to Amsterdam latency and RTT](/docs/network/latency/pairs/mow-ams-rtt): 37.6 ms * [Moscow to London latency and RTT](/docs/network/latency/pairs/mow-lon-rtt): 42.4 ms * [Moscow to Paris latency and RTT](/docs/network/latency/pairs/mow-par-rtt): 44 ms **Fastest routes arriving at São Paulo (GRU)** * [Ashburn to São Paulo latency and RTT](/docs/network/latency/pairs/iad-gru-rtt): 101.4 ms * [New York to São Paulo latency and RTT](/docs/network/latency/pairs/nyc-gru-rtt): 106.9 ms * [Miami to São Paulo latency and RTT](/docs/network/latency/pairs/mia-gru-rtt): 128.7 ms * [Los Angeles to São Paulo latency and RTT](/docs/network/latency/pairs/lax-gru-rtt): 131.3 ms * [Seattle to São Paulo latency and RTT](/docs/network/latency/pairs/sea-gru-rtt): 155.7 ms **Same corridor (Europe → South America)** * [London to São Paulo latency and RTT](/docs/network/latency/pairs/lon-gru-rtt): 170.7 ms * [Amsterdam to São Paulo latency and RTT](/docs/network/latency/pairs/ams-gru-rtt): 175.5 ms * [Paris to São Paulo latency and RTT](/docs/network/latency/pairs/par-gru-rtt): 175.8 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/mow-gru.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Moscow → São Paulo round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/mow-gru-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Moscow, Russia → Hong Kong RTT 🇷🇺 **Moscow, Russia (MOW)** → 🇭🇰 **Hong Kong (HKG)** Measurement round: `2026-08-16T04:07:28Z`. Moscow to Hong Kong averaged 118.6 ms across 50 replies. The observed range was 116.41 to 127.8 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 2.24 ms, or 1.9% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------------------- | | RTT | **118.6 ms** | | Jitter | **1.64 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **70.05 ms** | | Fiber Efficiency | **59.1%** | | Latency Tier | Good | | Source Region | [Europe](/docs/network/latency/regions/europe) | | Destination Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Moscow](https://www.geonames.org/524901) (55.75204, 37.61781) and [Hong Kong](https://www.geonames.org/1819729) (22.27832, 114.17469). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **7,153.6 km** | | Vacuum RTT floor | **47.72 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **70.05 ms** | | Low-latency fiber material floor | **69.77 ms** | | Engineering floor (5% path allowance) | **73.57 ms** | | Research 1.33× mapped-fiber reference | **93.17 ms** | | Estimated unamplified path loss | **1502.3 dB** | | Transparent optical spans / inline amplifiers | **94 / 93** | | Published RTT inflation over fiber floor | **1.69×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **116.41 ms** | | Average RTT | **118.6 ms** | | Maximum RTT | **127.8 ms** | | Standard deviation | **2.24 ms** | | Stdev / average | **1.9%** | ## Route Context * Departure PoP: [Moscow (MOW)](/docs/network/latency/mow-moscow) * Destination PoP: [Hong Kong (HKG)](/docs/network/latency/hkg-hong-kong) * Region overview: [Europe](/docs/network/latency/regions/europe) → [Asia Pacific](/docs/network/latency/regions/asia-pacific) * Corridor overview: [Europe to Asia Pacific](/docs/network/latency/regions/europe-to-asia-pacific) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Hong Kong to Moscow latency and RTT](/docs/network/latency/pairs/hkg-mow-rtt): 118.2 ms **Fastest routes departing Moscow (MOW)** * [Moscow to Berlin latency and RTT](/docs/network/latency/pairs/mow-ber-rtt): 27.5 ms * [Moscow to Frankfurt latency and RTT](/docs/network/latency/pairs/mow-fra-rtt): 35.9 ms * [Moscow to Amsterdam latency and RTT](/docs/network/latency/pairs/mow-ams-rtt): 37.6 ms * [Moscow to London latency and RTT](/docs/network/latency/pairs/mow-lon-rtt): 42.4 ms * [Moscow to Paris latency and RTT](/docs/network/latency/pairs/mow-par-rtt): 44 ms **Fastest routes arriving at Hong Kong (HKG)** * [Taipei to Hong Kong latency and RTT](/docs/network/latency/pairs/tpe-hkg-rtt): 15.5 ms * [Singapore to Hong Kong latency and RTT](/docs/network/latency/pairs/sin-hkg-rtt): 30.8 ms * [Tokyo to Hong Kong latency and RTT](/docs/network/latency/pairs/tyo-hkg-rtt): 44.8 ms * [Seattle to Hong Kong latency and RTT](/docs/network/latency/pairs/sea-hkg-rtt): 130.4 ms * [Sydney to Hong Kong latency and RTT](/docs/network/latency/pairs/syd-hkg-rtt): 135.7 ms **Same corridor (Europe → Asia Pacific)** * [Moscow to Taipei latency and RTT](/docs/network/latency/pairs/mow-tpe-rtt): 132.1 ms * [Marseille to Singapore latency and RTT](/docs/network/latency/pairs/mrs-sin-rtt): 139.6 ms * [Berlin to Hong Kong latency and RTT](/docs/network/latency/pairs/ber-hkg-rtt): 144.9 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/mow-hkg.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Moscow → Hong Kong round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/mow-hkg-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Moscow, Russia → Ashburn, USA RTT 🇷🇺 **Moscow, Russia (MOW)** → 🇺🇸 **Ashburn, USA (IAD)** Measurement round: `2026-08-16T04:07:28Z`. Moscow to Ashburn averaged 113.9 ms across 50 replies. The observed range was 107.82 to 129.74 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 5.37 ms, or 4.7% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **113.9 ms** | | Jitter | **4.1 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **76.88 ms** | | Fiber Efficiency | **67.5%** | | Latency Tier | Good | | Source Region | [Europe](/docs/network/latency/regions/europe) | | Destination Region | [North America](/docs/network/latency/regions/north-america) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Moscow](https://www.geonames.org/524901) (55.75204, 37.61781) and [Ashburn](https://www.geonames.org/4744870) (39.04372, -77.48749). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **7,850.6 km** | | Vacuum RTT floor | **52.37 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **76.88 ms** | | Low-latency fiber material floor | **76.57 ms** | | Engineering floor (5% path allowance) | **80.73 ms** | | Research 1.33× mapped-fiber reference | **102.25 ms** | | Estimated unamplified path loss | **1648.6 dB** | | Transparent optical spans / inline amplifiers | **104 / 103** | | Published RTT inflation over fiber floor | **1.48×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **107.82 ms** | | Average RTT | **113.9 ms** | | Maximum RTT | **129.74 ms** | | Standard deviation | **5.37 ms** | | Stdev / average | **4.7%** | ## Route Context * Departure PoP: [Moscow (MOW)](/docs/network/latency/mow-moscow) * Destination PoP: [Ashburn (IAD)](/docs/network/latency/iad-ashburn) * Region overview: [Europe](/docs/network/latency/regions/europe) → [North America](/docs/network/latency/regions/north-america) * Corridor overview: [Europe to North America](/docs/network/latency/regions/europe-to-north-america) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Ashburn to Moscow latency and RTT](/docs/network/latency/pairs/iad-mow-rtt): 112.8 ms **Fastest routes departing Moscow (MOW)** * [Moscow to Berlin latency and RTT](/docs/network/latency/pairs/mow-ber-rtt): 27.5 ms * [Moscow to Frankfurt latency and RTT](/docs/network/latency/pairs/mow-fra-rtt): 35.9 ms * [Moscow to Amsterdam latency and RTT](/docs/network/latency/pairs/mow-ams-rtt): 37.6 ms * [Moscow to London latency and RTT](/docs/network/latency/pairs/mow-lon-rtt): 42.4 ms * [Moscow to Paris latency and RTT](/docs/network/latency/pairs/mow-par-rtt): 44 ms **Fastest routes arriving at Ashburn (IAD)** * [New York to Ashburn latency and RTT](/docs/network/latency/pairs/nyc-iad-rtt): 6 ms * [Miami to Ashburn latency and RTT](/docs/network/latency/pairs/mia-iad-rtt): 27.3 ms * [Los Angeles to Ashburn latency and RTT](/docs/network/latency/pairs/lax-iad-rtt): 60.4 ms * [Seattle to Ashburn latency and RTT](/docs/network/latency/pairs/sea-iad-rtt): 61.5 ms * [London to Ashburn latency and RTT](/docs/network/latency/pairs/lon-iad-rtt): 71 ms **Same corridor (Europe → North America)** * [London to New York latency and RTT](/docs/network/latency/pairs/lon-nyc-rtt): 63.8 ms * [Amsterdam to New York latency and RTT](/docs/network/latency/pairs/ams-nyc-rtt): 68.6 ms * [Paris to New York latency and RTT](/docs/network/latency/pairs/par-nyc-rtt): 68.9 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/mow-iad.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Moscow → Ashburn round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/mow-iad-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Moscow, Russia → Johannesburg, South Africa RTT 🇷🇺 **Moscow, Russia (MOW)** → 🇿🇦 **Johannesburg, South Africa (JNB)** Measurement round: `2026-08-16T04:07:28Z`. Moscow to Johannesburg averaged 201.9 ms, compared with a 89.37 ms city-centre fiber reference. The difference cannot be assigned to a particular cable or carrier from RTT alone. Standard deviation was 7.94 ms, or 3.9% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------- | | RTT | **201.9 ms** | | Jitter | **7.49 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **89.37 ms** | | Fiber Efficiency | **44.3%** | | Latency Tier | Fair | | Source Region | [Europe](/docs/network/latency/regions/europe) | | Destination Region | [Africa](/docs/network/latency/regions/africa) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Moscow](https://www.geonames.org/524901) (55.75204, 37.61781) and [Johannesburg](https://www.geonames.org/993800) (-26.20227, 28.04363). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **9,125.7 km** | | Vacuum RTT floor | **60.88 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **89.37 ms** | | Low-latency fiber material floor | **89.01 ms** | | Engineering floor (5% path allowance) | **93.85 ms** | | Research 1.33× mapped-fiber reference | **118.86 ms** | | Estimated unamplified path loss | **1916.4 dB** | | Transparent optical spans / inline amplifiers | **120 / 119** | | Published RTT inflation over fiber floor | **2.26×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **192.01 ms** | | Average RTT | **201.9 ms** | | Maximum RTT | **232.17 ms** | | Standard deviation | **7.94 ms** | | Stdev / average | **3.9%** | ## Route Context * Departure PoP: [Moscow (MOW)](/docs/network/latency/mow-moscow) * Destination PoP: [Johannesburg (JNB)](/docs/network/latency/jnb-johannesburg) * Region overview: [Europe](/docs/network/latency/regions/europe) → [Africa](/docs/network/latency/regions/africa) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Johannesburg to Moscow latency and RTT](/docs/network/latency/pairs/jnb-mow-rtt): 200.7 ms **Fastest routes departing Moscow (MOW)** * [Moscow to Berlin latency and RTT](/docs/network/latency/pairs/mow-ber-rtt): 27.5 ms * [Moscow to Frankfurt latency and RTT](/docs/network/latency/pairs/mow-fra-rtt): 35.9 ms * [Moscow to Amsterdam latency and RTT](/docs/network/latency/pairs/mow-ams-rtt): 37.6 ms * [Moscow to London latency and RTT](/docs/network/latency/pairs/mow-lon-rtt): 42.4 ms * [Moscow to Paris latency and RTT](/docs/network/latency/pairs/mow-par-rtt): 44 ms **Fastest routes arriving at Johannesburg (JNB)** * [Amsterdam to Johannesburg latency and RTT](/docs/network/latency/pairs/ams-jnb-rtt): 171.9 ms * [Berlin to Johannesburg latency and RTT](/docs/network/latency/pairs/ber-jnb-rtt): 172.1 ms * [London to Johannesburg latency and RTT](/docs/network/latency/pairs/lon-jnb-rtt): 179.6 ms * [Paris to Johannesburg latency and RTT](/docs/network/latency/pairs/par-jnb-rtt): 180.9 ms * [Frankfurt to Johannesburg latency and RTT](/docs/network/latency/pairs/fra-jnb-rtt): 188.4 ms **Same corridor (Europe → Africa)** * [Marseille to Johannesburg latency and RTT](/docs/network/latency/pairs/mrs-jnb-rtt): 198.5 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/mow-jnb.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Moscow → Johannesburg round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/mow-jnb-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Moscow, Russia → Los Angeles, USA RTT 🇷🇺 **Moscow, Russia (MOW)** → 🇺🇸 **Los Angeles, USA (LAX)** Measurement round: `2026-08-16T04:07:28Z`. Moscow to Los Angeles averaged 177.7 ms across 50 replies. The observed range was 173.78 to 191.21 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 3.21 ms, or 1.8% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **177.7 ms** | | Jitter | **2.25 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **95.9 ms** | | Fiber Efficiency | **54%** | | Latency Tier | Fair | | Source Region | [Europe](/docs/network/latency/regions/europe) | | Destination Region | [North America](/docs/network/latency/regions/north-america) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Moscow](https://www.geonames.org/524901) (55.75204, 37.61781) and [Los Angeles](https://www.geonames.org/5368361) (34.05223, -118.24368). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **9,793.2 km** | | Vacuum RTT floor | **65.33 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **95.9 ms** | | Low-latency fiber material floor | **95.52 ms** | | Engineering floor (5% path allowance) | **100.71 ms** | | Research 1.33× mapped-fiber reference | **127.55 ms** | | Estimated unamplified path loss | **2056.6 dB** | | Transparent optical spans / inline amplifiers | **129 / 128** | | Published RTT inflation over fiber floor | **1.85×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **173.78 ms** | | Average RTT | **177.7 ms** | | Maximum RTT | **191.21 ms** | | Standard deviation | **3.21 ms** | | Stdev / average | **1.8%** | ## Route Context * Departure PoP: [Moscow (MOW)](/docs/network/latency/mow-moscow) * Destination PoP: [Los Angeles (LAX)](/docs/network/latency/lax-los-angeles) * Region overview: [Europe](/docs/network/latency/regions/europe) → [North America](/docs/network/latency/regions/north-america) * Corridor overview: [Europe to North America](/docs/network/latency/regions/europe-to-north-america) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Los Angeles to Moscow latency and RTT](/docs/network/latency/pairs/lax-mow-rtt): 177 ms **Fastest routes departing Moscow (MOW)** * [Moscow to Berlin latency and RTT](/docs/network/latency/pairs/mow-ber-rtt): 27.5 ms * [Moscow to Frankfurt latency and RTT](/docs/network/latency/pairs/mow-fra-rtt): 35.9 ms * [Moscow to Amsterdam latency and RTT](/docs/network/latency/pairs/mow-ams-rtt): 37.6 ms * [Moscow to London latency and RTT](/docs/network/latency/pairs/mow-lon-rtt): 42.4 ms * [Moscow to Paris latency and RTT](/docs/network/latency/pairs/mow-par-rtt): 44 ms **Fastest routes arriving at Los Angeles (LAX)** * [Seattle to Los Angeles latency and RTT](/docs/network/latency/pairs/sea-lax-rtt): 25.9 ms * [Miami to Los Angeles latency and RTT](/docs/network/latency/pairs/mia-lax-rtt): 56.8 ms * [New York to Los Angeles latency and RTT](/docs/network/latency/pairs/nyc-lax-rtt): 58.9 ms * [Ashburn to Los Angeles latency and RTT](/docs/network/latency/pairs/iad-lax-rtt): 60.7 ms * [Tokyo to Los Angeles latency and RTT](/docs/network/latency/pairs/tyo-lax-rtt): 101.2 ms **Same corridor (Europe → North America)** * [London to New York latency and RTT](/docs/network/latency/pairs/lon-nyc-rtt): 63.8 ms * [Amsterdam to New York latency and RTT](/docs/network/latency/pairs/ams-nyc-rtt): 68.6 ms * [Paris to New York latency and RTT](/docs/network/latency/pairs/par-nyc-rtt): 68.9 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/mow-lax.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Moscow → Los Angeles round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/mow-lax-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Moscow, Russia → London, UK RTT 🇷🇺 **Moscow, Russia (MOW)** → 🇬🇧 **London, UK (LON)** Measurement round: `2026-08-16T04:07:28Z`. Moscow to London averaged 42.4 ms across 50 replies. The observed range was 40.38 to 48.72 ms; these figures describe one probe round, not a long-term service target. This is number 4 of 19 outgoing routes by average RTT. That ranking compares different destinations and is not adjusted for distance. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------- | | RTT | **42.4 ms** | | Jitter | **1.31 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **24.56 ms** | | Fiber Efficiency | **57.9%** | | Latency Tier | Excellent | | Source Region | [Europe](/docs/network/latency/regions/europe) | | Destination Region | [Europe](/docs/network/latency/regions/europe) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Moscow](https://www.geonames.org/524901) (55.75204, 37.61781) and [London](https://www.geonames.org/2643743) (51.50853, -0.12574). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **2,508.5 km** | | Vacuum RTT floor | **16.73 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **24.56 ms** | | Low-latency fiber material floor | **24.47 ms** | | Engineering floor (5% path allowance) | **25.8 ms** | | Research 1.33× mapped-fiber reference | **32.67 ms** | | Estimated unamplified path loss | **526.8 dB** | | Transparent optical spans / inline amplifiers | **33 / 32** | | Published RTT inflation over fiber floor | **1.73×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------ | | Minimum RTT | **40.38 ms** | | Average RTT | **42.4 ms** | | Maximum RTT | **48.72 ms** | | Standard deviation | **1.83 ms** | | Stdev / average | **4.3%** | ## Route Context * Departure PoP: [Moscow (MOW)](/docs/network/latency/mow-moscow) * Destination PoP: [London (LON)](/docs/network/latency/lon-london) * Region overview: [Europe](/docs/network/latency/regions/europe) → [Europe](/docs/network/latency/regions/europe) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [London to Moscow latency and RTT](/docs/network/latency/pairs/lon-mow-rtt): 43.5 ms **Fastest routes departing Moscow (MOW)** * [Moscow to Berlin latency and RTT](/docs/network/latency/pairs/mow-ber-rtt): 27.5 ms * [Moscow to Frankfurt latency and RTT](/docs/network/latency/pairs/mow-fra-rtt): 35.9 ms * [Moscow to Amsterdam latency and RTT](/docs/network/latency/pairs/mow-ams-rtt): 37.6 ms * [Moscow to Paris latency and RTT](/docs/network/latency/pairs/mow-par-rtt): 44 ms * [Moscow to Marseille latency and RTT](/docs/network/latency/pairs/mow-mrs-rtt): 49.9 ms **Fastest routes arriving at London (LON)** * [Amsterdam to London latency and RTT](/docs/network/latency/pairs/ams-lon-rtt): 5.2 ms * [Paris to London latency and RTT](/docs/network/latency/pairs/par-lon-rtt): 6.4 ms * [Frankfurt to London latency and RTT](/docs/network/latency/pairs/fra-lon-rtt): 12.9 ms * [Berlin to London latency and RTT](/docs/network/latency/pairs/ber-lon-rtt): 15.2 ms * [Marseille to London latency and RTT](/docs/network/latency/pairs/mrs-lon-rtt): 18.9 ms **Same corridor (Europe → Europe)** * [London to Amsterdam latency and RTT](/docs/network/latency/pairs/lon-ams-rtt): 5.2 ms * [Amsterdam to Frankfurt latency and RTT](/docs/network/latency/pairs/ams-fra-rtt): 5.9 ms * [Frankfurt to Amsterdam latency and RTT](/docs/network/latency/pairs/fra-ams-rtt): 5.9 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/mow-lon.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Moscow → London round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/mow-lon-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Moscow, Russia → Melbourne, Australia RTT 🇷🇺 **Moscow, Russia (MOW)** → 🇦🇺 **Melbourne, Australia (MEL)** Measurement round: `2026-08-16T04:07:28Z`. Moscow to Melbourne averaged 235.9 ms across 50 replies. The observed range was 223.93 to 262.01 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 8.78 ms, or 3.7% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------------------- | | RTT | **235.9 ms** | | Jitter | **7.29 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **141.13 ms** | | Fiber Efficiency | **59.8%** | | Latency Tier | Fair | | Source Region | [Europe](/docs/network/latency/regions/europe) | | Destination Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Moscow](https://www.geonames.org/524901) (55.75204, 37.61781) and [Melbourne](https://www.geonames.org/2158177) (-37.81400, 144.96332). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | --------------- | | WGS-84 geodesic distance | **14,412.1 km** | | Vacuum RTT floor | **96.15 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **141.13 ms** | | Low-latency fiber material floor | **140.57 ms** | | Engineering floor (5% path allowance) | **148.21 ms** | | Research 1.33× mapped-fiber reference | **187.71 ms** | | Estimated unamplified path loss | **3026.5 dB** | | Transparent optical spans / inline amplifiers | **190 / 189** | | Published RTT inflation over fiber floor | **1.67×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **223.93 ms** | | Average RTT | **235.9 ms** | | Maximum RTT | **262.01 ms** | | Standard deviation | **8.78 ms** | | Stdev / average | **3.7%** | ## Route Context * Departure PoP: [Moscow (MOW)](/docs/network/latency/mow-moscow) * Destination PoP: [Melbourne (MEL)](/docs/network/latency/mel-melbourne) * Region overview: [Europe](/docs/network/latency/regions/europe) → [Asia Pacific](/docs/network/latency/regions/asia-pacific) * Corridor overview: [Europe to Asia Pacific](/docs/network/latency/regions/europe-to-asia-pacific) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Melbourne to Moscow latency and RTT](/docs/network/latency/pairs/mel-mow-rtt): 236.1 ms **Fastest routes departing Moscow (MOW)** * [Moscow to Berlin latency and RTT](/docs/network/latency/pairs/mow-ber-rtt): 27.5 ms * [Moscow to Frankfurt latency and RTT](/docs/network/latency/pairs/mow-fra-rtt): 35.9 ms * [Moscow to Amsterdam latency and RTT](/docs/network/latency/pairs/mow-ams-rtt): 37.6 ms * [Moscow to London latency and RTT](/docs/network/latency/pairs/mow-lon-rtt): 42.4 ms * [Moscow to Paris latency and RTT](/docs/network/latency/pairs/mow-par-rtt): 44 ms **Fastest routes arriving at Melbourne (MEL)** * [Sydney to Melbourne latency and RTT](/docs/network/latency/pairs/syd-mel-rtt): 9.8 ms * [Singapore to Melbourne latency and RTT](/docs/network/latency/pairs/sin-mel-rtt): 88.6 ms * [Tokyo to Melbourne latency and RTT](/docs/network/latency/pairs/tyo-mel-rtt): 113.9 ms * [Hong Kong to Melbourne latency and RTT](/docs/network/latency/pairs/hkg-mel-rtt): 138.4 ms * [Taipei to Melbourne latency and RTT](/docs/network/latency/pairs/tpe-mel-rtt): 142 ms **Same corridor (Europe → Asia Pacific)** * [Moscow to Hong Kong latency and RTT](/docs/network/latency/pairs/mow-hkg-rtt): 118.6 ms * [Moscow to Taipei latency and RTT](/docs/network/latency/pairs/mow-tpe-rtt): 132.1 ms * [Marseille to Singapore latency and RTT](/docs/network/latency/pairs/mrs-sin-rtt): 139.6 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/mow-mel.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Moscow → Melbourne round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/mow-mel-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Moscow, Russia → Miami, USA RTT 🇷🇺 **Moscow, Russia (MOW)** → 🇺🇸 **Miami, USA (MIA)** Measurement round: `2026-08-16T04:07:28Z`. Moscow to Miami averaged 144.5 ms across 50 replies. The observed range was 138.89 to 163.99 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 5.03 ms, or 3.5% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **144.5 ms** | | Jitter | **4.68 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **90.48 ms** | | Fiber Efficiency | **62.6%** | | Latency Tier | Good | | Source Region | [Europe](/docs/network/latency/regions/europe) | | Destination Region | [North America](/docs/network/latency/regions/north-america) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Moscow](https://www.geonames.org/524901) (55.75204, 37.61781) and [Miami](https://www.geonames.org/4164138) (25.77427, -80.19366). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **9,239.8 km** | | Vacuum RTT floor | **61.64 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **90.48 ms** | | Low-latency fiber material floor | **90.12 ms** | | Engineering floor (5% path allowance) | **95.02 ms** | | Research 1.33× mapped-fiber reference | **120.34 ms** | | Estimated unamplified path loss | **1940.4 dB** | | Transparent optical spans / inline amplifiers | **122 / 121** | | Published RTT inflation over fiber floor | **1.6×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **138.89 ms** | | Average RTT | **144.5 ms** | | Maximum RTT | **163.99 ms** | | Standard deviation | **5.03 ms** | | Stdev / average | **3.5%** | ## Route Context * Departure PoP: [Moscow (MOW)](/docs/network/latency/mow-moscow) * Destination PoP: [Miami (MIA)](/docs/network/latency/mia-miami) * Region overview: [Europe](/docs/network/latency/regions/europe) → [North America](/docs/network/latency/regions/north-america) * Corridor overview: [Europe to North America](/docs/network/latency/regions/europe-to-north-america) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Miami to Moscow latency and RTT](/docs/network/latency/pairs/mia-mow-rtt): 145.5 ms **Fastest routes departing Moscow (MOW)** * [Moscow to Berlin latency and RTT](/docs/network/latency/pairs/mow-ber-rtt): 27.5 ms * [Moscow to Frankfurt latency and RTT](/docs/network/latency/pairs/mow-fra-rtt): 35.9 ms * [Moscow to Amsterdam latency and RTT](/docs/network/latency/pairs/mow-ams-rtt): 37.6 ms * [Moscow to London latency and RTT](/docs/network/latency/pairs/mow-lon-rtt): 42.4 ms * [Moscow to Paris latency and RTT](/docs/network/latency/pairs/mow-par-rtt): 44 ms **Fastest routes arriving at Miami (MIA)** * [Ashburn to Miami latency and RTT](/docs/network/latency/pairs/iad-mia-rtt): 27.4 ms * [New York to Miami latency and RTT](/docs/network/latency/pairs/nyc-mia-rtt): 32.9 ms * [Los Angeles to Miami latency and RTT](/docs/network/latency/pairs/lax-mia-rtt): 57.3 ms * [São Paulo to Miami latency and RTT](/docs/network/latency/pairs/gru-mia-rtt): 74 ms * [Seattle to Miami latency and RTT](/docs/network/latency/pairs/sea-mia-rtt): 81.7 ms **Same corridor (Europe → North America)** * [London to New York latency and RTT](/docs/network/latency/pairs/lon-nyc-rtt): 63.8 ms * [Amsterdam to New York latency and RTT](/docs/network/latency/pairs/ams-nyc-rtt): 68.6 ms * [Paris to New York latency and RTT](/docs/network/latency/pairs/par-nyc-rtt): 68.9 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/mow-mia.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Moscow → Miami round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/mow-mia-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Moscow, Russia → Marseille, France RTT 🇷🇺 **Moscow, Russia (MOW)** → 🇫🇷 **Marseille, France (MRS)** Measurement round: `2026-08-16T04:07:28Z`. Moscow to Marseille averaged 49.9 ms across 50 replies. The observed range was 47.94 to 53.35 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 1.46 ms, or 2.9% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------- | | RTT | **49.9 ms** | | Jitter | **1.46 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **26.23 ms** | | Fiber Efficiency | **52.6%** | | Latency Tier | Excellent | | Source Region | [Europe](/docs/network/latency/regions/europe) | | Destination Region | [Europe](/docs/network/latency/regions/europe) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Moscow](https://www.geonames.org/524901) (55.75204, 37.61781) and [Marseille](https://www.geonames.org/2995469) (43.29695, 5.38107). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **2,678.2 km** | | Vacuum RTT floor | **17.87 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **26.23 ms** | | Low-latency fiber material floor | **26.12 ms** | | Engineering floor (5% path allowance) | **27.54 ms** | | Research 1.33× mapped-fiber reference | **34.88 ms** | | Estimated unamplified path loss | **562.4 dB** | | Transparent optical spans / inline amplifiers | **36 / 35** | | Published RTT inflation over fiber floor | **1.9×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------ | | Minimum RTT | **47.94 ms** | | Average RTT | **49.9 ms** | | Maximum RTT | **53.35 ms** | | Standard deviation | **1.46 ms** | | Stdev / average | **2.9%** | ## Route Context * Departure PoP: [Moscow (MOW)](/docs/network/latency/mow-moscow) * Destination PoP: [Marseille (MRS)](/docs/network/latency/mrs-marseille) * Region overview: [Europe](/docs/network/latency/regions/europe) → [Europe](/docs/network/latency/regions/europe) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Marseille to Moscow latency and RTT](/docs/network/latency/pairs/mrs-mow-rtt): 49.6 ms **Fastest routes departing Moscow (MOW)** * [Moscow to Berlin latency and RTT](/docs/network/latency/pairs/mow-ber-rtt): 27.5 ms * [Moscow to Frankfurt latency and RTT](/docs/network/latency/pairs/mow-fra-rtt): 35.9 ms * [Moscow to Amsterdam latency and RTT](/docs/network/latency/pairs/mow-ams-rtt): 37.6 ms * [Moscow to London latency and RTT](/docs/network/latency/pairs/mow-lon-rtt): 42.4 ms * [Moscow to Paris latency and RTT](/docs/network/latency/pairs/mow-par-rtt): 44 ms **Fastest routes arriving at Marseille (MRS)** * [Paris to Marseille latency and RTT](/docs/network/latency/pairs/par-mrs-rtt): 8.9 ms * [Frankfurt to Marseille latency and RTT](/docs/network/latency/pairs/fra-mrs-rtt): 16.4 ms * [London to Marseille latency and RTT](/docs/network/latency/pairs/lon-mrs-rtt): 17.9 ms * [Amsterdam to Marseille latency and RTT](/docs/network/latency/pairs/ams-mrs-rtt): 19.6 ms * [Berlin to Marseille latency and RTT](/docs/network/latency/pairs/ber-mrs-rtt): 20.2 ms **Same corridor (Europe → Europe)** * [Amsterdam to London latency and RTT](/docs/network/latency/pairs/ams-lon-rtt): 5.2 ms * [London to Amsterdam latency and RTT](/docs/network/latency/pairs/lon-ams-rtt): 5.2 ms * [Amsterdam to Frankfurt latency and RTT](/docs/network/latency/pairs/ams-fra-rtt): 5.9 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/mow-mrs.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Moscow → Marseille round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/mow-mrs-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Moscow, Russia → New York, USA RTT 🇷🇺 **Moscow, Russia (MOW)** → 🇺🇸 **New York, USA (NYC)** Measurement round: `2026-08-16T04:07:28Z`. Moscow to New York averaged 107.5 ms across 50 replies. The observed range was 103.9 to 121.54 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 3.46 ms, or 3.2% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **107.5 ms** | | Jitter | **2.69 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **73.75 ms** | | Fiber Efficiency | **68.6%** | | Latency Tier | Good | | Source Region | [Europe](/docs/network/latency/regions/europe) | | Destination Region | [North America](/docs/network/latency/regions/north-america) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Moscow](https://www.geonames.org/524901) (55.75204, 37.61781) and [New York City](https://www.geonames.org/5128581) (40.71427, -74.00597). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **7,531.3 km** | | Vacuum RTT floor | **50.24 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **73.75 ms** | | Low-latency fiber material floor | **73.46 ms** | | Engineering floor (5% path allowance) | **77.45 ms** | | Research 1.33× mapped-fiber reference | **98.09 ms** | | Estimated unamplified path loss | **1581.6 dB** | | Transparent optical spans / inline amplifiers | **99 / 98** | | Published RTT inflation over fiber floor | **1.46×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **103.9 ms** | | Average RTT | **107.5 ms** | | Maximum RTT | **121.54 ms** | | Standard deviation | **3.46 ms** | | Stdev / average | **3.2%** | ## Route Context * Departure PoP: [Moscow (MOW)](/docs/network/latency/mow-moscow) * Destination PoP: [New York (NYC)](/docs/network/latency/nyc-new-york) * Region overview: [Europe](/docs/network/latency/regions/europe) → [North America](/docs/network/latency/regions/north-america) * Corridor overview: [Europe to North America](/docs/network/latency/regions/europe-to-north-america) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [New York to Moscow latency and RTT](/docs/network/latency/pairs/nyc-mow-rtt): 107 ms **Fastest routes departing Moscow (MOW)** * [Moscow to Berlin latency and RTT](/docs/network/latency/pairs/mow-ber-rtt): 27.5 ms * [Moscow to Frankfurt latency and RTT](/docs/network/latency/pairs/mow-fra-rtt): 35.9 ms * [Moscow to Amsterdam latency and RTT](/docs/network/latency/pairs/mow-ams-rtt): 37.6 ms * [Moscow to London latency and RTT](/docs/network/latency/pairs/mow-lon-rtt): 42.4 ms * [Moscow to Paris latency and RTT](/docs/network/latency/pairs/mow-par-rtt): 44 ms **Fastest routes arriving at New York (NYC)** * [Ashburn to New York latency and RTT](/docs/network/latency/pairs/iad-nyc-rtt): 6 ms * [Miami to New York latency and RTT](/docs/network/latency/pairs/mia-nyc-rtt): 33.3 ms * [Los Angeles to New York latency and RTT](/docs/network/latency/pairs/lax-nyc-rtt): 57.9 ms * [Seattle to New York latency and RTT](/docs/network/latency/pairs/sea-nyc-rtt): 59.5 ms * [London to New York latency and RTT](/docs/network/latency/pairs/lon-nyc-rtt): 63.8 ms **Same corridor (Europe → North America)** * [Amsterdam to New York latency and RTT](/docs/network/latency/pairs/ams-nyc-rtt): 68.6 ms * [Paris to New York latency and RTT](/docs/network/latency/pairs/par-nyc-rtt): 68.9 ms * [London to Ashburn latency and RTT](/docs/network/latency/pairs/lon-iad-rtt): 71 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/mow-nyc.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Moscow → New York round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/mow-nyc-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Moscow, Russia → Paris, France RTT 🇷🇺 **Moscow, Russia (MOW)** → 🇫🇷 **Paris, France (PAR)** Measurement round: `2026-08-16T04:07:28Z`. Moscow to Paris averaged 44 ms across 50 replies. The observed range was 42.02 to 49.11 ms; these figures describe one probe round, not a long-term service target. This is number 5 of 19 outgoing routes by average RTT. That ranking compares different destinations and is not adjusted for distance. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------- | | RTT | **44 ms** | | Jitter | **1.37 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **24.42 ms** | | Fiber Efficiency | **55.5%** | | Latency Tier | Excellent | | Source Region | [Europe](/docs/network/latency/regions/europe) | | Destination Region | [Europe](/docs/network/latency/regions/europe) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Moscow](https://www.geonames.org/524901) (55.75204, 37.61781) and [Paris](https://www.geonames.org/2988507) (48.85341, 2.34880). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **2,494.1 km** | | Vacuum RTT floor | **16.64 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **24.42 ms** | | Low-latency fiber material floor | **24.33 ms** | | Engineering floor (5% path allowance) | **25.65 ms** | | Research 1.33× mapped-fiber reference | **32.48 ms** | | Estimated unamplified path loss | **523.8 dB** | | Transparent optical spans / inline amplifiers | **33 / 32** | | Published RTT inflation over fiber floor | **1.8×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------ | | Minimum RTT | **42.02 ms** | | Average RTT | **44 ms** | | Maximum RTT | **49.11 ms** | | Standard deviation | **1.64 ms** | | Stdev / average | **3.7%** | ## Route Context * Departure PoP: [Moscow (MOW)](/docs/network/latency/mow-moscow) * Destination PoP: [Paris (PAR)](/docs/network/latency/par-paris) * Region overview: [Europe](/docs/network/latency/regions/europe) → [Europe](/docs/network/latency/regions/europe) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Paris to Moscow latency and RTT](/docs/network/latency/pairs/par-mow-rtt): 44.7 ms **Fastest routes departing Moscow (MOW)** * [Moscow to Berlin latency and RTT](/docs/network/latency/pairs/mow-ber-rtt): 27.5 ms * [Moscow to Frankfurt latency and RTT](/docs/network/latency/pairs/mow-fra-rtt): 35.9 ms * [Moscow to Amsterdam latency and RTT](/docs/network/latency/pairs/mow-ams-rtt): 37.6 ms * [Moscow to London latency and RTT](/docs/network/latency/pairs/mow-lon-rtt): 42.4 ms * [Moscow to Marseille latency and RTT](/docs/network/latency/pairs/mow-mrs-rtt): 49.9 ms **Fastest routes arriving at Paris (PAR)** * [London to Paris latency and RTT](/docs/network/latency/pairs/lon-par-rtt): 6.4 ms * [Amsterdam to Paris latency and RTT](/docs/network/latency/pairs/ams-par-rtt): 7 ms * [Frankfurt to Paris latency and RTT](/docs/network/latency/pairs/fra-par-rtt): 7.6 ms * [Marseille to Paris latency and RTT](/docs/network/latency/pairs/mrs-par-rtt): 8.9 ms * [Berlin to Paris latency and RTT](/docs/network/latency/pairs/ber-par-rtt): 15.6 ms **Same corridor (Europe → Europe)** * [Amsterdam to London latency and RTT](/docs/network/latency/pairs/ams-lon-rtt): 5.2 ms * [London to Amsterdam latency and RTT](/docs/network/latency/pairs/lon-ams-rtt): 5.2 ms * [Amsterdam to Frankfurt latency and RTT](/docs/network/latency/pairs/ams-fra-rtt): 5.9 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/mow-par.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Moscow → Paris round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/mow-par-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Moscow, Russia → Seattle, USA RTT 🇷🇺 **Moscow, Russia (MOW)** → 🇺🇸 **Seattle, USA (SEA)** Measurement round: `2026-08-16T04:07:28Z`. Moscow to Seattle averaged 164.8 ms, compared with a 82.23 ms city-centre fiber reference. The difference cannot be assigned to a particular cable or carrier from RTT alone. Standard deviation was 7.9 ms, or 4.8% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **164.8 ms** | | Jitter | **6.37 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **82.23 ms** | | Fiber Efficiency | **49.9%** | | Latency Tier | Fair | | Source Region | [Europe](/docs/network/latency/regions/europe) | | Destination Region | [North America](/docs/network/latency/regions/north-america) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Moscow](https://www.geonames.org/524901) (55.75204, 37.61781) and [Seattle](https://www.geonames.org/5809844) (47.60621, -122.33207). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **8,396.9 km** | | Vacuum RTT floor | **56.02 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **82.23 ms** | | Low-latency fiber material floor | **81.9 ms** | | Engineering floor (5% path allowance) | **86.35 ms** | | Research 1.33× mapped-fiber reference | **109.36 ms** | | Estimated unamplified path loss | **1763.3 dB** | | Transparent optical spans / inline amplifiers | **111 / 110** | | Published RTT inflation over fiber floor | **2×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **155.86 ms** | | Average RTT | **164.8 ms** | | Maximum RTT | **193.38 ms** | | Standard deviation | **7.9 ms** | | Stdev / average | **4.8%** | ## Route Context * Departure PoP: [Moscow (MOW)](/docs/network/latency/mow-moscow) * Destination PoP: [Seattle (SEA)](/docs/network/latency/sea-seattle) * Region overview: [Europe](/docs/network/latency/regions/europe) → [North America](/docs/network/latency/regions/north-america) * Corridor overview: [Europe to North America](/docs/network/latency/regions/europe-to-north-america) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Seattle to Moscow latency and RTT](/docs/network/latency/pairs/sea-mow-rtt): 164.6 ms **Fastest routes departing Moscow (MOW)** * [Moscow to Berlin latency and RTT](/docs/network/latency/pairs/mow-ber-rtt): 27.5 ms * [Moscow to Frankfurt latency and RTT](/docs/network/latency/pairs/mow-fra-rtt): 35.9 ms * [Moscow to Amsterdam latency and RTT](/docs/network/latency/pairs/mow-ams-rtt): 37.6 ms * [Moscow to London latency and RTT](/docs/network/latency/pairs/mow-lon-rtt): 42.4 ms * [Moscow to Paris latency and RTT](/docs/network/latency/pairs/mow-par-rtt): 44 ms **Fastest routes arriving at Seattle (SEA)** * [Los Angeles to Seattle latency and RTT](/docs/network/latency/pairs/lax-sea-rtt): 26.9 ms * [New York to Seattle latency and RTT](/docs/network/latency/pairs/nyc-sea-rtt): 58.7 ms * [Ashburn to Seattle latency and RTT](/docs/network/latency/pairs/iad-sea-rtt): 62.3 ms * [Miami to Seattle latency and RTT](/docs/network/latency/pairs/mia-sea-rtt): 81.9 ms * [Tokyo to Seattle latency and RTT](/docs/network/latency/pairs/tyo-sea-rtt): 84.8 ms **Same corridor (Europe → North America)** * [London to New York latency and RTT](/docs/network/latency/pairs/lon-nyc-rtt): 63.8 ms * [Amsterdam to New York latency and RTT](/docs/network/latency/pairs/ams-nyc-rtt): 68.6 ms * [Paris to New York latency and RTT](/docs/network/latency/pairs/par-nyc-rtt): 68.9 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/mow-sea.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Moscow → Seattle round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/mow-sea-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Moscow, Russia → Singapore RTT 🇷🇺 **Moscow, Russia (MOW)** → 🇸🇬 **Singapore (SIN)** Measurement round: `2026-08-16T04:07:28Z`. Moscow to Singapore averaged 147.3 ms across 50 replies. The observed range was 144.41 to 156.92 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 2.32 ms, or 1.6% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------------------- | | RTT | **147.3 ms** | | Jitter | **1.99 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **82.49 ms** | | Fiber Efficiency | **56%** | | Latency Tier | Good | | Source Region | [Europe](/docs/network/latency/regions/europe) | | Destination Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Moscow](https://www.geonames.org/524901) (55.75204, 37.61781) and [Singapore](https://www.geonames.org/1880252) (1.28967, 103.85007). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **8,423.2 km** | | Vacuum RTT floor | **56.19 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **82.49 ms** | | Low-latency fiber material floor | **82.15 ms** | | Engineering floor (5% path allowance) | **86.62 ms** | | Research 1.33× mapped-fiber reference | **109.71 ms** | | Estimated unamplified path loss | **1768.9 dB** | | Transparent optical spans / inline amplifiers | **111 / 110** | | Published RTT inflation over fiber floor | **1.79×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **144.41 ms** | | Average RTT | **147.3 ms** | | Maximum RTT | **156.92 ms** | | Standard deviation | **2.32 ms** | | Stdev / average | **1.6%** | ## Route Context * Departure PoP: [Moscow (MOW)](/docs/network/latency/mow-moscow) * Destination PoP: [Singapore (SIN)](/docs/network/latency/sin-singapore) * Region overview: [Europe](/docs/network/latency/regions/europe) → [Asia Pacific](/docs/network/latency/regions/asia-pacific) * Corridor overview: [Europe to Asia Pacific](/docs/network/latency/regions/europe-to-asia-pacific) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Singapore to Moscow latency and RTT](/docs/network/latency/pairs/sin-mow-rtt): 147.7 ms **Fastest routes departing Moscow (MOW)** * [Moscow to Berlin latency and RTT](/docs/network/latency/pairs/mow-ber-rtt): 27.5 ms * [Moscow to Frankfurt latency and RTT](/docs/network/latency/pairs/mow-fra-rtt): 35.9 ms * [Moscow to Amsterdam latency and RTT](/docs/network/latency/pairs/mow-ams-rtt): 37.6 ms * [Moscow to London latency and RTT](/docs/network/latency/pairs/mow-lon-rtt): 42.4 ms * [Moscow to Paris latency and RTT](/docs/network/latency/pairs/mow-par-rtt): 44 ms **Fastest routes arriving at Singapore (SIN)** * [Hong Kong to Singapore latency and RTT](/docs/network/latency/pairs/hkg-sin-rtt): 31.5 ms * [Taipei to Singapore latency and RTT](/docs/network/latency/pairs/tpe-sin-rtt): 44.3 ms * [Tokyo to Singapore latency and RTT](/docs/network/latency/pairs/tyo-sin-rtt): 69.4 ms * [Melbourne to Singapore latency and RTT](/docs/network/latency/pairs/mel-sin-rtt): 88.4 ms * [Sydney to Singapore latency and RTT](/docs/network/latency/pairs/syd-sin-rtt): 94.5 ms **Same corridor (Europe → Asia Pacific)** * [Moscow to Hong Kong latency and RTT](/docs/network/latency/pairs/mow-hkg-rtt): 118.6 ms * [Moscow to Taipei latency and RTT](/docs/network/latency/pairs/mow-tpe-rtt): 132.1 ms * [Marseille to Singapore latency and RTT](/docs/network/latency/pairs/mrs-sin-rtt): 139.6 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/mow-sin.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Moscow → Singapore round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/mow-sin-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Moscow, Russia → Sydney, Australia RTT 🇷🇺 **Moscow, Russia (MOW)** → 🇦🇺 **Sydney, Australia (SYD)** Measurement round: `2026-08-16T04:07:28Z`. Moscow to Sydney averaged 241.8 ms across 50 replies. The observed range was 231.04 to 265.67 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 7.65 ms, or 3.2% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------------------- | | RTT | **241.8 ms** | | Jitter | **5.98 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **141.84 ms** | | Fiber Efficiency | **58.7%** | | Latency Tier | Fair | | Source Region | [Europe](/docs/network/latency/regions/europe) | | Destination Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Moscow](https://www.geonames.org/524901) (55.75204, 37.61781) and [Sydney](https://www.geonames.org/2147714) (-33.86785, 151.20732). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | --------------- | | WGS-84 geodesic distance | **14,483.9 km** | | Vacuum RTT floor | **96.63 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **141.84 ms** | | Low-latency fiber material floor | **141.27 ms** | | Engineering floor (5% path allowance) | **148.95 ms** | | Research 1.33× mapped-fiber reference | **188.64 ms** | | Estimated unamplified path loss | **3041.6 dB** | | Transparent optical spans / inline amplifiers | **191 / 190** | | Published RTT inflation over fiber floor | **1.7×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **231.04 ms** | | Average RTT | **241.8 ms** | | Maximum RTT | **265.67 ms** | | Standard deviation | **7.65 ms** | | Stdev / average | **3.2%** | ## Route Context * Departure PoP: [Moscow (MOW)](/docs/network/latency/mow-moscow) * Destination PoP: [Sydney (SYD)](/docs/network/latency/syd-sydney) * Region overview: [Europe](/docs/network/latency/regions/europe) → [Asia Pacific](/docs/network/latency/regions/asia-pacific) * Corridor overview: [Europe to Asia Pacific](/docs/network/latency/regions/europe-to-asia-pacific) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Sydney to Moscow latency and RTT](/docs/network/latency/pairs/syd-mow-rtt): 242.2 ms **Fastest routes departing Moscow (MOW)** * [Moscow to Berlin latency and RTT](/docs/network/latency/pairs/mow-ber-rtt): 27.5 ms * [Moscow to Frankfurt latency and RTT](/docs/network/latency/pairs/mow-fra-rtt): 35.9 ms * [Moscow to Amsterdam latency and RTT](/docs/network/latency/pairs/mow-ams-rtt): 37.6 ms * [Moscow to London latency and RTT](/docs/network/latency/pairs/mow-lon-rtt): 42.4 ms * [Moscow to Paris latency and RTT](/docs/network/latency/pairs/mow-par-rtt): 44 ms **Fastest routes arriving at Sydney (SYD)** * [Melbourne to Sydney latency and RTT](/docs/network/latency/pairs/mel-syd-rtt): 9.8 ms * [Singapore to Sydney latency and RTT](/docs/network/latency/pairs/sin-syd-rtt): 94.5 ms * [Tokyo to Sydney latency and RTT](/docs/network/latency/pairs/tyo-syd-rtt): 101.9 ms * [Taipei to Sydney latency and RTT](/docs/network/latency/pairs/tpe-syd-rtt): 133.2 ms * [Hong Kong to Sydney latency and RTT](/docs/network/latency/pairs/hkg-syd-rtt): 137.2 ms **Same corridor (Europe → Asia Pacific)** * [Moscow to Hong Kong latency and RTT](/docs/network/latency/pairs/mow-hkg-rtt): 118.6 ms * [Moscow to Taipei latency and RTT](/docs/network/latency/pairs/mow-tpe-rtt): 132.1 ms * [Marseille to Singapore latency and RTT](/docs/network/latency/pairs/mrs-sin-rtt): 139.6 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/mow-syd.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Moscow → Sydney round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/mow-syd-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Moscow, Russia → Taipei, Taiwan RTT 🇷🇺 **Moscow, Russia (MOW)** → 🇹🇼 **Taipei, Taiwan (TPE)** Measurement round: `2026-08-16T04:07:28Z`. Moscow to Taipei averaged 132.1 ms across 50 replies. The observed range was 126.75 to 145.67 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 4.52 ms, or 3.4% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------------------- | | RTT | **132.1 ms** | | Jitter | **4.42 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **72.16 ms** | | Fiber Efficiency | **54.6%** | | Latency Tier | Good | | Source Region | [Europe](/docs/network/latency/regions/europe) | | Destination Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Moscow](https://www.geonames.org/524901) (55.75204, 37.61781) and [Taipei](https://www.geonames.org/1668341) (25.05306, 121.52639). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | ------------- | | WGS-84 geodesic distance | **7,369 km** | | Vacuum RTT floor | **49.16 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **72.16 ms** | | Low-latency fiber material floor | **71.87 ms** | | Engineering floor (5% path allowance) | **75.78 ms** | | Research 1.33× mapped-fiber reference | **95.98 ms** | | Estimated unamplified path loss | **1547.5 dB** | | Transparent optical spans / inline amplifiers | **97 / 96** | | Published RTT inflation over fiber floor | **1.83×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **126.75 ms** | | Average RTT | **132.1 ms** | | Maximum RTT | **145.67 ms** | | Standard deviation | **4.52 ms** | | Stdev / average | **3.4%** | ## Route Context * Departure PoP: [Moscow (MOW)](/docs/network/latency/mow-moscow) * Destination PoP: [Taipei (TPE)](/docs/network/latency/tpe-taipei) * Region overview: [Europe](/docs/network/latency/regions/europe) → [Asia Pacific](/docs/network/latency/regions/asia-pacific) * Corridor overview: [Europe to Asia Pacific](/docs/network/latency/regions/europe-to-asia-pacific) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Taipei to Moscow latency and RTT](/docs/network/latency/pairs/tpe-mow-rtt): 132.7 ms **Fastest routes departing Moscow (MOW)** * [Moscow to Berlin latency and RTT](/docs/network/latency/pairs/mow-ber-rtt): 27.5 ms * [Moscow to Frankfurt latency and RTT](/docs/network/latency/pairs/mow-fra-rtt): 35.9 ms * [Moscow to Amsterdam latency and RTT](/docs/network/latency/pairs/mow-ams-rtt): 37.6 ms * [Moscow to London latency and RTT](/docs/network/latency/pairs/mow-lon-rtt): 42.4 ms * [Moscow to Paris latency and RTT](/docs/network/latency/pairs/mow-par-rtt): 44 ms **Fastest routes arriving at Taipei (TPE)** * [Hong Kong to Taipei latency and RTT](/docs/network/latency/pairs/hkg-tpe-rtt): 14.7 ms * [Tokyo to Taipei latency and RTT](/docs/network/latency/pairs/tyo-tpe-rtt): 32.3 ms * [Singapore to Taipei latency and RTT](/docs/network/latency/pairs/sin-tpe-rtt): 45.8 ms * [Seattle to Taipei latency and RTT](/docs/network/latency/pairs/sea-tpe-rtt): 114.7 ms * [Sydney to Taipei latency and RTT](/docs/network/latency/pairs/syd-tpe-rtt): 131.7 ms **Same corridor (Europe → Asia Pacific)** * [Moscow to Hong Kong latency and RTT](/docs/network/latency/pairs/mow-hkg-rtt): 118.6 ms * [Marseille to Singapore latency and RTT](/docs/network/latency/pairs/mrs-sin-rtt): 139.6 ms * [Berlin to Hong Kong latency and RTT](/docs/network/latency/pairs/ber-hkg-rtt): 144.9 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/mow-tpe.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Moscow → Taipei round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/mow-tpe-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Moscow, Russia → Tokyo, Japan RTT 🇷🇺 **Moscow, Russia (MOW)** → 🇯🇵 **Tokyo, Japan (TYO)** Measurement round: `2026-08-16T04:07:28Z`. Moscow to Tokyo averaged 162.2 ms, compared with a 73.42 ms city-centre fiber reference. The difference cannot be assigned to a particular cable or carrier from RTT alone. Standard deviation was 6.91 ms, or 4.3% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------------------- | | RTT | **162.2 ms** | | Jitter | **5.65 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **73.42 ms** | | Fiber Efficiency | **45.3%** | | Latency Tier | Fair | | Source Region | [Europe](/docs/network/latency/regions/europe) | | Destination Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Moscow](https://www.geonames.org/524901) (55.75204, 37.61781) and [Tokyo](https://www.geonames.org/1850147) (35.68950, 139.69171). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | ------------- | | WGS-84 geodesic distance | **7,497 km** | | Vacuum RTT floor | **50.01 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **73.42 ms** | | Low-latency fiber material floor | **73.12 ms** | | Engineering floor (5% path allowance) | **77.1 ms** | | Research 1.33× mapped-fiber reference | **97.64 ms** | | Estimated unamplified path loss | **1574.4 dB** | | Transparent optical spans / inline amplifiers | **99 / 98** | | Published RTT inflation over fiber floor | **2.21×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **152.69 ms** | | Average RTT | **162.2 ms** | | Maximum RTT | **185.56 ms** | | Standard deviation | **6.91 ms** | | Stdev / average | **4.3%** | ## Route Context * Departure PoP: [Moscow (MOW)](/docs/network/latency/mow-moscow) * Destination PoP: [Tokyo (TYO)](/docs/network/latency/tyo-tokyo) * Region overview: [Europe](/docs/network/latency/regions/europe) → [Asia Pacific](/docs/network/latency/regions/asia-pacific) * Corridor overview: [Europe to Asia Pacific](/docs/network/latency/regions/europe-to-asia-pacific) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Tokyo to Moscow latency and RTT](/docs/network/latency/pairs/tyo-mow-rtt): 161.7 ms **Fastest routes departing Moscow (MOW)** * [Moscow to Berlin latency and RTT](/docs/network/latency/pairs/mow-ber-rtt): 27.5 ms * [Moscow to Frankfurt latency and RTT](/docs/network/latency/pairs/mow-fra-rtt): 35.9 ms * [Moscow to Amsterdam latency and RTT](/docs/network/latency/pairs/mow-ams-rtt): 37.6 ms * [Moscow to London latency and RTT](/docs/network/latency/pairs/mow-lon-rtt): 42.4 ms * [Moscow to Paris latency and RTT](/docs/network/latency/pairs/mow-par-rtt): 44 ms **Fastest routes arriving at Tokyo (TYO)** * [Taipei to Tokyo latency and RTT](/docs/network/latency/pairs/tpe-tyo-rtt): 31.9 ms * [Hong Kong to Tokyo latency and RTT](/docs/network/latency/pairs/hkg-tyo-rtt): 44.9 ms * [Singapore to Tokyo latency and RTT](/docs/network/latency/pairs/sin-tyo-rtt): 68.9 ms * [Seattle to Tokyo latency and RTT](/docs/network/latency/pairs/sea-tyo-rtt): 85 ms * [Los Angeles to Tokyo latency and RTT](/docs/network/latency/pairs/lax-tyo-rtt): 101.2 ms **Same corridor (Europe → Asia Pacific)** * [Moscow to Hong Kong latency and RTT](/docs/network/latency/pairs/mow-hkg-rtt): 118.6 ms * [Moscow to Taipei latency and RTT](/docs/network/latency/pairs/mow-tpe-rtt): 132.1 ms * [Marseille to Singapore latency and RTT](/docs/network/latency/pairs/mrs-sin-rtt): 139.6 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/mow-tyo.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Moscow → Tokyo round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/mow-tyo-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Marseille, France → Amsterdam, Netherlands RTT 🇫🇷 **Marseille, France (MRS)** → 🇳🇱 **Amsterdam, Netherlands (AMS)** Measurement round: `2026-08-16T04:07:28Z`. Marseille to Amsterdam averaged 20.3 ms, compared with a 9.89 ms city-centre fiber reference. The difference cannot be assigned to a particular cable or carrier from RTT alone. This is number 4 of 19 outgoing routes by average RTT. That ranking compares different destinations and is not adjusted for distance. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------- | | RTT | **20.3 ms** | | Jitter | **0.48 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **9.89 ms** | | Fiber Efficiency | **48.7%** | | Latency Tier | Ultra-Low | | Source Region | [Europe](/docs/network/latency/regions/europe) | | Destination Region | [Europe](/docs/network/latency/regions/europe) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Marseille](https://www.geonames.org/2995469) (43.29695, 5.38107) and [Amsterdam](https://www.geonames.org/2759794) (52.37403, 4.88969). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **1,009.9 km** | | Vacuum RTT floor | **6.74 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **9.89 ms** | | Low-latency fiber material floor | **9.85 ms** | | Engineering floor (5% path allowance) | **10.39 ms** | | Research 1.33× mapped-fiber reference | **13.15 ms** | | Estimated unamplified path loss | **212.1 dB** | | Transparent optical spans / inline amplifiers | **14 / 13** | | Published RTT inflation over fiber floor | **2.05×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------ | | Minimum RTT | **19.62 ms** | | Average RTT | **20.3 ms** | | Maximum RTT | **22.75 ms** | | Standard deviation | **0.64 ms** | | Stdev / average | **3.2%** | ## Route Context * Departure PoP: [Marseille (MRS)](/docs/network/latency/mrs-marseille) * Destination PoP: [Amsterdam (AMS)](/docs/network/latency/ams-amsterdam) * Region overview: [Europe](/docs/network/latency/regions/europe) → [Europe](/docs/network/latency/regions/europe) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Amsterdam to Marseille latency and RTT](/docs/network/latency/pairs/ams-mrs-rtt): 19.6 ms **Fastest routes departing Marseille (MRS)** * [Marseille to Paris latency and RTT](/docs/network/latency/pairs/mrs-par-rtt): 8.9 ms * [Marseille to Frankfurt latency and RTT](/docs/network/latency/pairs/mrs-fra-rtt): 16 ms * [Marseille to London latency and RTT](/docs/network/latency/pairs/mrs-lon-rtt): 18.9 ms * [Marseille to Berlin latency and RTT](/docs/network/latency/pairs/mrs-ber-rtt): 20.6 ms * [Marseille to Moscow latency and RTT](/docs/network/latency/pairs/mrs-mow-rtt): 49.6 ms **Fastest routes arriving at Amsterdam (AMS)** * [London to Amsterdam latency and RTT](/docs/network/latency/pairs/lon-ams-rtt): 5.2 ms * [Frankfurt to Amsterdam latency and RTT](/docs/network/latency/pairs/fra-ams-rtt): 5.9 ms * [Paris to Amsterdam latency and RTT](/docs/network/latency/pairs/par-ams-rtt): 7 ms * [Berlin to Amsterdam latency and RTT](/docs/network/latency/pairs/ber-ams-rtt): 7.2 ms * [Moscow to Amsterdam latency and RTT](/docs/network/latency/pairs/mow-ams-rtt): 37.6 ms **Same corridor (Europe → Europe)** * [Amsterdam to London latency and RTT](/docs/network/latency/pairs/ams-lon-rtt): 5.2 ms * [Amsterdam to Frankfurt latency and RTT](/docs/network/latency/pairs/ams-fra-rtt): 5.9 ms * [Berlin to Frankfurt latency and RTT](/docs/network/latency/pairs/ber-fra-rtt): 6.1 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/mrs-ams.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Marseille → Amsterdam round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/mrs-ams-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Marseille, France → Berlin, Germany RTT 🇫🇷 **Marseille, France (MRS)** → 🇩🇪 **Berlin, Germany (BER)** Measurement round: `2026-08-16T04:07:28Z`. Marseille to Berlin averaged 20.6 ms across 50 replies. The observed range was 19.99 to 21.74 ms; these figures describe one probe round, not a long-term service target. This is number 5 of 19 outgoing routes by average RTT. That ranking compares different destinations and is not adjusted for distance. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------- | | RTT | **20.6 ms** | | Jitter | **0.3 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **11.62 ms** | | Fiber Efficiency | **56.4%** | | Latency Tier | Ultra-Low | | Source Region | [Europe](/docs/network/latency/regions/europe) | | Destination Region | [Europe](/docs/network/latency/regions/europe) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Marseille](https://www.geonames.org/2995469) (43.29695, 5.38107) and [Berlin](https://www.geonames.org/2950159) (52.52437, 13.41053). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | ------------ | | WGS-84 geodesic distance | **1,187 km** | | Vacuum RTT floor | **7.92 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **11.62 ms** | | Low-latency fiber material floor | **11.58 ms** | | Engineering floor (5% path allowance) | **12.21 ms** | | Research 1.33× mapped-fiber reference | **15.46 ms** | | Estimated unamplified path loss | **249.3 dB** | | Transparent optical spans / inline amplifiers | **16 / 15** | | Published RTT inflation over fiber floor | **1.77×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------ | | Minimum RTT | **19.99 ms** | | Average RTT | **20.6 ms** | | Maximum RTT | **21.74 ms** | | Standard deviation | **0.35 ms** | | Stdev / average | **1.7%** | ## Route Context * Departure PoP: [Marseille (MRS)](/docs/network/latency/mrs-marseille) * Destination PoP: [Berlin (BER)](/docs/network/latency/ber-berlin) * Region overview: [Europe](/docs/network/latency/regions/europe) → [Europe](/docs/network/latency/regions/europe) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Berlin to Marseille latency and RTT](/docs/network/latency/pairs/ber-mrs-rtt): 20.2 ms **Fastest routes departing Marseille (MRS)** * [Marseille to Paris latency and RTT](/docs/network/latency/pairs/mrs-par-rtt): 8.9 ms * [Marseille to Frankfurt latency and RTT](/docs/network/latency/pairs/mrs-fra-rtt): 16 ms * [Marseille to London latency and RTT](/docs/network/latency/pairs/mrs-lon-rtt): 18.9 ms * [Marseille to Amsterdam latency and RTT](/docs/network/latency/pairs/mrs-ams-rtt): 20.3 ms * [Marseille to Moscow latency and RTT](/docs/network/latency/pairs/mrs-mow-rtt): 49.6 ms **Fastest routes arriving at Berlin (BER)** * [Frankfurt to Berlin latency and RTT](/docs/network/latency/pairs/fra-ber-rtt): 6.1 ms * [Amsterdam to Berlin latency and RTT](/docs/network/latency/pairs/ams-ber-rtt): 7.2 ms * [Paris to Berlin latency and RTT](/docs/network/latency/pairs/par-ber-rtt): 15.5 ms * [London to Berlin latency and RTT](/docs/network/latency/pairs/lon-ber-rtt): 16.2 ms * [Moscow to Berlin latency and RTT](/docs/network/latency/pairs/mow-ber-rtt): 27.5 ms **Same corridor (Europe → Europe)** * [Amsterdam to London latency and RTT](/docs/network/latency/pairs/ams-lon-rtt): 5.2 ms * [London to Amsterdam latency and RTT](/docs/network/latency/pairs/lon-ams-rtt): 5.2 ms * [Amsterdam to Frankfurt latency and RTT](/docs/network/latency/pairs/ams-fra-rtt): 5.9 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/mrs-ber.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Marseille → Berlin round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/mrs-ber-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Marseille, France → Frankfurt, Germany RTT 🇫🇷 **Marseille, France (MRS)** → 🇩🇪 **Frankfurt, Germany (FRA)** Measurement round: `2026-08-16T04:07:28Z`. Marseille to Frankfurt averaged 16 ms, compared with a 7.82 ms city-centre fiber reference. The difference cannot be assigned to a particular cable or carrier from RTT alone. This is number 2 of 19 outgoing routes by average RTT. That ranking compares different destinations and is not adjusted for distance. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------- | | RTT | **16 ms** | | Jitter | **0.59 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **7.82 ms** | | Fiber Efficiency | **48.9%** | | Latency Tier | Ultra-Low | | Source Region | [Europe](/docs/network/latency/regions/europe) | | Destination Region | [Europe](/docs/network/latency/regions/europe) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Marseille](https://www.geonames.org/2995469) (43.29695, 5.38107) and [Frankfurt am Main](https://www.geonames.org/2925533) (50.11552, 8.68417). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | ------------ | | WGS-84 geodesic distance | **798.8 km** | | Vacuum RTT floor | **5.33 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **7.82 ms** | | Low-latency fiber material floor | **7.79 ms** | | Engineering floor (5% path allowance) | **8.21 ms** | | Research 1.33× mapped-fiber reference | **10.4 ms** | | Estimated unamplified path loss | **167.7 dB** | | Transparent optical spans / inline amplifiers | **11 / 10** | | Published RTT inflation over fiber floor | **2.05×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------ | | Minimum RTT | **15.12 ms** | | Average RTT | **16 ms** | | Maximum RTT | **18.91 ms** | | Standard deviation | **0.79 ms** | | Stdev / average | **4.9%** | ## Route Context * Departure PoP: [Marseille (MRS)](/docs/network/latency/mrs-marseille) * Destination PoP: [Frankfurt (FRA)](/docs/network/latency/fra-frankfurt) * Region overview: [Europe](/docs/network/latency/regions/europe) → [Europe](/docs/network/latency/regions/europe) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Frankfurt to Marseille latency and RTT](/docs/network/latency/pairs/fra-mrs-rtt): 16.4 ms **Fastest routes departing Marseille (MRS)** * [Marseille to Paris latency and RTT](/docs/network/latency/pairs/mrs-par-rtt): 8.9 ms * [Marseille to London latency and RTT](/docs/network/latency/pairs/mrs-lon-rtt): 18.9 ms * [Marseille to Amsterdam latency and RTT](/docs/network/latency/pairs/mrs-ams-rtt): 20.3 ms * [Marseille to Berlin latency and RTT](/docs/network/latency/pairs/mrs-ber-rtt): 20.6 ms * [Marseille to Moscow latency and RTT](/docs/network/latency/pairs/mrs-mow-rtt): 49.6 ms **Fastest routes arriving at Frankfurt (FRA)** * [Amsterdam to Frankfurt latency and RTT](/docs/network/latency/pairs/ams-fra-rtt): 5.9 ms * [Berlin to Frankfurt latency and RTT](/docs/network/latency/pairs/ber-fra-rtt): 6.1 ms * [Paris to Frankfurt latency and RTT](/docs/network/latency/pairs/par-fra-rtt): 7.6 ms * [London to Frankfurt latency and RTT](/docs/network/latency/pairs/lon-fra-rtt): 13.6 ms * [Moscow to Frankfurt latency and RTT](/docs/network/latency/pairs/mow-fra-rtt): 35.9 ms **Same corridor (Europe → Europe)** * [Amsterdam to London latency and RTT](/docs/network/latency/pairs/ams-lon-rtt): 5.2 ms * [London to Amsterdam latency and RTT](/docs/network/latency/pairs/lon-ams-rtt): 5.2 ms * [Frankfurt to Amsterdam latency and RTT](/docs/network/latency/pairs/fra-ams-rtt): 5.9 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/mrs-fra.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Marseille → Frankfurt round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/mrs-fra-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Marseille, France → São Paulo, Brazil RTT 🇫🇷 **Marseille, France (MRS)** → 🇧🇷 **São Paulo, Brazil (GRU)** Measurement round: `2026-08-16T04:07:28Z`. Marseille to São Paulo averaged 186.7 ms, compared with a 89.24 ms city-centre fiber reference. The difference cannot be assigned to a particular cable or carrier from RTT alone. Standard deviation was 7.56 ms, or 4.0% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **186.7 ms** | | Jitter | **6.61 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **89.24 ms** | | Fiber Efficiency | **47.8%** | | Latency Tier | Fair | | Source Region | [Europe](/docs/network/latency/regions/europe) | | Destination Region | [South America](/docs/network/latency/regions/south-america) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Marseille](https://www.geonames.org/2995469) (43.29695, 5.38107) and [São Paulo](https://www.geonames.org/3448439) (-23.54750, -46.63611). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | ------------- | | WGS-84 geodesic distance | **9,113 km** | | Vacuum RTT floor | **60.8 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **89.24 ms** | | Low-latency fiber material floor | **88.88 ms** | | Engineering floor (5% path allowance) | **93.72 ms** | | Research 1.33× mapped-fiber reference | **118.69 ms** | | Estimated unamplified path loss | **1913.7 dB** | | Transparent optical spans / inline amplifiers | **120 / 119** | | Published RTT inflation over fiber floor | **2.09×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **176.61 ms** | | Average RTT | **186.7 ms** | | Maximum RTT | **215.07 ms** | | Standard deviation | **7.56 ms** | | Stdev / average | **4.0%** | ## Route Context * Departure PoP: [Marseille (MRS)](/docs/network/latency/mrs-marseille) * Destination PoP: [São Paulo (GRU)](/docs/network/latency/gru-sao-paulo) * Region overview: [Europe](/docs/network/latency/regions/europe) → [South America](/docs/network/latency/regions/south-america) * Corridor overview: [Europe to South America](/docs/network/latency/regions/europe-to-south-america) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [São Paulo to Marseille latency and RTT](/docs/network/latency/pairs/gru-mrs-rtt): 187.7 ms **Fastest routes departing Marseille (MRS)** * [Marseille to Paris latency and RTT](/docs/network/latency/pairs/mrs-par-rtt): 8.9 ms * [Marseille to Frankfurt latency and RTT](/docs/network/latency/pairs/mrs-fra-rtt): 16 ms * [Marseille to London latency and RTT](/docs/network/latency/pairs/mrs-lon-rtt): 18.9 ms * [Marseille to Amsterdam latency and RTT](/docs/network/latency/pairs/mrs-ams-rtt): 20.3 ms * [Marseille to Berlin latency and RTT](/docs/network/latency/pairs/mrs-ber-rtt): 20.6 ms **Fastest routes arriving at São Paulo (GRU)** * [Ashburn to São Paulo latency and RTT](/docs/network/latency/pairs/iad-gru-rtt): 101.4 ms * [New York to São Paulo latency and RTT](/docs/network/latency/pairs/nyc-gru-rtt): 106.9 ms * [Miami to São Paulo latency and RTT](/docs/network/latency/pairs/mia-gru-rtt): 128.7 ms * [Los Angeles to São Paulo latency and RTT](/docs/network/latency/pairs/lax-gru-rtt): 131.3 ms * [Seattle to São Paulo latency and RTT](/docs/network/latency/pairs/sea-gru-rtt): 155.7 ms **Same corridor (Europe → South America)** * [London to São Paulo latency and RTT](/docs/network/latency/pairs/lon-gru-rtt): 170.7 ms * [Amsterdam to São Paulo latency and RTT](/docs/network/latency/pairs/ams-gru-rtt): 175.5 ms * [Paris to São Paulo latency and RTT](/docs/network/latency/pairs/par-gru-rtt): 175.8 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/mrs-gru.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Marseille → São Paulo round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/mrs-gru-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Marseille, France → Hong Kong RTT 🇫🇷 **Marseille, France (MRS)** → 🇭🇰 **Hong Kong (HKG)** Measurement round: `2026-08-16T04:07:28Z`. Marseille to Hong Kong averaged 166.4 ms across 50 replies. The observed range was 161.4 to 176.84 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 3.5 ms, or 2.1% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------------------- | | RTT | **166.4 ms** | | Jitter | **2.92 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **95.48 ms** | | Fiber Efficiency | **57.4%** | | Latency Tier | Fair | | Source Region | [Europe](/docs/network/latency/regions/europe) | | Destination Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Marseille](https://www.geonames.org/2995469) (43.29695, 5.38107) and [Hong Kong](https://www.geonames.org/1819729) (22.27832, 114.17469). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **9,750.5 km** | | Vacuum RTT floor | **65.05 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **95.48 ms** | | Low-latency fiber material floor | **95.1 ms** | | Engineering floor (5% path allowance) | **100.27 ms** | | Research 1.33× mapped-fiber reference | **126.99 ms** | | Estimated unamplified path loss | **2047.6 dB** | | Transparent optical spans / inline amplifiers | **128 / 127** | | Published RTT inflation over fiber floor | **1.74×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **161.4 ms** | | Average RTT | **166.4 ms** | | Maximum RTT | **176.84 ms** | | Standard deviation | **3.5 ms** | | Stdev / average | **2.1%** | ## Route Context * Departure PoP: [Marseille (MRS)](/docs/network/latency/mrs-marseille) * Destination PoP: [Hong Kong (HKG)](/docs/network/latency/hkg-hong-kong) * Region overview: [Europe](/docs/network/latency/regions/europe) → [Asia Pacific](/docs/network/latency/regions/asia-pacific) * Corridor overview: [Europe to Asia Pacific](/docs/network/latency/regions/europe-to-asia-pacific) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Hong Kong to Marseille latency and RTT](/docs/network/latency/pairs/hkg-mrs-rtt): 165.4 ms **Fastest routes departing Marseille (MRS)** * [Marseille to Paris latency and RTT](/docs/network/latency/pairs/mrs-par-rtt): 8.9 ms * [Marseille to Frankfurt latency and RTT](/docs/network/latency/pairs/mrs-fra-rtt): 16 ms * [Marseille to London latency and RTT](/docs/network/latency/pairs/mrs-lon-rtt): 18.9 ms * [Marseille to Amsterdam latency and RTT](/docs/network/latency/pairs/mrs-ams-rtt): 20.3 ms * [Marseille to Berlin latency and RTT](/docs/network/latency/pairs/mrs-ber-rtt): 20.6 ms **Fastest routes arriving at Hong Kong (HKG)** * [Taipei to Hong Kong latency and RTT](/docs/network/latency/pairs/tpe-hkg-rtt): 15.5 ms * [Singapore to Hong Kong latency and RTT](/docs/network/latency/pairs/sin-hkg-rtt): 30.8 ms * [Tokyo to Hong Kong latency and RTT](/docs/network/latency/pairs/tyo-hkg-rtt): 44.8 ms * [Moscow to Hong Kong latency and RTT](/docs/network/latency/pairs/mow-hkg-rtt): 118.6 ms * [Seattle to Hong Kong latency and RTT](/docs/network/latency/pairs/sea-hkg-rtt): 130.4 ms **Same corridor (Europe → Asia Pacific)** * [Moscow to Taipei latency and RTT](/docs/network/latency/pairs/mow-tpe-rtt): 132.1 ms * [Marseille to Singapore latency and RTT](/docs/network/latency/pairs/mrs-sin-rtt): 139.6 ms * [Berlin to Hong Kong latency and RTT](/docs/network/latency/pairs/ber-hkg-rtt): 144.9 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/mrs-hkg.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Marseille → Hong Kong round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/mrs-hkg-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Marseille, France → Ashburn, USA RTT 🇫🇷 **Marseille, France (MRS)** → 🇺🇸 **Ashburn, USA (IAD)** Measurement round: `2026-08-16T04:07:28Z`. Marseille to Ashburn averaged 85.7 ms across 50 replies. The observed range was 82.1 to 94.47 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 2.63 ms, or 3.1% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **85.7 ms** | | Jitter | **1.95 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **65.34 ms** | | Fiber Efficiency | **76.2%** | | Latency Tier | Good | | Source Region | [Europe](/docs/network/latency/regions/europe) | | Destination Region | [North America](/docs/network/latency/regions/north-america) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Marseille](https://www.geonames.org/2995469) (43.29695, 5.38107) and [Ashburn](https://www.geonames.org/4744870) (39.04372, -77.48749). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **6,672.7 km** | | Vacuum RTT floor | **44.52 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **65.34 ms** | | Low-latency fiber material floor | **65.08 ms** | | Engineering floor (5% path allowance) | **68.62 ms** | | Research 1.33× mapped-fiber reference | **86.91 ms** | | Estimated unamplified path loss | **1401.3 dB** | | Transparent optical spans / inline amplifiers | **88 / 87** | | Published RTT inflation over fiber floor | **1.31×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------ | | Minimum RTT | **82.1 ms** | | Average RTT | **85.7 ms** | | Maximum RTT | **94.47 ms** | | Standard deviation | **2.63 ms** | | Stdev / average | **3.1%** | ## Route Context * Departure PoP: [Marseille (MRS)](/docs/network/latency/mrs-marseille) * Destination PoP: [Ashburn (IAD)](/docs/network/latency/iad-ashburn) * Region overview: [Europe](/docs/network/latency/regions/europe) → [North America](/docs/network/latency/regions/north-america) * Corridor overview: [Europe to North America](/docs/network/latency/regions/europe-to-north-america) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Ashburn to Marseille latency and RTT](/docs/network/latency/pairs/iad-mrs-rtt): 86.5 ms **Fastest routes departing Marseille (MRS)** * [Marseille to Paris latency and RTT](/docs/network/latency/pairs/mrs-par-rtt): 8.9 ms * [Marseille to Frankfurt latency and RTT](/docs/network/latency/pairs/mrs-fra-rtt): 16 ms * [Marseille to London latency and RTT](/docs/network/latency/pairs/mrs-lon-rtt): 18.9 ms * [Marseille to Amsterdam latency and RTT](/docs/network/latency/pairs/mrs-ams-rtt): 20.3 ms * [Marseille to Berlin latency and RTT](/docs/network/latency/pairs/mrs-ber-rtt): 20.6 ms **Fastest routes arriving at Ashburn (IAD)** * [New York to Ashburn latency and RTT](/docs/network/latency/pairs/nyc-iad-rtt): 6 ms * [Miami to Ashburn latency and RTT](/docs/network/latency/pairs/mia-iad-rtt): 27.3 ms * [Los Angeles to Ashburn latency and RTT](/docs/network/latency/pairs/lax-iad-rtt): 60.4 ms * [Seattle to Ashburn latency and RTT](/docs/network/latency/pairs/sea-iad-rtt): 61.5 ms * [London to Ashburn latency and RTT](/docs/network/latency/pairs/lon-iad-rtt): 71 ms **Same corridor (Europe → North America)** * [London to New York latency and RTT](/docs/network/latency/pairs/lon-nyc-rtt): 63.8 ms * [Amsterdam to New York latency and RTT](/docs/network/latency/pairs/ams-nyc-rtt): 68.6 ms * [Paris to New York latency and RTT](/docs/network/latency/pairs/par-nyc-rtt): 68.9 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/mrs-iad.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Marseille → Ashburn round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/mrs-iad-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Marseille, France → Johannesburg, South Africa RTT 🇫🇷 **Marseille, France (MRS)** → 🇿🇦 **Johannesburg, South Africa (JNB)** Measurement round: `2026-08-16T04:07:28Z`. Marseille to Johannesburg averaged 198.5 ms, compared with a 78.71 ms city-centre fiber reference. The difference cannot be assigned to a particular cable or carrier from RTT alone. Standard deviation was 3.18 ms, or 1.6% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------- | | RTT | **198.5 ms** | | Jitter | **2.5 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **78.71 ms** | | Fiber Efficiency | **39.7%** | | Latency Tier | Fair | | Source Region | [Europe](/docs/network/latency/regions/europe) | | Destination Region | [Africa](/docs/network/latency/regions/africa) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Marseille](https://www.geonames.org/2995469) (43.29695, 5.38107) and [Johannesburg](https://www.geonames.org/993800) (-26.20227, 28.04363). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **8,037.1 km** | | Vacuum RTT floor | **53.62 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **78.71 ms** | | Low-latency fiber material floor | **78.39 ms** | | Engineering floor (5% path allowance) | **82.65 ms** | | Research 1.33× mapped-fiber reference | **104.68 ms** | | Estimated unamplified path loss | **1687.8 dB** | | Transparent optical spans / inline amplifiers | **106 / 105** | | Published RTT inflation over fiber floor | **2.52×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **194.33 ms** | | Average RTT | **198.5 ms** | | Maximum RTT | **207.86 ms** | | Standard deviation | **3.18 ms** | | Stdev / average | **1.6%** | ## Route Context * Departure PoP: [Marseille (MRS)](/docs/network/latency/mrs-marseille) * Destination PoP: [Johannesburg (JNB)](/docs/network/latency/jnb-johannesburg) * Region overview: [Europe](/docs/network/latency/regions/europe) → [Africa](/docs/network/latency/regions/africa) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Johannesburg to Marseille latency and RTT](/docs/network/latency/pairs/jnb-mrs-rtt): 172 ms **Fastest routes departing Marseille (MRS)** * [Marseille to Paris latency and RTT](/docs/network/latency/pairs/mrs-par-rtt): 8.9 ms * [Marseille to Frankfurt latency and RTT](/docs/network/latency/pairs/mrs-fra-rtt): 16 ms * [Marseille to London latency and RTT](/docs/network/latency/pairs/mrs-lon-rtt): 18.9 ms * [Marseille to Amsterdam latency and RTT](/docs/network/latency/pairs/mrs-ams-rtt): 20.3 ms * [Marseille to Berlin latency and RTT](/docs/network/latency/pairs/mrs-ber-rtt): 20.6 ms **Fastest routes arriving at Johannesburg (JNB)** * [Amsterdam to Johannesburg latency and RTT](/docs/network/latency/pairs/ams-jnb-rtt): 171.9 ms * [Berlin to Johannesburg latency and RTT](/docs/network/latency/pairs/ber-jnb-rtt): 172.1 ms * [London to Johannesburg latency and RTT](/docs/network/latency/pairs/lon-jnb-rtt): 179.6 ms * [Paris to Johannesburg latency and RTT](/docs/network/latency/pairs/par-jnb-rtt): 180.9 ms * [Frankfurt to Johannesburg latency and RTT](/docs/network/latency/pairs/fra-jnb-rtt): 188.4 ms **Same corridor (Europe → Africa)** * [Moscow to Johannesburg latency and RTT](/docs/network/latency/pairs/mow-jnb-rtt): 201.9 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/mrs-jnb.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Marseille → Johannesburg round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/mrs-jnb-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Marseille, France → Los Angeles, USA RTT 🇫🇷 **Marseille, France (MRS)** → 🇺🇸 **Los Angeles, USA (LAX)** Measurement round: `2026-08-16T04:07:28Z`. Marseille to Los Angeles averaged 144.5 ms across 50 replies. The observed range was 140.24 to 154.02 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 3.04 ms, or 2.1% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **144.5 ms** | | Jitter | **2.51 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **95.09 ms** | | Fiber Efficiency | **65.8%** | | Latency Tier | Good | | Source Region | [Europe](/docs/network/latency/regions/europe) | | Destination Region | [North America](/docs/network/latency/regions/north-america) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Marseille](https://www.geonames.org/2995469) (43.29695, 5.38107) and [Los Angeles](https://www.geonames.org/5368361) (34.05223, -118.24368). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **9,710.4 km** | | Vacuum RTT floor | **64.78 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **95.09 ms** | | Low-latency fiber material floor | **94.71 ms** | | Engineering floor (5% path allowance) | **99.86 ms** | | Research 1.33× mapped-fiber reference | **126.47 ms** | | Estimated unamplified path loss | **2039.2 dB** | | Transparent optical spans / inline amplifiers | **128 / 127** | | Published RTT inflation over fiber floor | **1.52×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **140.24 ms** | | Average RTT | **144.5 ms** | | Maximum RTT | **154.02 ms** | | Standard deviation | **3.04 ms** | | Stdev / average | **2.1%** | ## Route Context * Departure PoP: [Marseille (MRS)](/docs/network/latency/mrs-marseille) * Destination PoP: [Los Angeles (LAX)](/docs/network/latency/lax-los-angeles) * Region overview: [Europe](/docs/network/latency/regions/europe) → [North America](/docs/network/latency/regions/north-america) * Corridor overview: [Europe to North America](/docs/network/latency/regions/europe-to-north-america) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Los Angeles to Marseille latency and RTT](/docs/network/latency/pairs/lax-mrs-rtt): 143.9 ms **Fastest routes departing Marseille (MRS)** * [Marseille to Paris latency and RTT](/docs/network/latency/pairs/mrs-par-rtt): 8.9 ms * [Marseille to Frankfurt latency and RTT](/docs/network/latency/pairs/mrs-fra-rtt): 16 ms * [Marseille to London latency and RTT](/docs/network/latency/pairs/mrs-lon-rtt): 18.9 ms * [Marseille to Amsterdam latency and RTT](/docs/network/latency/pairs/mrs-ams-rtt): 20.3 ms * [Marseille to Berlin latency and RTT](/docs/network/latency/pairs/mrs-ber-rtt): 20.6 ms **Fastest routes arriving at Los Angeles (LAX)** * [Seattle to Los Angeles latency and RTT](/docs/network/latency/pairs/sea-lax-rtt): 25.9 ms * [Miami to Los Angeles latency and RTT](/docs/network/latency/pairs/mia-lax-rtt): 56.8 ms * [New York to Los Angeles latency and RTT](/docs/network/latency/pairs/nyc-lax-rtt): 58.9 ms * [Ashburn to Los Angeles latency and RTT](/docs/network/latency/pairs/iad-lax-rtt): 60.7 ms * [Tokyo to Los Angeles latency and RTT](/docs/network/latency/pairs/tyo-lax-rtt): 101.2 ms **Same corridor (Europe → North America)** * [London to New York latency and RTT](/docs/network/latency/pairs/lon-nyc-rtt): 63.8 ms * [Amsterdam to New York latency and RTT](/docs/network/latency/pairs/ams-nyc-rtt): 68.6 ms * [Paris to New York latency and RTT](/docs/network/latency/pairs/par-nyc-rtt): 68.9 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/mrs-lax.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Marseille → Los Angeles round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/mrs-lax-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Marseille, France → London, UK RTT 🇫🇷 **Marseille, France (MRS)** → 🇬🇧 **London, UK (LON)** Measurement round: `2026-08-16T04:07:28Z`. Marseille to London averaged 18.9 ms across 50 replies. The observed range was 18.41 to 20.09 ms; these figures describe one probe round, not a long-term service target. This is number 3 of 19 outgoing routes by average RTT. That ranking compares different destinations and is not adjusted for distance. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------- | | RTT | **18.9 ms** | | Jitter | **0.34 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **9.82 ms** | | Fiber Efficiency | **51.9%** | | Latency Tier | Ultra-Low | | Source Region | [Europe](/docs/network/latency/regions/europe) | | Destination Region | [Europe](/docs/network/latency/regions/europe) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Marseille](https://www.geonames.org/2995469) (43.29695, 5.38107) and [London](https://www.geonames.org/2643743) (51.50853, -0.12574). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **1,002.4 km** | | Vacuum RTT floor | **6.69 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **9.82 ms** | | Low-latency fiber material floor | **9.78 ms** | | Engineering floor (5% path allowance) | **10.31 ms** | | Research 1.33× mapped-fiber reference | **13.06 ms** | | Estimated unamplified path loss | **210.5 dB** | | Transparent optical spans / inline amplifiers | **14 / 13** | | Published RTT inflation over fiber floor | **1.93×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------ | | Minimum RTT | **18.41 ms** | | Average RTT | **18.9 ms** | | Maximum RTT | **20.09 ms** | | Standard deviation | **0.42 ms** | | Stdev / average | **2.2%** | ## Route Context * Departure PoP: [Marseille (MRS)](/docs/network/latency/mrs-marseille) * Destination PoP: [London (LON)](/docs/network/latency/lon-london) * Region overview: [Europe](/docs/network/latency/regions/europe) → [Europe](/docs/network/latency/regions/europe) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [London to Marseille latency and RTT](/docs/network/latency/pairs/lon-mrs-rtt): 17.9 ms **Fastest routes departing Marseille (MRS)** * [Marseille to Paris latency and RTT](/docs/network/latency/pairs/mrs-par-rtt): 8.9 ms * [Marseille to Frankfurt latency and RTT](/docs/network/latency/pairs/mrs-fra-rtt): 16 ms * [Marseille to Amsterdam latency and RTT](/docs/network/latency/pairs/mrs-ams-rtt): 20.3 ms * [Marseille to Berlin latency and RTT](/docs/network/latency/pairs/mrs-ber-rtt): 20.6 ms * [Marseille to Moscow latency and RTT](/docs/network/latency/pairs/mrs-mow-rtt): 49.6 ms **Fastest routes arriving at London (LON)** * [Amsterdam to London latency and RTT](/docs/network/latency/pairs/ams-lon-rtt): 5.2 ms * [Paris to London latency and RTT](/docs/network/latency/pairs/par-lon-rtt): 6.4 ms * [Frankfurt to London latency and RTT](/docs/network/latency/pairs/fra-lon-rtt): 12.9 ms * [Berlin to London latency and RTT](/docs/network/latency/pairs/ber-lon-rtt): 15.2 ms * [Moscow to London latency and RTT](/docs/network/latency/pairs/mow-lon-rtt): 42.4 ms **Same corridor (Europe → Europe)** * [London to Amsterdam latency and RTT](/docs/network/latency/pairs/lon-ams-rtt): 5.2 ms * [Amsterdam to Frankfurt latency and RTT](/docs/network/latency/pairs/ams-fra-rtt): 5.9 ms * [Frankfurt to Amsterdam latency and RTT](/docs/network/latency/pairs/fra-ams-rtt): 5.9 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/mrs-lon.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Marseille → London round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/mrs-lon-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Marseille, France → Melbourne, Australia RTT 🇫🇷 **Marseille, France (MRS)** → 🇦🇺 **Melbourne, Australia (MEL)** Measurement round: `2026-08-16T04:07:28Z`. Marseille to Melbourne averaged 234.2 ms across 50 replies. The observed range was 225.73 to 253.61 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 7.53 ms, or 3.2% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------------------- | | RTT | **234.2 ms** | | Jitter | **5.82 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **162.36 ms** | | Fiber Efficiency | **69.3%** | | Latency Tier | Fair | | Source Region | [Europe](/docs/network/latency/regions/europe) | | Destination Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Marseille](https://www.geonames.org/2995469) (43.29695, 5.38107) and [Melbourne](https://www.geonames.org/2158177) (-37.81400, 144.96332). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | --------------- | | WGS-84 geodesic distance | **16,579.2 km** | | Vacuum RTT floor | **110.6 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **162.36 ms** | | Low-latency fiber material floor | **161.7 ms** | | Engineering floor (5% path allowance) | **170.5 ms** | | Research 1.33× mapped-fiber reference | **215.93 ms** | | Estimated unamplified path loss | **3481.6 dB** | | Transparent optical spans / inline amplifiers | **218 / 217** | | Published RTT inflation over fiber floor | **1.44×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **225.73 ms** | | Average RTT | **234.2 ms** | | Maximum RTT | **253.61 ms** | | Standard deviation | **7.53 ms** | | Stdev / average | **3.2%** | ## Route Context * Departure PoP: [Marseille (MRS)](/docs/network/latency/mrs-marseille) * Destination PoP: [Melbourne (MEL)](/docs/network/latency/mel-melbourne) * Region overview: [Europe](/docs/network/latency/regions/europe) → [Asia Pacific](/docs/network/latency/regions/asia-pacific) * Corridor overview: [Europe to Asia Pacific](/docs/network/latency/regions/europe-to-asia-pacific) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Melbourne to Marseille latency and RTT](/docs/network/latency/pairs/mel-mrs-rtt): 233.5 ms **Fastest routes departing Marseille (MRS)** * [Marseille to Paris latency and RTT](/docs/network/latency/pairs/mrs-par-rtt): 8.9 ms * [Marseille to Frankfurt latency and RTT](/docs/network/latency/pairs/mrs-fra-rtt): 16 ms * [Marseille to London latency and RTT](/docs/network/latency/pairs/mrs-lon-rtt): 18.9 ms * [Marseille to Amsterdam latency and RTT](/docs/network/latency/pairs/mrs-ams-rtt): 20.3 ms * [Marseille to Berlin latency and RTT](/docs/network/latency/pairs/mrs-ber-rtt): 20.6 ms **Fastest routes arriving at Melbourne (MEL)** * [Sydney to Melbourne latency and RTT](/docs/network/latency/pairs/syd-mel-rtt): 9.8 ms * [Singapore to Melbourne latency and RTT](/docs/network/latency/pairs/sin-mel-rtt): 88.6 ms * [Tokyo to Melbourne latency and RTT](/docs/network/latency/pairs/tyo-mel-rtt): 113.9 ms * [Hong Kong to Melbourne latency and RTT](/docs/network/latency/pairs/hkg-mel-rtt): 138.4 ms * [Taipei to Melbourne latency and RTT](/docs/network/latency/pairs/tpe-mel-rtt): 142 ms **Same corridor (Europe → Asia Pacific)** * [Moscow to Hong Kong latency and RTT](/docs/network/latency/pairs/mow-hkg-rtt): 118.6 ms * [Moscow to Taipei latency and RTT](/docs/network/latency/pairs/mow-tpe-rtt): 132.1 ms * [Marseille to Singapore latency and RTT](/docs/network/latency/pairs/mrs-sin-rtt): 139.6 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/mrs-mel.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Marseille → Melbourne round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/mrs-mel-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Marseille, France → Miami, USA RTT 🇫🇷 **Marseille, France (MRS)** → 🇺🇸 **Miami, USA (MIA)** Measurement round: `2026-08-16T04:07:28Z`. Marseille to Miami averaged 112.4 ms across 50 replies. The observed range was 105.38 to 129.43 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 5.13 ms, or 4.6% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **112.4 ms** | | Jitter | **4.52 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **75.92 ms** | | Fiber Efficiency | **67.5%** | | Latency Tier | Good | | Source Region | [Europe](/docs/network/latency/regions/europe) | | Destination Region | [North America](/docs/network/latency/regions/north-america) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Marseille](https://www.geonames.org/2995469) (43.29695, 5.38107) and [Miami](https://www.geonames.org/4164138) (25.77427, -80.19366). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **7,752.4 km** | | Vacuum RTT floor | **51.72 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **75.92 ms** | | Low-latency fiber material floor | **75.61 ms** | | Engineering floor (5% path allowance) | **79.72 ms** | | Research 1.33× mapped-fiber reference | **100.97 ms** | | Estimated unamplified path loss | **1628 dB** | | Transparent optical spans / inline amplifiers | **102 / 101** | | Published RTT inflation over fiber floor | **1.48×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **105.38 ms** | | Average RTT | **112.4 ms** | | Maximum RTT | **129.43 ms** | | Standard deviation | **5.13 ms** | | Stdev / average | **4.6%** | ## Route Context * Departure PoP: [Marseille (MRS)](/docs/network/latency/mrs-marseille) * Destination PoP: [Miami (MIA)](/docs/network/latency/mia-miami) * Region overview: [Europe](/docs/network/latency/regions/europe) → [North America](/docs/network/latency/regions/north-america) * Corridor overview: [Europe to North America](/docs/network/latency/regions/europe-to-north-america) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Miami to Marseille latency and RTT](/docs/network/latency/pairs/mia-mrs-rtt): 113.7 ms **Fastest routes departing Marseille (MRS)** * [Marseille to Paris latency and RTT](/docs/network/latency/pairs/mrs-par-rtt): 8.9 ms * [Marseille to Frankfurt latency and RTT](/docs/network/latency/pairs/mrs-fra-rtt): 16 ms * [Marseille to London latency and RTT](/docs/network/latency/pairs/mrs-lon-rtt): 18.9 ms * [Marseille to Amsterdam latency and RTT](/docs/network/latency/pairs/mrs-ams-rtt): 20.3 ms * [Marseille to Berlin latency and RTT](/docs/network/latency/pairs/mrs-ber-rtt): 20.6 ms **Fastest routes arriving at Miami (MIA)** * [Ashburn to Miami latency and RTT](/docs/network/latency/pairs/iad-mia-rtt): 27.4 ms * [New York to Miami latency and RTT](/docs/network/latency/pairs/nyc-mia-rtt): 32.9 ms * [Los Angeles to Miami latency and RTT](/docs/network/latency/pairs/lax-mia-rtt): 57.3 ms * [São Paulo to Miami latency and RTT](/docs/network/latency/pairs/gru-mia-rtt): 74 ms * [Seattle to Miami latency and RTT](/docs/network/latency/pairs/sea-mia-rtt): 81.7 ms **Same corridor (Europe → North America)** * [London to New York latency and RTT](/docs/network/latency/pairs/lon-nyc-rtt): 63.8 ms * [Amsterdam to New York latency and RTT](/docs/network/latency/pairs/ams-nyc-rtt): 68.6 ms * [Paris to New York latency and RTT](/docs/network/latency/pairs/par-nyc-rtt): 68.9 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/mrs-mia.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Marseille → Miami round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/mrs-mia-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Marseille, France → Moscow, Russia RTT 🇫🇷 **Marseille, France (MRS)** → 🇷🇺 **Moscow, Russia (MOW)** Measurement round: `2026-08-16T04:07:28Z`. Marseille to Moscow averaged 49.6 ms across 50 replies. The observed range was 47.26 to 58 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 2.18 ms, or 4.4% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------- | | RTT | **49.6 ms** | | Jitter | **1.92 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **26.23 ms** | | Fiber Efficiency | **52.9%** | | Latency Tier | Excellent | | Source Region | [Europe](/docs/network/latency/regions/europe) | | Destination Region | [Europe](/docs/network/latency/regions/europe) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Marseille](https://www.geonames.org/2995469) (43.29695, 5.38107) and [Moscow](https://www.geonames.org/524901) (55.75204, 37.61781). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **2,678.2 km** | | Vacuum RTT floor | **17.87 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **26.23 ms** | | Low-latency fiber material floor | **26.12 ms** | | Engineering floor (5% path allowance) | **27.54 ms** | | Research 1.33× mapped-fiber reference | **34.88 ms** | | Estimated unamplified path loss | **562.4 dB** | | Transparent optical spans / inline amplifiers | **36 / 35** | | Published RTT inflation over fiber floor | **1.89×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------ | | Minimum RTT | **47.26 ms** | | Average RTT | **49.6 ms** | | Maximum RTT | **58 ms** | | Standard deviation | **2.18 ms** | | Stdev / average | **4.4%** | ## Route Context * Departure PoP: [Marseille (MRS)](/docs/network/latency/mrs-marseille) * Destination PoP: [Moscow (MOW)](/docs/network/latency/mow-moscow) * Region overview: [Europe](/docs/network/latency/regions/europe) → [Europe](/docs/network/latency/regions/europe) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Moscow to Marseille latency and RTT](/docs/network/latency/pairs/mow-mrs-rtt): 49.9 ms **Fastest routes departing Marseille (MRS)** * [Marseille to Paris latency and RTT](/docs/network/latency/pairs/mrs-par-rtt): 8.9 ms * [Marseille to Frankfurt latency and RTT](/docs/network/latency/pairs/mrs-fra-rtt): 16 ms * [Marseille to London latency and RTT](/docs/network/latency/pairs/mrs-lon-rtt): 18.9 ms * [Marseille to Amsterdam latency and RTT](/docs/network/latency/pairs/mrs-ams-rtt): 20.3 ms * [Marseille to Berlin latency and RTT](/docs/network/latency/pairs/mrs-ber-rtt): 20.6 ms **Fastest routes arriving at Moscow (MOW)** * [Berlin to Moscow latency and RTT](/docs/network/latency/pairs/ber-mow-rtt): 28.6 ms * [Frankfurt to Moscow latency and RTT](/docs/network/latency/pairs/fra-mow-rtt): 34.7 ms * [Amsterdam to Moscow latency and RTT](/docs/network/latency/pairs/ams-mow-rtt): 38.1 ms * [London to Moscow latency and RTT](/docs/network/latency/pairs/lon-mow-rtt): 43.5 ms * [Paris to Moscow latency and RTT](/docs/network/latency/pairs/par-mow-rtt): 44.7 ms **Same corridor (Europe → Europe)** * [Amsterdam to London latency and RTT](/docs/network/latency/pairs/ams-lon-rtt): 5.2 ms * [London to Amsterdam latency and RTT](/docs/network/latency/pairs/lon-ams-rtt): 5.2 ms * [Amsterdam to Frankfurt latency and RTT](/docs/network/latency/pairs/ams-fra-rtt): 5.9 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/mrs-mow.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Marseille → Moscow round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/mrs-mow-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Marseille, France → New York, USA RTT 🇫🇷 **Marseille, France (MRS)** → 🇺🇸 **New York, USA (NYC)** Measurement round: `2026-08-16T04:07:28Z`. Marseille to New York averaged 79.8 ms across 50 replies. The observed range was 75.53 to 89.73 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 3.65 ms, or 4.6% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **79.8 ms** | | Jitter | **3.14 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **61.91 ms** | | Fiber Efficiency | **77.6%** | | Latency Tier | Excellent | | Source Region | [Europe](/docs/network/latency/regions/europe) | | Destination Region | [North America](/docs/network/latency/regions/north-america) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Marseille](https://www.geonames.org/2995469) (43.29695, 5.38107) and [New York City](https://www.geonames.org/5128581) (40.71427, -74.00597). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **6,321.9 km** | | Vacuum RTT floor | **42.18 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **61.91 ms** | | Low-latency fiber material floor | **61.66 ms** | | Engineering floor (5% path allowance) | **65.01 ms** | | Research 1.33× mapped-fiber reference | **82.34 ms** | | Estimated unamplified path loss | **1327.6 dB** | | Transparent optical spans / inline amplifiers | **83 / 82** | | Published RTT inflation over fiber floor | **1.29×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------ | | Minimum RTT | **75.53 ms** | | Average RTT | **79.8 ms** | | Maximum RTT | **89.73 ms** | | Standard deviation | **3.65 ms** | | Stdev / average | **4.6%** | ## Route Context * Departure PoP: [Marseille (MRS)](/docs/network/latency/mrs-marseille) * Destination PoP: [New York (NYC)](/docs/network/latency/nyc-new-york) * Region overview: [Europe](/docs/network/latency/regions/europe) → [North America](/docs/network/latency/regions/north-america) * Corridor overview: [Europe to North America](/docs/network/latency/regions/europe-to-north-america) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [New York to Marseille latency and RTT](/docs/network/latency/pairs/nyc-mrs-rtt): 80.2 ms **Fastest routes departing Marseille (MRS)** * [Marseille to Paris latency and RTT](/docs/network/latency/pairs/mrs-par-rtt): 8.9 ms * [Marseille to Frankfurt latency and RTT](/docs/network/latency/pairs/mrs-fra-rtt): 16 ms * [Marseille to London latency and RTT](/docs/network/latency/pairs/mrs-lon-rtt): 18.9 ms * [Marseille to Amsterdam latency and RTT](/docs/network/latency/pairs/mrs-ams-rtt): 20.3 ms * [Marseille to Berlin latency and RTT](/docs/network/latency/pairs/mrs-ber-rtt): 20.6 ms **Fastest routes arriving at New York (NYC)** * [Ashburn to New York latency and RTT](/docs/network/latency/pairs/iad-nyc-rtt): 6 ms * [Miami to New York latency and RTT](/docs/network/latency/pairs/mia-nyc-rtt): 33.3 ms * [Los Angeles to New York latency and RTT](/docs/network/latency/pairs/lax-nyc-rtt): 57.9 ms * [Seattle to New York latency and RTT](/docs/network/latency/pairs/sea-nyc-rtt): 59.5 ms * [London to New York latency and RTT](/docs/network/latency/pairs/lon-nyc-rtt): 63.8 ms **Same corridor (Europe → North America)** * [Amsterdam to New York latency and RTT](/docs/network/latency/pairs/ams-nyc-rtt): 68.6 ms * [Paris to New York latency and RTT](/docs/network/latency/pairs/par-nyc-rtt): 68.9 ms * [London to Ashburn latency and RTT](/docs/network/latency/pairs/lon-iad-rtt): 71 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/mrs-nyc.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Marseille → New York round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/mrs-nyc-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Marseille, France → Paris, France RTT 🇫🇷 **Marseille, France (MRS)** → 🇫🇷 **Paris, France (PAR)** Measurement round: `2026-08-16T04:07:28Z`. Marseille to Paris averaged 8.9 ms across 50 replies. The observed range was 8.72 to 9.67 ms; these figures describe one probe round, not a long-term service target. This is number 1 of 19 outgoing routes by average RTT. That ranking compares different destinations and is not adjusted for distance. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------- | | RTT | **8.9 ms** | | Jitter | **0.13 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **6.47 ms** | | Fiber Efficiency | **72.7%** | | Latency Tier | Ultra-Low | | Source Region | [Europe](/docs/network/latency/regions/europe) | | Destination Region | [Europe](/docs/network/latency/regions/europe) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Marseille](https://www.geonames.org/2995469) (43.29695, 5.38107) and [Paris](https://www.geonames.org/2988507) (48.85341, 2.34880). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | ------------ | | WGS-84 geodesic distance | **660.5 km** | | Vacuum RTT floor | **4.41 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **6.47 ms** | | Low-latency fiber material floor | **6.44 ms** | | Engineering floor (5% path allowance) | **6.79 ms** | | Research 1.33× mapped-fiber reference | **8.6 ms** | | Estimated unamplified path loss | **138.7 dB** | | Transparent optical spans / inline amplifiers | **9 / 8** | | Published RTT inflation over fiber floor | **1.38×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ----------- | | Minimum RTT | **8.72 ms** | | Average RTT | **8.9 ms** | | Maximum RTT | **9.67 ms** | | Standard deviation | **0.16 ms** | | Stdev / average | **1.8%** | ## Route Context * Departure PoP: [Marseille (MRS)](/docs/network/latency/mrs-marseille) * Destination PoP: [Paris (PAR)](/docs/network/latency/par-paris) * Region overview: [Europe](/docs/network/latency/regions/europe) → [Europe](/docs/network/latency/regions/europe) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Paris to Marseille latency and RTT](/docs/network/latency/pairs/par-mrs-rtt): 8.9 ms **Fastest routes departing Marseille (MRS)** * [Marseille to Frankfurt latency and RTT](/docs/network/latency/pairs/mrs-fra-rtt): 16 ms * [Marseille to London latency and RTT](/docs/network/latency/pairs/mrs-lon-rtt): 18.9 ms * [Marseille to Amsterdam latency and RTT](/docs/network/latency/pairs/mrs-ams-rtt): 20.3 ms * [Marseille to Berlin latency and RTT](/docs/network/latency/pairs/mrs-ber-rtt): 20.6 ms * [Marseille to Moscow latency and RTT](/docs/network/latency/pairs/mrs-mow-rtt): 49.6 ms **Fastest routes arriving at Paris (PAR)** * [London to Paris latency and RTT](/docs/network/latency/pairs/lon-par-rtt): 6.4 ms * [Amsterdam to Paris latency and RTT](/docs/network/latency/pairs/ams-par-rtt): 7 ms * [Frankfurt to Paris latency and RTT](/docs/network/latency/pairs/fra-par-rtt): 7.6 ms * [Berlin to Paris latency and RTT](/docs/network/latency/pairs/ber-par-rtt): 15.6 ms * [Moscow to Paris latency and RTT](/docs/network/latency/pairs/mow-par-rtt): 44 ms **Same corridor (Europe → Europe)** * [Amsterdam to London latency and RTT](/docs/network/latency/pairs/ams-lon-rtt): 5.2 ms * [London to Amsterdam latency and RTT](/docs/network/latency/pairs/lon-ams-rtt): 5.2 ms * [Amsterdam to Frankfurt latency and RTT](/docs/network/latency/pairs/ams-fra-rtt): 5.9 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/mrs-par.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Marseille → Paris round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/mrs-par-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Marseille, France → Seattle, USA RTT 🇫🇷 **Marseille, France (MRS)** → 🇺🇸 **Seattle, USA (SEA)** Measurement round: `2026-08-16T04:07:28Z`. Marseille to Seattle averaged 141.7 ms across 50 replies. The observed range was 135.29 to 163.72 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 5.67 ms, or 4.0% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **141.7 ms** | | Jitter | **4.7 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **85.27 ms** | | Fiber Efficiency | **60.2%** | | Latency Tier | Good | | Source Region | [Europe](/docs/network/latency/regions/europe) | | Destination Region | [North America](/docs/network/latency/regions/north-america) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Marseille](https://www.geonames.org/2995469) (43.29695, 5.38107) and [Seattle](https://www.geonames.org/5809844) (47.60621, -122.33207). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **8,707.5 km** | | Vacuum RTT floor | **58.09 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **85.27 ms** | | Low-latency fiber material floor | **84.93 ms** | | Engineering floor (5% path allowance) | **89.55 ms** | | Research 1.33× mapped-fiber reference | **113.41 ms** | | Estimated unamplified path loss | **1828.6 dB** | | Transparent optical spans / inline amplifiers | **115 / 114** | | Published RTT inflation over fiber floor | **1.66×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **135.29 ms** | | Average RTT | **141.7 ms** | | Maximum RTT | **163.72 ms** | | Standard deviation | **5.67 ms** | | Stdev / average | **4.0%** | ## Route Context * Departure PoP: [Marseille (MRS)](/docs/network/latency/mrs-marseille) * Destination PoP: [Seattle (SEA)](/docs/network/latency/sea-seattle) * Region overview: [Europe](/docs/network/latency/regions/europe) → [North America](/docs/network/latency/regions/north-america) * Corridor overview: [Europe to North America](/docs/network/latency/regions/europe-to-north-america) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Seattle to Marseille latency and RTT](/docs/network/latency/pairs/sea-mrs-rtt): 141.2 ms **Fastest routes departing Marseille (MRS)** * [Marseille to Paris latency and RTT](/docs/network/latency/pairs/mrs-par-rtt): 8.9 ms * [Marseille to Frankfurt latency and RTT](/docs/network/latency/pairs/mrs-fra-rtt): 16 ms * [Marseille to London latency and RTT](/docs/network/latency/pairs/mrs-lon-rtt): 18.9 ms * [Marseille to Amsterdam latency and RTT](/docs/network/latency/pairs/mrs-ams-rtt): 20.3 ms * [Marseille to Berlin latency and RTT](/docs/network/latency/pairs/mrs-ber-rtt): 20.6 ms **Fastest routes arriving at Seattle (SEA)** * [Los Angeles to Seattle latency and RTT](/docs/network/latency/pairs/lax-sea-rtt): 26.9 ms * [New York to Seattle latency and RTT](/docs/network/latency/pairs/nyc-sea-rtt): 58.7 ms * [Ashburn to Seattle latency and RTT](/docs/network/latency/pairs/iad-sea-rtt): 62.3 ms * [Miami to Seattle latency and RTT](/docs/network/latency/pairs/mia-sea-rtt): 81.9 ms * [Tokyo to Seattle latency and RTT](/docs/network/latency/pairs/tyo-sea-rtt): 84.8 ms **Same corridor (Europe → North America)** * [London to New York latency and RTT](/docs/network/latency/pairs/lon-nyc-rtt): 63.8 ms * [Amsterdam to New York latency and RTT](/docs/network/latency/pairs/ams-nyc-rtt): 68.6 ms * [Paris to New York latency and RTT](/docs/network/latency/pairs/par-nyc-rtt): 68.9 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/mrs-sea.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Marseille → Seattle round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/mrs-sea-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Marseille, France → Singapore RTT 🇫🇷 **Marseille, France (MRS)** → 🇸🇬 **Singapore (SIN)** Measurement round: `2026-08-16T04:07:28Z`. Marseille to Singapore averaged 139.6 ms across 50 replies. The observed range was 132.36 to 163.5 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 6.56 ms, or 4.7% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------------------- | | RTT | **139.6 ms** | | Jitter | **4.16 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **103.79 ms** | | Fiber Efficiency | **74.3%** | | Latency Tier | Good | | Source Region | [Europe](/docs/network/latency/regions/europe) | | Destination Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Marseille](https://www.geonames.org/2995469) (43.29695, 5.38107) and [Singapore](https://www.geonames.org/1880252) (1.28967, 103.85007). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | --------------- | | WGS-84 geodesic distance | **10,598.5 km** | | Vacuum RTT floor | **70.71 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **103.79 ms** | | Low-latency fiber material floor | **103.37 ms** | | Engineering floor (5% path allowance) | **108.99 ms** | | Research 1.33× mapped-fiber reference | **138.04 ms** | | Estimated unamplified path loss | **2225.7 dB** | | Transparent optical spans / inline amplifiers | **140 / 139** | | Published RTT inflation over fiber floor | **1.35×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **132.36 ms** | | Average RTT | **139.6 ms** | | Maximum RTT | **163.5 ms** | | Standard deviation | **6.56 ms** | | Stdev / average | **4.7%** | ## Route Context * Departure PoP: [Marseille (MRS)](/docs/network/latency/mrs-marseille) * Destination PoP: [Singapore (SIN)](/docs/network/latency/sin-singapore) * Region overview: [Europe](/docs/network/latency/regions/europe) → [Asia Pacific](/docs/network/latency/regions/asia-pacific) * Corridor overview: [Europe to Asia Pacific](/docs/network/latency/regions/europe-to-asia-pacific) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Singapore to Marseille latency and RTT](/docs/network/latency/pairs/sin-mrs-rtt): 140.6 ms **Fastest routes departing Marseille (MRS)** * [Marseille to Paris latency and RTT](/docs/network/latency/pairs/mrs-par-rtt): 8.9 ms * [Marseille to Frankfurt latency and RTT](/docs/network/latency/pairs/mrs-fra-rtt): 16 ms * [Marseille to London latency and RTT](/docs/network/latency/pairs/mrs-lon-rtt): 18.9 ms * [Marseille to Amsterdam latency and RTT](/docs/network/latency/pairs/mrs-ams-rtt): 20.3 ms * [Marseille to Berlin latency and RTT](/docs/network/latency/pairs/mrs-ber-rtt): 20.6 ms **Fastest routes arriving at Singapore (SIN)** * [Hong Kong to Singapore latency and RTT](/docs/network/latency/pairs/hkg-sin-rtt): 31.5 ms * [Taipei to Singapore latency and RTT](/docs/network/latency/pairs/tpe-sin-rtt): 44.3 ms * [Tokyo to Singapore latency and RTT](/docs/network/latency/pairs/tyo-sin-rtt): 69.4 ms * [Melbourne to Singapore latency and RTT](/docs/network/latency/pairs/mel-sin-rtt): 88.4 ms * [Sydney to Singapore latency and RTT](/docs/network/latency/pairs/syd-sin-rtt): 94.5 ms **Same corridor (Europe → Asia Pacific)** * [Moscow to Hong Kong latency and RTT](/docs/network/latency/pairs/mow-hkg-rtt): 118.6 ms * [Moscow to Taipei latency and RTT](/docs/network/latency/pairs/mow-tpe-rtt): 132.1 ms * [Berlin to Hong Kong latency and RTT](/docs/network/latency/pairs/ber-hkg-rtt): 144.9 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/mrs-sin.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Marseille → Singapore round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/mrs-sin-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Marseille, France → Sydney, Australia RTT 🇫🇷 **Marseille, France (MRS)** → 🇦🇺 **Sydney, Australia (SYD)** Measurement round: `2026-08-16T04:07:28Z`. Marseille to Sydney averaged 240.6 ms across 50 replies. The observed range was 231.48 to 266.22 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 8.19 ms, or 3.4% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------------------- | | RTT | **240.6 ms** | | Jitter | **6.59 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **165.41 ms** | | Fiber Efficiency | **68.7%** | | Latency Tier | Fair | | Source Region | [Europe](/docs/network/latency/regions/europe) | | Destination Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Marseille](https://www.geonames.org/2995469) (43.29695, 5.38107) and [Sydney](https://www.geonames.org/2147714) (-33.86785, 151.20732). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | ------------- | | WGS-84 geodesic distance | **16,891 km** | | Vacuum RTT floor | **112.68 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **165.41 ms** | | Low-latency fiber material floor | **164.74 ms** | | Engineering floor (5% path allowance) | **173.7 ms** | | Research 1.33× mapped-fiber reference | **219.99 ms** | | Estimated unamplified path loss | **3547.1 dB** | | Transparent optical spans / inline amplifiers | **222 / 221** | | Published RTT inflation over fiber floor | **1.45×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **231.48 ms** | | Average RTT | **240.6 ms** | | Maximum RTT | **266.22 ms** | | Standard deviation | **8.19 ms** | | Stdev / average | **3.4%** | ## Route Context * Departure PoP: [Marseille (MRS)](/docs/network/latency/mrs-marseille) * Destination PoP: [Sydney (SYD)](/docs/network/latency/syd-sydney) * Region overview: [Europe](/docs/network/latency/regions/europe) → [Asia Pacific](/docs/network/latency/regions/asia-pacific) * Corridor overview: [Europe to Asia Pacific](/docs/network/latency/regions/europe-to-asia-pacific) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Sydney to Marseille latency and RTT](/docs/network/latency/pairs/syd-mrs-rtt): 240.5 ms **Fastest routes departing Marseille (MRS)** * [Marseille to Paris latency and RTT](/docs/network/latency/pairs/mrs-par-rtt): 8.9 ms * [Marseille to Frankfurt latency and RTT](/docs/network/latency/pairs/mrs-fra-rtt): 16 ms * [Marseille to London latency and RTT](/docs/network/latency/pairs/mrs-lon-rtt): 18.9 ms * [Marseille to Amsterdam latency and RTT](/docs/network/latency/pairs/mrs-ams-rtt): 20.3 ms * [Marseille to Berlin latency and RTT](/docs/network/latency/pairs/mrs-ber-rtt): 20.6 ms **Fastest routes arriving at Sydney (SYD)** * [Melbourne to Sydney latency and RTT](/docs/network/latency/pairs/mel-syd-rtt): 9.8 ms * [Singapore to Sydney latency and RTT](/docs/network/latency/pairs/sin-syd-rtt): 94.5 ms * [Tokyo to Sydney latency and RTT](/docs/network/latency/pairs/tyo-syd-rtt): 101.9 ms * [Taipei to Sydney latency and RTT](/docs/network/latency/pairs/tpe-syd-rtt): 133.2 ms * [Hong Kong to Sydney latency and RTT](/docs/network/latency/pairs/hkg-syd-rtt): 137.2 ms **Same corridor (Europe → Asia Pacific)** * [Moscow to Hong Kong latency and RTT](/docs/network/latency/pairs/mow-hkg-rtt): 118.6 ms * [Moscow to Taipei latency and RTT](/docs/network/latency/pairs/mow-tpe-rtt): 132.1 ms * [Marseille to Singapore latency and RTT](/docs/network/latency/pairs/mrs-sin-rtt): 139.6 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/mrs-syd.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Marseille → Sydney round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/mrs-syd-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Marseille, France → Taipei, Taiwan RTT 🇫🇷 **Marseille, France (MRS)** → 🇹🇼 **Taipei, Taiwan (TPE)** Measurement round: `2026-08-16T04:07:28Z`. Marseille to Taipei averaged 180.9 ms across 50 replies. The observed range was 172.41 to 200.88 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 6.92 ms, or 3.8% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------------------- | | RTT | **180.9 ms** | | Jitter | **5.8 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **98.2 ms** | | Fiber Efficiency | **54.3%** | | Latency Tier | Fair | | Source Region | [Europe](/docs/network/latency/regions/europe) | | Destination Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Marseille](https://www.geonames.org/2995469) (43.29695, 5.38107) and [Taipei](https://www.geonames.org/1668341) (25.05306, 121.52639). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | --------------- | | WGS-84 geodesic distance | **10,027.3 km** | | Vacuum RTT floor | **66.9 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **98.2 ms** | | Low-latency fiber material floor | **97.8 ms** | | Engineering floor (5% path allowance) | **103.12 ms** | | Research 1.33× mapped-fiber reference | **130.6 ms** | | Estimated unamplified path loss | **2105.7 dB** | | Transparent optical spans / inline amplifiers | **132 / 131** | | Published RTT inflation over fiber floor | **1.84×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **172.41 ms** | | Average RTT | **180.9 ms** | | Maximum RTT | **200.88 ms** | | Standard deviation | **6.92 ms** | | Stdev / average | **3.8%** | ## Route Context * Departure PoP: [Marseille (MRS)](/docs/network/latency/mrs-marseille) * Destination PoP: [Taipei (TPE)](/docs/network/latency/tpe-taipei) * Region overview: [Europe](/docs/network/latency/regions/europe) → [Asia Pacific](/docs/network/latency/regions/asia-pacific) * Corridor overview: [Europe to Asia Pacific](/docs/network/latency/regions/europe-to-asia-pacific) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Taipei to Marseille latency and RTT](/docs/network/latency/pairs/tpe-mrs-rtt): 179.5 ms **Fastest routes departing Marseille (MRS)** * [Marseille to Paris latency and RTT](/docs/network/latency/pairs/mrs-par-rtt): 8.9 ms * [Marseille to Frankfurt latency and RTT](/docs/network/latency/pairs/mrs-fra-rtt): 16 ms * [Marseille to London latency and RTT](/docs/network/latency/pairs/mrs-lon-rtt): 18.9 ms * [Marseille to Amsterdam latency and RTT](/docs/network/latency/pairs/mrs-ams-rtt): 20.3 ms * [Marseille to Berlin latency and RTT](/docs/network/latency/pairs/mrs-ber-rtt): 20.6 ms **Fastest routes arriving at Taipei (TPE)** * [Hong Kong to Taipei latency and RTT](/docs/network/latency/pairs/hkg-tpe-rtt): 14.7 ms * [Tokyo to Taipei latency and RTT](/docs/network/latency/pairs/tyo-tpe-rtt): 32.3 ms * [Singapore to Taipei latency and RTT](/docs/network/latency/pairs/sin-tpe-rtt): 45.8 ms * [Seattle to Taipei latency and RTT](/docs/network/latency/pairs/sea-tpe-rtt): 114.7 ms * [Sydney to Taipei latency and RTT](/docs/network/latency/pairs/syd-tpe-rtt): 131.7 ms **Same corridor (Europe → Asia Pacific)** * [Moscow to Hong Kong latency and RTT](/docs/network/latency/pairs/mow-hkg-rtt): 118.6 ms * [Moscow to Taipei latency and RTT](/docs/network/latency/pairs/mow-tpe-rtt): 132.1 ms * [Marseille to Singapore latency and RTT](/docs/network/latency/pairs/mrs-sin-rtt): 139.6 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/mrs-tpe.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Marseille → Taipei round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/mrs-tpe-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Marseille, France → Tokyo, Japan RTT 🇫🇷 **Marseille, France (MRS)** → 🇯🇵 **Tokyo, Japan (TYO)** Measurement round: `2026-08-16T04:07:28Z`. Marseille to Tokyo averaged 203.8 ms, compared with a 99.03 ms city-centre fiber reference. The difference cannot be assigned to a particular cable or carrier from RTT alone. Standard deviation was 8.55 ms, or 4.2% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------------------- | | RTT | **203.8 ms** | | Jitter | **7.35 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **99.03 ms** | | Fiber Efficiency | **48.6%** | | Latency Tier | Fair | | Source Region | [Europe](/docs/network/latency/regions/europe) | | Destination Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Marseille](https://www.geonames.org/2995469) (43.29695, 5.38107) and [Tokyo](https://www.geonames.org/1850147) (35.68950, 139.69171). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | --------------- | | WGS-84 geodesic distance | **10,112.8 km** | | Vacuum RTT floor | **67.47 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **99.03 ms** | | Low-latency fiber material floor | **98.63 ms** | | Engineering floor (5% path allowance) | **104 ms** | | Research 1.33× mapped-fiber reference | **131.71 ms** | | Estimated unamplified path loss | **2123.7 dB** | | Transparent optical spans / inline amplifiers | **133 / 132** | | Published RTT inflation over fiber floor | **2.06×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **194.59 ms** | | Average RTT | **203.8 ms** | | Maximum RTT | **241.85 ms** | | Standard deviation | **8.55 ms** | | Stdev / average | **4.2%** | ## Route Context * Departure PoP: [Marseille (MRS)](/docs/network/latency/mrs-marseille) * Destination PoP: [Tokyo (TYO)](/docs/network/latency/tyo-tokyo) * Region overview: [Europe](/docs/network/latency/regions/europe) → [Asia Pacific](/docs/network/latency/regions/asia-pacific) * Corridor overview: [Europe to Asia Pacific](/docs/network/latency/regions/europe-to-asia-pacific) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Tokyo to Marseille latency and RTT](/docs/network/latency/pairs/tyo-mrs-rtt): 205.6 ms **Fastest routes departing Marseille (MRS)** * [Marseille to Paris latency and RTT](/docs/network/latency/pairs/mrs-par-rtt): 8.9 ms * [Marseille to Frankfurt latency and RTT](/docs/network/latency/pairs/mrs-fra-rtt): 16 ms * [Marseille to London latency and RTT](/docs/network/latency/pairs/mrs-lon-rtt): 18.9 ms * [Marseille to Amsterdam latency and RTT](/docs/network/latency/pairs/mrs-ams-rtt): 20.3 ms * [Marseille to Berlin latency and RTT](/docs/network/latency/pairs/mrs-ber-rtt): 20.6 ms **Fastest routes arriving at Tokyo (TYO)** * [Taipei to Tokyo latency and RTT](/docs/network/latency/pairs/tpe-tyo-rtt): 31.9 ms * [Hong Kong to Tokyo latency and RTT](/docs/network/latency/pairs/hkg-tyo-rtt): 44.9 ms * [Singapore to Tokyo latency and RTT](/docs/network/latency/pairs/sin-tyo-rtt): 68.9 ms * [Seattle to Tokyo latency and RTT](/docs/network/latency/pairs/sea-tyo-rtt): 85 ms * [Los Angeles to Tokyo latency and RTT](/docs/network/latency/pairs/lax-tyo-rtt): 101.2 ms **Same corridor (Europe → Asia Pacific)** * [Moscow to Hong Kong latency and RTT](/docs/network/latency/pairs/mow-hkg-rtt): 118.6 ms * [Moscow to Taipei latency and RTT](/docs/network/latency/pairs/mow-tpe-rtt): 132.1 ms * [Marseille to Singapore latency and RTT](/docs/network/latency/pairs/mrs-sin-rtt): 139.6 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/mrs-tyo.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Marseille → Tokyo round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/mrs-tyo-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # New York, USA → Amsterdam, Netherlands RTT 🇺🇸 **New York, USA (NYC)** → 🇳🇱 **Amsterdam, Netherlands (AMS)** Measurement round: `2026-08-16T04:07:28Z`. New York to Amsterdam averaged 68.7 ms across 50 replies. The observed range was 66.97 to 74.13 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 1.38 ms, or 2.0% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **68.7 ms** | | Jitter | **1.26 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **57.56 ms** | | Fiber Efficiency | **83.8%** | | Latency Tier | Excellent | | Source Region | [North America](/docs/network/latency/regions/north-america) | | Destination Region | [Europe](/docs/network/latency/regions/europe) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [New York City](https://www.geonames.org/5128581) (40.71427, -74.00597) and [Amsterdam](https://www.geonames.org/2759794) (52.37403, 4.88969). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **5,877.8 km** | | Vacuum RTT floor | **39.21 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **57.56 ms** | | Low-latency fiber material floor | **57.33 ms** | | Engineering floor (5% path allowance) | **60.45 ms** | | Research 1.33× mapped-fiber reference | **76.56 ms** | | Estimated unamplified path loss | **1234.3 dB** | | Transparent optical spans / inline amplifiers | **78 / 77** | | Published RTT inflation over fiber floor | **1.19×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------ | | Minimum RTT | **66.97 ms** | | Average RTT | **68.7 ms** | | Maximum RTT | **74.13 ms** | | Standard deviation | **1.38 ms** | | Stdev / average | **2.0%** | ## Route Context * Departure PoP: [New York (NYC)](/docs/network/latency/nyc-new-york) * Destination PoP: [Amsterdam (AMS)](/docs/network/latency/ams-amsterdam) * Region overview: [North America](/docs/network/latency/regions/north-america) → [Europe](/docs/network/latency/regions/europe) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Amsterdam to New York latency and RTT](/docs/network/latency/pairs/ams-nyc-rtt): 68.6 ms **Fastest routes departing New York (NYC)** * [New York to Ashburn latency and RTT](/docs/network/latency/pairs/nyc-iad-rtt): 6 ms * [New York to Miami latency and RTT](/docs/network/latency/pairs/nyc-mia-rtt): 32.9 ms * [New York to Seattle latency and RTT](/docs/network/latency/pairs/nyc-sea-rtt): 58.7 ms * [New York to Los Angeles latency and RTT](/docs/network/latency/pairs/nyc-lax-rtt): 58.9 ms * [New York to London latency and RTT](/docs/network/latency/pairs/nyc-lon-rtt): 63.5 ms **Fastest routes arriving at Amsterdam (AMS)** * [London to Amsterdam latency and RTT](/docs/network/latency/pairs/lon-ams-rtt): 5.2 ms * [Frankfurt to Amsterdam latency and RTT](/docs/network/latency/pairs/fra-ams-rtt): 5.9 ms * [Paris to Amsterdam latency and RTT](/docs/network/latency/pairs/par-ams-rtt): 7 ms * [Berlin to Amsterdam latency and RTT](/docs/network/latency/pairs/ber-ams-rtt): 7.2 ms * [Marseille to Amsterdam latency and RTT](/docs/network/latency/pairs/mrs-ams-rtt): 20.3 ms **Same corridor (North America → Europe)** * [New York to Paris latency and RTT](/docs/network/latency/pairs/nyc-par-rtt): 68.9 ms * [Ashburn to London latency and RTT](/docs/network/latency/pairs/iad-lon-rtt): 70.3 ms * [New York to Frankfurt latency and RTT](/docs/network/latency/pairs/nyc-fra-rtt): 73.9 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/nyc-ams.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, New York → Amsterdam round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/nyc-ams-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # New York, USA → Berlin, Germany RTT 🇺🇸 **New York, USA (NYC)** → 🇩🇪 **Berlin, Germany (BER)** Measurement round: `2026-08-16T04:07:28Z`. New York to Berlin averaged 78.8 ms across 50 replies. The observed range was 77.15 to 84.9 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 1.59 ms, or 2.0% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **78.8 ms** | | Jitter | **1.13 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **62.7 ms** | | Fiber Efficiency | **79.6%** | | Latency Tier | Excellent | | Source Region | [North America](/docs/network/latency/regions/north-america) | | Destination Region | [Europe](/docs/network/latency/regions/europe) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [New York City](https://www.geonames.org/5128581) (40.71427, -74.00597) and [Berlin](https://www.geonames.org/2950159) (52.52437, 13.41053). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **6,402.4 km** | | Vacuum RTT floor | **42.71 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **62.7 ms** | | Low-latency fiber material floor | **62.45 ms** | | Engineering floor (5% path allowance) | **65.84 ms** | | Research 1.33× mapped-fiber reference | **83.39 ms** | | Estimated unamplified path loss | **1344.5 dB** | | Transparent optical spans / inline amplifiers | **85 / 84** | | Published RTT inflation over fiber floor | **1.26×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------ | | Minimum RTT | **77.15 ms** | | Average RTT | **78.8 ms** | | Maximum RTT | **84.9 ms** | | Standard deviation | **1.59 ms** | | Stdev / average | **2.0%** | ## Route Context * Departure PoP: [New York (NYC)](/docs/network/latency/nyc-new-york) * Destination PoP: [Berlin (BER)](/docs/network/latency/ber-berlin) * Region overview: [North America](/docs/network/latency/regions/north-america) → [Europe](/docs/network/latency/regions/europe) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Berlin to New York latency and RTT](/docs/network/latency/pairs/ber-nyc-rtt): 79.6 ms **Fastest routes departing New York (NYC)** * [New York to Ashburn latency and RTT](/docs/network/latency/pairs/nyc-iad-rtt): 6 ms * [New York to Miami latency and RTT](/docs/network/latency/pairs/nyc-mia-rtt): 32.9 ms * [New York to Seattle latency and RTT](/docs/network/latency/pairs/nyc-sea-rtt): 58.7 ms * [New York to Los Angeles latency and RTT](/docs/network/latency/pairs/nyc-lax-rtt): 58.9 ms * [New York to London latency and RTT](/docs/network/latency/pairs/nyc-lon-rtt): 63.5 ms **Fastest routes arriving at Berlin (BER)** * [Frankfurt to Berlin latency and RTT](/docs/network/latency/pairs/fra-ber-rtt): 6.1 ms * [Amsterdam to Berlin latency and RTT](/docs/network/latency/pairs/ams-ber-rtt): 7.2 ms * [Paris to Berlin latency and RTT](/docs/network/latency/pairs/par-ber-rtt): 15.5 ms * [London to Berlin latency and RTT](/docs/network/latency/pairs/lon-ber-rtt): 16.2 ms * [Marseille to Berlin latency and RTT](/docs/network/latency/pairs/mrs-ber-rtt): 20.6 ms **Same corridor (North America → Europe)** * [New York to Amsterdam latency and RTT](/docs/network/latency/pairs/nyc-ams-rtt): 68.7 ms * [New York to Paris latency and RTT](/docs/network/latency/pairs/nyc-par-rtt): 68.9 ms * [Ashburn to London latency and RTT](/docs/network/latency/pairs/iad-lon-rtt): 70.3 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/nyc-ber.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, New York → Berlin round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/nyc-ber-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # New York, USA → Frankfurt, Germany RTT 🇺🇸 **New York, USA (NYC)** → 🇩🇪 **Frankfurt, Germany (FRA)** Measurement round: `2026-08-16T04:07:28Z`. New York to Frankfurt averaged 73.9 ms across 50 replies. The observed range was 71.94 to 81.47 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 1.98 ms, or 2.7% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **73.9 ms** | | Jitter | **1.42 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **60.9 ms** | | Fiber Efficiency | **82.4%** | | Latency Tier | Excellent | | Source Region | [North America](/docs/network/latency/regions/north-america) | | Destination Region | [Europe](/docs/network/latency/regions/europe) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [New York City](https://www.geonames.org/5128581) (40.71427, -74.00597) and [Frankfurt am Main](https://www.geonames.org/2925533) (50.11552, 8.68417). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **6,219.3 km** | | Vacuum RTT floor | **41.49 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **60.9 ms** | | Low-latency fiber material floor | **60.66 ms** | | Engineering floor (5% path allowance) | **63.96 ms** | | Research 1.33× mapped-fiber reference | **81 ms** | | Estimated unamplified path loss | **1306 dB** | | Transparent optical spans / inline amplifiers | **82 / 81** | | Published RTT inflation over fiber floor | **1.21×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------ | | Minimum RTT | **71.94 ms** | | Average RTT | **73.9 ms** | | Maximum RTT | **81.47 ms** | | Standard deviation | **1.98 ms** | | Stdev / average | **2.7%** | ## Route Context * Departure PoP: [New York (NYC)](/docs/network/latency/nyc-new-york) * Destination PoP: [Frankfurt (FRA)](/docs/network/latency/fra-frankfurt) * Region overview: [North America](/docs/network/latency/regions/north-america) → [Europe](/docs/network/latency/regions/europe) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Frankfurt to New York latency and RTT](/docs/network/latency/pairs/fra-nyc-rtt): 74.4 ms **Fastest routes departing New York (NYC)** * [New York to Ashburn latency and RTT](/docs/network/latency/pairs/nyc-iad-rtt): 6 ms * [New York to Miami latency and RTT](/docs/network/latency/pairs/nyc-mia-rtt): 32.9 ms * [New York to Seattle latency and RTT](/docs/network/latency/pairs/nyc-sea-rtt): 58.7 ms * [New York to Los Angeles latency and RTT](/docs/network/latency/pairs/nyc-lax-rtt): 58.9 ms * [New York to London latency and RTT](/docs/network/latency/pairs/nyc-lon-rtt): 63.5 ms **Fastest routes arriving at Frankfurt (FRA)** * [Amsterdam to Frankfurt latency and RTT](/docs/network/latency/pairs/ams-fra-rtt): 5.9 ms * [Berlin to Frankfurt latency and RTT](/docs/network/latency/pairs/ber-fra-rtt): 6.1 ms * [Paris to Frankfurt latency and RTT](/docs/network/latency/pairs/par-fra-rtt): 7.6 ms * [London to Frankfurt latency and RTT](/docs/network/latency/pairs/lon-fra-rtt): 13.6 ms * [Marseille to Frankfurt latency and RTT](/docs/network/latency/pairs/mrs-fra-rtt): 16 ms **Same corridor (North America → Europe)** * [New York to Amsterdam latency and RTT](/docs/network/latency/pairs/nyc-ams-rtt): 68.7 ms * [New York to Paris latency and RTT](/docs/network/latency/pairs/nyc-par-rtt): 68.9 ms * [Ashburn to London latency and RTT](/docs/network/latency/pairs/iad-lon-rtt): 70.3 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/nyc-fra.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, New York → Frankfurt round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/nyc-fra-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # New York, USA → São Paulo, Brazil RTT 🇺🇸 **New York, USA (NYC)** → 🇧🇷 **São Paulo, Brazil (GRU)** Measurement round: `2026-08-16T04:07:28Z`. New York to São Paulo averaged 106.9 ms across 50 replies. The observed range was 104.25 to 117.31 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 2.21 ms, or 2.1% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **106.9 ms** | | Jitter | **1.84 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **74.99 ms** | | Fiber Efficiency | **70.1%** | | Latency Tier | Good | | Source Region | [North America](/docs/network/latency/regions/north-america) | | Destination Region | [South America](/docs/network/latency/regions/south-america) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [New York City](https://www.geonames.org/5128581) (40.71427, -74.00597) and [São Paulo](https://www.geonames.org/3448439) (-23.54750, -46.63611). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **7,657.6 km** | | Vacuum RTT floor | **51.09 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **74.99 ms** | | Low-latency fiber material floor | **74.69 ms** | | Engineering floor (5% path allowance) | **78.75 ms** | | Research 1.33× mapped-fiber reference | **99.74 ms** | | Estimated unamplified path loss | **1608.1 dB** | | Transparent optical spans / inline amplifiers | **101 / 100** | | Published RTT inflation over fiber floor | **1.43×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **104.25 ms** | | Average RTT | **106.9 ms** | | Maximum RTT | **117.31 ms** | | Standard deviation | **2.21 ms** | | Stdev / average | **2.1%** | ## Route Context * Departure PoP: [New York (NYC)](/docs/network/latency/nyc-new-york) * Destination PoP: [São Paulo (GRU)](/docs/network/latency/gru-sao-paulo) * Region overview: [North America](/docs/network/latency/regions/north-america) → [South America](/docs/network/latency/regions/south-america) * Corridor overview: [North America to South America](/docs/network/latency/regions/north-america-to-south-america) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [São Paulo to New York latency and RTT](/docs/network/latency/pairs/gru-nyc-rtt): 122 ms **Fastest routes departing New York (NYC)** * [New York to Ashburn latency and RTT](/docs/network/latency/pairs/nyc-iad-rtt): 6 ms * [New York to Miami latency and RTT](/docs/network/latency/pairs/nyc-mia-rtt): 32.9 ms * [New York to Seattle latency and RTT](/docs/network/latency/pairs/nyc-sea-rtt): 58.7 ms * [New York to Los Angeles latency and RTT](/docs/network/latency/pairs/nyc-lax-rtt): 58.9 ms * [New York to London latency and RTT](/docs/network/latency/pairs/nyc-lon-rtt): 63.5 ms **Fastest routes arriving at São Paulo (GRU)** * [Ashburn to São Paulo latency and RTT](/docs/network/latency/pairs/iad-gru-rtt): 101.4 ms * [Miami to São Paulo latency and RTT](/docs/network/latency/pairs/mia-gru-rtt): 128.7 ms * [Los Angeles to São Paulo latency and RTT](/docs/network/latency/pairs/lax-gru-rtt): 131.3 ms * [Seattle to São Paulo latency and RTT](/docs/network/latency/pairs/sea-gru-rtt): 155.7 ms * [London to São Paulo latency and RTT](/docs/network/latency/pairs/lon-gru-rtt): 170.7 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/nyc-gru.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, New York → São Paulo round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/nyc-gru-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # New York, USA → Hong Kong RTT 🇺🇸 **New York, USA (NYC)** → 🇭🇰 **Hong Kong (HKG)** Measurement round: `2026-08-16T04:07:28Z`. New York to Hong Kong averaged 187.1 ms across 50 replies. The observed range was 178.46 to 215.39 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 7.38 ms, or 3.9% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **187.1 ms** | | Jitter | **6.29 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **127.11 ms** | | Fiber Efficiency | **67.9%** | | Latency Tier | Fair | | Source Region | [North America](/docs/network/latency/regions/north-america) | | Destination Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [New York City](https://www.geonames.org/5128581) (40.71427, -74.00597) and [Hong Kong](https://www.geonames.org/1819729) (22.27832, 114.17469). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | --------------- | | WGS-84 geodesic distance | **12,979.6 km** | | Vacuum RTT floor | **86.59 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **127.11 ms** | | Low-latency fiber material floor | **126.6 ms** | | Engineering floor (5% path allowance) | **133.48 ms** | | Research 1.33× mapped-fiber reference | **169.05 ms** | | Estimated unamplified path loss | **2725.7 dB** | | Transparent optical spans / inline amplifiers | **171 / 170** | | Published RTT inflation over fiber floor | **1.47×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **178.46 ms** | | Average RTT | **187.1 ms** | | Maximum RTT | **215.39 ms** | | Standard deviation | **7.38 ms** | | Stdev / average | **3.9%** | ## Route Context * Departure PoP: [New York (NYC)](/docs/network/latency/nyc-new-york) * Destination PoP: [Hong Kong (HKG)](/docs/network/latency/hkg-hong-kong) * Region overview: [North America](/docs/network/latency/regions/north-america) → [Asia Pacific](/docs/network/latency/regions/asia-pacific) * Corridor overview: [North America to Asia Pacific](/docs/network/latency/regions/north-america-to-asia-pacific) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Hong Kong to New York latency and RTT](/docs/network/latency/pairs/hkg-nyc-rtt): 184.7 ms **Fastest routes departing New York (NYC)** * [New York to Ashburn latency and RTT](/docs/network/latency/pairs/nyc-iad-rtt): 6 ms * [New York to Miami latency and RTT](/docs/network/latency/pairs/nyc-mia-rtt): 32.9 ms * [New York to Seattle latency and RTT](/docs/network/latency/pairs/nyc-sea-rtt): 58.7 ms * [New York to Los Angeles latency and RTT](/docs/network/latency/pairs/nyc-lax-rtt): 58.9 ms * [New York to London latency and RTT](/docs/network/latency/pairs/nyc-lon-rtt): 63.5 ms **Fastest routes arriving at Hong Kong (HKG)** * [Taipei to Hong Kong latency and RTT](/docs/network/latency/pairs/tpe-hkg-rtt): 15.5 ms * [Singapore to Hong Kong latency and RTT](/docs/network/latency/pairs/sin-hkg-rtt): 30.8 ms * [Tokyo to Hong Kong latency and RTT](/docs/network/latency/pairs/tyo-hkg-rtt): 44.8 ms * [Moscow to Hong Kong latency and RTT](/docs/network/latency/pairs/mow-hkg-rtt): 118.6 ms * [Seattle to Hong Kong latency and RTT](/docs/network/latency/pairs/sea-hkg-rtt): 130.4 ms **Same corridor (North America → Asia Pacific)** * [Seattle to Tokyo latency and RTT](/docs/network/latency/pairs/sea-tyo-rtt): 85 ms * [Los Angeles to Tokyo latency and RTT](/docs/network/latency/pairs/lax-tyo-rtt): 101.2 ms * [Seattle to Taipei latency and RTT](/docs/network/latency/pairs/sea-tpe-rtt): 114.7 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/nyc-hkg.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, New York → Hong Kong round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/nyc-hkg-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # New York, USA → Ashburn, USA RTT 🇺🇸 **New York, USA (NYC)** → 🇺🇸 **Ashburn, USA (IAD)** Measurement round: `2026-08-16T04:07:28Z`. New York to Ashburn averaged 6 ms across 50 replies. The observed range was 5.58 to 7.18 ms; these figures describe one probe round, not a long-term service target. This is number 1 of 19 outgoing routes by average RTT. That ranking compares different destinations and is not adjusted for distance. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **6 ms** | | Jitter | **0.26 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **3.44 ms** | | Fiber Efficiency | **57.3%** | | Latency Tier | Ultra-Low | | Source Region | [North America](/docs/network/latency/regions/north-america) | | Destination Region | [North America](/docs/network/latency/regions/north-america) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [New York City](https://www.geonames.org/5128581) (40.71427, -74.00597) and [Ashburn](https://www.geonames.org/4744870) (39.04372, -77.48749). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | ------------ | | WGS-84 geodesic distance | **350.8 km** | | Vacuum RTT floor | **2.34 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **3.44 ms** | | Low-latency fiber material floor | **3.42 ms** | | Engineering floor (5% path allowance) | **3.61 ms** | | Research 1.33× mapped-fiber reference | **4.57 ms** | | Estimated unamplified path loss | **73.7 dB** | | Transparent optical spans / inline amplifiers | **5 / 4** | | Published RTT inflation over fiber floor | **1.75×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ----------- | | Minimum RTT | **5.58 ms** | | Average RTT | **6 ms** | | Maximum RTT | **7.18 ms** | | Standard deviation | **0.3 ms** | | Stdev / average | **5.0%** | ## Route Context * Departure PoP: [New York (NYC)](/docs/network/latency/nyc-new-york) * Destination PoP: [Ashburn (IAD)](/docs/network/latency/iad-ashburn) * Region overview: [North America](/docs/network/latency/regions/north-america) → [North America](/docs/network/latency/regions/north-america) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Ashburn to New York latency and RTT](/docs/network/latency/pairs/iad-nyc-rtt): 6 ms **Fastest routes departing New York (NYC)** * [New York to Miami latency and RTT](/docs/network/latency/pairs/nyc-mia-rtt): 32.9 ms * [New York to Seattle latency and RTT](/docs/network/latency/pairs/nyc-sea-rtt): 58.7 ms * [New York to Los Angeles latency and RTT](/docs/network/latency/pairs/nyc-lax-rtt): 58.9 ms * [New York to London latency and RTT](/docs/network/latency/pairs/nyc-lon-rtt): 63.5 ms * [New York to Amsterdam latency and RTT](/docs/network/latency/pairs/nyc-ams-rtt): 68.7 ms **Fastest routes arriving at Ashburn (IAD)** * [Miami to Ashburn latency and RTT](/docs/network/latency/pairs/mia-iad-rtt): 27.3 ms * [Los Angeles to Ashburn latency and RTT](/docs/network/latency/pairs/lax-iad-rtt): 60.4 ms * [Seattle to Ashburn latency and RTT](/docs/network/latency/pairs/sea-iad-rtt): 61.5 ms * [London to Ashburn latency and RTT](/docs/network/latency/pairs/lon-iad-rtt): 71 ms * [Paris to Ashburn latency and RTT](/docs/network/latency/pairs/par-iad-rtt): 75.5 ms **Same corridor (North America → North America)** * [Seattle to Los Angeles latency and RTT](/docs/network/latency/pairs/sea-lax-rtt): 25.9 ms * [Los Angeles to Seattle latency and RTT](/docs/network/latency/pairs/lax-sea-rtt): 26.9 ms * [Ashburn to Miami latency and RTT](/docs/network/latency/pairs/iad-mia-rtt): 27.4 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/nyc-iad.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, New York → Ashburn round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/nyc-iad-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # New York, USA → Johannesburg, South Africa RTT 🇺🇸 **New York, USA (NYC)** → 🇿🇦 **Johannesburg, South Africa (JNB)** Measurement round: `2026-08-16T04:07:28Z`. New York to Johannesburg averaged 221.5 ms across 50 replies. The observed range was 209.09 to 252.44 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 9.18 ms, or 4.1% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **221.5 ms** | | Jitter | **8.44 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **125.67 ms** | | Fiber Efficiency | **56.7%** | | Latency Tier | Fair | | Source Region | [North America](/docs/network/latency/regions/north-america) | | Destination Region | [Africa](/docs/network/latency/regions/africa) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [New York City](https://www.geonames.org/5128581) (40.71427, -74.00597) and [Johannesburg](https://www.geonames.org/993800) (-26.20227, 28.04363). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | --------------- | | WGS-84 geodesic distance | **12,832.9 km** | | Vacuum RTT floor | **85.61 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **125.67 ms** | | Low-latency fiber material floor | **125.16 ms** | | Engineering floor (5% path allowance) | **131.97 ms** | | Research 1.33× mapped-fiber reference | **167.14 ms** | | Estimated unamplified path loss | **2694.9 dB** | | Transparent optical spans / inline amplifiers | **169 / 168** | | Published RTT inflation over fiber floor | **1.76×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **209.09 ms** | | Average RTT | **221.5 ms** | | Maximum RTT | **252.44 ms** | | Standard deviation | **9.18 ms** | | Stdev / average | **4.1%** | ## Route Context * Departure PoP: [New York (NYC)](/docs/network/latency/nyc-new-york) * Destination PoP: [Johannesburg (JNB)](/docs/network/latency/jnb-johannesburg) * Region overview: [North America](/docs/network/latency/regions/north-america) → [Africa](/docs/network/latency/regions/africa) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Johannesburg to New York latency and RTT](/docs/network/latency/pairs/jnb-nyc-rtt): 221.8 ms **Fastest routes departing New York (NYC)** * [New York to Ashburn latency and RTT](/docs/network/latency/pairs/nyc-iad-rtt): 6 ms * [New York to Miami latency and RTT](/docs/network/latency/pairs/nyc-mia-rtt): 32.9 ms * [New York to Seattle latency and RTT](/docs/network/latency/pairs/nyc-sea-rtt): 58.7 ms * [New York to Los Angeles latency and RTT](/docs/network/latency/pairs/nyc-lax-rtt): 58.9 ms * [New York to London latency and RTT](/docs/network/latency/pairs/nyc-lon-rtt): 63.5 ms **Fastest routes arriving at Johannesburg (JNB)** * [Amsterdam to Johannesburg latency and RTT](/docs/network/latency/pairs/ams-jnb-rtt): 171.9 ms * [Berlin to Johannesburg latency and RTT](/docs/network/latency/pairs/ber-jnb-rtt): 172.1 ms * [London to Johannesburg latency and RTT](/docs/network/latency/pairs/lon-jnb-rtt): 179.6 ms * [Paris to Johannesburg latency and RTT](/docs/network/latency/pairs/par-jnb-rtt): 180.9 ms * [Frankfurt to Johannesburg latency and RTT](/docs/network/latency/pairs/fra-jnb-rtt): 188.4 ms **Same corridor (North America → Africa)** * [Ashburn to Johannesburg latency and RTT](/docs/network/latency/pairs/iad-jnb-rtt): 228.3 ms * [Miami to Johannesburg latency and RTT](/docs/network/latency/pairs/mia-jnb-rtt): 259.8 ms * [Seattle to Johannesburg latency and RTT](/docs/network/latency/pairs/sea-jnb-rtt): 282.1 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/nyc-jnb.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, New York → Johannesburg round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/nyc-jnb-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # New York, USA → Los Angeles, USA RTT 🇺🇸 **New York, USA (NYC)** → 🇺🇸 **Los Angeles, USA (LAX)** Measurement round: `2026-08-16T04:07:28Z`. New York to Los Angeles averaged 58.9 ms across 50 replies. The observed range was 56.67 to 64.24 ms; these figures describe one probe round, not a long-term service target. This is number 4 of 19 outgoing routes by average RTT. That ranking compares different destinations and is not adjusted for distance. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **58.9 ms** | | Jitter | **1.55 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **38.63 ms** | | Fiber Efficiency | **65.6%** | | Latency Tier | Excellent | | Source Region | [North America](/docs/network/latency/regions/north-america) | | Destination Region | [North America](/docs/network/latency/regions/north-america) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [New York City](https://www.geonames.org/5128581) (40.71427, -74.00597) and [Los Angeles](https://www.geonames.org/5368361) (34.05223, -118.24368). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **3,944.4 km** | | Vacuum RTT floor | **26.31 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **38.63 ms** | | Low-latency fiber material floor | **38.47 ms** | | Engineering floor (5% path allowance) | **40.56 ms** | | Research 1.33× mapped-fiber reference | **51.37 ms** | | Estimated unamplified path loss | **828.3 dB** | | Transparent optical spans / inline amplifiers | **52 / 51** | | Published RTT inflation over fiber floor | **1.52×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------ | | Minimum RTT | **56.67 ms** | | Average RTT | **58.9 ms** | | Maximum RTT | **64.24 ms** | | Standard deviation | **2.02 ms** | | Stdev / average | **3.4%** | ## Route Context * Departure PoP: [New York (NYC)](/docs/network/latency/nyc-new-york) * Destination PoP: [Los Angeles (LAX)](/docs/network/latency/lax-los-angeles) * Region overview: [North America](/docs/network/latency/regions/north-america) → [North America](/docs/network/latency/regions/north-america) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Los Angeles to New York latency and RTT](/docs/network/latency/pairs/lax-nyc-rtt): 57.9 ms **Fastest routes departing New York (NYC)** * [New York to Ashburn latency and RTT](/docs/network/latency/pairs/nyc-iad-rtt): 6 ms * [New York to Miami latency and RTT](/docs/network/latency/pairs/nyc-mia-rtt): 32.9 ms * [New York to Seattle latency and RTT](/docs/network/latency/pairs/nyc-sea-rtt): 58.7 ms * [New York to London latency and RTT](/docs/network/latency/pairs/nyc-lon-rtt): 63.5 ms * [New York to Amsterdam latency and RTT](/docs/network/latency/pairs/nyc-ams-rtt): 68.7 ms **Fastest routes arriving at Los Angeles (LAX)** * [Seattle to Los Angeles latency and RTT](/docs/network/latency/pairs/sea-lax-rtt): 25.9 ms * [Miami to Los Angeles latency and RTT](/docs/network/latency/pairs/mia-lax-rtt): 56.8 ms * [Ashburn to Los Angeles latency and RTT](/docs/network/latency/pairs/iad-lax-rtt): 60.7 ms * [Tokyo to Los Angeles latency and RTT](/docs/network/latency/pairs/tyo-lax-rtt): 101.2 ms * [London to Los Angeles latency and RTT](/docs/network/latency/pairs/lon-lax-rtt): 127.5 ms **Same corridor (North America → North America)** * [Ashburn to New York latency and RTT](/docs/network/latency/pairs/iad-nyc-rtt): 6 ms * [Los Angeles to Seattle latency and RTT](/docs/network/latency/pairs/lax-sea-rtt): 26.9 ms * [Miami to Ashburn latency and RTT](/docs/network/latency/pairs/mia-iad-rtt): 27.3 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/nyc-lax.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, New York → Los Angeles round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/nyc-lax-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # New York, USA → London, UK RTT 🇺🇸 **New York, USA (NYC)** → 🇬🇧 **London, UK (LON)** Measurement round: `2026-08-16T04:07:28Z`. New York to London averaged 63.5 ms across 50 replies. The observed range was 61.91 to 68.19 ms; these figures describe one probe round, not a long-term service target. This is number 5 of 19 outgoing routes by average RTT. That ranking compares different destinations and is not adjusted for distance. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **63.5 ms** | | Jitter | **1.15 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **54.69 ms** | | Fiber Efficiency | **86.1%** | | Latency Tier | Excellent | | Source Region | [North America](/docs/network/latency/regions/north-america) | | Destination Region | [Europe](/docs/network/latency/regions/europe) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [New York City](https://www.geonames.org/5128581) (40.71427, -74.00597) and [London](https://www.geonames.org/2643743) (51.50853, -0.12574). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **5,585.2 km** | | Vacuum RTT floor | **37.26 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **54.69 ms** | | Low-latency fiber material floor | **54.48 ms** | | Engineering floor (5% path allowance) | **57.44 ms** | | Research 1.33× mapped-fiber reference | **72.74 ms** | | Estimated unamplified path loss | **1172.9 dB** | | Transparent optical spans / inline amplifiers | **74 / 73** | | Published RTT inflation over fiber floor | **1.16×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------ | | Minimum RTT | **61.91 ms** | | Average RTT | **63.5 ms** | | Maximum RTT | **68.19 ms** | | Standard deviation | **1.32 ms** | | Stdev / average | **2.1%** | ## Route Context * Departure PoP: [New York (NYC)](/docs/network/latency/nyc-new-york) * Destination PoP: [London (LON)](/docs/network/latency/lon-london) * Region overview: [North America](/docs/network/latency/regions/north-america) → [Europe](/docs/network/latency/regions/europe) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [London to New York latency and RTT](/docs/network/latency/pairs/lon-nyc-rtt): 63.8 ms **Fastest routes departing New York (NYC)** * [New York to Ashburn latency and RTT](/docs/network/latency/pairs/nyc-iad-rtt): 6 ms * [New York to Miami latency and RTT](/docs/network/latency/pairs/nyc-mia-rtt): 32.9 ms * [New York to Seattle latency and RTT](/docs/network/latency/pairs/nyc-sea-rtt): 58.7 ms * [New York to Los Angeles latency and RTT](/docs/network/latency/pairs/nyc-lax-rtt): 58.9 ms * [New York to Amsterdam latency and RTT](/docs/network/latency/pairs/nyc-ams-rtt): 68.7 ms **Fastest routes arriving at London (LON)** * [Amsterdam to London latency and RTT](/docs/network/latency/pairs/ams-lon-rtt): 5.2 ms * [Paris to London latency and RTT](/docs/network/latency/pairs/par-lon-rtt): 6.4 ms * [Frankfurt to London latency and RTT](/docs/network/latency/pairs/fra-lon-rtt): 12.9 ms * [Berlin to London latency and RTT](/docs/network/latency/pairs/ber-lon-rtt): 15.2 ms * [Marseille to London latency and RTT](/docs/network/latency/pairs/mrs-lon-rtt): 18.9 ms **Same corridor (North America → Europe)** * [New York to Paris latency and RTT](/docs/network/latency/pairs/nyc-par-rtt): 68.9 ms * [Ashburn to London latency and RTT](/docs/network/latency/pairs/iad-lon-rtt): 70.3 ms * [New York to Frankfurt latency and RTT](/docs/network/latency/pairs/nyc-fra-rtt): 73.9 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/nyc-lon.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, New York → London round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/nyc-lon-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # New York, USA → Melbourne, Australia RTT 🇺🇸 **New York, USA (NYC)** → 🇦🇺 **Melbourne, Australia (MEL)** Measurement round: `2026-08-16T04:07:28Z`. New York to Melbourne averaged 205.1 ms across 50 replies. The observed range was 196.41 to 229.31 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 6.74 ms, or 3.3% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **205.1 ms** | | Jitter | **6.21 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **163.27 ms** | | Fiber Efficiency | **79.6%** | | Latency Tier | Fair | | Source Region | [North America](/docs/network/latency/regions/north-america) | | Destination Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [New York City](https://www.geonames.org/5128581) (40.71427, -74.00597) and [Melbourne](https://www.geonames.org/2158177) (-37.81400, 144.96332). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | ------------- | | WGS-84 geodesic distance | **16,672 km** | | Vacuum RTT floor | **111.22 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **163.27 ms** | | Low-latency fiber material floor | **162.61 ms** | | Engineering floor (5% path allowance) | **171.45 ms** | | Research 1.33× mapped-fiber reference | **217.14 ms** | | Estimated unamplified path loss | **3501.1 dB** | | Transparent optical spans / inline amplifiers | **219 / 218** | | Published RTT inflation over fiber floor | **1.26×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **196.41 ms** | | Average RTT | **205.1 ms** | | Maximum RTT | **229.31 ms** | | Standard deviation | **6.74 ms** | | Stdev / average | **3.3%** | ## Route Context * Departure PoP: [New York (NYC)](/docs/network/latency/nyc-new-york) * Destination PoP: [Melbourne (MEL)](/docs/network/latency/mel-melbourne) * Region overview: [North America](/docs/network/latency/regions/north-america) → [Asia Pacific](/docs/network/latency/regions/asia-pacific) * Corridor overview: [North America to Asia Pacific](/docs/network/latency/regions/north-america-to-asia-pacific) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Melbourne to New York latency and RTT](/docs/network/latency/pairs/mel-nyc-rtt): 204.2 ms **Fastest routes departing New York (NYC)** * [New York to Ashburn latency and RTT](/docs/network/latency/pairs/nyc-iad-rtt): 6 ms * [New York to Miami latency and RTT](/docs/network/latency/pairs/nyc-mia-rtt): 32.9 ms * [New York to Seattle latency and RTT](/docs/network/latency/pairs/nyc-sea-rtt): 58.7 ms * [New York to Los Angeles latency and RTT](/docs/network/latency/pairs/nyc-lax-rtt): 58.9 ms * [New York to London latency and RTT](/docs/network/latency/pairs/nyc-lon-rtt): 63.5 ms **Fastest routes arriving at Melbourne (MEL)** * [Sydney to Melbourne latency and RTT](/docs/network/latency/pairs/syd-mel-rtt): 9.8 ms * [Singapore to Melbourne latency and RTT](/docs/network/latency/pairs/sin-mel-rtt): 88.6 ms * [Tokyo to Melbourne latency and RTT](/docs/network/latency/pairs/tyo-mel-rtt): 113.9 ms * [Hong Kong to Melbourne latency and RTT](/docs/network/latency/pairs/hkg-mel-rtt): 138.4 ms * [Taipei to Melbourne latency and RTT](/docs/network/latency/pairs/tpe-mel-rtt): 142 ms **Same corridor (North America → Asia Pacific)** * [Seattle to Tokyo latency and RTT](/docs/network/latency/pairs/sea-tyo-rtt): 85 ms * [Los Angeles to Tokyo latency and RTT](/docs/network/latency/pairs/lax-tyo-rtt): 101.2 ms * [Seattle to Taipei latency and RTT](/docs/network/latency/pairs/sea-tpe-rtt): 114.7 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/nyc-mel.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, New York → Melbourne round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/nyc-mel-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # New York, USA → Miami, USA RTT 🇺🇸 **New York, USA (NYC)** → 🇺🇸 **Miami, USA (MIA)** Measurement round: `2026-08-16T04:07:28Z`. New York to Miami averaged 32.9 ms across 50 replies. The observed range was 31.54 to 39.09 ms; these figures describe one probe round, not a long-term service target. This is number 2 of 19 outgoing routes by average RTT. That ranking compares different destinations and is not adjusted for distance. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **32.9 ms** | | Jitter | **0.86 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **17.17 ms** | | Fiber Efficiency | **52.2%** | | Latency Tier | Excellent | | Source Region | [North America](/docs/network/latency/regions/north-america) | | Destination Region | [North America](/docs/network/latency/regions/north-america) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [New York City](https://www.geonames.org/5128581) (40.71427, -74.00597) and [Miami](https://www.geonames.org/4164138) (25.77427, -80.19366). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **1,753.2 km** | | Vacuum RTT floor | **11.7 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **17.17 ms** | | Low-latency fiber material floor | **17.1 ms** | | Engineering floor (5% path allowance) | **18.03 ms** | | Research 1.33× mapped-fiber reference | **22.83 ms** | | Estimated unamplified path loss | **368.2 dB** | | Transparent optical spans / inline amplifiers | **24 / 23** | | Published RTT inflation over fiber floor | **1.92×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------ | | Minimum RTT | **31.54 ms** | | Average RTT | **32.9 ms** | | Maximum RTT | **39.09 ms** | | Standard deviation | **1.24 ms** | | Stdev / average | **3.8%** | ## Route Context * Departure PoP: [New York (NYC)](/docs/network/latency/nyc-new-york) * Destination PoP: [Miami (MIA)](/docs/network/latency/mia-miami) * Region overview: [North America](/docs/network/latency/regions/north-america) → [North America](/docs/network/latency/regions/north-america) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Miami to New York latency and RTT](/docs/network/latency/pairs/mia-nyc-rtt): 33.3 ms **Fastest routes departing New York (NYC)** * [New York to Ashburn latency and RTT](/docs/network/latency/pairs/nyc-iad-rtt): 6 ms * [New York to Seattle latency and RTT](/docs/network/latency/pairs/nyc-sea-rtt): 58.7 ms * [New York to Los Angeles latency and RTT](/docs/network/latency/pairs/nyc-lax-rtt): 58.9 ms * [New York to London latency and RTT](/docs/network/latency/pairs/nyc-lon-rtt): 63.5 ms * [New York to Amsterdam latency and RTT](/docs/network/latency/pairs/nyc-ams-rtt): 68.7 ms **Fastest routes arriving at Miami (MIA)** * [Ashburn to Miami latency and RTT](/docs/network/latency/pairs/iad-mia-rtt): 27.4 ms * [Los Angeles to Miami latency and RTT](/docs/network/latency/pairs/lax-mia-rtt): 57.3 ms * [São Paulo to Miami latency and RTT](/docs/network/latency/pairs/gru-mia-rtt): 74 ms * [Seattle to Miami latency and RTT](/docs/network/latency/pairs/sea-mia-rtt): 81.7 ms * [London to Miami latency and RTT](/docs/network/latency/pairs/lon-mia-rtt): 101.9 ms **Same corridor (North America → North America)** * [Ashburn to New York latency and RTT](/docs/network/latency/pairs/iad-nyc-rtt): 6 ms * [Seattle to Los Angeles latency and RTT](/docs/network/latency/pairs/sea-lax-rtt): 25.9 ms * [Los Angeles to Seattle latency and RTT](/docs/network/latency/pairs/lax-sea-rtt): 26.9 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/nyc-mia.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, New York → Miami round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/nyc-mia-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # New York, USA → Moscow, Russia RTT 🇺🇸 **New York, USA (NYC)** → 🇷🇺 **Moscow, Russia (MOW)** Measurement round: `2026-08-16T04:07:28Z`. New York to Moscow averaged 107 ms across 50 replies. The observed range was 101.46 to 119.28 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 4.26 ms, or 4.0% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **107 ms** | | Jitter | **3.73 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **73.75 ms** | | Fiber Efficiency | **68.9%** | | Latency Tier | Good | | Source Region | [North America](/docs/network/latency/regions/north-america) | | Destination Region | [Europe](/docs/network/latency/regions/europe) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [New York City](https://www.geonames.org/5128581) (40.71427, -74.00597) and [Moscow](https://www.geonames.org/524901) (55.75204, 37.61781). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **7,531.3 km** | | Vacuum RTT floor | **50.24 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **73.75 ms** | | Low-latency fiber material floor | **73.46 ms** | | Engineering floor (5% path allowance) | **77.45 ms** | | Research 1.33× mapped-fiber reference | **98.09 ms** | | Estimated unamplified path loss | **1581.6 dB** | | Transparent optical spans / inline amplifiers | **99 / 98** | | Published RTT inflation over fiber floor | **1.45×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **101.46 ms** | | Average RTT | **107 ms** | | Maximum RTT | **119.28 ms** | | Standard deviation | **4.26 ms** | | Stdev / average | **4.0%** | ## Route Context * Departure PoP: [New York (NYC)](/docs/network/latency/nyc-new-york) * Destination PoP: [Moscow (MOW)](/docs/network/latency/mow-moscow) * Region overview: [North America](/docs/network/latency/regions/north-america) → [Europe](/docs/network/latency/regions/europe) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Moscow to New York latency and RTT](/docs/network/latency/pairs/mow-nyc-rtt): 107.5 ms **Fastest routes departing New York (NYC)** * [New York to Ashburn latency and RTT](/docs/network/latency/pairs/nyc-iad-rtt): 6 ms * [New York to Miami latency and RTT](/docs/network/latency/pairs/nyc-mia-rtt): 32.9 ms * [New York to Seattle latency and RTT](/docs/network/latency/pairs/nyc-sea-rtt): 58.7 ms * [New York to Los Angeles latency and RTT](/docs/network/latency/pairs/nyc-lax-rtt): 58.9 ms * [New York to London latency and RTT](/docs/network/latency/pairs/nyc-lon-rtt): 63.5 ms **Fastest routes arriving at Moscow (MOW)** * [Berlin to Moscow latency and RTT](/docs/network/latency/pairs/ber-mow-rtt): 28.6 ms * [Frankfurt to Moscow latency and RTT](/docs/network/latency/pairs/fra-mow-rtt): 34.7 ms * [Amsterdam to Moscow latency and RTT](/docs/network/latency/pairs/ams-mow-rtt): 38.1 ms * [London to Moscow latency and RTT](/docs/network/latency/pairs/lon-mow-rtt): 43.5 ms * [Paris to Moscow latency and RTT](/docs/network/latency/pairs/par-mow-rtt): 44.7 ms **Same corridor (North America → Europe)** * [New York to Amsterdam latency and RTT](/docs/network/latency/pairs/nyc-ams-rtt): 68.7 ms * [New York to Paris latency and RTT](/docs/network/latency/pairs/nyc-par-rtt): 68.9 ms * [Ashburn to London latency and RTT](/docs/network/latency/pairs/iad-lon-rtt): 70.3 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/nyc-mow.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, New York → Moscow round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/nyc-mow-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # New York, USA → Marseille, France RTT 🇺🇸 **New York, USA (NYC)** → 🇫🇷 **Marseille, France (MRS)** Measurement round: `2026-08-16T04:07:28Z`. New York to Marseille averaged 80.2 ms across 50 replies. The observed range was 78.94 to 85.06 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 1.32 ms, or 1.6% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **80.2 ms** | | Jitter | **0.93 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **61.91 ms** | | Fiber Efficiency | **77.2%** | | Latency Tier | Good | | Source Region | [North America](/docs/network/latency/regions/north-america) | | Destination Region | [Europe](/docs/network/latency/regions/europe) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [New York City](https://www.geonames.org/5128581) (40.71427, -74.00597) and [Marseille](https://www.geonames.org/2995469) (43.29695, 5.38107). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **6,321.9 km** | | Vacuum RTT floor | **42.18 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **61.91 ms** | | Low-latency fiber material floor | **61.66 ms** | | Engineering floor (5% path allowance) | **65.01 ms** | | Research 1.33× mapped-fiber reference | **82.34 ms** | | Estimated unamplified path loss | **1327.6 dB** | | Transparent optical spans / inline amplifiers | **83 / 82** | | Published RTT inflation over fiber floor | **1.3×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------ | | Minimum RTT | **78.94 ms** | | Average RTT | **80.2 ms** | | Maximum RTT | **85.06 ms** | | Standard deviation | **1.32 ms** | | Stdev / average | **1.6%** | ## Route Context * Departure PoP: [New York (NYC)](/docs/network/latency/nyc-new-york) * Destination PoP: [Marseille (MRS)](/docs/network/latency/mrs-marseille) * Region overview: [North America](/docs/network/latency/regions/north-america) → [Europe](/docs/network/latency/regions/europe) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Marseille to New York latency and RTT](/docs/network/latency/pairs/mrs-nyc-rtt): 79.8 ms **Fastest routes departing New York (NYC)** * [New York to Ashburn latency and RTT](/docs/network/latency/pairs/nyc-iad-rtt): 6 ms * [New York to Miami latency and RTT](/docs/network/latency/pairs/nyc-mia-rtt): 32.9 ms * [New York to Seattle latency and RTT](/docs/network/latency/pairs/nyc-sea-rtt): 58.7 ms * [New York to Los Angeles latency and RTT](/docs/network/latency/pairs/nyc-lax-rtt): 58.9 ms * [New York to London latency and RTT](/docs/network/latency/pairs/nyc-lon-rtt): 63.5 ms **Fastest routes arriving at Marseille (MRS)** * [Paris to Marseille latency and RTT](/docs/network/latency/pairs/par-mrs-rtt): 8.9 ms * [Frankfurt to Marseille latency and RTT](/docs/network/latency/pairs/fra-mrs-rtt): 16.4 ms * [London to Marseille latency and RTT](/docs/network/latency/pairs/lon-mrs-rtt): 17.9 ms * [Amsterdam to Marseille latency and RTT](/docs/network/latency/pairs/ams-mrs-rtt): 19.6 ms * [Berlin to Marseille latency and RTT](/docs/network/latency/pairs/ber-mrs-rtt): 20.2 ms **Same corridor (North America → Europe)** * [New York to Amsterdam latency and RTT](/docs/network/latency/pairs/nyc-ams-rtt): 68.7 ms * [New York to Paris latency and RTT](/docs/network/latency/pairs/nyc-par-rtt): 68.9 ms * [Ashburn to London latency and RTT](/docs/network/latency/pairs/iad-lon-rtt): 70.3 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/nyc-mrs.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, New York → Marseille round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/nyc-mrs-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # New York, USA → Paris, France RTT 🇺🇸 **New York, USA (NYC)** → 🇫🇷 **Paris, France (PAR)** Measurement round: `2026-08-16T04:07:28Z`. New York to Paris averaged 68.9 ms across 50 replies. The observed range was 67.48 to 73.26 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 1.31 ms, or 1.9% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **68.9 ms** | | Jitter | **1.05 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **57.31 ms** | | Fiber Efficiency | **83.2%** | | Latency Tier | Excellent | | Source Region | [North America](/docs/network/latency/regions/north-america) | | Destination Region | [Europe](/docs/network/latency/regions/europe) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [New York City](https://www.geonames.org/5128581) (40.71427, -74.00597) and [Paris](https://www.geonames.org/2988507) (48.85341, 2.34880). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **5,852.7 km** | | Vacuum RTT floor | **39.05 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **57.31 ms** | | Low-latency fiber material floor | **57.08 ms** | | Engineering floor (5% path allowance) | **60.19 ms** | | Research 1.33× mapped-fiber reference | **76.23 ms** | | Estimated unamplified path loss | **1229.1 dB** | | Transparent optical spans / inline amplifiers | **77 / 76** | | Published RTT inflation over fiber floor | **1.2×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------ | | Minimum RTT | **67.48 ms** | | Average RTT | **68.9 ms** | | Maximum RTT | **73.26 ms** | | Standard deviation | **1.31 ms** | | Stdev / average | **1.9%** | ## Route Context * Departure PoP: [New York (NYC)](/docs/network/latency/nyc-new-york) * Destination PoP: [Paris (PAR)](/docs/network/latency/par-paris) * Region overview: [North America](/docs/network/latency/regions/north-america) → [Europe](/docs/network/latency/regions/europe) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Paris to New York latency and RTT](/docs/network/latency/pairs/par-nyc-rtt): 68.9 ms **Fastest routes departing New York (NYC)** * [New York to Ashburn latency and RTT](/docs/network/latency/pairs/nyc-iad-rtt): 6 ms * [New York to Miami latency and RTT](/docs/network/latency/pairs/nyc-mia-rtt): 32.9 ms * [New York to Seattle latency and RTT](/docs/network/latency/pairs/nyc-sea-rtt): 58.7 ms * [New York to Los Angeles latency and RTT](/docs/network/latency/pairs/nyc-lax-rtt): 58.9 ms * [New York to London latency and RTT](/docs/network/latency/pairs/nyc-lon-rtt): 63.5 ms **Fastest routes arriving at Paris (PAR)** * [London to Paris latency and RTT](/docs/network/latency/pairs/lon-par-rtt): 6.4 ms * [Amsterdam to Paris latency and RTT](/docs/network/latency/pairs/ams-par-rtt): 7 ms * [Frankfurt to Paris latency and RTT](/docs/network/latency/pairs/fra-par-rtt): 7.6 ms * [Marseille to Paris latency and RTT](/docs/network/latency/pairs/mrs-par-rtt): 8.9 ms * [Berlin to Paris latency and RTT](/docs/network/latency/pairs/ber-par-rtt): 15.6 ms **Same corridor (North America → Europe)** * [New York to Amsterdam latency and RTT](/docs/network/latency/pairs/nyc-ams-rtt): 68.7 ms * [Ashburn to London latency and RTT](/docs/network/latency/pairs/iad-lon-rtt): 70.3 ms * [New York to Frankfurt latency and RTT](/docs/network/latency/pairs/nyc-fra-rtt): 73.9 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/nyc-par.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, New York → Paris round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/nyc-par-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # New York, USA → Seattle, USA RTT 🇺🇸 **New York, USA (NYC)** → 🇺🇸 **Seattle, USA (SEA)** Measurement round: `2026-08-16T04:07:28Z`. New York to Seattle averaged 58.7 ms across 50 replies. The observed range was 56.64 to 67.85 ms; these figures describe one probe round, not a long-term service target. This is number 3 of 19 outgoing routes by average RTT. That ranking compares different destinations and is not adjusted for distance. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **58.7 ms** | | Jitter | **1.34 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **37.95 ms** | | Fiber Efficiency | **64.7%** | | Latency Tier | Excellent | | Source Region | [North America](/docs/network/latency/regions/north-america) | | Destination Region | [North America](/docs/network/latency/regions/north-america) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [New York City](https://www.geonames.org/5128581) (40.71427, -74.00597) and [Seattle](https://www.geonames.org/5809844) (47.60621, -122.33207). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **3,875.5 km** | | Vacuum RTT floor | **25.85 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **37.95 ms** | | Low-latency fiber material floor | **37.8 ms** | | Engineering floor (5% path allowance) | **39.85 ms** | | Research 1.33× mapped-fiber reference | **50.48 ms** | | Estimated unamplified path loss | **813.9 dB** | | Transparent optical spans / inline amplifiers | **51 / 50** | | Published RTT inflation over fiber floor | **1.55×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------ | | Minimum RTT | **56.64 ms** | | Average RTT | **58.7 ms** | | Maximum RTT | **67.85 ms** | | Standard deviation | **2.09 ms** | | Stdev / average | **3.6%** | ## Route Context * Departure PoP: [New York (NYC)](/docs/network/latency/nyc-new-york) * Destination PoP: [Seattle (SEA)](/docs/network/latency/sea-seattle) * Region overview: [North America](/docs/network/latency/regions/north-america) → [North America](/docs/network/latency/regions/north-america) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Seattle to New York latency and RTT](/docs/network/latency/pairs/sea-nyc-rtt): 59.5 ms **Fastest routes departing New York (NYC)** * [New York to Ashburn latency and RTT](/docs/network/latency/pairs/nyc-iad-rtt): 6 ms * [New York to Miami latency and RTT](/docs/network/latency/pairs/nyc-mia-rtt): 32.9 ms * [New York to Los Angeles latency and RTT](/docs/network/latency/pairs/nyc-lax-rtt): 58.9 ms * [New York to London latency and RTT](/docs/network/latency/pairs/nyc-lon-rtt): 63.5 ms * [New York to Amsterdam latency and RTT](/docs/network/latency/pairs/nyc-ams-rtt): 68.7 ms **Fastest routes arriving at Seattle (SEA)** * [Los Angeles to Seattle latency and RTT](/docs/network/latency/pairs/lax-sea-rtt): 26.9 ms * [Ashburn to Seattle latency and RTT](/docs/network/latency/pairs/iad-sea-rtt): 62.3 ms * [Miami to Seattle latency and RTT](/docs/network/latency/pairs/mia-sea-rtt): 81.9 ms * [Tokyo to Seattle latency and RTT](/docs/network/latency/pairs/tyo-sea-rtt): 84.8 ms * [Taipei to Seattle latency and RTT](/docs/network/latency/pairs/tpe-sea-rtt): 115.8 ms **Same corridor (North America → North America)** * [Ashburn to New York latency and RTT](/docs/network/latency/pairs/iad-nyc-rtt): 6 ms * [Seattle to Los Angeles latency and RTT](/docs/network/latency/pairs/sea-lax-rtt): 25.9 ms * [Miami to Ashburn latency and RTT](/docs/network/latency/pairs/mia-iad-rtt): 27.3 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/nyc-sea.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, New York → Seattle round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/nyc-sea-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # New York, USA → Singapore RTT 🇺🇸 **New York, USA (NYC)** → 🇸🇬 **Singapore (SIN)** Measurement round: `2026-08-16T04:07:28Z`. New York to Singapore averaged 208.7 ms across 50 replies. The observed range was 197.34 to 232.34 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 8.03 ms, or 3.8% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **208.7 ms** | | Jitter | **7.87 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **150.29 ms** | | Fiber Efficiency | **72%** | | Latency Tier | Fair | | Source Region | [North America](/docs/network/latency/regions/north-america) | | Destination Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [New York City](https://www.geonames.org/5128581) (40.71427, -74.00597) and [Singapore](https://www.geonames.org/1880252) (1.28967, 103.85007). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | --------------- | | WGS-84 geodesic distance | **15,347.5 km** | | Vacuum RTT floor | **102.39 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **150.29 ms** | | Low-latency fiber material floor | **149.69 ms** | | Engineering floor (5% path allowance) | **157.83 ms** | | Research 1.33× mapped-fiber reference | **199.89 ms** | | Estimated unamplified path loss | **3223 dB** | | Transparent optical spans / inline amplifiers | **202 / 201** | | Published RTT inflation over fiber floor | **1.39×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **197.34 ms** | | Average RTT | **208.7 ms** | | Maximum RTT | **232.34 ms** | | Standard deviation | **8.03 ms** | | Stdev / average | **3.8%** | ## Route Context * Departure PoP: [New York (NYC)](/docs/network/latency/nyc-new-york) * Destination PoP: [Singapore (SIN)](/docs/network/latency/sin-singapore) * Region overview: [North America](/docs/network/latency/regions/north-america) → [Asia Pacific](/docs/network/latency/regions/asia-pacific) * Corridor overview: [North America to Asia Pacific](/docs/network/latency/regions/north-america-to-asia-pacific) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Singapore to New York latency and RTT](/docs/network/latency/pairs/sin-nyc-rtt): 209.3 ms **Fastest routes departing New York (NYC)** * [New York to Ashburn latency and RTT](/docs/network/latency/pairs/nyc-iad-rtt): 6 ms * [New York to Miami latency and RTT](/docs/network/latency/pairs/nyc-mia-rtt): 32.9 ms * [New York to Seattle latency and RTT](/docs/network/latency/pairs/nyc-sea-rtt): 58.7 ms * [New York to Los Angeles latency and RTT](/docs/network/latency/pairs/nyc-lax-rtt): 58.9 ms * [New York to London latency and RTT](/docs/network/latency/pairs/nyc-lon-rtt): 63.5 ms **Fastest routes arriving at Singapore (SIN)** * [Hong Kong to Singapore latency and RTT](/docs/network/latency/pairs/hkg-sin-rtt): 31.5 ms * [Taipei to Singapore latency and RTT](/docs/network/latency/pairs/tpe-sin-rtt): 44.3 ms * [Tokyo to Singapore latency and RTT](/docs/network/latency/pairs/tyo-sin-rtt): 69.4 ms * [Melbourne to Singapore latency and RTT](/docs/network/latency/pairs/mel-sin-rtt): 88.4 ms * [Sydney to Singapore latency and RTT](/docs/network/latency/pairs/syd-sin-rtt): 94.5 ms **Same corridor (North America → Asia Pacific)** * [Seattle to Tokyo latency and RTT](/docs/network/latency/pairs/sea-tyo-rtt): 85 ms * [Los Angeles to Tokyo latency and RTT](/docs/network/latency/pairs/lax-tyo-rtt): 101.2 ms * [Seattle to Taipei latency and RTT](/docs/network/latency/pairs/sea-tpe-rtt): 114.7 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/nyc-sin.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, New York → Singapore round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/nyc-sin-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # New York, USA → Sydney, Australia RTT 🇺🇸 **New York, USA (NYC)** → 🇦🇺 **Sydney, Australia (SYD)** Measurement round: `2026-08-16T04:07:28Z`. New York to Sydney averaged 192.6 ms across 50 replies. The observed range was 185.89 to 212.11 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 5.61 ms, or 2.9% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **192.6 ms** | | Jitter | **4.03 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **156.57 ms** | | Fiber Efficiency | **81.3%** | | Latency Tier | Fair | | Source Region | [North America](/docs/network/latency/regions/north-america) | | Destination Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [New York City](https://www.geonames.org/5128581) (40.71427, -74.00597) and [Sydney](https://www.geonames.org/2147714) (-33.86785, 151.20732). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | --------------- | | WGS-84 geodesic distance | **15,988.1 km** | | Vacuum RTT floor | **106.66 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **156.57 ms** | | Low-latency fiber material floor | **155.94 ms** | | Engineering floor (5% path allowance) | **164.42 ms** | | Research 1.33× mapped-fiber reference | **208.24 ms** | | Estimated unamplified path loss | **3357.5 dB** | | Transparent optical spans / inline amplifiers | **210 / 209** | | Published RTT inflation over fiber floor | **1.23×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **185.89 ms** | | Average RTT | **192.6 ms** | | Maximum RTT | **212.11 ms** | | Standard deviation | **5.61 ms** | | Stdev / average | **2.9%** | ## Route Context * Departure PoP: [New York (NYC)](/docs/network/latency/nyc-new-york) * Destination PoP: [Sydney (SYD)](/docs/network/latency/syd-sydney) * Region overview: [North America](/docs/network/latency/regions/north-america) → [Asia Pacific](/docs/network/latency/regions/asia-pacific) * Corridor overview: [North America to Asia Pacific](/docs/network/latency/regions/north-america-to-asia-pacific) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Sydney to New York latency and RTT](/docs/network/latency/pairs/syd-nyc-rtt): 191.9 ms **Fastest routes departing New York (NYC)** * [New York to Ashburn latency and RTT](/docs/network/latency/pairs/nyc-iad-rtt): 6 ms * [New York to Miami latency and RTT](/docs/network/latency/pairs/nyc-mia-rtt): 32.9 ms * [New York to Seattle latency and RTT](/docs/network/latency/pairs/nyc-sea-rtt): 58.7 ms * [New York to Los Angeles latency and RTT](/docs/network/latency/pairs/nyc-lax-rtt): 58.9 ms * [New York to London latency and RTT](/docs/network/latency/pairs/nyc-lon-rtt): 63.5 ms **Fastest routes arriving at Sydney (SYD)** * [Melbourne to Sydney latency and RTT](/docs/network/latency/pairs/mel-syd-rtt): 9.8 ms * [Singapore to Sydney latency and RTT](/docs/network/latency/pairs/sin-syd-rtt): 94.5 ms * [Tokyo to Sydney latency and RTT](/docs/network/latency/pairs/tyo-syd-rtt): 101.9 ms * [Taipei to Sydney latency and RTT](/docs/network/latency/pairs/tpe-syd-rtt): 133.2 ms * [Hong Kong to Sydney latency and RTT](/docs/network/latency/pairs/hkg-syd-rtt): 137.2 ms **Same corridor (North America → Asia Pacific)** * [Seattle to Tokyo latency and RTT](/docs/network/latency/pairs/sea-tyo-rtt): 85 ms * [Los Angeles to Tokyo latency and RTT](/docs/network/latency/pairs/lax-tyo-rtt): 101.2 ms * [Seattle to Taipei latency and RTT](/docs/network/latency/pairs/sea-tpe-rtt): 114.7 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/nyc-syd.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, New York → Sydney round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/nyc-syd-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # New York, USA → Taipei, Taiwan RTT 🇺🇸 **New York, USA (NYC)** → 🇹🇼 **Taipei, Taiwan (TPE)** Measurement round: `2026-08-16T04:07:28Z`. New York to Taipei averaged 173.4 ms across 50 replies. The observed range was 169.05 to 186.13 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 3.35 ms, or 1.9% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **173.4 ms** | | Jitter | **3.23 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **122.89 ms** | | Fiber Efficiency | **70.9%** | | Latency Tier | Fair | | Source Region | [North America](/docs/network/latency/regions/north-america) | | Destination Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [New York City](https://www.geonames.org/5128581) (40.71427, -74.00597) and [Taipei](https://www.geonames.org/1668341) (25.05306, 121.52639). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | --------------- | | WGS-84 geodesic distance | **12,549.1 km** | | Vacuum RTT floor | **83.72 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **122.89 ms** | | Low-latency fiber material floor | **122.4 ms** | | Engineering floor (5% path allowance) | **129.05 ms** | | Research 1.33× mapped-fiber reference | **163.44 ms** | | Estimated unamplified path loss | **2635.3 dB** | | Transparent optical spans / inline amplifiers | **165 / 164** | | Published RTT inflation over fiber floor | **1.41×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **169.05 ms** | | Average RTT | **173.4 ms** | | Maximum RTT | **186.13 ms** | | Standard deviation | **3.35 ms** | | Stdev / average | **1.9%** | ## Route Context * Departure PoP: [New York (NYC)](/docs/network/latency/nyc-new-york) * Destination PoP: [Taipei (TPE)](/docs/network/latency/tpe-taipei) * Region overview: [North America](/docs/network/latency/regions/north-america) → [Asia Pacific](/docs/network/latency/regions/asia-pacific) * Corridor overview: [North America to Asia Pacific](/docs/network/latency/regions/north-america-to-asia-pacific) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Taipei to New York latency and RTT](/docs/network/latency/pairs/tpe-nyc-rtt): 174.2 ms **Fastest routes departing New York (NYC)** * [New York to Ashburn latency and RTT](/docs/network/latency/pairs/nyc-iad-rtt): 6 ms * [New York to Miami latency and RTT](/docs/network/latency/pairs/nyc-mia-rtt): 32.9 ms * [New York to Seattle latency and RTT](/docs/network/latency/pairs/nyc-sea-rtt): 58.7 ms * [New York to Los Angeles latency and RTT](/docs/network/latency/pairs/nyc-lax-rtt): 58.9 ms * [New York to London latency and RTT](/docs/network/latency/pairs/nyc-lon-rtt): 63.5 ms **Fastest routes arriving at Taipei (TPE)** * [Hong Kong to Taipei latency and RTT](/docs/network/latency/pairs/hkg-tpe-rtt): 14.7 ms * [Tokyo to Taipei latency and RTT](/docs/network/latency/pairs/tyo-tpe-rtt): 32.3 ms * [Singapore to Taipei latency and RTT](/docs/network/latency/pairs/sin-tpe-rtt): 45.8 ms * [Seattle to Taipei latency and RTT](/docs/network/latency/pairs/sea-tpe-rtt): 114.7 ms * [Sydney to Taipei latency and RTT](/docs/network/latency/pairs/syd-tpe-rtt): 131.7 ms **Same corridor (North America → Asia Pacific)** * [Seattle to Tokyo latency and RTT](/docs/network/latency/pairs/sea-tyo-rtt): 85 ms * [Los Angeles to Tokyo latency and RTT](/docs/network/latency/pairs/lax-tyo-rtt): 101.2 ms * [Seattle to Hong Kong latency and RTT](/docs/network/latency/pairs/sea-hkg-rtt): 130.4 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/nyc-tpe.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, New York → Taipei round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/nyc-tpe-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # New York, USA → Tokyo, Japan RTT 🇺🇸 **New York, USA (NYC)** → 🇯🇵 **Tokyo, Japan (TYO)** Measurement round: `2026-08-16T04:07:28Z`. New York to Tokyo averaged 143.1 ms across 50 replies. The observed range was 136.69 to 171.34 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 6.58 ms, or 4.6% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **143.1 ms** | | Jitter | **4.7 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **106.47 ms** | | Fiber Efficiency | **74.4%** | | Latency Tier | Good | | Source Region | [North America](/docs/network/latency/regions/north-america) | | Destination Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [New York City](https://www.geonames.org/5128581) (40.71427, -74.00597) and [Tokyo](https://www.geonames.org/1850147) (35.68950, 139.69171). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | --------------- | | WGS-84 geodesic distance | **10,872.7 km** | | Vacuum RTT floor | **72.53 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **106.47 ms** | | Low-latency fiber material floor | **106.05 ms** | | Engineering floor (5% path allowance) | **111.81 ms** | | Research 1.33× mapped-fiber reference | **141.61 ms** | | Estimated unamplified path loss | **2283.3 dB** | | Transparent optical spans / inline amplifiers | **143 / 142** | | Published RTT inflation over fiber floor | **1.34×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **136.69 ms** | | Average RTT | **143.1 ms** | | Maximum RTT | **171.34 ms** | | Standard deviation | **6.58 ms** | | Stdev / average | **4.6%** | ## Route Context * Departure PoP: [New York (NYC)](/docs/network/latency/nyc-new-york) * Destination PoP: [Tokyo (TYO)](/docs/network/latency/tyo-tokyo) * Region overview: [North America](/docs/network/latency/regions/north-america) → [Asia Pacific](/docs/network/latency/regions/asia-pacific) * Corridor overview: [North America to Asia Pacific](/docs/network/latency/regions/north-america-to-asia-pacific) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Tokyo to New York latency and RTT](/docs/network/latency/pairs/tyo-nyc-rtt): 144.6 ms **Fastest routes departing New York (NYC)** * [New York to Ashburn latency and RTT](/docs/network/latency/pairs/nyc-iad-rtt): 6 ms * [New York to Miami latency and RTT](/docs/network/latency/pairs/nyc-mia-rtt): 32.9 ms * [New York to Seattle latency and RTT](/docs/network/latency/pairs/nyc-sea-rtt): 58.7 ms * [New York to Los Angeles latency and RTT](/docs/network/latency/pairs/nyc-lax-rtt): 58.9 ms * [New York to London latency and RTT](/docs/network/latency/pairs/nyc-lon-rtt): 63.5 ms **Fastest routes arriving at Tokyo (TYO)** * [Taipei to Tokyo latency and RTT](/docs/network/latency/pairs/tpe-tyo-rtt): 31.9 ms * [Hong Kong to Tokyo latency and RTT](/docs/network/latency/pairs/hkg-tyo-rtt): 44.9 ms * [Singapore to Tokyo latency and RTT](/docs/network/latency/pairs/sin-tyo-rtt): 68.9 ms * [Seattle to Tokyo latency and RTT](/docs/network/latency/pairs/sea-tyo-rtt): 85 ms * [Los Angeles to Tokyo latency and RTT](/docs/network/latency/pairs/lax-tyo-rtt): 101.2 ms **Same corridor (North America → Asia Pacific)** * [Seattle to Taipei latency and RTT](/docs/network/latency/pairs/sea-tpe-rtt): 114.7 ms * [Seattle to Hong Kong latency and RTT](/docs/network/latency/pairs/sea-hkg-rtt): 130.4 ms * [Los Angeles to Taipei latency and RTT](/docs/network/latency/pairs/lax-tpe-rtt): 132.9 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/nyc-tyo.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, New York → Tokyo round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/nyc-tyo-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Paris, France → Amsterdam, Netherlands RTT 🇫🇷 **Paris, France (PAR)** → 🇳🇱 **Amsterdam, Netherlands (AMS)** Measurement round: `2026-08-16T04:07:28Z`. Paris to Amsterdam averaged 7 ms across 50 replies. The observed range was 6.65 to 8.5 ms; these figures describe one probe round, not a long-term service target. This is number 2 of 19 outgoing routes by average RTT. That ranking compares different destinations and is not adjusted for distance. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------- | | RTT | **7 ms** | | Jitter | **0.24 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **4.22 ms** | | Fiber Efficiency | **60.3%** | | Latency Tier | Ultra-Low | | Source Region | [Europe](/docs/network/latency/regions/europe) | | Destination Region | [Europe](/docs/network/latency/regions/europe) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Paris](https://www.geonames.org/2988507) (48.85341, 2.34880) and [Amsterdam](https://www.geonames.org/2759794) (52.37403, 4.88969). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | ------------ | | WGS-84 geodesic distance | **430.9 km** | | Vacuum RTT floor | **2.87 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **4.22 ms** | | Low-latency fiber material floor | **4.2 ms** | | Engineering floor (5% path allowance) | **4.43 ms** | | Research 1.33× mapped-fiber reference | **5.61 ms** | | Estimated unamplified path loss | **90.5 dB** | | Transparent optical spans / inline amplifiers | **6 / 5** | | Published RTT inflation over fiber floor | **1.66×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ----------- | | Minimum RTT | **6.65 ms** | | Average RTT | **7 ms** | | Maximum RTT | **8.5 ms** | | Standard deviation | **0.33 ms** | | Stdev / average | **4.7%** | ## Route Context * Departure PoP: [Paris (PAR)](/docs/network/latency/par-paris) * Destination PoP: [Amsterdam (AMS)](/docs/network/latency/ams-amsterdam) * Region overview: [Europe](/docs/network/latency/regions/europe) → [Europe](/docs/network/latency/regions/europe) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Amsterdam to Paris latency and RTT](/docs/network/latency/pairs/ams-par-rtt): 7 ms **Fastest routes departing Paris (PAR)** * [Paris to London latency and RTT](/docs/network/latency/pairs/par-lon-rtt): 6.4 ms * [Paris to Frankfurt latency and RTT](/docs/network/latency/pairs/par-fra-rtt): 7.6 ms * [Paris to Marseille latency and RTT](/docs/network/latency/pairs/par-mrs-rtt): 8.9 ms * [Paris to Berlin latency and RTT](/docs/network/latency/pairs/par-ber-rtt): 15.5 ms * [Paris to Moscow latency and RTT](/docs/network/latency/pairs/par-mow-rtt): 44.7 ms **Fastest routes arriving at Amsterdam (AMS)** * [London to Amsterdam latency and RTT](/docs/network/latency/pairs/lon-ams-rtt): 5.2 ms * [Frankfurt to Amsterdam latency and RTT](/docs/network/latency/pairs/fra-ams-rtt): 5.9 ms * [Berlin to Amsterdam latency and RTT](/docs/network/latency/pairs/ber-ams-rtt): 7.2 ms * [Marseille to Amsterdam latency and RTT](/docs/network/latency/pairs/mrs-ams-rtt): 20.3 ms * [Moscow to Amsterdam latency and RTT](/docs/network/latency/pairs/mow-ams-rtt): 37.6 ms **Same corridor (Europe → Europe)** * [Amsterdam to London latency and RTT](/docs/network/latency/pairs/ams-lon-rtt): 5.2 ms * [Amsterdam to Frankfurt latency and RTT](/docs/network/latency/pairs/ams-fra-rtt): 5.9 ms * [Berlin to Frankfurt latency and RTT](/docs/network/latency/pairs/ber-fra-rtt): 6.1 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/par-ams.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Paris → Amsterdam round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/par-ams-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Paris, France → Berlin, Germany RTT 🇫🇷 **Paris, France (PAR)** → 🇩🇪 **Berlin, Germany (BER)** Measurement round: `2026-08-16T04:07:28Z`. Paris to Berlin averaged 15.5 ms across 50 replies. The observed range was 14.87 to 17.28 ms; these figures describe one probe round, not a long-term service target. This is number 5 of 19 outgoing routes by average RTT. That ranking compares different destinations and is not adjusted for distance. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------- | | RTT | **15.5 ms** | | Jitter | **0.43 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **8.62 ms** | | Fiber Efficiency | **55.6%** | | Latency Tier | Ultra-Low | | Source Region | [Europe](/docs/network/latency/regions/europe) | | Destination Region | [Europe](/docs/network/latency/regions/europe) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Paris](https://www.geonames.org/2988507) (48.85341, 2.34880) and [Berlin](https://www.geonames.org/2950159) (52.52437, 13.41053). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | ------------ | | WGS-84 geodesic distance | **880.6 km** | | Vacuum RTT floor | **5.87 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **8.62 ms** | | Low-latency fiber material floor | **8.59 ms** | | Engineering floor (5% path allowance) | **9.06 ms** | | Research 1.33× mapped-fiber reference | **11.47 ms** | | Estimated unamplified path loss | **184.9 dB** | | Transparent optical spans / inline amplifiers | **12 / 11** | | Published RTT inflation over fiber floor | **1.8×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------ | | Minimum RTT | **14.87 ms** | | Average RTT | **15.5 ms** | | Maximum RTT | **17.28 ms** | | Standard deviation | **0.57 ms** | | Stdev / average | **3.7%** | ## Route Context * Departure PoP: [Paris (PAR)](/docs/network/latency/par-paris) * Destination PoP: [Berlin (BER)](/docs/network/latency/ber-berlin) * Region overview: [Europe](/docs/network/latency/regions/europe) → [Europe](/docs/network/latency/regions/europe) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Berlin to Paris latency and RTT](/docs/network/latency/pairs/ber-par-rtt): 15.6 ms **Fastest routes departing Paris (PAR)** * [Paris to London latency and RTT](/docs/network/latency/pairs/par-lon-rtt): 6.4 ms * [Paris to Amsterdam latency and RTT](/docs/network/latency/pairs/par-ams-rtt): 7 ms * [Paris to Frankfurt latency and RTT](/docs/network/latency/pairs/par-fra-rtt): 7.6 ms * [Paris to Marseille latency and RTT](/docs/network/latency/pairs/par-mrs-rtt): 8.9 ms * [Paris to Moscow latency and RTT](/docs/network/latency/pairs/par-mow-rtt): 44.7 ms **Fastest routes arriving at Berlin (BER)** * [Frankfurt to Berlin latency and RTT](/docs/network/latency/pairs/fra-ber-rtt): 6.1 ms * [Amsterdam to Berlin latency and RTT](/docs/network/latency/pairs/ams-ber-rtt): 7.2 ms * [London to Berlin latency and RTT](/docs/network/latency/pairs/lon-ber-rtt): 16.2 ms * [Marseille to Berlin latency and RTT](/docs/network/latency/pairs/mrs-ber-rtt): 20.6 ms * [Moscow to Berlin latency and RTT](/docs/network/latency/pairs/mow-ber-rtt): 27.5 ms **Same corridor (Europe → Europe)** * [Amsterdam to London latency and RTT](/docs/network/latency/pairs/ams-lon-rtt): 5.2 ms * [London to Amsterdam latency and RTT](/docs/network/latency/pairs/lon-ams-rtt): 5.2 ms * [Amsterdam to Frankfurt latency and RTT](/docs/network/latency/pairs/ams-fra-rtt): 5.9 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/par-ber.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Paris → Berlin round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/par-ber-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Paris, France → Frankfurt, Germany RTT 🇫🇷 **Paris, France (PAR)** → 🇩🇪 **Frankfurt, Germany (FRA)** Measurement round: `2026-08-16T04:07:28Z`. Paris to Frankfurt averaged 7.6 ms across 50 replies. The observed range was 7.35 to 8.42 ms; these figures describe one probe round, not a long-term service target. This is number 3 of 19 outgoing routes by average RTT. That ranking compares different destinations and is not adjusted for distance. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------- | | RTT | **7.6 ms** | | Jitter | **0.19 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **4.7 ms** | | Fiber Efficiency | **61.8%** | | Latency Tier | Ultra-Low | | Source Region | [Europe](/docs/network/latency/regions/europe) | | Destination Region | [Europe](/docs/network/latency/regions/europe) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Paris](https://www.geonames.org/2988507) (48.85341, 2.34880) and [Frankfurt am Main](https://www.geonames.org/2925533) (50.11552, 8.68417). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | ------------ | | WGS-84 geodesic distance | **479.9 km** | | Vacuum RTT floor | **3.2 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **4.7 ms** | | Low-latency fiber material floor | **4.68 ms** | | Engineering floor (5% path allowance) | **4.93 ms** | | Research 1.33× mapped-fiber reference | **6.25 ms** | | Estimated unamplified path loss | **100.8 dB** | | Transparent optical spans / inline amplifiers | **7 / 6** | | Published RTT inflation over fiber floor | **1.62×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ----------- | | Minimum RTT | **7.35 ms** | | Average RTT | **7.6 ms** | | Maximum RTT | **8.42 ms** | | Standard deviation | **0.22 ms** | | Stdev / average | **2.9%** | ## Route Context * Departure PoP: [Paris (PAR)](/docs/network/latency/par-paris) * Destination PoP: [Frankfurt (FRA)](/docs/network/latency/fra-frankfurt) * Region overview: [Europe](/docs/network/latency/regions/europe) → [Europe](/docs/network/latency/regions/europe) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Frankfurt to Paris latency and RTT](/docs/network/latency/pairs/fra-par-rtt): 7.6 ms **Fastest routes departing Paris (PAR)** * [Paris to London latency and RTT](/docs/network/latency/pairs/par-lon-rtt): 6.4 ms * [Paris to Amsterdam latency and RTT](/docs/network/latency/pairs/par-ams-rtt): 7 ms * [Paris to Marseille latency and RTT](/docs/network/latency/pairs/par-mrs-rtt): 8.9 ms * [Paris to Berlin latency and RTT](/docs/network/latency/pairs/par-ber-rtt): 15.5 ms * [Paris to Moscow latency and RTT](/docs/network/latency/pairs/par-mow-rtt): 44.7 ms **Fastest routes arriving at Frankfurt (FRA)** * [Amsterdam to Frankfurt latency and RTT](/docs/network/latency/pairs/ams-fra-rtt): 5.9 ms * [Berlin to Frankfurt latency and RTT](/docs/network/latency/pairs/ber-fra-rtt): 6.1 ms * [London to Frankfurt latency and RTT](/docs/network/latency/pairs/lon-fra-rtt): 13.6 ms * [Marseille to Frankfurt latency and RTT](/docs/network/latency/pairs/mrs-fra-rtt): 16 ms * [Moscow to Frankfurt latency and RTT](/docs/network/latency/pairs/mow-fra-rtt): 35.9 ms **Same corridor (Europe → Europe)** * [Amsterdam to London latency and RTT](/docs/network/latency/pairs/ams-lon-rtt): 5.2 ms * [London to Amsterdam latency and RTT](/docs/network/latency/pairs/lon-ams-rtt): 5.2 ms * [Frankfurt to Amsterdam latency and RTT](/docs/network/latency/pairs/fra-ams-rtt): 5.9 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/par-fra.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Paris → Frankfurt round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/par-fra-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Paris, France → São Paulo, Brazil RTT 🇫🇷 **Paris, France (PAR)** → 🇧🇷 **São Paulo, Brazil (GRU)** Measurement round: `2026-08-16T04:07:28Z`. Paris to São Paulo averaged 175.8 ms across 50 replies. The observed range was 173.09 to 186.7 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 2.78 ms, or 1.6% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **175.8 ms** | | Jitter | **2.5 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **91.83 ms** | | Fiber Efficiency | **52.2%** | | Latency Tier | Fair | | Source Region | [Europe](/docs/network/latency/regions/europe) | | Destination Region | [South America](/docs/network/latency/regions/south-america) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Paris](https://www.geonames.org/2988507) (48.85341, 2.34880) and [São Paulo](https://www.geonames.org/3448439) (-23.54750, -46.63611). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **9,377.8 km** | | Vacuum RTT floor | **62.56 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **91.83 ms** | | Low-latency fiber material floor | **91.47 ms** | | Engineering floor (5% path allowance) | **96.44 ms** | | Research 1.33× mapped-fiber reference | **122.14 ms** | | Estimated unamplified path loss | **1969.3 dB** | | Transparent optical spans / inline amplifiers | **124 / 123** | | Published RTT inflation over fiber floor | **1.91×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **173.09 ms** | | Average RTT | **175.8 ms** | | Maximum RTT | **186.7 ms** | | Standard deviation | **2.78 ms** | | Stdev / average | **1.6%** | ## Route Context * Departure PoP: [Paris (PAR)](/docs/network/latency/par-paris) * Destination PoP: [São Paulo (GRU)](/docs/network/latency/gru-sao-paulo) * Region overview: [Europe](/docs/network/latency/regions/europe) → [South America](/docs/network/latency/regions/south-america) * Corridor overview: [Europe to South America](/docs/network/latency/regions/europe-to-south-america) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [São Paulo to Paris latency and RTT](/docs/network/latency/pairs/gru-par-rtt): 180.2 ms **Fastest routes departing Paris (PAR)** * [Paris to London latency and RTT](/docs/network/latency/pairs/par-lon-rtt): 6.4 ms * [Paris to Amsterdam latency and RTT](/docs/network/latency/pairs/par-ams-rtt): 7 ms * [Paris to Frankfurt latency and RTT](/docs/network/latency/pairs/par-fra-rtt): 7.6 ms * [Paris to Marseille latency and RTT](/docs/network/latency/pairs/par-mrs-rtt): 8.9 ms * [Paris to Berlin latency and RTT](/docs/network/latency/pairs/par-ber-rtt): 15.5 ms **Fastest routes arriving at São Paulo (GRU)** * [Ashburn to São Paulo latency and RTT](/docs/network/latency/pairs/iad-gru-rtt): 101.4 ms * [New York to São Paulo latency and RTT](/docs/network/latency/pairs/nyc-gru-rtt): 106.9 ms * [Miami to São Paulo latency and RTT](/docs/network/latency/pairs/mia-gru-rtt): 128.7 ms * [Los Angeles to São Paulo latency and RTT](/docs/network/latency/pairs/lax-gru-rtt): 131.3 ms * [Seattle to São Paulo latency and RTT](/docs/network/latency/pairs/sea-gru-rtt): 155.7 ms **Same corridor (Europe → South America)** * [London to São Paulo latency and RTT](/docs/network/latency/pairs/lon-gru-rtt): 170.7 ms * [Amsterdam to São Paulo latency and RTT](/docs/network/latency/pairs/ams-gru-rtt): 175.5 ms * [Frankfurt to São Paulo latency and RTT](/docs/network/latency/pairs/fra-gru-rtt): 181.3 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/par-gru.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Paris → São Paulo round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/par-gru-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Paris, France → Hong Kong RTT 🇫🇷 **Paris, France (PAR)** → 🇭🇰 **Hong Kong (HKG)** Measurement round: `2026-08-16T04:07:28Z`. Paris to Hong Kong averaged 160.1 ms across 50 replies. The observed range was 155.89 to 182.08 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 4.45 ms, or 2.8% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------------------- | | RTT | **160.1 ms** | | Jitter | **2.89 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **94.48 ms** | | Fiber Efficiency | **59%** | | Latency Tier | Fair | | Source Region | [Europe](/docs/network/latency/regions/europe) | | Destination Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Paris](https://www.geonames.org/2988507) (48.85341, 2.34880) and [Hong Kong](https://www.geonames.org/1819729) (22.27832, 114.17469). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **9,647.6 km** | | Vacuum RTT floor | **64.36 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **94.48 ms** | | Low-latency fiber material floor | **94.1 ms** | | Engineering floor (5% path allowance) | **99.21 ms** | | Research 1.33× mapped-fiber reference | **125.65 ms** | | Estimated unamplified path loss | **2026 dB** | | Transparent optical spans / inline amplifiers | **127 / 126** | | Published RTT inflation over fiber floor | **1.69×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **155.89 ms** | | Average RTT | **160.1 ms** | | Maximum RTT | **182.08 ms** | | Standard deviation | **4.45 ms** | | Stdev / average | **2.8%** | ## Route Context * Departure PoP: [Paris (PAR)](/docs/network/latency/par-paris) * Destination PoP: [Hong Kong (HKG)](/docs/network/latency/hkg-hong-kong) * Region overview: [Europe](/docs/network/latency/regions/europe) → [Asia Pacific](/docs/network/latency/regions/asia-pacific) * Corridor overview: [Europe to Asia Pacific](/docs/network/latency/regions/europe-to-asia-pacific) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Hong Kong to Paris latency and RTT](/docs/network/latency/pairs/hkg-par-rtt): 159.4 ms **Fastest routes departing Paris (PAR)** * [Paris to London latency and RTT](/docs/network/latency/pairs/par-lon-rtt): 6.4 ms * [Paris to Amsterdam latency and RTT](/docs/network/latency/pairs/par-ams-rtt): 7 ms * [Paris to Frankfurt latency and RTT](/docs/network/latency/pairs/par-fra-rtt): 7.6 ms * [Paris to Marseille latency and RTT](/docs/network/latency/pairs/par-mrs-rtt): 8.9 ms * [Paris to Berlin latency and RTT](/docs/network/latency/pairs/par-ber-rtt): 15.5 ms **Fastest routes arriving at Hong Kong (HKG)** * [Taipei to Hong Kong latency and RTT](/docs/network/latency/pairs/tpe-hkg-rtt): 15.5 ms * [Singapore to Hong Kong latency and RTT](/docs/network/latency/pairs/sin-hkg-rtt): 30.8 ms * [Tokyo to Hong Kong latency and RTT](/docs/network/latency/pairs/tyo-hkg-rtt): 44.8 ms * [Moscow to Hong Kong latency and RTT](/docs/network/latency/pairs/mow-hkg-rtt): 118.6 ms * [Seattle to Hong Kong latency and RTT](/docs/network/latency/pairs/sea-hkg-rtt): 130.4 ms **Same corridor (Europe → Asia Pacific)** * [Moscow to Taipei latency and RTT](/docs/network/latency/pairs/mow-tpe-rtt): 132.1 ms * [Marseille to Singapore latency and RTT](/docs/network/latency/pairs/mrs-sin-rtt): 139.6 ms * [Berlin to Hong Kong latency and RTT](/docs/network/latency/pairs/ber-hkg-rtt): 144.9 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/par-hkg.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Paris → Hong Kong round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/par-hkg-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Paris, France → Ashburn, USA RTT 🇫🇷 **Paris, France (PAR)** → 🇺🇸 **Ashburn, USA (IAD)** Measurement round: `2026-08-16T04:07:28Z`. Paris to Ashburn averaged 75.5 ms across 50 replies. The observed range was 72.75 to 85.2 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 2.46 ms, or 3.3% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **75.5 ms** | | Jitter | **2.26 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **60.73 ms** | | Fiber Efficiency | **80.4%** | | Latency Tier | Excellent | | Source Region | [Europe](/docs/network/latency/regions/europe) | | Destination Region | [North America](/docs/network/latency/regions/north-america) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Paris](https://www.geonames.org/2988507) (48.85341, 2.34880) and [Ashburn](https://www.geonames.org/4744870) (39.04372, -77.48749). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | ------------- | | WGS-84 geodesic distance | **6,202 km** | | Vacuum RTT floor | **41.38 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **60.73 ms** | | Low-latency fiber material floor | **60.49 ms** | | Engineering floor (5% path allowance) | **63.78 ms** | | Research 1.33× mapped-fiber reference | **80.78 ms** | | Estimated unamplified path loss | **1302.4 dB** | | Transparent optical spans / inline amplifiers | **82 / 81** | | Published RTT inflation over fiber floor | **1.24×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------ | | Minimum RTT | **72.75 ms** | | Average RTT | **75.5 ms** | | Maximum RTT | **85.2 ms** | | Standard deviation | **2.46 ms** | | Stdev / average | **3.3%** | ## Route Context * Departure PoP: [Paris (PAR)](/docs/network/latency/par-paris) * Destination PoP: [Ashburn (IAD)](/docs/network/latency/iad-ashburn) * Region overview: [Europe](/docs/network/latency/regions/europe) → [North America](/docs/network/latency/regions/north-america) * Corridor overview: [Europe to North America](/docs/network/latency/regions/europe-to-north-america) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Ashburn to Paris latency and RTT](/docs/network/latency/pairs/iad-par-rtt): 76.1 ms **Fastest routes departing Paris (PAR)** * [Paris to London latency and RTT](/docs/network/latency/pairs/par-lon-rtt): 6.4 ms * [Paris to Amsterdam latency and RTT](/docs/network/latency/pairs/par-ams-rtt): 7 ms * [Paris to Frankfurt latency and RTT](/docs/network/latency/pairs/par-fra-rtt): 7.6 ms * [Paris to Marseille latency and RTT](/docs/network/latency/pairs/par-mrs-rtt): 8.9 ms * [Paris to Berlin latency and RTT](/docs/network/latency/pairs/par-ber-rtt): 15.5 ms **Fastest routes arriving at Ashburn (IAD)** * [New York to Ashburn latency and RTT](/docs/network/latency/pairs/nyc-iad-rtt): 6 ms * [Miami to Ashburn latency and RTT](/docs/network/latency/pairs/mia-iad-rtt): 27.3 ms * [Los Angeles to Ashburn latency and RTT](/docs/network/latency/pairs/lax-iad-rtt): 60.4 ms * [Seattle to Ashburn latency and RTT](/docs/network/latency/pairs/sea-iad-rtt): 61.5 ms * [London to Ashburn latency and RTT](/docs/network/latency/pairs/lon-iad-rtt): 71 ms **Same corridor (Europe → North America)** * [London to New York latency and RTT](/docs/network/latency/pairs/lon-nyc-rtt): 63.8 ms * [Amsterdam to New York latency and RTT](/docs/network/latency/pairs/ams-nyc-rtt): 68.6 ms * [Paris to New York latency and RTT](/docs/network/latency/pairs/par-nyc-rtt): 68.9 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/par-iad.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Paris → Ashburn round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/par-iad-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Paris, France → Johannesburg, South Africa RTT 🇫🇷 **Paris, France (PAR)** → 🇿🇦 **Johannesburg, South Africa (JNB)** Measurement round: `2026-08-16T04:07:28Z`. Paris to Johannesburg averaged 180.9 ms, compared with a 85.17 ms city-centre fiber reference. The difference cannot be assigned to a particular cable or carrier from RTT alone. Standard deviation was 3.41 ms, or 1.9% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------- | | RTT | **180.9 ms** | | Jitter | **3.26 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **85.17 ms** | | Fiber Efficiency | **47.1%** | | Latency Tier | Fair | | Source Region | [Europe](/docs/network/latency/regions/europe) | | Destination Region | [Africa](/docs/network/latency/regions/africa) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Paris](https://www.geonames.org/2988507) (48.85341, 2.34880) and [Johannesburg](https://www.geonames.org/993800) (-26.20227, 28.04363). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **8,697.3 km** | | Vacuum RTT floor | **58.02 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **85.17 ms** | | Low-latency fiber material floor | **84.83 ms** | | Engineering floor (5% path allowance) | **89.44 ms** | | Research 1.33× mapped-fiber reference | **113.28 ms** | | Estimated unamplified path loss | **1826.4 dB** | | Transparent optical spans / inline amplifiers | **115 / 114** | | Published RTT inflation over fiber floor | **2.12×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **176.39 ms** | | Average RTT | **180.9 ms** | | Maximum RTT | **195.6 ms** | | Standard deviation | **3.41 ms** | | Stdev / average | **1.9%** | ## Route Context * Departure PoP: [Paris (PAR)](/docs/network/latency/par-paris) * Destination PoP: [Johannesburg (JNB)](/docs/network/latency/jnb-johannesburg) * Region overview: [Europe](/docs/network/latency/regions/europe) → [Africa](/docs/network/latency/regions/africa) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Johannesburg to Paris latency and RTT](/docs/network/latency/pairs/jnb-par-rtt): 164.4 ms **Fastest routes departing Paris (PAR)** * [Paris to London latency and RTT](/docs/network/latency/pairs/par-lon-rtt): 6.4 ms * [Paris to Amsterdam latency and RTT](/docs/network/latency/pairs/par-ams-rtt): 7 ms * [Paris to Frankfurt latency and RTT](/docs/network/latency/pairs/par-fra-rtt): 7.6 ms * [Paris to Marseille latency and RTT](/docs/network/latency/pairs/par-mrs-rtt): 8.9 ms * [Paris to Berlin latency and RTT](/docs/network/latency/pairs/par-ber-rtt): 15.5 ms **Fastest routes arriving at Johannesburg (JNB)** * [Amsterdam to Johannesburg latency and RTT](/docs/network/latency/pairs/ams-jnb-rtt): 171.9 ms * [Berlin to Johannesburg latency and RTT](/docs/network/latency/pairs/ber-jnb-rtt): 172.1 ms * [London to Johannesburg latency and RTT](/docs/network/latency/pairs/lon-jnb-rtt): 179.6 ms * [Frankfurt to Johannesburg latency and RTT](/docs/network/latency/pairs/fra-jnb-rtt): 188.4 ms * [Marseille to Johannesburg latency and RTT](/docs/network/latency/pairs/mrs-jnb-rtt): 198.5 ms **Same corridor (Europe → Africa)** * [Moscow to Johannesburg latency and RTT](/docs/network/latency/pairs/mow-jnb-rtt): 201.9 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/par-jnb.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Paris → Johannesburg round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/par-jnb-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Paris, France → Los Angeles, USA RTT 🇫🇷 **Paris, France (PAR)** → 🇺🇸 **Los Angeles, USA (LAX)** Measurement round: `2026-08-16T04:07:28Z`. Paris to Los Angeles averaged 136.5 ms across 50 replies. The observed range was 130.92 to 152.23 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 5.13 ms, or 3.8% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **136.5 ms** | | Jitter | **3.55 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **89.18 ms** | | Fiber Efficiency | **65.3%** | | Latency Tier | Good | | Source Region | [Europe](/docs/network/latency/regions/europe) | | Destination Region | [North America](/docs/network/latency/regions/north-america) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Paris](https://www.geonames.org/2988507) (48.85341, 2.34880) and [Los Angeles](https://www.geonames.org/5368361) (34.05223, -118.24368). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **9,107.2 km** | | Vacuum RTT floor | **60.76 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **89.18 ms** | | Low-latency fiber material floor | **88.83 ms** | | Engineering floor (5% path allowance) | **93.66 ms** | | Research 1.33× mapped-fiber reference | **118.62 ms** | | Estimated unamplified path loss | **1912.5 dB** | | Transparent optical spans / inline amplifiers | **120 / 119** | | Published RTT inflation over fiber floor | **1.53×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **130.92 ms** | | Average RTT | **136.5 ms** | | Maximum RTT | **152.23 ms** | | Standard deviation | **5.13 ms** | | Stdev / average | **3.8%** | ## Route Context * Departure PoP: [Paris (PAR)](/docs/network/latency/par-paris) * Destination PoP: [Los Angeles (LAX)](/docs/network/latency/lax-los-angeles) * Region overview: [Europe](/docs/network/latency/regions/europe) → [North America](/docs/network/latency/regions/north-america) * Corridor overview: [Europe to North America](/docs/network/latency/regions/europe-to-north-america) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Los Angeles to Paris latency and RTT](/docs/network/latency/pairs/lax-par-rtt): 135.9 ms **Fastest routes departing Paris (PAR)** * [Paris to London latency and RTT](/docs/network/latency/pairs/par-lon-rtt): 6.4 ms * [Paris to Amsterdam latency and RTT](/docs/network/latency/pairs/par-ams-rtt): 7 ms * [Paris to Frankfurt latency and RTT](/docs/network/latency/pairs/par-fra-rtt): 7.6 ms * [Paris to Marseille latency and RTT](/docs/network/latency/pairs/par-mrs-rtt): 8.9 ms * [Paris to Berlin latency and RTT](/docs/network/latency/pairs/par-ber-rtt): 15.5 ms **Fastest routes arriving at Los Angeles (LAX)** * [Seattle to Los Angeles latency and RTT](/docs/network/latency/pairs/sea-lax-rtt): 25.9 ms * [Miami to Los Angeles latency and RTT](/docs/network/latency/pairs/mia-lax-rtt): 56.8 ms * [New York to Los Angeles latency and RTT](/docs/network/latency/pairs/nyc-lax-rtt): 58.9 ms * [Ashburn to Los Angeles latency and RTT](/docs/network/latency/pairs/iad-lax-rtt): 60.7 ms * [Tokyo to Los Angeles latency and RTT](/docs/network/latency/pairs/tyo-lax-rtt): 101.2 ms **Same corridor (Europe → North America)** * [London to New York latency and RTT](/docs/network/latency/pairs/lon-nyc-rtt): 63.8 ms * [Amsterdam to New York latency and RTT](/docs/network/latency/pairs/ams-nyc-rtt): 68.6 ms * [Paris to New York latency and RTT](/docs/network/latency/pairs/par-nyc-rtt): 68.9 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/par-lax.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Paris → Los Angeles round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/par-lax-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Paris, France → London, UK RTT 🇫🇷 **Paris, France (PAR)** → 🇬🇧 **London, UK (LON)** Measurement round: `2026-08-16T04:07:28Z`. Paris to London averaged 6.4 ms across 50 replies. The observed range was 6.03 to 7.35 ms; these figures describe one probe round, not a long-term service target. This is number 1 of 19 outgoing routes by average RTT. That ranking compares different destinations and is not adjusted for distance. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------- | | RTT | **6.4 ms** | | Jitter | **0.2 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **3.37 ms** | | Fiber Efficiency | **52.7%** | | Latency Tier | Ultra-Low | | Source Region | [Europe](/docs/network/latency/regions/europe) | | Destination Region | [Europe](/docs/network/latency/regions/europe) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Paris](https://www.geonames.org/2988507) (48.85341, 2.34880) and [London](https://www.geonames.org/2643743) (51.50853, -0.12574). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | ------------ | | WGS-84 geodesic distance | **344.1 km** | | Vacuum RTT floor | **2.3 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **3.37 ms** | | Low-latency fiber material floor | **3.36 ms** | | Engineering floor (5% path allowance) | **3.54 ms** | | Research 1.33× mapped-fiber reference | **4.48 ms** | | Estimated unamplified path loss | **72.3 dB** | | Transparent optical spans / inline amplifiers | **5 / 4** | | Published RTT inflation over fiber floor | **1.9×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ----------- | | Minimum RTT | **6.03 ms** | | Average RTT | **6.4 ms** | | Maximum RTT | **7.35 ms** | | Standard deviation | **0.26 ms** | | Stdev / average | **4.1%** | ## Route Context * Departure PoP: [Paris (PAR)](/docs/network/latency/par-paris) * Destination PoP: [London (LON)](/docs/network/latency/lon-london) * Region overview: [Europe](/docs/network/latency/regions/europe) → [Europe](/docs/network/latency/regions/europe) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [London to Paris latency and RTT](/docs/network/latency/pairs/lon-par-rtt): 6.4 ms **Fastest routes departing Paris (PAR)** * [Paris to Amsterdam latency and RTT](/docs/network/latency/pairs/par-ams-rtt): 7 ms * [Paris to Frankfurt latency and RTT](/docs/network/latency/pairs/par-fra-rtt): 7.6 ms * [Paris to Marseille latency and RTT](/docs/network/latency/pairs/par-mrs-rtt): 8.9 ms * [Paris to Berlin latency and RTT](/docs/network/latency/pairs/par-ber-rtt): 15.5 ms * [Paris to Moscow latency and RTT](/docs/network/latency/pairs/par-mow-rtt): 44.7 ms **Fastest routes arriving at London (LON)** * [Amsterdam to London latency and RTT](/docs/network/latency/pairs/ams-lon-rtt): 5.2 ms * [Frankfurt to London latency and RTT](/docs/network/latency/pairs/fra-lon-rtt): 12.9 ms * [Berlin to London latency and RTT](/docs/network/latency/pairs/ber-lon-rtt): 15.2 ms * [Marseille to London latency and RTT](/docs/network/latency/pairs/mrs-lon-rtt): 18.9 ms * [Moscow to London latency and RTT](/docs/network/latency/pairs/mow-lon-rtt): 42.4 ms **Same corridor (Europe → Europe)** * [London to Amsterdam latency and RTT](/docs/network/latency/pairs/lon-ams-rtt): 5.2 ms * [Amsterdam to Frankfurt latency and RTT](/docs/network/latency/pairs/ams-fra-rtt): 5.9 ms * [Frankfurt to Amsterdam latency and RTT](/docs/network/latency/pairs/fra-ams-rtt): 5.9 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/par-lon.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Paris → London round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/par-lon-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Paris, France → Melbourne, Australia RTT 🇫🇷 **Paris, France (PAR)** → 🇦🇺 **Melbourne, Australia (MEL)** Measurement round: `2026-08-16T04:07:28Z`. Paris to Melbourne averaged 246.4 ms across 50 replies. The observed range was 232.44 to 296.68 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 11.89 ms, or 4.8% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------------------- | | RTT | **246.4 ms** | | Jitter | **11.49 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **164.4 ms** | | Fiber Efficiency | **66.7%** | | Latency Tier | Fair | | Source Region | [Europe](/docs/network/latency/regions/europe) | | Destination Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Paris](https://www.geonames.org/2988507) (48.85341, 2.34880) and [Melbourne](https://www.geonames.org/2158177) (-37.81400, 144.96332). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | --------------- | | WGS-84 geodesic distance | **16,787.7 km** | | Vacuum RTT floor | **112 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **164.4 ms** | | Low-latency fiber material floor | **163.74 ms** | | Engineering floor (5% path allowance) | **172.64 ms** | | Research 1.33× mapped-fiber reference | **218.65 ms** | | Estimated unamplified path loss | **3525.4 dB** | | Transparent optical spans / inline amplifiers | **221 / 220** | | Published RTT inflation over fiber floor | **1.5×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **232.44 ms** | | Average RTT | **246.4 ms** | | Maximum RTT | **296.68 ms** | | Standard deviation | **11.89 ms** | | Stdev / average | **4.8%** | ## Route Context * Departure PoP: [Paris (PAR)](/docs/network/latency/par-paris) * Destination PoP: [Melbourne (MEL)](/docs/network/latency/mel-melbourne) * Region overview: [Europe](/docs/network/latency/regions/europe) → [Asia Pacific](/docs/network/latency/regions/asia-pacific) * Corridor overview: [Europe to Asia Pacific](/docs/network/latency/regions/europe-to-asia-pacific) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Melbourne to Paris latency and RTT](/docs/network/latency/pairs/mel-par-rtt): 246.5 ms **Fastest routes departing Paris (PAR)** * [Paris to London latency and RTT](/docs/network/latency/pairs/par-lon-rtt): 6.4 ms * [Paris to Amsterdam latency and RTT](/docs/network/latency/pairs/par-ams-rtt): 7 ms * [Paris to Frankfurt latency and RTT](/docs/network/latency/pairs/par-fra-rtt): 7.6 ms * [Paris to Marseille latency and RTT](/docs/network/latency/pairs/par-mrs-rtt): 8.9 ms * [Paris to Berlin latency and RTT](/docs/network/latency/pairs/par-ber-rtt): 15.5 ms **Fastest routes arriving at Melbourne (MEL)** * [Sydney to Melbourne latency and RTT](/docs/network/latency/pairs/syd-mel-rtt): 9.8 ms * [Singapore to Melbourne latency and RTT](/docs/network/latency/pairs/sin-mel-rtt): 88.6 ms * [Tokyo to Melbourne latency and RTT](/docs/network/latency/pairs/tyo-mel-rtt): 113.9 ms * [Hong Kong to Melbourne latency and RTT](/docs/network/latency/pairs/hkg-mel-rtt): 138.4 ms * [Taipei to Melbourne latency and RTT](/docs/network/latency/pairs/tpe-mel-rtt): 142 ms **Same corridor (Europe → Asia Pacific)** * [Moscow to Hong Kong latency and RTT](/docs/network/latency/pairs/mow-hkg-rtt): 118.6 ms * [Moscow to Taipei latency and RTT](/docs/network/latency/pairs/mow-tpe-rtt): 132.1 ms * [Marseille to Singapore latency and RTT](/docs/network/latency/pairs/mrs-sin-rtt): 139.6 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/par-mel.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Paris → Melbourne round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/par-mel-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Paris, France → Miami, USA RTT 🇫🇷 **Paris, France (PAR)** → 🇺🇸 **Miami, USA (MIA)** Measurement round: `2026-08-16T04:07:28Z`. Paris to Miami averaged 106.8 ms across 50 replies. The observed range was 101.15 to 125.38 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 5.28 ms, or 4.9% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **106.8 ms** | | Jitter | **4.08 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **72.16 ms** | | Fiber Efficiency | **67.6%** | | Latency Tier | Good | | Source Region | [Europe](/docs/network/latency/regions/europe) | | Destination Region | [North America](/docs/network/latency/regions/north-america) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Paris](https://www.geonames.org/2988507) (48.85341, 2.34880) and [Miami](https://www.geonames.org/4164138) (25.77427, -80.19366). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **7,368.8 km** | | Vacuum RTT floor | **49.16 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **72.16 ms** | | Low-latency fiber material floor | **71.87 ms** | | Engineering floor (5% path allowance) | **75.78 ms** | | Research 1.33× mapped-fiber reference | **95.97 ms** | | Estimated unamplified path loss | **1547.4 dB** | | Transparent optical spans / inline amplifiers | **97 / 96** | | Published RTT inflation over fiber floor | **1.48×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **101.15 ms** | | Average RTT | **106.8 ms** | | Maximum RTT | **125.38 ms** | | Standard deviation | **5.28 ms** | | Stdev / average | **4.9%** | ## Route Context * Departure PoP: [Paris (PAR)](/docs/network/latency/par-paris) * Destination PoP: [Miami (MIA)](/docs/network/latency/mia-miami) * Region overview: [Europe](/docs/network/latency/regions/europe) → [North America](/docs/network/latency/regions/north-america) * Corridor overview: [Europe to North America](/docs/network/latency/regions/europe-to-north-america) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Miami to Paris latency and RTT](/docs/network/latency/pairs/mia-par-rtt): 106.2 ms **Fastest routes departing Paris (PAR)** * [Paris to London latency and RTT](/docs/network/latency/pairs/par-lon-rtt): 6.4 ms * [Paris to Amsterdam latency and RTT](/docs/network/latency/pairs/par-ams-rtt): 7 ms * [Paris to Frankfurt latency and RTT](/docs/network/latency/pairs/par-fra-rtt): 7.6 ms * [Paris to Marseille latency and RTT](/docs/network/latency/pairs/par-mrs-rtt): 8.9 ms * [Paris to Berlin latency and RTT](/docs/network/latency/pairs/par-ber-rtt): 15.5 ms **Fastest routes arriving at Miami (MIA)** * [Ashburn to Miami latency and RTT](/docs/network/latency/pairs/iad-mia-rtt): 27.4 ms * [New York to Miami latency and RTT](/docs/network/latency/pairs/nyc-mia-rtt): 32.9 ms * [Los Angeles to Miami latency and RTT](/docs/network/latency/pairs/lax-mia-rtt): 57.3 ms * [São Paulo to Miami latency and RTT](/docs/network/latency/pairs/gru-mia-rtt): 74 ms * [Seattle to Miami latency and RTT](/docs/network/latency/pairs/sea-mia-rtt): 81.7 ms **Same corridor (Europe → North America)** * [London to New York latency and RTT](/docs/network/latency/pairs/lon-nyc-rtt): 63.8 ms * [Amsterdam to New York latency and RTT](/docs/network/latency/pairs/ams-nyc-rtt): 68.6 ms * [Paris to New York latency and RTT](/docs/network/latency/pairs/par-nyc-rtt): 68.9 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/par-mia.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Paris → Miami round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/par-mia-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Paris, France → Moscow, Russia RTT 🇫🇷 **Paris, France (PAR)** → 🇷🇺 **Moscow, Russia (MOW)** Measurement round: `2026-08-16T04:07:28Z`. Paris to Moscow averaged 44.7 ms across 50 replies. The observed range was 43.04 to 51.46 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 1.56 ms, or 3.5% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------- | | RTT | **44.7 ms** | | Jitter | **1.18 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **24.42 ms** | | Fiber Efficiency | **54.6%** | | Latency Tier | Excellent | | Source Region | [Europe](/docs/network/latency/regions/europe) | | Destination Region | [Europe](/docs/network/latency/regions/europe) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Paris](https://www.geonames.org/2988507) (48.85341, 2.34880) and [Moscow](https://www.geonames.org/524901) (55.75204, 37.61781). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **2,494.1 km** | | Vacuum RTT floor | **16.64 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **24.42 ms** | | Low-latency fiber material floor | **24.33 ms** | | Engineering floor (5% path allowance) | **25.65 ms** | | Research 1.33× mapped-fiber reference | **32.48 ms** | | Estimated unamplified path loss | **523.8 dB** | | Transparent optical spans / inline amplifiers | **33 / 32** | | Published RTT inflation over fiber floor | **1.83×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------ | | Minimum RTT | **43.04 ms** | | Average RTT | **44.7 ms** | | Maximum RTT | **51.46 ms** | | Standard deviation | **1.56 ms** | | Stdev / average | **3.5%** | ## Route Context * Departure PoP: [Paris (PAR)](/docs/network/latency/par-paris) * Destination PoP: [Moscow (MOW)](/docs/network/latency/mow-moscow) * Region overview: [Europe](/docs/network/latency/regions/europe) → [Europe](/docs/network/latency/regions/europe) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Moscow to Paris latency and RTT](/docs/network/latency/pairs/mow-par-rtt): 44 ms **Fastest routes departing Paris (PAR)** * [Paris to London latency and RTT](/docs/network/latency/pairs/par-lon-rtt): 6.4 ms * [Paris to Amsterdam latency and RTT](/docs/network/latency/pairs/par-ams-rtt): 7 ms * [Paris to Frankfurt latency and RTT](/docs/network/latency/pairs/par-fra-rtt): 7.6 ms * [Paris to Marseille latency and RTT](/docs/network/latency/pairs/par-mrs-rtt): 8.9 ms * [Paris to Berlin latency and RTT](/docs/network/latency/pairs/par-ber-rtt): 15.5 ms **Fastest routes arriving at Moscow (MOW)** * [Berlin to Moscow latency and RTT](/docs/network/latency/pairs/ber-mow-rtt): 28.6 ms * [Frankfurt to Moscow latency and RTT](/docs/network/latency/pairs/fra-mow-rtt): 34.7 ms * [Amsterdam to Moscow latency and RTT](/docs/network/latency/pairs/ams-mow-rtt): 38.1 ms * [London to Moscow latency and RTT](/docs/network/latency/pairs/lon-mow-rtt): 43.5 ms * [Marseille to Moscow latency and RTT](/docs/network/latency/pairs/mrs-mow-rtt): 49.6 ms **Same corridor (Europe → Europe)** * [Amsterdam to London latency and RTT](/docs/network/latency/pairs/ams-lon-rtt): 5.2 ms * [London to Amsterdam latency and RTT](/docs/network/latency/pairs/lon-ams-rtt): 5.2 ms * [Amsterdam to Frankfurt latency and RTT](/docs/network/latency/pairs/ams-fra-rtt): 5.9 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/par-mow.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Paris → Moscow round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/par-mow-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Paris, France → Marseille, France RTT 🇫🇷 **Paris, France (PAR)** → 🇫🇷 **Marseille, France (MRS)** Measurement round: `2026-08-16T04:07:28Z`. Paris to Marseille averaged 8.9 ms across 50 replies. The observed range was 8.59 to 9.69 ms; these figures describe one probe round, not a long-term service target. This is number 4 of 19 outgoing routes by average RTT. That ranking compares different destinations and is not adjusted for distance. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------- | | RTT | **8.9 ms** | | Jitter | **0.25 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **6.47 ms** | | Fiber Efficiency | **72.7%** | | Latency Tier | Ultra-Low | | Source Region | [Europe](/docs/network/latency/regions/europe) | | Destination Region | [Europe](/docs/network/latency/regions/europe) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Paris](https://www.geonames.org/2988507) (48.85341, 2.34880) and [Marseille](https://www.geonames.org/2995469) (43.29695, 5.38107). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | ------------ | | WGS-84 geodesic distance | **660.5 km** | | Vacuum RTT floor | **4.41 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **6.47 ms** | | Low-latency fiber material floor | **6.44 ms** | | Engineering floor (5% path allowance) | **6.79 ms** | | Research 1.33× mapped-fiber reference | **8.6 ms** | | Estimated unamplified path loss | **138.7 dB** | | Transparent optical spans / inline amplifiers | **9 / 8** | | Published RTT inflation over fiber floor | **1.38×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ----------- | | Minimum RTT | **8.59 ms** | | Average RTT | **8.9 ms** | | Maximum RTT | **9.69 ms** | | Standard deviation | **0.26 ms** | | Stdev / average | **2.9%** | ## Route Context * Departure PoP: [Paris (PAR)](/docs/network/latency/par-paris) * Destination PoP: [Marseille (MRS)](/docs/network/latency/mrs-marseille) * Region overview: [Europe](/docs/network/latency/regions/europe) → [Europe](/docs/network/latency/regions/europe) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Marseille to Paris latency and RTT](/docs/network/latency/pairs/mrs-par-rtt): 8.9 ms **Fastest routes departing Paris (PAR)** * [Paris to London latency and RTT](/docs/network/latency/pairs/par-lon-rtt): 6.4 ms * [Paris to Amsterdam latency and RTT](/docs/network/latency/pairs/par-ams-rtt): 7 ms * [Paris to Frankfurt latency and RTT](/docs/network/latency/pairs/par-fra-rtt): 7.6 ms * [Paris to Berlin latency and RTT](/docs/network/latency/pairs/par-ber-rtt): 15.5 ms * [Paris to Moscow latency and RTT](/docs/network/latency/pairs/par-mow-rtt): 44.7 ms **Fastest routes arriving at Marseille (MRS)** * [Frankfurt to Marseille latency and RTT](/docs/network/latency/pairs/fra-mrs-rtt): 16.4 ms * [London to Marseille latency and RTT](/docs/network/latency/pairs/lon-mrs-rtt): 17.9 ms * [Amsterdam to Marseille latency and RTT](/docs/network/latency/pairs/ams-mrs-rtt): 19.6 ms * [Berlin to Marseille latency and RTT](/docs/network/latency/pairs/ber-mrs-rtt): 20.2 ms * [Moscow to Marseille latency and RTT](/docs/network/latency/pairs/mow-mrs-rtt): 49.9 ms **Same corridor (Europe → Europe)** * [Amsterdam to London latency and RTT](/docs/network/latency/pairs/ams-lon-rtt): 5.2 ms * [London to Amsterdam latency and RTT](/docs/network/latency/pairs/lon-ams-rtt): 5.2 ms * [Amsterdam to Frankfurt latency and RTT](/docs/network/latency/pairs/ams-fra-rtt): 5.9 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/par-mrs.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Paris → Marseille round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/par-mrs-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Paris, France → New York, USA RTT 🇫🇷 **Paris, France (PAR)** → 🇺🇸 **New York, USA (NYC)** Measurement round: `2026-08-16T04:07:28Z`. Paris to New York averaged 68.9 ms across 50 replies. The observed range was 66.76 to 73.07 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 1.55 ms, or 2.2% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **68.9 ms** | | Jitter | **1.37 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **57.31 ms** | | Fiber Efficiency | **83.2%** | | Latency Tier | Excellent | | Source Region | [Europe](/docs/network/latency/regions/europe) | | Destination Region | [North America](/docs/network/latency/regions/north-america) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Paris](https://www.geonames.org/2988507) (48.85341, 2.34880) and [New York City](https://www.geonames.org/5128581) (40.71427, -74.00597). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **5,852.7 km** | | Vacuum RTT floor | **39.05 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **57.31 ms** | | Low-latency fiber material floor | **57.08 ms** | | Engineering floor (5% path allowance) | **60.19 ms** | | Research 1.33× mapped-fiber reference | **76.23 ms** | | Estimated unamplified path loss | **1229.1 dB** | | Transparent optical spans / inline amplifiers | **77 / 76** | | Published RTT inflation over fiber floor | **1.2×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------ | | Minimum RTT | **66.76 ms** | | Average RTT | **68.9 ms** | | Maximum RTT | **73.07 ms** | | Standard deviation | **1.55 ms** | | Stdev / average | **2.2%** | ## Route Context * Departure PoP: [Paris (PAR)](/docs/network/latency/par-paris) * Destination PoP: [New York (NYC)](/docs/network/latency/nyc-new-york) * Region overview: [Europe](/docs/network/latency/regions/europe) → [North America](/docs/network/latency/regions/north-america) * Corridor overview: [Europe to North America](/docs/network/latency/regions/europe-to-north-america) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [New York to Paris latency and RTT](/docs/network/latency/pairs/nyc-par-rtt): 68.9 ms **Fastest routes departing Paris (PAR)** * [Paris to London latency and RTT](/docs/network/latency/pairs/par-lon-rtt): 6.4 ms * [Paris to Amsterdam latency and RTT](/docs/network/latency/pairs/par-ams-rtt): 7 ms * [Paris to Frankfurt latency and RTT](/docs/network/latency/pairs/par-fra-rtt): 7.6 ms * [Paris to Marseille latency and RTT](/docs/network/latency/pairs/par-mrs-rtt): 8.9 ms * [Paris to Berlin latency and RTT](/docs/network/latency/pairs/par-ber-rtt): 15.5 ms **Fastest routes arriving at New York (NYC)** * [Ashburn to New York latency and RTT](/docs/network/latency/pairs/iad-nyc-rtt): 6 ms * [Miami to New York latency and RTT](/docs/network/latency/pairs/mia-nyc-rtt): 33.3 ms * [Los Angeles to New York latency and RTT](/docs/network/latency/pairs/lax-nyc-rtt): 57.9 ms * [Seattle to New York latency and RTT](/docs/network/latency/pairs/sea-nyc-rtt): 59.5 ms * [London to New York latency and RTT](/docs/network/latency/pairs/lon-nyc-rtt): 63.8 ms **Same corridor (Europe → North America)** * [Amsterdam to New York latency and RTT](/docs/network/latency/pairs/ams-nyc-rtt): 68.6 ms * [London to Ashburn latency and RTT](/docs/network/latency/pairs/lon-iad-rtt): 71 ms * [Frankfurt to New York latency and RTT](/docs/network/latency/pairs/fra-nyc-rtt): 74.4 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/par-nyc.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Paris → New York round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/par-nyc-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Paris, France → Seattle, USA RTT 🇫🇷 **Paris, France (PAR)** → 🇺🇸 **Seattle, USA (SEA)** Measurement round: `2026-08-16T04:07:28Z`. Paris to Seattle averaged 128 ms across 50 replies. The observed range was 123.76 to 137.5 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 3.51 ms, or 2.7% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **128 ms** | | Jitter | **2.63 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **78.98 ms** | | Fiber Efficiency | **61.7%** | | Latency Tier | Good | | Source Region | [Europe](/docs/network/latency/regions/europe) | | Destination Region | [North America](/docs/network/latency/regions/north-america) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Paris](https://www.geonames.org/2988507) (48.85341, 2.34880) and [Seattle](https://www.geonames.org/5809844) (47.60621, -122.33207). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **8,065.2 km** | | Vacuum RTT floor | **53.8 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **78.98 ms** | | Low-latency fiber material floor | **78.66 ms** | | Engineering floor (5% path allowance) | **82.94 ms** | | Research 1.33× mapped-fiber reference | **105.04 ms** | | Estimated unamplified path loss | **1693.7 dB** | | Transparent optical spans / inline amplifiers | **106 / 105** | | Published RTT inflation over fiber floor | **1.62×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **123.76 ms** | | Average RTT | **128 ms** | | Maximum RTT | **137.5 ms** | | Standard deviation | **3.51 ms** | | Stdev / average | **2.7%** | ## Route Context * Departure PoP: [Paris (PAR)](/docs/network/latency/par-paris) * Destination PoP: [Seattle (SEA)](/docs/network/latency/sea-seattle) * Region overview: [Europe](/docs/network/latency/regions/europe) → [North America](/docs/network/latency/regions/north-america) * Corridor overview: [Europe to North America](/docs/network/latency/regions/europe-to-north-america) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Seattle to Paris latency and RTT](/docs/network/latency/pairs/sea-par-rtt): 129.2 ms **Fastest routes departing Paris (PAR)** * [Paris to London latency and RTT](/docs/network/latency/pairs/par-lon-rtt): 6.4 ms * [Paris to Amsterdam latency and RTT](/docs/network/latency/pairs/par-ams-rtt): 7 ms * [Paris to Frankfurt latency and RTT](/docs/network/latency/pairs/par-fra-rtt): 7.6 ms * [Paris to Marseille latency and RTT](/docs/network/latency/pairs/par-mrs-rtt): 8.9 ms * [Paris to Berlin latency and RTT](/docs/network/latency/pairs/par-ber-rtt): 15.5 ms **Fastest routes arriving at Seattle (SEA)** * [Los Angeles to Seattle latency and RTT](/docs/network/latency/pairs/lax-sea-rtt): 26.9 ms * [New York to Seattle latency and RTT](/docs/network/latency/pairs/nyc-sea-rtt): 58.7 ms * [Ashburn to Seattle latency and RTT](/docs/network/latency/pairs/iad-sea-rtt): 62.3 ms * [Miami to Seattle latency and RTT](/docs/network/latency/pairs/mia-sea-rtt): 81.9 ms * [Tokyo to Seattle latency and RTT](/docs/network/latency/pairs/tyo-sea-rtt): 84.8 ms **Same corridor (Europe → North America)** * [London to New York latency and RTT](/docs/network/latency/pairs/lon-nyc-rtt): 63.8 ms * [Amsterdam to New York latency and RTT](/docs/network/latency/pairs/ams-nyc-rtt): 68.6 ms * [Paris to New York latency and RTT](/docs/network/latency/pairs/par-nyc-rtt): 68.9 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/par-sea.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Paris → Seattle round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/par-sea-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Paris, France → Singapore RTT 🇫🇷 **Paris, France (PAR)** → 🇸🇬 **Singapore (SIN)** Measurement round: `2026-08-16T04:07:28Z`. Paris to Singapore averaged 150.3 ms across 50 replies. The observed range was 144.03 to 164.23 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 4.53 ms, or 3.0% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------------------- | | RTT | **150.3 ms** | | Jitter | **4.83 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **105.19 ms** | | Fiber Efficiency | **70%** | | Latency Tier | Fair | | Source Region | [Europe](/docs/network/latency/regions/europe) | | Destination Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Paris](https://www.geonames.org/2988507) (48.85341, 2.34880) and [Singapore](https://www.geonames.org/1880252) (1.28967, 103.85007). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | --------------- | | WGS-84 geodesic distance | **10,741.6 km** | | Vacuum RTT floor | **71.66 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **105.19 ms** | | Low-latency fiber material floor | **104.77 ms** | | Engineering floor (5% path allowance) | **110.46 ms** | | Research 1.33× mapped-fiber reference | **139.9 ms** | | Estimated unamplified path loss | **2255.7 dB** | | Transparent optical spans / inline amplifiers | **141 / 140** | | Published RTT inflation over fiber floor | **1.43×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **144.03 ms** | | Average RTT | **150.3 ms** | | Maximum RTT | **164.23 ms** | | Standard deviation | **4.53 ms** | | Stdev / average | **3.0%** | ## Route Context * Departure PoP: [Paris (PAR)](/docs/network/latency/par-paris) * Destination PoP: [Singapore (SIN)](/docs/network/latency/sin-singapore) * Region overview: [Europe](/docs/network/latency/regions/europe) → [Asia Pacific](/docs/network/latency/regions/asia-pacific) * Corridor overview: [Europe to Asia Pacific](/docs/network/latency/regions/europe-to-asia-pacific) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Singapore to Paris latency and RTT](/docs/network/latency/pairs/sin-par-rtt): 151.4 ms **Fastest routes departing Paris (PAR)** * [Paris to London latency and RTT](/docs/network/latency/pairs/par-lon-rtt): 6.4 ms * [Paris to Amsterdam latency and RTT](/docs/network/latency/pairs/par-ams-rtt): 7 ms * [Paris to Frankfurt latency and RTT](/docs/network/latency/pairs/par-fra-rtt): 7.6 ms * [Paris to Marseille latency and RTT](/docs/network/latency/pairs/par-mrs-rtt): 8.9 ms * [Paris to Berlin latency and RTT](/docs/network/latency/pairs/par-ber-rtt): 15.5 ms **Fastest routes arriving at Singapore (SIN)** * [Hong Kong to Singapore latency and RTT](/docs/network/latency/pairs/hkg-sin-rtt): 31.5 ms * [Taipei to Singapore latency and RTT](/docs/network/latency/pairs/tpe-sin-rtt): 44.3 ms * [Tokyo to Singapore latency and RTT](/docs/network/latency/pairs/tyo-sin-rtt): 69.4 ms * [Melbourne to Singapore latency and RTT](/docs/network/latency/pairs/mel-sin-rtt): 88.4 ms * [Sydney to Singapore latency and RTT](/docs/network/latency/pairs/syd-sin-rtt): 94.5 ms **Same corridor (Europe → Asia Pacific)** * [Moscow to Hong Kong latency and RTT](/docs/network/latency/pairs/mow-hkg-rtt): 118.6 ms * [Moscow to Taipei latency and RTT](/docs/network/latency/pairs/mow-tpe-rtt): 132.1 ms * [Marseille to Singapore latency and RTT](/docs/network/latency/pairs/mrs-sin-rtt): 139.6 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/par-sin.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Paris → Singapore round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/par-sin-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Paris, France → Sydney, Australia RTT 🇫🇷 **Paris, France (PAR)** → 🇦🇺 **Sydney, Australia (SYD)** Measurement round: `2026-08-16T04:07:28Z`. Paris to Sydney averaged 249.7 ms across 50 replies. The observed range was 240.64 to 279.37 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 9.37 ms, or 3.8% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------------------- | | RTT | **249.7 ms** | | Jitter | **7.49 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **166.06 ms** | | Fiber Efficiency | **66.5%** | | Latency Tier | Fair | | Source Region | [Europe](/docs/network/latency/regions/europe) | | Destination Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Paris](https://www.geonames.org/2988507) (48.85341, 2.34880) and [Sydney](https://www.geonames.org/2147714) (-33.86785, 151.20732). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | --------------- | | WGS-84 geodesic distance | **16,957.7 km** | | Vacuum RTT floor | **113.13 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **166.06 ms** | | Low-latency fiber material floor | **165.4 ms** | | Engineering floor (5% path allowance) | **174.39 ms** | | Research 1.33× mapped-fiber reference | **220.86 ms** | | Estimated unamplified path loss | **3561.1 dB** | | Transparent optical spans / inline amplifiers | **223 / 222** | | Published RTT inflation over fiber floor | **1.5×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **240.64 ms** | | Average RTT | **249.7 ms** | | Maximum RTT | **279.37 ms** | | Standard deviation | **9.37 ms** | | Stdev / average | **3.8%** | ## Route Context * Departure PoP: [Paris (PAR)](/docs/network/latency/par-paris) * Destination PoP: [Sydney (SYD)](/docs/network/latency/syd-sydney) * Region overview: [Europe](/docs/network/latency/regions/europe) → [Asia Pacific](/docs/network/latency/regions/asia-pacific) * Corridor overview: [Europe to Asia Pacific](/docs/network/latency/regions/europe-to-asia-pacific) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Sydney to Paris latency and RTT](/docs/network/latency/pairs/syd-par-rtt): 250.3 ms **Fastest routes departing Paris (PAR)** * [Paris to London latency and RTT](/docs/network/latency/pairs/par-lon-rtt): 6.4 ms * [Paris to Amsterdam latency and RTT](/docs/network/latency/pairs/par-ams-rtt): 7 ms * [Paris to Frankfurt latency and RTT](/docs/network/latency/pairs/par-fra-rtt): 7.6 ms * [Paris to Marseille latency and RTT](/docs/network/latency/pairs/par-mrs-rtt): 8.9 ms * [Paris to Berlin latency and RTT](/docs/network/latency/pairs/par-ber-rtt): 15.5 ms **Fastest routes arriving at Sydney (SYD)** * [Melbourne to Sydney latency and RTT](/docs/network/latency/pairs/mel-syd-rtt): 9.8 ms * [Singapore to Sydney latency and RTT](/docs/network/latency/pairs/sin-syd-rtt): 94.5 ms * [Tokyo to Sydney latency and RTT](/docs/network/latency/pairs/tyo-syd-rtt): 101.9 ms * [Taipei to Sydney latency and RTT](/docs/network/latency/pairs/tpe-syd-rtt): 133.2 ms * [Hong Kong to Sydney latency and RTT](/docs/network/latency/pairs/hkg-syd-rtt): 137.2 ms **Same corridor (Europe → Asia Pacific)** * [Moscow to Hong Kong latency and RTT](/docs/network/latency/pairs/mow-hkg-rtt): 118.6 ms * [Moscow to Taipei latency and RTT](/docs/network/latency/pairs/mow-tpe-rtt): 132.1 ms * [Marseille to Singapore latency and RTT](/docs/network/latency/pairs/mrs-sin-rtt): 139.6 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/par-syd.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Paris → Sydney round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/par-syd-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Paris, France → Taipei, Taiwan RTT 🇫🇷 **Paris, France (PAR)** → 🇹🇼 **Taipei, Taiwan (TPE)** Measurement round: `2026-08-16T04:07:28Z`. Paris to Taipei averaged 171.9 ms across 50 replies. The observed range was 163.74 to 192.89 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 7.09 ms, or 4.1% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------------------- | | RTT | **171.9 ms** | | Jitter | **5.24 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **96.42 ms** | | Fiber Efficiency | **56.1%** | | Latency Tier | Fair | | Source Region | [Europe](/docs/network/latency/regions/europe) | | Destination Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Paris](https://www.geonames.org/2988507) (48.85341, 2.34880) and [Taipei](https://www.geonames.org/1668341) (25.05306, 121.52639). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **9,846.2 km** | | Vacuum RTT floor | **65.69 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **96.42 ms** | | Low-latency fiber material floor | **96.03 ms** | | Engineering floor (5% path allowance) | **101.26 ms** | | Research 1.33× mapped-fiber reference | **128.24 ms** | | Estimated unamplified path loss | **2067.7 dB** | | Transparent optical spans / inline amplifiers | **130 / 129** | | Published RTT inflation over fiber floor | **1.78×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **163.74 ms** | | Average RTT | **171.9 ms** | | Maximum RTT | **192.89 ms** | | Standard deviation | **7.09 ms** | | Stdev / average | **4.1%** | ## Route Context * Departure PoP: [Paris (PAR)](/docs/network/latency/par-paris) * Destination PoP: [Taipei (TPE)](/docs/network/latency/tpe-taipei) * Region overview: [Europe](/docs/network/latency/regions/europe) → [Asia Pacific](/docs/network/latency/regions/asia-pacific) * Corridor overview: [Europe to Asia Pacific](/docs/network/latency/regions/europe-to-asia-pacific) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Taipei to Paris latency and RTT](/docs/network/latency/pairs/tpe-par-rtt): 173.8 ms **Fastest routes departing Paris (PAR)** * [Paris to London latency and RTT](/docs/network/latency/pairs/par-lon-rtt): 6.4 ms * [Paris to Amsterdam latency and RTT](/docs/network/latency/pairs/par-ams-rtt): 7 ms * [Paris to Frankfurt latency and RTT](/docs/network/latency/pairs/par-fra-rtt): 7.6 ms * [Paris to Marseille latency and RTT](/docs/network/latency/pairs/par-mrs-rtt): 8.9 ms * [Paris to Berlin latency and RTT](/docs/network/latency/pairs/par-ber-rtt): 15.5 ms **Fastest routes arriving at Taipei (TPE)** * [Hong Kong to Taipei latency and RTT](/docs/network/latency/pairs/hkg-tpe-rtt): 14.7 ms * [Tokyo to Taipei latency and RTT](/docs/network/latency/pairs/tyo-tpe-rtt): 32.3 ms * [Singapore to Taipei latency and RTT](/docs/network/latency/pairs/sin-tpe-rtt): 45.8 ms * [Seattle to Taipei latency and RTT](/docs/network/latency/pairs/sea-tpe-rtt): 114.7 ms * [Sydney to Taipei latency and RTT](/docs/network/latency/pairs/syd-tpe-rtt): 131.7 ms **Same corridor (Europe → Asia Pacific)** * [Moscow to Hong Kong latency and RTT](/docs/network/latency/pairs/mow-hkg-rtt): 118.6 ms * [Moscow to Taipei latency and RTT](/docs/network/latency/pairs/mow-tpe-rtt): 132.1 ms * [Marseille to Singapore latency and RTT](/docs/network/latency/pairs/mrs-sin-rtt): 139.6 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/par-tpe.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Paris → Taipei round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/par-tpe-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Paris, France → Tokyo, Japan RTT 🇫🇷 **Paris, France (PAR)** → 🇯🇵 **Tokyo, Japan (TYO)** Measurement round: `2026-08-16T04:07:28Z`. Paris to Tokyo averaged 204.7 ms, compared with a 95.34 ms city-centre fiber reference. The difference cannot be assigned to a particular cable or carrier from RTT alone. Standard deviation was 9.12 ms, or 4.5% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------------------- | | RTT | **204.7 ms** | | Jitter | **7.97 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **95.34 ms** | | Fiber Efficiency | **46.6%** | | Latency Tier | Fair | | Source Region | [Europe](/docs/network/latency/regions/europe) | | Destination Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Paris](https://www.geonames.org/2988507) (48.85341, 2.34880) and [Tokyo](https://www.geonames.org/1850147) (35.68950, 139.69171). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **9,736.1 km** | | Vacuum RTT floor | **64.95 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **95.34 ms** | | Low-latency fiber material floor | **94.96 ms** | | Engineering floor (5% path allowance) | **100.12 ms** | | Research 1.33× mapped-fiber reference | **126.81 ms** | | Estimated unamplified path loss | **2044.6 dB** | | Transparent optical spans / inline amplifiers | **128 / 127** | | Published RTT inflation over fiber floor | **2.15×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **194.79 ms** | | Average RTT | **204.7 ms** | | Maximum RTT | **241.63 ms** | | Standard deviation | **9.12 ms** | | Stdev / average | **4.5%** | ## Route Context * Departure PoP: [Paris (PAR)](/docs/network/latency/par-paris) * Destination PoP: [Tokyo (TYO)](/docs/network/latency/tyo-tokyo) * Region overview: [Europe](/docs/network/latency/regions/europe) → [Asia Pacific](/docs/network/latency/regions/asia-pacific) * Corridor overview: [Europe to Asia Pacific](/docs/network/latency/regions/europe-to-asia-pacific) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Tokyo to Paris latency and RTT](/docs/network/latency/pairs/tyo-par-rtt): 204.8 ms **Fastest routes departing Paris (PAR)** * [Paris to London latency and RTT](/docs/network/latency/pairs/par-lon-rtt): 6.4 ms * [Paris to Amsterdam latency and RTT](/docs/network/latency/pairs/par-ams-rtt): 7 ms * [Paris to Frankfurt latency and RTT](/docs/network/latency/pairs/par-fra-rtt): 7.6 ms * [Paris to Marseille latency and RTT](/docs/network/latency/pairs/par-mrs-rtt): 8.9 ms * [Paris to Berlin latency and RTT](/docs/network/latency/pairs/par-ber-rtt): 15.5 ms **Fastest routes arriving at Tokyo (TYO)** * [Taipei to Tokyo latency and RTT](/docs/network/latency/pairs/tpe-tyo-rtt): 31.9 ms * [Hong Kong to Tokyo latency and RTT](/docs/network/latency/pairs/hkg-tyo-rtt): 44.9 ms * [Singapore to Tokyo latency and RTT](/docs/network/latency/pairs/sin-tyo-rtt): 68.9 ms * [Seattle to Tokyo latency and RTT](/docs/network/latency/pairs/sea-tyo-rtt): 85 ms * [Los Angeles to Tokyo latency and RTT](/docs/network/latency/pairs/lax-tyo-rtt): 101.2 ms **Same corridor (Europe → Asia Pacific)** * [Moscow to Hong Kong latency and RTT](/docs/network/latency/pairs/mow-hkg-rtt): 118.6 ms * [Moscow to Taipei latency and RTT](/docs/network/latency/pairs/mow-tpe-rtt): 132.1 ms * [Marseille to Singapore latency and RTT](/docs/network/latency/pairs/mrs-sin-rtt): 139.6 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/par-tyo.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Paris → Tokyo round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/par-tyo-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Seattle, USA → Amsterdam, Netherlands RTT 🇺🇸 **Seattle, USA (SEA)** → 🇳🇱 **Amsterdam, Netherlands (AMS)** Measurement round: `2026-08-16T04:07:28Z`. Seattle to Amsterdam averaged 128.6 ms across 50 replies. The observed range was 126.47 to 138.71 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 2.06 ms, or 1.6% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **128.6 ms** | | Jitter | **1.28 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **76.86 ms** | | Fiber Efficiency | **59.8%** | | Latency Tier | Good | | Source Region | [North America](/docs/network/latency/regions/north-america) | | Destination Region | [Europe](/docs/network/latency/regions/europe) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Seattle](https://www.geonames.org/5809844) (47.60621, -122.33207) and [Amsterdam](https://www.geonames.org/2759794) (52.37403, 4.88969). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **7,848.5 km** | | Vacuum RTT floor | **52.36 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **76.86 ms** | | Low-latency fiber material floor | **76.55 ms** | | Engineering floor (5% path allowance) | **80.71 ms** | | Research 1.33× mapped-fiber reference | **102.22 ms** | | Estimated unamplified path loss | **1648.2 dB** | | Transparent optical spans / inline amplifiers | **104 / 103** | | Published RTT inflation over fiber floor | **1.67×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **126.47 ms** | | Average RTT | **128.6 ms** | | Maximum RTT | **138.71 ms** | | Standard deviation | **2.06 ms** | | Stdev / average | **1.6%** | ## Route Context * Departure PoP: [Seattle (SEA)](/docs/network/latency/sea-seattle) * Destination PoP: [Amsterdam (AMS)](/docs/network/latency/ams-amsterdam) * Region overview: [North America](/docs/network/latency/regions/north-america) → [Europe](/docs/network/latency/regions/europe) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Amsterdam to Seattle latency and RTT](/docs/network/latency/pairs/ams-sea-rtt): 130.2 ms **Fastest routes departing Seattle (SEA)** * [Seattle to Los Angeles latency and RTT](/docs/network/latency/pairs/sea-lax-rtt): 25.9 ms * [Seattle to New York latency and RTT](/docs/network/latency/pairs/sea-nyc-rtt): 59.5 ms * [Seattle to Ashburn latency and RTT](/docs/network/latency/pairs/sea-iad-rtt): 61.5 ms * [Seattle to Miami latency and RTT](/docs/network/latency/pairs/sea-mia-rtt): 81.7 ms * [Seattle to Tokyo latency and RTT](/docs/network/latency/pairs/sea-tyo-rtt): 85 ms **Fastest routes arriving at Amsterdam (AMS)** * [London to Amsterdam latency and RTT](/docs/network/latency/pairs/lon-ams-rtt): 5.2 ms * [Frankfurt to Amsterdam latency and RTT](/docs/network/latency/pairs/fra-ams-rtt): 5.9 ms * [Paris to Amsterdam latency and RTT](/docs/network/latency/pairs/par-ams-rtt): 7 ms * [Berlin to Amsterdam latency and RTT](/docs/network/latency/pairs/ber-ams-rtt): 7.2 ms * [Marseille to Amsterdam latency and RTT](/docs/network/latency/pairs/mrs-ams-rtt): 20.3 ms **Same corridor (North America → Europe)** * [New York to London latency and RTT](/docs/network/latency/pairs/nyc-lon-rtt): 63.5 ms * [New York to Amsterdam latency and RTT](/docs/network/latency/pairs/nyc-ams-rtt): 68.7 ms * [New York to Paris latency and RTT](/docs/network/latency/pairs/nyc-par-rtt): 68.9 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/sea-ams.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Seattle → Amsterdam round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/sea-ams-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Seattle, USA → Berlin, Germany RTT 🇺🇸 **Seattle, USA (SEA)** → 🇩🇪 **Berlin, Germany (BER)** Measurement round: `2026-08-16T04:07:28Z`. Seattle to Berlin averaged 140.9 ms across 50 replies. The observed range was 132.69 to 157.15 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 6.52 ms, or 4.6% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **140.9 ms** | | Jitter | **5.89 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **79.73 ms** | | Fiber Efficiency | **56.6%** | | Latency Tier | Good | | Source Region | [North America](/docs/network/latency/regions/north-america) | | Destination Region | [Europe](/docs/network/latency/regions/europe) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Seattle](https://www.geonames.org/5809844) (47.60621, -122.33207) and [Berlin](https://www.geonames.org/2950159) (52.52437, 13.41053). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **8,142.2 km** | | Vacuum RTT floor | **54.32 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **79.73 ms** | | Low-latency fiber material floor | **79.41 ms** | | Engineering floor (5% path allowance) | **83.73 ms** | | Research 1.33× mapped-fiber reference | **106.05 ms** | | Estimated unamplified path loss | **1709.9 dB** | | Transparent optical spans / inline amplifiers | **107 / 106** | | Published RTT inflation over fiber floor | **1.77×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **132.69 ms** | | Average RTT | **140.9 ms** | | Maximum RTT | **157.15 ms** | | Standard deviation | **6.52 ms** | | Stdev / average | **4.6%** | ## Route Context * Departure PoP: [Seattle (SEA)](/docs/network/latency/sea-seattle) * Destination PoP: [Berlin (BER)](/docs/network/latency/ber-berlin) * Region overview: [North America](/docs/network/latency/regions/north-america) → [Europe](/docs/network/latency/regions/europe) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Berlin to Seattle latency and RTT](/docs/network/latency/pairs/ber-sea-rtt): 139.7 ms **Fastest routes departing Seattle (SEA)** * [Seattle to Los Angeles latency and RTT](/docs/network/latency/pairs/sea-lax-rtt): 25.9 ms * [Seattle to New York latency and RTT](/docs/network/latency/pairs/sea-nyc-rtt): 59.5 ms * [Seattle to Ashburn latency and RTT](/docs/network/latency/pairs/sea-iad-rtt): 61.5 ms * [Seattle to Miami latency and RTT](/docs/network/latency/pairs/sea-mia-rtt): 81.7 ms * [Seattle to Tokyo latency and RTT](/docs/network/latency/pairs/sea-tyo-rtt): 85 ms **Fastest routes arriving at Berlin (BER)** * [Frankfurt to Berlin latency and RTT](/docs/network/latency/pairs/fra-ber-rtt): 6.1 ms * [Amsterdam to Berlin latency and RTT](/docs/network/latency/pairs/ams-ber-rtt): 7.2 ms * [Paris to Berlin latency and RTT](/docs/network/latency/pairs/par-ber-rtt): 15.5 ms * [London to Berlin latency and RTT](/docs/network/latency/pairs/lon-ber-rtt): 16.2 ms * [Marseille to Berlin latency and RTT](/docs/network/latency/pairs/mrs-ber-rtt): 20.6 ms **Same corridor (North America → Europe)** * [New York to London latency and RTT](/docs/network/latency/pairs/nyc-lon-rtt): 63.5 ms * [New York to Amsterdam latency and RTT](/docs/network/latency/pairs/nyc-ams-rtt): 68.7 ms * [New York to Paris latency and RTT](/docs/network/latency/pairs/nyc-par-rtt): 68.9 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/sea-ber.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Seattle → Berlin round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/sea-ber-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Seattle, USA → Frankfurt, Germany RTT 🇺🇸 **Seattle, USA (SEA)** → 🇩🇪 **Frankfurt, Germany (FRA)** Measurement round: `2026-08-16T04:07:28Z`. Seattle to Frankfurt averaged 133.2 ms across 50 replies. The observed range was 130.2 to 141.75 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 2.72 ms, or 2.0% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **133.2 ms** | | Jitter | **2.46 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **80.34 ms** | | Fiber Efficiency | **60.3%** | | Latency Tier | Good | | Source Region | [North America](/docs/network/latency/regions/north-america) | | Destination Region | [Europe](/docs/network/latency/regions/europe) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Seattle](https://www.geonames.org/5809844) (47.60621, -122.33207) and [Frankfurt am Main](https://www.geonames.org/2925533) (50.11552, 8.68417). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | ------------- | | WGS-84 geodesic distance | **8,204 km** | | Vacuum RTT floor | **54.73 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **80.34 ms** | | Low-latency fiber material floor | **80.02 ms** | | Engineering floor (5% path allowance) | **84.37 ms** | | Research 1.33× mapped-fiber reference | **106.85 ms** | | Estimated unamplified path loss | **1722.8 dB** | | Transparent optical spans / inline amplifiers | **108 / 107** | | Published RTT inflation over fiber floor | **1.66×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **130.2 ms** | | Average RTT | **133.2 ms** | | Maximum RTT | **141.75 ms** | | Standard deviation | **2.72 ms** | | Stdev / average | **2.0%** | ## Route Context * Departure PoP: [Seattle (SEA)](/docs/network/latency/sea-seattle) * Destination PoP: [Frankfurt (FRA)](/docs/network/latency/fra-frankfurt) * Region overview: [North America](/docs/network/latency/regions/north-america) → [Europe](/docs/network/latency/regions/europe) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Frankfurt to Seattle latency and RTT](/docs/network/latency/pairs/fra-sea-rtt): 135.2 ms **Fastest routes departing Seattle (SEA)** * [Seattle to Los Angeles latency and RTT](/docs/network/latency/pairs/sea-lax-rtt): 25.9 ms * [Seattle to New York latency and RTT](/docs/network/latency/pairs/sea-nyc-rtt): 59.5 ms * [Seattle to Ashburn latency and RTT](/docs/network/latency/pairs/sea-iad-rtt): 61.5 ms * [Seattle to Miami latency and RTT](/docs/network/latency/pairs/sea-mia-rtt): 81.7 ms * [Seattle to Tokyo latency and RTT](/docs/network/latency/pairs/sea-tyo-rtt): 85 ms **Fastest routes arriving at Frankfurt (FRA)** * [Amsterdam to Frankfurt latency and RTT](/docs/network/latency/pairs/ams-fra-rtt): 5.9 ms * [Berlin to Frankfurt latency and RTT](/docs/network/latency/pairs/ber-fra-rtt): 6.1 ms * [Paris to Frankfurt latency and RTT](/docs/network/latency/pairs/par-fra-rtt): 7.6 ms * [London to Frankfurt latency and RTT](/docs/network/latency/pairs/lon-fra-rtt): 13.6 ms * [Marseille to Frankfurt latency and RTT](/docs/network/latency/pairs/mrs-fra-rtt): 16 ms **Same corridor (North America → Europe)** * [New York to London latency and RTT](/docs/network/latency/pairs/nyc-lon-rtt): 63.5 ms * [New York to Amsterdam latency and RTT](/docs/network/latency/pairs/nyc-ams-rtt): 68.7 ms * [New York to Paris latency and RTT](/docs/network/latency/pairs/nyc-par-rtt): 68.9 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/sea-fra.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Seattle → Frankfurt round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/sea-fra-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Seattle, USA → São Paulo, Brazil RTT 🇺🇸 **Seattle, USA (SEA)** → 🇧🇷 **São Paulo, Brazil (GRU)** Measurement round: `2026-08-16T04:07:28Z`. Seattle to São Paulo averaged 155.7 ms across 50 replies. The observed range was 152.75 to 163 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 2.37 ms, or 1.5% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **155.7 ms** | | Jitter | **2.3 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **106.76 ms** | | Fiber Efficiency | **68.6%** | | Latency Tier | Fair | | Source Region | [North America](/docs/network/latency/regions/north-america) | | Destination Region | [South America](/docs/network/latency/regions/south-america) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Seattle](https://www.geonames.org/5809844) (47.60621, -122.33207) and [São Paulo](https://www.geonames.org/3448439) (-23.54750, -46.63611). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | --------------- | | WGS-84 geodesic distance | **10,902.3 km** | | Vacuum RTT floor | **72.73 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **106.76 ms** | | Low-latency fiber material floor | **106.33 ms** | | Engineering floor (5% path allowance) | **112.12 ms** | | Research 1.33× mapped-fiber reference | **142 ms** | | Estimated unamplified path loss | **2289.5 dB** | | Transparent optical spans / inline amplifiers | **144 / 143** | | Published RTT inflation over fiber floor | **1.46×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **152.75 ms** | | Average RTT | **155.7 ms** | | Maximum RTT | **163 ms** | | Standard deviation | **2.37 ms** | | Stdev / average | **1.5%** | ## Route Context * Departure PoP: [Seattle (SEA)](/docs/network/latency/sea-seattle) * Destination PoP: [São Paulo (GRU)](/docs/network/latency/gru-sao-paulo) * Region overview: [North America](/docs/network/latency/regions/north-america) → [South America](/docs/network/latency/regions/south-america) * Corridor overview: [North America to South America](/docs/network/latency/regions/north-america-to-south-america) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [São Paulo to Seattle latency and RTT](/docs/network/latency/pairs/gru-sea-rtt): 155.9 ms **Fastest routes departing Seattle (SEA)** * [Seattle to Los Angeles latency and RTT](/docs/network/latency/pairs/sea-lax-rtt): 25.9 ms * [Seattle to New York latency and RTT](/docs/network/latency/pairs/sea-nyc-rtt): 59.5 ms * [Seattle to Ashburn latency and RTT](/docs/network/latency/pairs/sea-iad-rtt): 61.5 ms * [Seattle to Miami latency and RTT](/docs/network/latency/pairs/sea-mia-rtt): 81.7 ms * [Seattle to Tokyo latency and RTT](/docs/network/latency/pairs/sea-tyo-rtt): 85 ms **Fastest routes arriving at São Paulo (GRU)** * [Ashburn to São Paulo latency and RTT](/docs/network/latency/pairs/iad-gru-rtt): 101.4 ms * [New York to São Paulo latency and RTT](/docs/network/latency/pairs/nyc-gru-rtt): 106.9 ms * [Miami to São Paulo latency and RTT](/docs/network/latency/pairs/mia-gru-rtt): 128.7 ms * [Los Angeles to São Paulo latency and RTT](/docs/network/latency/pairs/lax-gru-rtt): 131.3 ms * [London to São Paulo latency and RTT](/docs/network/latency/pairs/lon-gru-rtt): 170.7 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/sea-gru.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Seattle → São Paulo round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/sea-gru-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Seattle, USA → Hong Kong RTT 🇺🇸 **Seattle, USA (SEA)** → 🇭🇰 **Hong Kong (HKG)** Measurement round: `2026-08-16T04:07:28Z`. Seattle to Hong Kong averaged 130.4 ms across 50 replies. The observed range was 125.05 to 145.66 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 4.85 ms, or 3.7% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **130.4 ms** | | Jitter | **4.02 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **102.2 ms** | | Fiber Efficiency | **78.4%** | | Latency Tier | Good | | Source Region | [North America](/docs/network/latency/regions/north-america) | | Destination Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Seattle](https://www.geonames.org/5809844) (47.60621, -122.33207) and [Hong Kong](https://www.geonames.org/1819729) (22.27832, 114.17469). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | --------------- | | WGS-84 geodesic distance | **10,435.9 km** | | Vacuum RTT floor | **69.62 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **102.2 ms** | | Low-latency fiber material floor | **101.79 ms** | | Engineering floor (5% path allowance) | **107.32 ms** | | Research 1.33× mapped-fiber reference | **135.92 ms** | | Estimated unamplified path loss | **2191.5 dB** | | Transparent optical spans / inline amplifiers | **137 / 136** | | Published RTT inflation over fiber floor | **1.28×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **125.05 ms** | | Average RTT | **130.4 ms** | | Maximum RTT | **145.66 ms** | | Standard deviation | **4.85 ms** | | Stdev / average | **3.7%** | ## Route Context * Departure PoP: [Seattle (SEA)](/docs/network/latency/sea-seattle) * Destination PoP: [Hong Kong (HKG)](/docs/network/latency/hkg-hong-kong) * Region overview: [North America](/docs/network/latency/regions/north-america) → [Asia Pacific](/docs/network/latency/regions/asia-pacific) * Corridor overview: [North America to Asia Pacific](/docs/network/latency/regions/north-america-to-asia-pacific) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Hong Kong to Seattle latency and RTT](/docs/network/latency/pairs/hkg-sea-rtt): 131.6 ms **Fastest routes departing Seattle (SEA)** * [Seattle to Los Angeles latency and RTT](/docs/network/latency/pairs/sea-lax-rtt): 25.9 ms * [Seattle to New York latency and RTT](/docs/network/latency/pairs/sea-nyc-rtt): 59.5 ms * [Seattle to Ashburn latency and RTT](/docs/network/latency/pairs/sea-iad-rtt): 61.5 ms * [Seattle to Miami latency and RTT](/docs/network/latency/pairs/sea-mia-rtt): 81.7 ms * [Seattle to Tokyo latency and RTT](/docs/network/latency/pairs/sea-tyo-rtt): 85 ms **Fastest routes arriving at Hong Kong (HKG)** * [Taipei to Hong Kong latency and RTT](/docs/network/latency/pairs/tpe-hkg-rtt): 15.5 ms * [Singapore to Hong Kong latency and RTT](/docs/network/latency/pairs/sin-hkg-rtt): 30.8 ms * [Tokyo to Hong Kong latency and RTT](/docs/network/latency/pairs/tyo-hkg-rtt): 44.8 ms * [Moscow to Hong Kong latency and RTT](/docs/network/latency/pairs/mow-hkg-rtt): 118.6 ms * [Sydney to Hong Kong latency and RTT](/docs/network/latency/pairs/syd-hkg-rtt): 135.7 ms **Same corridor (North America → Asia Pacific)** * [Los Angeles to Tokyo latency and RTT](/docs/network/latency/pairs/lax-tyo-rtt): 101.2 ms * [Seattle to Taipei latency and RTT](/docs/network/latency/pairs/sea-tpe-rtt): 114.7 ms * [Los Angeles to Taipei latency and RTT](/docs/network/latency/pairs/lax-tpe-rtt): 132.9 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/sea-hkg.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Seattle → Hong Kong round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/sea-hkg-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Seattle, USA → Ashburn, USA RTT 🇺🇸 **Seattle, USA (SEA)** → 🇺🇸 **Ashburn, USA (IAD)** Measurement round: `2026-08-16T04:07:28Z`. Seattle to Ashburn averaged 61.5 ms across 50 replies. The observed range was 57.92 to 70.85 ms; these figures describe one probe round, not a long-term service target. This is number 3 of 19 outgoing routes by average RTT. That ranking compares different destinations and is not adjusted for distance. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **61.5 ms** | | Jitter | **2.57 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **36.27 ms** | | Fiber Efficiency | **59%** | | Latency Tier | Excellent | | Source Region | [North America](/docs/network/latency/regions/north-america) | | Destination Region | [North America](/docs/network/latency/regions/north-america) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Seattle](https://www.geonames.org/5809844) (47.60621, -122.33207) and [Ashburn](https://www.geonames.org/4744870) (39.04372, -77.48749). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | ------------ | | WGS-84 geodesic distance | **3,704 km** | | Vacuum RTT floor | **24.71 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **36.27 ms** | | Low-latency fiber material floor | **36.13 ms** | | Engineering floor (5% path allowance) | **38.09 ms** | | Research 1.33× mapped-fiber reference | **48.24 ms** | | Estimated unamplified path loss | **777.8 dB** | | Transparent optical spans / inline amplifiers | **49 / 48** | | Published RTT inflation over fiber floor | **1.7×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------ | | Minimum RTT | **57.92 ms** | | Average RTT | **61.5 ms** | | Maximum RTT | **70.85 ms** | | Standard deviation | **2.96 ms** | | Stdev / average | **4.8%** | ## Route Context * Departure PoP: [Seattle (SEA)](/docs/network/latency/sea-seattle) * Destination PoP: [Ashburn (IAD)](/docs/network/latency/iad-ashburn) * Region overview: [North America](/docs/network/latency/regions/north-america) → [North America](/docs/network/latency/regions/north-america) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Ashburn to Seattle latency and RTT](/docs/network/latency/pairs/iad-sea-rtt): 62.3 ms **Fastest routes departing Seattle (SEA)** * [Seattle to Los Angeles latency and RTT](/docs/network/latency/pairs/sea-lax-rtt): 25.9 ms * [Seattle to New York latency and RTT](/docs/network/latency/pairs/sea-nyc-rtt): 59.5 ms * [Seattle to Miami latency and RTT](/docs/network/latency/pairs/sea-mia-rtt): 81.7 ms * [Seattle to Tokyo latency and RTT](/docs/network/latency/pairs/sea-tyo-rtt): 85 ms * [Seattle to Taipei latency and RTT](/docs/network/latency/pairs/sea-tpe-rtt): 114.7 ms **Fastest routes arriving at Ashburn (IAD)** * [New York to Ashburn latency and RTT](/docs/network/latency/pairs/nyc-iad-rtt): 6 ms * [Miami to Ashburn latency and RTT](/docs/network/latency/pairs/mia-iad-rtt): 27.3 ms * [Los Angeles to Ashburn latency and RTT](/docs/network/latency/pairs/lax-iad-rtt): 60.4 ms * [London to Ashburn latency and RTT](/docs/network/latency/pairs/lon-iad-rtt): 71 ms * [Paris to Ashburn latency and RTT](/docs/network/latency/pairs/par-iad-rtt): 75.5 ms **Same corridor (North America → North America)** * [Ashburn to New York latency and RTT](/docs/network/latency/pairs/iad-nyc-rtt): 6 ms * [Los Angeles to Seattle latency and RTT](/docs/network/latency/pairs/lax-sea-rtt): 26.9 ms * [Ashburn to Miami latency and RTT](/docs/network/latency/pairs/iad-mia-rtt): 27.4 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/sea-iad.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Seattle → Ashburn round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/sea-iad-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Seattle, USA → Johannesburg, South Africa RTT 🇺🇸 **Seattle, USA (SEA)** → 🇿🇦 **Johannesburg, South Africa (JNB)** Measurement round: `2026-08-16T04:07:28Z`. Seattle to Johannesburg averaged 282.1 ms across 50 replies. The observed range was 273.42 to 300.55 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 6.9 ms, or 2.4% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **282.1 ms** | | Jitter | **5.55 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **161.61 ms** | | Fiber Efficiency | **57.3%** | | Latency Tier | High | | Source Region | [North America](/docs/network/latency/regions/north-america) | | Destination Region | [Africa](/docs/network/latency/regions/africa) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Seattle](https://www.geonames.org/5809844) (47.60621, -122.33207) and [Johannesburg](https://www.geonames.org/993800) (-26.20227, 28.04363). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | --------------- | | WGS-84 geodesic distance | **16,502.7 km** | | Vacuum RTT floor | **110.09 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **161.61 ms** | | Low-latency fiber material floor | **160.96 ms** | | Engineering floor (5% path allowance) | **169.71 ms** | | Research 1.33× mapped-fiber reference | **214.94 ms** | | Estimated unamplified path loss | **3465.6 dB** | | Transparent optical spans / inline amplifiers | **217 / 216** | | Published RTT inflation over fiber floor | **1.75×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **273.42 ms** | | Average RTT | **282.1 ms** | | Maximum RTT | **300.55 ms** | | Standard deviation | **6.9 ms** | | Stdev / average | **2.4%** | ## Route Context * Departure PoP: [Seattle (SEA)](/docs/network/latency/sea-seattle) * Destination PoP: [Johannesburg (JNB)](/docs/network/latency/jnb-johannesburg) * Region overview: [North America](/docs/network/latency/regions/north-america) → [Africa](/docs/network/latency/regions/africa) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Johannesburg to Seattle latency and RTT](/docs/network/latency/pairs/jnb-sea-rtt): 281.2 ms **Fastest routes departing Seattle (SEA)** * [Seattle to Los Angeles latency and RTT](/docs/network/latency/pairs/sea-lax-rtt): 25.9 ms * [Seattle to New York latency and RTT](/docs/network/latency/pairs/sea-nyc-rtt): 59.5 ms * [Seattle to Ashburn latency and RTT](/docs/network/latency/pairs/sea-iad-rtt): 61.5 ms * [Seattle to Miami latency and RTT](/docs/network/latency/pairs/sea-mia-rtt): 81.7 ms * [Seattle to Tokyo latency and RTT](/docs/network/latency/pairs/sea-tyo-rtt): 85 ms **Fastest routes arriving at Johannesburg (JNB)** * [Amsterdam to Johannesburg latency and RTT](/docs/network/latency/pairs/ams-jnb-rtt): 171.9 ms * [Berlin to Johannesburg latency and RTT](/docs/network/latency/pairs/ber-jnb-rtt): 172.1 ms * [London to Johannesburg latency and RTT](/docs/network/latency/pairs/lon-jnb-rtt): 179.6 ms * [Paris to Johannesburg latency and RTT](/docs/network/latency/pairs/par-jnb-rtt): 180.9 ms * [Frankfurt to Johannesburg latency and RTT](/docs/network/latency/pairs/fra-jnb-rtt): 188.4 ms **Same corridor (North America → Africa)** * [New York to Johannesburg latency and RTT](/docs/network/latency/pairs/nyc-jnb-rtt): 221.5 ms * [Ashburn to Johannesburg latency and RTT](/docs/network/latency/pairs/iad-jnb-rtt): 228.3 ms * [Miami to Johannesburg latency and RTT](/docs/network/latency/pairs/mia-jnb-rtt): 259.8 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/sea-jnb.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Seattle → Johannesburg round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/sea-jnb-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Seattle, USA → Los Angeles, USA RTT 🇺🇸 **Seattle, USA (SEA)** → 🇺🇸 **Los Angeles, USA (LAX)** Measurement round: `2026-08-16T04:07:28Z`. Seattle to Los Angeles averaged 25.9 ms across 50 replies. The observed range was 25.44 to 27.56 ms; these figures describe one probe round, not a long-term service target. This is number 1 of 19 outgoing routes by average RTT. That ranking compares different destinations and is not adjusted for distance. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **25.9 ms** | | Jitter | **0.33 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **15.12 ms** | | Fiber Efficiency | **58.4%** | | Latency Tier | Ultra-Low | | Source Region | [North America](/docs/network/latency/regions/north-america) | | Destination Region | [North America](/docs/network/latency/regions/north-america) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Seattle](https://www.geonames.org/5809844) (47.60621, -122.33207) and [Los Angeles](https://www.geonames.org/5368361) (34.05223, -118.24368). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **1,543.6 km** | | Vacuum RTT floor | **10.3 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **15.12 ms** | | Low-latency fiber material floor | **15.06 ms** | | Engineering floor (5% path allowance) | **15.87 ms** | | Research 1.33× mapped-fiber reference | **20.1 ms** | | Estimated unamplified path loss | **324.2 dB** | | Transparent optical spans / inline amplifiers | **21 / 20** | | Published RTT inflation over fiber floor | **1.71×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------ | | Minimum RTT | **25.44 ms** | | Average RTT | **25.9 ms** | | Maximum RTT | **27.56 ms** | | Standard deviation | **0.42 ms** | | Stdev / average | **1.6%** | ## Route Context * Departure PoP: [Seattle (SEA)](/docs/network/latency/sea-seattle) * Destination PoP: [Los Angeles (LAX)](/docs/network/latency/lax-los-angeles) * Region overview: [North America](/docs/network/latency/regions/north-america) → [North America](/docs/network/latency/regions/north-america) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Los Angeles to Seattle latency and RTT](/docs/network/latency/pairs/lax-sea-rtt): 26.9 ms **Fastest routes departing Seattle (SEA)** * [Seattle to New York latency and RTT](/docs/network/latency/pairs/sea-nyc-rtt): 59.5 ms * [Seattle to Ashburn latency and RTT](/docs/network/latency/pairs/sea-iad-rtt): 61.5 ms * [Seattle to Miami latency and RTT](/docs/network/latency/pairs/sea-mia-rtt): 81.7 ms * [Seattle to Tokyo latency and RTT](/docs/network/latency/pairs/sea-tyo-rtt): 85 ms * [Seattle to Taipei latency and RTT](/docs/network/latency/pairs/sea-tpe-rtt): 114.7 ms **Fastest routes arriving at Los Angeles (LAX)** * [Miami to Los Angeles latency and RTT](/docs/network/latency/pairs/mia-lax-rtt): 56.8 ms * [New York to Los Angeles latency and RTT](/docs/network/latency/pairs/nyc-lax-rtt): 58.9 ms * [Ashburn to Los Angeles latency and RTT](/docs/network/latency/pairs/iad-lax-rtt): 60.7 ms * [Tokyo to Los Angeles latency and RTT](/docs/network/latency/pairs/tyo-lax-rtt): 101.2 ms * [London to Los Angeles latency and RTT](/docs/network/latency/pairs/lon-lax-rtt): 127.5 ms **Same corridor (North America → North America)** * [Ashburn to New York latency and RTT](/docs/network/latency/pairs/iad-nyc-rtt): 6 ms * [New York to Ashburn latency and RTT](/docs/network/latency/pairs/nyc-iad-rtt): 6 ms * [Miami to Ashburn latency and RTT](/docs/network/latency/pairs/mia-iad-rtt): 27.3 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/sea-lax.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Seattle → Los Angeles round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/sea-lax-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Seattle, USA → London, UK RTT 🇺🇸 **Seattle, USA (SEA)** → 🇬🇧 **London, UK (LON)** Measurement round: `2026-08-16T04:07:28Z`. Seattle to London averaged 124.1 ms across 50 replies. The observed range was 118.94 to 137.2 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 3.96 ms, or 3.2% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **124.1 ms** | | Jitter | **3.75 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **75.63 ms** | | Fiber Efficiency | **60.9%** | | Latency Tier | Good | | Source Region | [North America](/docs/network/latency/regions/north-america) | | Destination Region | [Europe](/docs/network/latency/regions/europe) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Seattle](https://www.geonames.org/5809844) (47.60621, -122.33207) and [London](https://www.geonames.org/2643743) (51.50853, -0.12574). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **7,722.7 km** | | Vacuum RTT floor | **51.52 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **75.63 ms** | | Low-latency fiber material floor | **75.32 ms** | | Engineering floor (5% path allowance) | **79.42 ms** | | Research 1.33× mapped-fiber reference | **100.58 ms** | | Estimated unamplified path loss | **1621.8 dB** | | Transparent optical spans / inline amplifiers | **102 / 101** | | Published RTT inflation over fiber floor | **1.64×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **118.94 ms** | | Average RTT | **124.1 ms** | | Maximum RTT | **137.2 ms** | | Standard deviation | **3.96 ms** | | Stdev / average | **3.2%** | ## Route Context * Departure PoP: [Seattle (SEA)](/docs/network/latency/sea-seattle) * Destination PoP: [London (LON)](/docs/network/latency/lon-london) * Region overview: [North America](/docs/network/latency/regions/north-america) → [Europe](/docs/network/latency/regions/europe) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [London to Seattle latency and RTT](/docs/network/latency/pairs/lon-sea-rtt): 123.2 ms **Fastest routes departing Seattle (SEA)** * [Seattle to Los Angeles latency and RTT](/docs/network/latency/pairs/sea-lax-rtt): 25.9 ms * [Seattle to New York latency and RTT](/docs/network/latency/pairs/sea-nyc-rtt): 59.5 ms * [Seattle to Ashburn latency and RTT](/docs/network/latency/pairs/sea-iad-rtt): 61.5 ms * [Seattle to Miami latency and RTT](/docs/network/latency/pairs/sea-mia-rtt): 81.7 ms * [Seattle to Tokyo latency and RTT](/docs/network/latency/pairs/sea-tyo-rtt): 85 ms **Fastest routes arriving at London (LON)** * [Amsterdam to London latency and RTT](/docs/network/latency/pairs/ams-lon-rtt): 5.2 ms * [Paris to London latency and RTT](/docs/network/latency/pairs/par-lon-rtt): 6.4 ms * [Frankfurt to London latency and RTT](/docs/network/latency/pairs/fra-lon-rtt): 12.9 ms * [Berlin to London latency and RTT](/docs/network/latency/pairs/ber-lon-rtt): 15.2 ms * [Marseille to London latency and RTT](/docs/network/latency/pairs/mrs-lon-rtt): 18.9 ms **Same corridor (North America → Europe)** * [New York to London latency and RTT](/docs/network/latency/pairs/nyc-lon-rtt): 63.5 ms * [New York to Amsterdam latency and RTT](/docs/network/latency/pairs/nyc-ams-rtt): 68.7 ms * [New York to Paris latency and RTT](/docs/network/latency/pairs/nyc-par-rtt): 68.9 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/sea-lon.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Seattle → London round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/sea-lon-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Seattle, USA → Melbourne, Australia RTT 🇺🇸 **Seattle, USA (SEA)** → 🇦🇺 **Melbourne, Australia (MEL)** Measurement round: `2026-08-16T04:07:28Z`. Seattle to Melbourne averaged 170.1 ms across 50 replies. The observed range was 164.93 to 189.47 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 4.46 ms, or 2.6% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **170.1 ms** | | Jitter | **3.88 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **128.93 ms** | | Fiber Efficiency | **75.8%** | | Latency Tier | Fair | | Source Region | [North America](/docs/network/latency/regions/north-america) | | Destination Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Seattle](https://www.geonames.org/5809844) (47.60621, -122.33207) and [Melbourne](https://www.geonames.org/2158177) (-37.81400, 144.96332). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | ------------- | | WGS-84 geodesic distance | **13,166 km** | | Vacuum RTT floor | **87.83 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **128.93 ms** | | Low-latency fiber material floor | **128.41 ms** | | Engineering floor (5% path allowance) | **135.4 ms** | | Research 1.33× mapped-fiber reference | **171.48 ms** | | Estimated unamplified path loss | **2764.9 dB** | | Transparent optical spans / inline amplifiers | **173 / 172** | | Published RTT inflation over fiber floor | **1.32×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **164.93 ms** | | Average RTT | **170.1 ms** | | Maximum RTT | **189.47 ms** | | Standard deviation | **4.46 ms** | | Stdev / average | **2.6%** | ## Route Context * Departure PoP: [Seattle (SEA)](/docs/network/latency/sea-seattle) * Destination PoP: [Melbourne (MEL)](/docs/network/latency/mel-melbourne) * Region overview: [North America](/docs/network/latency/regions/north-america) → [Asia Pacific](/docs/network/latency/regions/asia-pacific) * Corridor overview: [North America to Asia Pacific](/docs/network/latency/regions/north-america-to-asia-pacific) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Melbourne to Seattle latency and RTT](/docs/network/latency/pairs/mel-sea-rtt): 172.1 ms **Fastest routes departing Seattle (SEA)** * [Seattle to Los Angeles latency and RTT](/docs/network/latency/pairs/sea-lax-rtt): 25.9 ms * [Seattle to New York latency and RTT](/docs/network/latency/pairs/sea-nyc-rtt): 59.5 ms * [Seattle to Ashburn latency and RTT](/docs/network/latency/pairs/sea-iad-rtt): 61.5 ms * [Seattle to Miami latency and RTT](/docs/network/latency/pairs/sea-mia-rtt): 81.7 ms * [Seattle to Tokyo latency and RTT](/docs/network/latency/pairs/sea-tyo-rtt): 85 ms **Fastest routes arriving at Melbourne (MEL)** * [Sydney to Melbourne latency and RTT](/docs/network/latency/pairs/syd-mel-rtt): 9.8 ms * [Singapore to Melbourne latency and RTT](/docs/network/latency/pairs/sin-mel-rtt): 88.6 ms * [Tokyo to Melbourne latency and RTT](/docs/network/latency/pairs/tyo-mel-rtt): 113.9 ms * [Hong Kong to Melbourne latency and RTT](/docs/network/latency/pairs/hkg-mel-rtt): 138.4 ms * [Taipei to Melbourne latency and RTT](/docs/network/latency/pairs/tpe-mel-rtt): 142 ms **Same corridor (North America → Asia Pacific)** * [Los Angeles to Tokyo latency and RTT](/docs/network/latency/pairs/lax-tyo-rtt): 101.2 ms * [Seattle to Taipei latency and RTT](/docs/network/latency/pairs/sea-tpe-rtt): 114.7 ms * [Seattle to Hong Kong latency and RTT](/docs/network/latency/pairs/sea-hkg-rtt): 130.4 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/sea-mel.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Seattle → Melbourne round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/sea-mel-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Seattle, USA → Miami, USA RTT 🇺🇸 **Seattle, USA (SEA)** → 🇺🇸 **Miami, USA (MIA)** Measurement round: `2026-08-16T04:07:28Z`. Seattle to Miami averaged 81.7 ms across 50 replies. The observed range was 79.89 to 89.57 ms; these figures describe one probe round, not a long-term service target. This is number 4 of 19 outgoing routes by average RTT. That ranking compares different destinations and is not adjusted for distance. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **81.7 ms** | | Jitter | **1.22 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **43.08 ms** | | Fiber Efficiency | **52.7%** | | Latency Tier | Good | | Source Region | [North America](/docs/network/latency/regions/north-america) | | Destination Region | [North America](/docs/network/latency/regions/north-america) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Seattle](https://www.geonames.org/5809844) (47.60621, -122.33207) and [Miami](https://www.geonames.org/4164138) (25.77427, -80.19366). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **4,399.4 km** | | Vacuum RTT floor | **29.35 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **43.08 ms** | | Low-latency fiber material floor | **42.91 ms** | | Engineering floor (5% path allowance) | **45.24 ms** | | Research 1.33× mapped-fiber reference | **57.3 ms** | | Estimated unamplified path loss | **923.9 dB** | | Transparent optical spans / inline amplifiers | **58 / 57** | | Published RTT inflation over fiber floor | **1.9×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------ | | Minimum RTT | **79.89 ms** | | Average RTT | **81.7 ms** | | Maximum RTT | **89.57 ms** | | Standard deviation | **1.94 ms** | | Stdev / average | **2.4%** | ## Route Context * Departure PoP: [Seattle (SEA)](/docs/network/latency/sea-seattle) * Destination PoP: [Miami (MIA)](/docs/network/latency/mia-miami) * Region overview: [North America](/docs/network/latency/regions/north-america) → [North America](/docs/network/latency/regions/north-america) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Miami to Seattle latency and RTT](/docs/network/latency/pairs/mia-sea-rtt): 81.9 ms **Fastest routes departing Seattle (SEA)** * [Seattle to Los Angeles latency and RTT](/docs/network/latency/pairs/sea-lax-rtt): 25.9 ms * [Seattle to New York latency and RTT](/docs/network/latency/pairs/sea-nyc-rtt): 59.5 ms * [Seattle to Ashburn latency and RTT](/docs/network/latency/pairs/sea-iad-rtt): 61.5 ms * [Seattle to Tokyo latency and RTT](/docs/network/latency/pairs/sea-tyo-rtt): 85 ms * [Seattle to Taipei latency and RTT](/docs/network/latency/pairs/sea-tpe-rtt): 114.7 ms **Fastest routes arriving at Miami (MIA)** * [Ashburn to Miami latency and RTT](/docs/network/latency/pairs/iad-mia-rtt): 27.4 ms * [New York to Miami latency and RTT](/docs/network/latency/pairs/nyc-mia-rtt): 32.9 ms * [Los Angeles to Miami latency and RTT](/docs/network/latency/pairs/lax-mia-rtt): 57.3 ms * [São Paulo to Miami latency and RTT](/docs/network/latency/pairs/gru-mia-rtt): 74 ms * [London to Miami latency and RTT](/docs/network/latency/pairs/lon-mia-rtt): 101.9 ms **Same corridor (North America → North America)** * [Ashburn to New York latency and RTT](/docs/network/latency/pairs/iad-nyc-rtt): 6 ms * [New York to Ashburn latency and RTT](/docs/network/latency/pairs/nyc-iad-rtt): 6 ms * [Los Angeles to Seattle latency and RTT](/docs/network/latency/pairs/lax-sea-rtt): 26.9 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/sea-mia.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Seattle → Miami round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/sea-mia-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Seattle, USA → Moscow, Russia RTT 🇺🇸 **Seattle, USA (SEA)** → 🇷🇺 **Moscow, Russia (MOW)** Measurement round: `2026-08-16T04:07:28Z`. Seattle to Moscow averaged 164.6 ms, compared with a 82.23 ms city-centre fiber reference. The difference cannot be assigned to a particular cable or carrier from RTT alone. Standard deviation was 7.23 ms, or 4.4% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **164.6 ms** | | Jitter | **5 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **82.23 ms** | | Fiber Efficiency | **50%** | | Latency Tier | Fair | | Source Region | [North America](/docs/network/latency/regions/north-america) | | Destination Region | [Europe](/docs/network/latency/regions/europe) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Seattle](https://www.geonames.org/5809844) (47.60621, -122.33207) and [Moscow](https://www.geonames.org/524901) (55.75204, 37.61781). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **8,396.9 km** | | Vacuum RTT floor | **56.02 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **82.23 ms** | | Low-latency fiber material floor | **81.9 ms** | | Engineering floor (5% path allowance) | **86.35 ms** | | Research 1.33× mapped-fiber reference | **109.36 ms** | | Estimated unamplified path loss | **1763.3 dB** | | Transparent optical spans / inline amplifiers | **111 / 110** | | Published RTT inflation over fiber floor | **2×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **158.38 ms** | | Average RTT | **164.6 ms** | | Maximum RTT | **192.7 ms** | | Standard deviation | **7.23 ms** | | Stdev / average | **4.4%** | ## Route Context * Departure PoP: [Seattle (SEA)](/docs/network/latency/sea-seattle) * Destination PoP: [Moscow (MOW)](/docs/network/latency/mow-moscow) * Region overview: [North America](/docs/network/latency/regions/north-america) → [Europe](/docs/network/latency/regions/europe) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Moscow to Seattle latency and RTT](/docs/network/latency/pairs/mow-sea-rtt): 164.8 ms **Fastest routes departing Seattle (SEA)** * [Seattle to Los Angeles latency and RTT](/docs/network/latency/pairs/sea-lax-rtt): 25.9 ms * [Seattle to New York latency and RTT](/docs/network/latency/pairs/sea-nyc-rtt): 59.5 ms * [Seattle to Ashburn latency and RTT](/docs/network/latency/pairs/sea-iad-rtt): 61.5 ms * [Seattle to Miami latency and RTT](/docs/network/latency/pairs/sea-mia-rtt): 81.7 ms * [Seattle to Tokyo latency and RTT](/docs/network/latency/pairs/sea-tyo-rtt): 85 ms **Fastest routes arriving at Moscow (MOW)** * [Berlin to Moscow latency and RTT](/docs/network/latency/pairs/ber-mow-rtt): 28.6 ms * [Frankfurt to Moscow latency and RTT](/docs/network/latency/pairs/fra-mow-rtt): 34.7 ms * [Amsterdam to Moscow latency and RTT](/docs/network/latency/pairs/ams-mow-rtt): 38.1 ms * [London to Moscow latency and RTT](/docs/network/latency/pairs/lon-mow-rtt): 43.5 ms * [Paris to Moscow latency and RTT](/docs/network/latency/pairs/par-mow-rtt): 44.7 ms **Same corridor (North America → Europe)** * [New York to London latency and RTT](/docs/network/latency/pairs/nyc-lon-rtt): 63.5 ms * [New York to Amsterdam latency and RTT](/docs/network/latency/pairs/nyc-ams-rtt): 68.7 ms * [New York to Paris latency and RTT](/docs/network/latency/pairs/nyc-par-rtt): 68.9 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/sea-mow.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Seattle → Moscow round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/sea-mow-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Seattle, USA → Marseille, France RTT 🇺🇸 **Seattle, USA (SEA)** → 🇫🇷 **Marseille, France (MRS)** Measurement round: `2026-08-16T04:07:28Z`. Seattle to Marseille averaged 141.2 ms across 50 replies. The observed range was 134.28 to 166.5 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 6.79 ms, or 4.8% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **141.2 ms** | | Jitter | **5.32 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **85.27 ms** | | Fiber Efficiency | **60.4%** | | Latency Tier | Good | | Source Region | [North America](/docs/network/latency/regions/north-america) | | Destination Region | [Europe](/docs/network/latency/regions/europe) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Seattle](https://www.geonames.org/5809844) (47.60621, -122.33207) and [Marseille](https://www.geonames.org/2995469) (43.29695, 5.38107). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **8,707.5 km** | | Vacuum RTT floor | **58.09 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **85.27 ms** | | Low-latency fiber material floor | **84.93 ms** | | Engineering floor (5% path allowance) | **89.55 ms** | | Research 1.33× mapped-fiber reference | **113.41 ms** | | Estimated unamplified path loss | **1828.6 dB** | | Transparent optical spans / inline amplifiers | **115 / 114** | | Published RTT inflation over fiber floor | **1.66×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **134.28 ms** | | Average RTT | **141.2 ms** | | Maximum RTT | **166.5 ms** | | Standard deviation | **6.79 ms** | | Stdev / average | **4.8%** | ## Route Context * Departure PoP: [Seattle (SEA)](/docs/network/latency/sea-seattle) * Destination PoP: [Marseille (MRS)](/docs/network/latency/mrs-marseille) * Region overview: [North America](/docs/network/latency/regions/north-america) → [Europe](/docs/network/latency/regions/europe) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Marseille to Seattle latency and RTT](/docs/network/latency/pairs/mrs-sea-rtt): 141.7 ms **Fastest routes departing Seattle (SEA)** * [Seattle to Los Angeles latency and RTT](/docs/network/latency/pairs/sea-lax-rtt): 25.9 ms * [Seattle to New York latency and RTT](/docs/network/latency/pairs/sea-nyc-rtt): 59.5 ms * [Seattle to Ashburn latency and RTT](/docs/network/latency/pairs/sea-iad-rtt): 61.5 ms * [Seattle to Miami latency and RTT](/docs/network/latency/pairs/sea-mia-rtt): 81.7 ms * [Seattle to Tokyo latency and RTT](/docs/network/latency/pairs/sea-tyo-rtt): 85 ms **Fastest routes arriving at Marseille (MRS)** * [Paris to Marseille latency and RTT](/docs/network/latency/pairs/par-mrs-rtt): 8.9 ms * [Frankfurt to Marseille latency and RTT](/docs/network/latency/pairs/fra-mrs-rtt): 16.4 ms * [London to Marseille latency and RTT](/docs/network/latency/pairs/lon-mrs-rtt): 17.9 ms * [Amsterdam to Marseille latency and RTT](/docs/network/latency/pairs/ams-mrs-rtt): 19.6 ms * [Berlin to Marseille latency and RTT](/docs/network/latency/pairs/ber-mrs-rtt): 20.2 ms **Same corridor (North America → Europe)** * [New York to London latency and RTT](/docs/network/latency/pairs/nyc-lon-rtt): 63.5 ms * [New York to Amsterdam latency and RTT](/docs/network/latency/pairs/nyc-ams-rtt): 68.7 ms * [New York to Paris latency and RTT](/docs/network/latency/pairs/nyc-par-rtt): 68.9 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/sea-mrs.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Seattle → Marseille round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/sea-mrs-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Seattle, USA → New York, USA RTT 🇺🇸 **Seattle, USA (SEA)** → 🇺🇸 **New York, USA (NYC)** Measurement round: `2026-08-16T04:07:28Z`. Seattle to New York averaged 59.5 ms across 50 replies. The observed range was 57.58 to 64.42 ms; these figures describe one probe round, not a long-term service target. This is number 2 of 19 outgoing routes by average RTT. That ranking compares different destinations and is not adjusted for distance. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **59.5 ms** | | Jitter | **1.21 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **37.95 ms** | | Fiber Efficiency | **63.8%** | | Latency Tier | Excellent | | Source Region | [North America](/docs/network/latency/regions/north-america) | | Destination Region | [North America](/docs/network/latency/regions/north-america) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Seattle](https://www.geonames.org/5809844) (47.60621, -122.33207) and [New York City](https://www.geonames.org/5128581) (40.71427, -74.00597). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **3,875.5 km** | | Vacuum RTT floor | **25.85 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **37.95 ms** | | Low-latency fiber material floor | **37.8 ms** | | Engineering floor (5% path allowance) | **39.85 ms** | | Research 1.33× mapped-fiber reference | **50.48 ms** | | Estimated unamplified path loss | **813.9 dB** | | Transparent optical spans / inline amplifiers | **51 / 50** | | Published RTT inflation over fiber floor | **1.57×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------ | | Minimum RTT | **57.58 ms** | | Average RTT | **59.5 ms** | | Maximum RTT | **64.42 ms** | | Standard deviation | **1.48 ms** | | Stdev / average | **2.5%** | ## Route Context * Departure PoP: [Seattle (SEA)](/docs/network/latency/sea-seattle) * Destination PoP: [New York (NYC)](/docs/network/latency/nyc-new-york) * Region overview: [North America](/docs/network/latency/regions/north-america) → [North America](/docs/network/latency/regions/north-america) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [New York to Seattle latency and RTT](/docs/network/latency/pairs/nyc-sea-rtt): 58.7 ms **Fastest routes departing Seattle (SEA)** * [Seattle to Los Angeles latency and RTT](/docs/network/latency/pairs/sea-lax-rtt): 25.9 ms * [Seattle to Ashburn latency and RTT](/docs/network/latency/pairs/sea-iad-rtt): 61.5 ms * [Seattle to Miami latency and RTT](/docs/network/latency/pairs/sea-mia-rtt): 81.7 ms * [Seattle to Tokyo latency and RTT](/docs/network/latency/pairs/sea-tyo-rtt): 85 ms * [Seattle to Taipei latency and RTT](/docs/network/latency/pairs/sea-tpe-rtt): 114.7 ms **Fastest routes arriving at New York (NYC)** * [Ashburn to New York latency and RTT](/docs/network/latency/pairs/iad-nyc-rtt): 6 ms * [Miami to New York latency and RTT](/docs/network/latency/pairs/mia-nyc-rtt): 33.3 ms * [Los Angeles to New York latency and RTT](/docs/network/latency/pairs/lax-nyc-rtt): 57.9 ms * [London to New York latency and RTT](/docs/network/latency/pairs/lon-nyc-rtt): 63.8 ms * [Amsterdam to New York latency and RTT](/docs/network/latency/pairs/ams-nyc-rtt): 68.6 ms **Same corridor (North America → North America)** * [New York to Ashburn latency and RTT](/docs/network/latency/pairs/nyc-iad-rtt): 6 ms * [Los Angeles to Seattle latency and RTT](/docs/network/latency/pairs/lax-sea-rtt): 26.9 ms * [Miami to Ashburn latency and RTT](/docs/network/latency/pairs/mia-iad-rtt): 27.3 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/sea-nyc.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Seattle → New York round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/sea-nyc-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Seattle, USA → Paris, France RTT 🇺🇸 **Seattle, USA (SEA)** → 🇫🇷 **Paris, France (PAR)** Measurement round: `2026-08-16T04:07:28Z`. Seattle to Paris averaged 129.2 ms across 50 replies. The observed range was 126.7 to 135.02 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 2.13 ms, or 1.6% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **129.2 ms** | | Jitter | **1.78 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **78.98 ms** | | Fiber Efficiency | **61.1%** | | Latency Tier | Good | | Source Region | [North America](/docs/network/latency/regions/north-america) | | Destination Region | [Europe](/docs/network/latency/regions/europe) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Seattle](https://www.geonames.org/5809844) (47.60621, -122.33207) and [Paris](https://www.geonames.org/2988507) (48.85341, 2.34880). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **8,065.2 km** | | Vacuum RTT floor | **53.8 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **78.98 ms** | | Low-latency fiber material floor | **78.66 ms** | | Engineering floor (5% path allowance) | **82.94 ms** | | Research 1.33× mapped-fiber reference | **105.04 ms** | | Estimated unamplified path loss | **1693.7 dB** | | Transparent optical spans / inline amplifiers | **106 / 105** | | Published RTT inflation over fiber floor | **1.64×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **126.7 ms** | | Average RTT | **129.2 ms** | | Maximum RTT | **135.02 ms** | | Standard deviation | **2.13 ms** | | Stdev / average | **1.6%** | ## Route Context * Departure PoP: [Seattle (SEA)](/docs/network/latency/sea-seattle) * Destination PoP: [Paris (PAR)](/docs/network/latency/par-paris) * Region overview: [North America](/docs/network/latency/regions/north-america) → [Europe](/docs/network/latency/regions/europe) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Paris to Seattle latency and RTT](/docs/network/latency/pairs/par-sea-rtt): 128 ms **Fastest routes departing Seattle (SEA)** * [Seattle to Los Angeles latency and RTT](/docs/network/latency/pairs/sea-lax-rtt): 25.9 ms * [Seattle to New York latency and RTT](/docs/network/latency/pairs/sea-nyc-rtt): 59.5 ms * [Seattle to Ashburn latency and RTT](/docs/network/latency/pairs/sea-iad-rtt): 61.5 ms * [Seattle to Miami latency and RTT](/docs/network/latency/pairs/sea-mia-rtt): 81.7 ms * [Seattle to Tokyo latency and RTT](/docs/network/latency/pairs/sea-tyo-rtt): 85 ms **Fastest routes arriving at Paris (PAR)** * [London to Paris latency and RTT](/docs/network/latency/pairs/lon-par-rtt): 6.4 ms * [Amsterdam to Paris latency and RTT](/docs/network/latency/pairs/ams-par-rtt): 7 ms * [Frankfurt to Paris latency and RTT](/docs/network/latency/pairs/fra-par-rtt): 7.6 ms * [Marseille to Paris latency and RTT](/docs/network/latency/pairs/mrs-par-rtt): 8.9 ms * [Berlin to Paris latency and RTT](/docs/network/latency/pairs/ber-par-rtt): 15.6 ms **Same corridor (North America → Europe)** * [New York to London latency and RTT](/docs/network/latency/pairs/nyc-lon-rtt): 63.5 ms * [New York to Amsterdam latency and RTT](/docs/network/latency/pairs/nyc-ams-rtt): 68.7 ms * [New York to Paris latency and RTT](/docs/network/latency/pairs/nyc-par-rtt): 68.9 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/sea-par.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Seattle → Paris round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/sea-par-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Seattle, USA → Singapore RTT 🇺🇸 **Seattle, USA (SEA)** → 🇸🇬 **Singapore (SIN)** Measurement round: `2026-08-16T04:07:28Z`. Seattle to Singapore averaged 151.1 ms across 50 replies. The observed range was 143.38 to 172.46 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 6.94 ms, or 4.6% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **151.1 ms** | | Jitter | **6.35 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **127.22 ms** | | Fiber Efficiency | **84.2%** | | Latency Tier | Fair | | Source Region | [North America](/docs/network/latency/regions/north-america) | | Destination Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Seattle](https://www.geonames.org/5809844) (47.60621, -122.33207) and [Singapore](https://www.geonames.org/1880252) (1.28967, 103.85007). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | --------------- | | WGS-84 geodesic distance | **12,991.1 km** | | Vacuum RTT floor | **86.67 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **127.22 ms** | | Low-latency fiber material floor | **126.71 ms** | | Engineering floor (5% path allowance) | **133.6 ms** | | Research 1.33× mapped-fiber reference | **169.2 ms** | | Estimated unamplified path loss | **2728.1 dB** | | Transparent optical spans / inline amplifiers | **171 / 170** | | Published RTT inflation over fiber floor | **1.19×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **143.38 ms** | | Average RTT | **151.1 ms** | | Maximum RTT | **172.46 ms** | | Standard deviation | **6.94 ms** | | Stdev / average | **4.6%** | ## Route Context * Departure PoP: [Seattle (SEA)](/docs/network/latency/sea-seattle) * Destination PoP: [Singapore (SIN)](/docs/network/latency/sin-singapore) * Region overview: [North America](/docs/network/latency/regions/north-america) → [Asia Pacific](/docs/network/latency/regions/asia-pacific) * Corridor overview: [North America to Asia Pacific](/docs/network/latency/regions/north-america-to-asia-pacific) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Singapore to Seattle latency and RTT](/docs/network/latency/pairs/sin-sea-rtt): 151.5 ms **Fastest routes departing Seattle (SEA)** * [Seattle to Los Angeles latency and RTT](/docs/network/latency/pairs/sea-lax-rtt): 25.9 ms * [Seattle to New York latency and RTT](/docs/network/latency/pairs/sea-nyc-rtt): 59.5 ms * [Seattle to Ashburn latency and RTT](/docs/network/latency/pairs/sea-iad-rtt): 61.5 ms * [Seattle to Miami latency and RTT](/docs/network/latency/pairs/sea-mia-rtt): 81.7 ms * [Seattle to Tokyo latency and RTT](/docs/network/latency/pairs/sea-tyo-rtt): 85 ms **Fastest routes arriving at Singapore (SIN)** * [Hong Kong to Singapore latency and RTT](/docs/network/latency/pairs/hkg-sin-rtt): 31.5 ms * [Taipei to Singapore latency and RTT](/docs/network/latency/pairs/tpe-sin-rtt): 44.3 ms * [Tokyo to Singapore latency and RTT](/docs/network/latency/pairs/tyo-sin-rtt): 69.4 ms * [Melbourne to Singapore latency and RTT](/docs/network/latency/pairs/mel-sin-rtt): 88.4 ms * [Sydney to Singapore latency and RTT](/docs/network/latency/pairs/syd-sin-rtt): 94.5 ms **Same corridor (North America → Asia Pacific)** * [Los Angeles to Tokyo latency and RTT](/docs/network/latency/pairs/lax-tyo-rtt): 101.2 ms * [Seattle to Taipei latency and RTT](/docs/network/latency/pairs/sea-tpe-rtt): 114.7 ms * [Seattle to Hong Kong latency and RTT](/docs/network/latency/pairs/sea-hkg-rtt): 130.4 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/sea-sin.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Seattle → Singapore round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/sea-sin-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Seattle, USA → Sydney, Australia RTT 🇺🇸 **Seattle, USA (SEA)** → 🇦🇺 **Sydney, Australia (SYD)** Measurement round: `2026-08-16T04:07:28Z`. Seattle to Sydney averaged 161.3 ms across 50 replies. The observed range was 157.6 to 177.48 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 3.32 ms, or 2.1% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **161.3 ms** | | Jitter | **2.66 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **121.96 ms** | | Fiber Efficiency | **75.6%** | | Latency Tier | Fair | | Source Region | [North America](/docs/network/latency/regions/north-america) | | Destination Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Seattle](https://www.geonames.org/5809844) (47.60621, -122.33207) and [Sydney](https://www.geonames.org/2147714) (-33.86785, 151.20732). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | --------------- | | WGS-84 geodesic distance | **12,453.7 km** | | Vacuum RTT floor | **83.08 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **121.96 ms** | | Low-latency fiber material floor | **121.47 ms** | | Engineering floor (5% path allowance) | **128.07 ms** | | Research 1.33× mapped-fiber reference | **162.2 ms** | | Estimated unamplified path loss | **2615.3 dB** | | Transparent optical spans / inline amplifiers | **164 / 163** | | Published RTT inflation over fiber floor | **1.32×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **157.6 ms** | | Average RTT | **161.3 ms** | | Maximum RTT | **177.48 ms** | | Standard deviation | **3.32 ms** | | Stdev / average | **2.1%** | ## Route Context * Departure PoP: [Seattle (SEA)](/docs/network/latency/sea-seattle) * Destination PoP: [Sydney (SYD)](/docs/network/latency/syd-sydney) * Region overview: [North America](/docs/network/latency/regions/north-america) → [Asia Pacific](/docs/network/latency/regions/asia-pacific) * Corridor overview: [North America to Asia Pacific](/docs/network/latency/regions/north-america-to-asia-pacific) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Sydney to Seattle latency and RTT](/docs/network/latency/pairs/syd-sea-rtt): 161.9 ms **Fastest routes departing Seattle (SEA)** * [Seattle to Los Angeles latency and RTT](/docs/network/latency/pairs/sea-lax-rtt): 25.9 ms * [Seattle to New York latency and RTT](/docs/network/latency/pairs/sea-nyc-rtt): 59.5 ms * [Seattle to Ashburn latency and RTT](/docs/network/latency/pairs/sea-iad-rtt): 61.5 ms * [Seattle to Miami latency and RTT](/docs/network/latency/pairs/sea-mia-rtt): 81.7 ms * [Seattle to Tokyo latency and RTT](/docs/network/latency/pairs/sea-tyo-rtt): 85 ms **Fastest routes arriving at Sydney (SYD)** * [Melbourne to Sydney latency and RTT](/docs/network/latency/pairs/mel-syd-rtt): 9.8 ms * [Singapore to Sydney latency and RTT](/docs/network/latency/pairs/sin-syd-rtt): 94.5 ms * [Tokyo to Sydney latency and RTT](/docs/network/latency/pairs/tyo-syd-rtt): 101.9 ms * [Taipei to Sydney latency and RTT](/docs/network/latency/pairs/tpe-syd-rtt): 133.2 ms * [Hong Kong to Sydney latency and RTT](/docs/network/latency/pairs/hkg-syd-rtt): 137.2 ms **Same corridor (North America → Asia Pacific)** * [Los Angeles to Tokyo latency and RTT](/docs/network/latency/pairs/lax-tyo-rtt): 101.2 ms * [Seattle to Taipei latency and RTT](/docs/network/latency/pairs/sea-tpe-rtt): 114.7 ms * [Seattle to Hong Kong latency and RTT](/docs/network/latency/pairs/sea-hkg-rtt): 130.4 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/sea-syd.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Seattle → Sydney round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/sea-syd-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Seattle, USA → Taipei, Taiwan RTT 🇺🇸 **Seattle, USA (SEA)** → 🇹🇼 **Taipei, Taiwan (TPE)** Measurement round: `2026-08-16T04:07:28Z`. Seattle to Taipei averaged 114.7 ms across 50 replies. The observed range was 109.75 to 137.83 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 5.07 ms, or 4.4% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **114.7 ms** | | Jitter | **4.49 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **95.46 ms** | | Fiber Efficiency | **83.2%** | | Latency Tier | Good | | Source Region | [North America](/docs/network/latency/regions/north-america) | | Destination Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Seattle](https://www.geonames.org/5809844) (47.60621, -122.33207) and [Taipei](https://www.geonames.org/1668341) (25.05306, 121.52639). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **9,748.1 km** | | Vacuum RTT floor | **65.03 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **95.46 ms** | | Low-latency fiber material floor | **95.08 ms** | | Engineering floor (5% path allowance) | **100.25 ms** | | Research 1.33× mapped-fiber reference | **126.96 ms** | | Estimated unamplified path loss | **2047.1 dB** | | Transparent optical spans / inline amplifiers | **128 / 127** | | Published RTT inflation over fiber floor | **1.2×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **109.75 ms** | | Average RTT | **114.7 ms** | | Maximum RTT | **137.83 ms** | | Standard deviation | **5.07 ms** | | Stdev / average | **4.4%** | ## Route Context * Departure PoP: [Seattle (SEA)](/docs/network/latency/sea-seattle) * Destination PoP: [Taipei (TPE)](/docs/network/latency/tpe-taipei) * Region overview: [North America](/docs/network/latency/regions/north-america) → [Asia Pacific](/docs/network/latency/regions/asia-pacific) * Corridor overview: [North America to Asia Pacific](/docs/network/latency/regions/north-america-to-asia-pacific) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Taipei to Seattle latency and RTT](/docs/network/latency/pairs/tpe-sea-rtt): 115.8 ms **Fastest routes departing Seattle (SEA)** * [Seattle to Los Angeles latency and RTT](/docs/network/latency/pairs/sea-lax-rtt): 25.9 ms * [Seattle to New York latency and RTT](/docs/network/latency/pairs/sea-nyc-rtt): 59.5 ms * [Seattle to Ashburn latency and RTT](/docs/network/latency/pairs/sea-iad-rtt): 61.5 ms * [Seattle to Miami latency and RTT](/docs/network/latency/pairs/sea-mia-rtt): 81.7 ms * [Seattle to Tokyo latency and RTT](/docs/network/latency/pairs/sea-tyo-rtt): 85 ms **Fastest routes arriving at Taipei (TPE)** * [Hong Kong to Taipei latency and RTT](/docs/network/latency/pairs/hkg-tpe-rtt): 14.7 ms * [Tokyo to Taipei latency and RTT](/docs/network/latency/pairs/tyo-tpe-rtt): 32.3 ms * [Singapore to Taipei latency and RTT](/docs/network/latency/pairs/sin-tpe-rtt): 45.8 ms * [Sydney to Taipei latency and RTT](/docs/network/latency/pairs/syd-tpe-rtt): 131.7 ms * [Moscow to Taipei latency and RTT](/docs/network/latency/pairs/mow-tpe-rtt): 132.1 ms **Same corridor (North America → Asia Pacific)** * [Los Angeles to Tokyo latency and RTT](/docs/network/latency/pairs/lax-tyo-rtt): 101.2 ms * [Seattle to Hong Kong latency and RTT](/docs/network/latency/pairs/sea-hkg-rtt): 130.4 ms * [Los Angeles to Taipei latency and RTT](/docs/network/latency/pairs/lax-tpe-rtt): 132.9 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/sea-tpe.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Seattle → Taipei round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/sea-tpe-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Seattle, USA → Tokyo, Japan RTT 🇺🇸 **Seattle, USA (SEA)** → 🇯🇵 **Tokyo, Japan (TYO)** Measurement round: `2026-08-16T04:07:28Z`. Seattle to Tokyo averaged 85 ms across 50 replies. The observed range was 81.03 to 97.43 ms; these figures describe one probe round, not a long-term service target. This is number 5 of 19 outgoing routes by average RTT. That ranking compares different destinations and is not adjusted for distance. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **85 ms** | | Jitter | **2.59 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **75.55 ms** | | Fiber Efficiency | **88.9%** | | Latency Tier | Good | | Source Region | [North America](/docs/network/latency/regions/north-america) | | Destination Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Seattle](https://www.geonames.org/5809844) (47.60621, -122.33207) and [Tokyo](https://www.geonames.org/1850147) (35.68950, 139.69171). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **7,715.2 km** | | Vacuum RTT floor | **51.47 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **75.55 ms** | | Low-latency fiber material floor | **75.25 ms** | | Engineering floor (5% path allowance) | **79.34 ms** | | Research 1.33× mapped-fiber reference | **100.49 ms** | | Estimated unamplified path loss | **1620.2 dB** | | Transparent optical spans / inline amplifiers | **102 / 101** | | Published RTT inflation over fiber floor | **1.13×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------ | | Minimum RTT | **81.03 ms** | | Average RTT | **85 ms** | | Maximum RTT | **97.43 ms** | | Standard deviation | **3.15 ms** | | Stdev / average | **3.7%** | ## Route Context * Departure PoP: [Seattle (SEA)](/docs/network/latency/sea-seattle) * Destination PoP: [Tokyo (TYO)](/docs/network/latency/tyo-tokyo) * Region overview: [North America](/docs/network/latency/regions/north-america) → [Asia Pacific](/docs/network/latency/regions/asia-pacific) * Corridor overview: [North America to Asia Pacific](/docs/network/latency/regions/north-america-to-asia-pacific) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Tokyo to Seattle latency and RTT](/docs/network/latency/pairs/tyo-sea-rtt): 84.8 ms **Fastest routes departing Seattle (SEA)** * [Seattle to Los Angeles latency and RTT](/docs/network/latency/pairs/sea-lax-rtt): 25.9 ms * [Seattle to New York latency and RTT](/docs/network/latency/pairs/sea-nyc-rtt): 59.5 ms * [Seattle to Ashburn latency and RTT](/docs/network/latency/pairs/sea-iad-rtt): 61.5 ms * [Seattle to Miami latency and RTT](/docs/network/latency/pairs/sea-mia-rtt): 81.7 ms * [Seattle to Taipei latency and RTT](/docs/network/latency/pairs/sea-tpe-rtt): 114.7 ms **Fastest routes arriving at Tokyo (TYO)** * [Taipei to Tokyo latency and RTT](/docs/network/latency/pairs/tpe-tyo-rtt): 31.9 ms * [Hong Kong to Tokyo latency and RTT](/docs/network/latency/pairs/hkg-tyo-rtt): 44.9 ms * [Singapore to Tokyo latency and RTT](/docs/network/latency/pairs/sin-tyo-rtt): 68.9 ms * [Los Angeles to Tokyo latency and RTT](/docs/network/latency/pairs/lax-tyo-rtt): 101.2 ms * [Sydney to Tokyo latency and RTT](/docs/network/latency/pairs/syd-tyo-rtt): 101.3 ms **Same corridor (North America → Asia Pacific)** * [Seattle to Hong Kong latency and RTT](/docs/network/latency/pairs/sea-hkg-rtt): 130.4 ms * [Los Angeles to Taipei latency and RTT](/docs/network/latency/pairs/lax-tpe-rtt): 132.9 ms * [Los Angeles to Sydney latency and RTT](/docs/network/latency/pairs/lax-syd-rtt): 137.6 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/sea-tyo.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Seattle → Tokyo round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/sea-tyo-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Singapore → Amsterdam, Netherlands RTT 🇸🇬 **Singapore (SIN)** → 🇳🇱 **Amsterdam, Netherlands (AMS)** Measurement round: `2026-08-16T04:07:28Z`. Singapore to Amsterdam averaged 155.8 ms across 50 replies. The observed range was 147.48 to 172.04 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 6.41 ms, or 4.1% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------------------- | | RTT | **155.8 ms** | | Jitter | **6.07 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **102.86 ms** | | Fiber Efficiency | **66%** | | Latency Tier | Fair | | Source Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | | Destination Region | [Europe](/docs/network/latency/regions/europe) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Singapore](https://www.geonames.org/1880252) (1.28967, 103.85007) and [Amsterdam](https://www.geonames.org/2759794) (52.37403, 4.88969). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | --------------- | | WGS-84 geodesic distance | **10,503.4 km** | | Vacuum RTT floor | **70.07 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **102.86 ms** | | Low-latency fiber material floor | **102.44 ms** | | Engineering floor (5% path allowance) | **108.01 ms** | | Research 1.33× mapped-fiber reference | **136.8 ms** | | Estimated unamplified path loss | **2205.7 dB** | | Transparent optical spans / inline amplifiers | **138 / 137** | | Published RTT inflation over fiber floor | **1.51×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **147.48 ms** | | Average RTT | **155.8 ms** | | Maximum RTT | **172.04 ms** | | Standard deviation | **6.41 ms** | | Stdev / average | **4.1%** | ## Route Context * Departure PoP: [Singapore (SIN)](/docs/network/latency/sin-singapore) * Destination PoP: [Amsterdam (AMS)](/docs/network/latency/ams-amsterdam) * Region overview: [Asia Pacific](/docs/network/latency/regions/asia-pacific) → [Europe](/docs/network/latency/regions/europe) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Amsterdam to Singapore latency and RTT](/docs/network/latency/pairs/ams-sin-rtt): 157.6 ms **Fastest routes departing Singapore (SIN)** * [Singapore to Hong Kong latency and RTT](/docs/network/latency/pairs/sin-hkg-rtt): 30.8 ms * [Singapore to Taipei latency and RTT](/docs/network/latency/pairs/sin-tpe-rtt): 45.8 ms * [Singapore to Tokyo latency and RTT](/docs/network/latency/pairs/sin-tyo-rtt): 68.9 ms * [Singapore to Melbourne latency and RTT](/docs/network/latency/pairs/sin-mel-rtt): 88.6 ms * [Singapore to Sydney latency and RTT](/docs/network/latency/pairs/sin-syd-rtt): 94.5 ms **Fastest routes arriving at Amsterdam (AMS)** * [London to Amsterdam latency and RTT](/docs/network/latency/pairs/lon-ams-rtt): 5.2 ms * [Frankfurt to Amsterdam latency and RTT](/docs/network/latency/pairs/fra-ams-rtt): 5.9 ms * [Paris to Amsterdam latency and RTT](/docs/network/latency/pairs/par-ams-rtt): 7 ms * [Berlin to Amsterdam latency and RTT](/docs/network/latency/pairs/ber-ams-rtt): 7.2 ms * [Marseille to Amsterdam latency and RTT](/docs/network/latency/pairs/mrs-ams-rtt): 20.3 ms **Same corridor (Asia Pacific → Europe)** * [Hong Kong to Moscow latency and RTT](/docs/network/latency/pairs/hkg-mow-rtt): 118.2 ms * [Taipei to Moscow latency and RTT](/docs/network/latency/pairs/tpe-mow-rtt): 132.7 ms * [Singapore to Marseille latency and RTT](/docs/network/latency/pairs/sin-mrs-rtt): 140.6 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/sin-ams.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Singapore → Amsterdam round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/sin-ams-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Singapore → Berlin, Germany RTT 🇸🇬 **Singapore (SIN)** → 🇩🇪 **Berlin, Germany (BER)** Measurement round: `2026-08-16T04:07:28Z`. Singapore to Berlin averaged 160.3 ms across 50 replies. The observed range was 152.4 to 193.36 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 7.96 ms, or 5.0% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------------------- | | RTT | **160.3 ms** | | Jitter | **5.38 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **97.19 ms** | | Fiber Efficiency | **60.6%** | | Latency Tier | Fair | | Source Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | | Destination Region | [Europe](/docs/network/latency/regions/europe) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Singapore](https://www.geonames.org/1880252) (1.28967, 103.85007) and [Berlin](https://www.geonames.org/2950159) (52.52437, 13.41053). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | ------------- | | WGS-84 geodesic distance | **9,925 km** | | Vacuum RTT floor | **66.21 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **97.19 ms** | | Low-latency fiber material floor | **96.8 ms** | | Engineering floor (5% path allowance) | **102.07 ms** | | Research 1.33× mapped-fiber reference | **129.27 ms** | | Estimated unamplified path loss | **2084.3 dB** | | Transparent optical spans / inline amplifiers | **131 / 130** | | Published RTT inflation over fiber floor | **1.65×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **152.4 ms** | | Average RTT | **160.3 ms** | | Maximum RTT | **193.36 ms** | | Standard deviation | **7.96 ms** | | Stdev / average | **5.0%** | ## Route Context * Departure PoP: [Singapore (SIN)](/docs/network/latency/sin-singapore) * Destination PoP: [Berlin (BER)](/docs/network/latency/ber-berlin) * Region overview: [Asia Pacific](/docs/network/latency/regions/asia-pacific) → [Europe](/docs/network/latency/regions/europe) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Berlin to Singapore latency and RTT](/docs/network/latency/pairs/ber-sin-rtt): 158.3 ms **Fastest routes departing Singapore (SIN)** * [Singapore to Hong Kong latency and RTT](/docs/network/latency/pairs/sin-hkg-rtt): 30.8 ms * [Singapore to Taipei latency and RTT](/docs/network/latency/pairs/sin-tpe-rtt): 45.8 ms * [Singapore to Tokyo latency and RTT](/docs/network/latency/pairs/sin-tyo-rtt): 68.9 ms * [Singapore to Melbourne latency and RTT](/docs/network/latency/pairs/sin-mel-rtt): 88.6 ms * [Singapore to Sydney latency and RTT](/docs/network/latency/pairs/sin-syd-rtt): 94.5 ms **Fastest routes arriving at Berlin (BER)** * [Frankfurt to Berlin latency and RTT](/docs/network/latency/pairs/fra-ber-rtt): 6.1 ms * [Amsterdam to Berlin latency and RTT](/docs/network/latency/pairs/ams-ber-rtt): 7.2 ms * [Paris to Berlin latency and RTT](/docs/network/latency/pairs/par-ber-rtt): 15.5 ms * [London to Berlin latency and RTT](/docs/network/latency/pairs/lon-ber-rtt): 16.2 ms * [Marseille to Berlin latency and RTT](/docs/network/latency/pairs/mrs-ber-rtt): 20.6 ms **Same corridor (Asia Pacific → Europe)** * [Hong Kong to Moscow latency and RTT](/docs/network/latency/pairs/hkg-mow-rtt): 118.2 ms * [Taipei to Moscow latency and RTT](/docs/network/latency/pairs/tpe-mow-rtt): 132.7 ms * [Singapore to Marseille latency and RTT](/docs/network/latency/pairs/sin-mrs-rtt): 140.6 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/sin-ber.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Singapore → Berlin round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/sin-ber-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Singapore → Frankfurt, Germany RTT 🇸🇬 **Singapore (SIN)** → 🇩🇪 **Frankfurt, Germany (FRA)** Measurement round: `2026-08-16T04:07:28Z`. Singapore to Frankfurt averaged 151.6 ms across 50 replies. The observed range was 145.11 to 167.35 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 5.04 ms, or 3.3% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------------------- | | RTT | **151.6 ms** | | Jitter | **4.7 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **100.56 ms** | | Fiber Efficiency | **66.3%** | | Latency Tier | Fair | | Source Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | | Destination Region | [Europe](/docs/network/latency/regions/europe) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Singapore](https://www.geonames.org/1880252) (1.28967, 103.85007) and [Frankfurt am Main](https://www.geonames.org/2925533) (50.11552, 8.68417). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | --------------- | | WGS-84 geodesic distance | **10,268.5 km** | | Vacuum RTT floor | **68.5 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **100.56 ms** | | Low-latency fiber material floor | **100.15 ms** | | Engineering floor (5% path allowance) | **105.6 ms** | | Research 1.33× mapped-fiber reference | **133.74 ms** | | Estimated unamplified path loss | **2156.4 dB** | | Transparent optical spans / inline amplifiers | **135 / 134** | | Published RTT inflation over fiber floor | **1.51×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **145.11 ms** | | Average RTT | **151.6 ms** | | Maximum RTT | **167.35 ms** | | Standard deviation | **5.04 ms** | | Stdev / average | **3.3%** | ## Route Context * Departure PoP: [Singapore (SIN)](/docs/network/latency/sin-singapore) * Destination PoP: [Frankfurt (FRA)](/docs/network/latency/fra-frankfurt) * Region overview: [Asia Pacific](/docs/network/latency/regions/asia-pacific) → [Europe](/docs/network/latency/regions/europe) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Frankfurt to Singapore latency and RTT](/docs/network/latency/pairs/fra-sin-rtt): 151.4 ms **Fastest routes departing Singapore (SIN)** * [Singapore to Hong Kong latency and RTT](/docs/network/latency/pairs/sin-hkg-rtt): 30.8 ms * [Singapore to Taipei latency and RTT](/docs/network/latency/pairs/sin-tpe-rtt): 45.8 ms * [Singapore to Tokyo latency and RTT](/docs/network/latency/pairs/sin-tyo-rtt): 68.9 ms * [Singapore to Melbourne latency and RTT](/docs/network/latency/pairs/sin-mel-rtt): 88.6 ms * [Singapore to Sydney latency and RTT](/docs/network/latency/pairs/sin-syd-rtt): 94.5 ms **Fastest routes arriving at Frankfurt (FRA)** * [Amsterdam to Frankfurt latency and RTT](/docs/network/latency/pairs/ams-fra-rtt): 5.9 ms * [Berlin to Frankfurt latency and RTT](/docs/network/latency/pairs/ber-fra-rtt): 6.1 ms * [Paris to Frankfurt latency and RTT](/docs/network/latency/pairs/par-fra-rtt): 7.6 ms * [London to Frankfurt latency and RTT](/docs/network/latency/pairs/lon-fra-rtt): 13.6 ms * [Marseille to Frankfurt latency and RTT](/docs/network/latency/pairs/mrs-fra-rtt): 16 ms **Same corridor (Asia Pacific → Europe)** * [Hong Kong to Moscow latency and RTT](/docs/network/latency/pairs/hkg-mow-rtt): 118.2 ms * [Taipei to Moscow latency and RTT](/docs/network/latency/pairs/tpe-mow-rtt): 132.7 ms * [Singapore to Marseille latency and RTT](/docs/network/latency/pairs/sin-mrs-rtt): 140.6 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/sin-fra.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Singapore → Frankfurt round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/sin-fra-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Singapore → São Paulo, Brazil RTT 🇸🇬 **Singapore (SIN)** → 🇧🇷 **São Paulo, Brazil (GRU)** Measurement round: `2026-08-16T04:07:28Z`. Singapore to São Paulo averaged 311.4 ms across 50 replies. The observed range was 306.39 to 332.38 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 5.29 ms, or 1.7% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **311.4 ms** | | Jitter | **3.96 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **156.67 ms** | | Fiber Efficiency | **50.3%** | | Latency Tier | High | | Source Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | | Destination Region | [South America](/docs/network/latency/regions/south-america) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Singapore](https://www.geonames.org/1880252) (1.28967, 103.85007) and [São Paulo](https://www.geonames.org/3448439) (-23.54750, -46.63611). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | --------------- | | WGS-84 geodesic distance | **15,998.2 km** | | Vacuum RTT floor | **106.73 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **156.67 ms** | | Low-latency fiber material floor | **156.04 ms** | | Engineering floor (5% path allowance) | **164.52 ms** | | Research 1.33× mapped-fiber reference | **208.37 ms** | | Estimated unamplified path loss | **3359.6 dB** | | Transparent optical spans / inline amplifiers | **210 / 209** | | Published RTT inflation over fiber floor | **1.99×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **306.39 ms** | | Average RTT | **311.4 ms** | | Maximum RTT | **332.38 ms** | | Standard deviation | **5.29 ms** | | Stdev / average | **1.7%** | ## Route Context * Departure PoP: [Singapore (SIN)](/docs/network/latency/sin-singapore) * Destination PoP: [São Paulo (GRU)](/docs/network/latency/gru-sao-paulo) * Region overview: [Asia Pacific](/docs/network/latency/regions/asia-pacific) → [South America](/docs/network/latency/regions/south-america) * Corridor overview: [Asia Pacific to South America](/docs/network/latency/regions/asia-pacific-to-south-america) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [São Paulo to Singapore latency and RTT](/docs/network/latency/pairs/gru-sin-rtt): 297.1 ms **Fastest routes departing Singapore (SIN)** * [Singapore to Hong Kong latency and RTT](/docs/network/latency/pairs/sin-hkg-rtt): 30.8 ms * [Singapore to Taipei latency and RTT](/docs/network/latency/pairs/sin-tpe-rtt): 45.8 ms * [Singapore to Tokyo latency and RTT](/docs/network/latency/pairs/sin-tyo-rtt): 68.9 ms * [Singapore to Melbourne latency and RTT](/docs/network/latency/pairs/sin-mel-rtt): 88.6 ms * [Singapore to Sydney latency and RTT](/docs/network/latency/pairs/sin-syd-rtt): 94.5 ms **Fastest routes arriving at São Paulo (GRU)** * [Ashburn to São Paulo latency and RTT](/docs/network/latency/pairs/iad-gru-rtt): 101.4 ms * [New York to São Paulo latency and RTT](/docs/network/latency/pairs/nyc-gru-rtt): 106.9 ms * [Miami to São Paulo latency and RTT](/docs/network/latency/pairs/mia-gru-rtt): 128.7 ms * [Los Angeles to São Paulo latency and RTT](/docs/network/latency/pairs/lax-gru-rtt): 131.3 ms * [Seattle to São Paulo latency and RTT](/docs/network/latency/pairs/sea-gru-rtt): 155.7 ms **Same corridor (Asia Pacific → South America)** * [Tokyo to São Paulo latency and RTT](/docs/network/latency/pairs/tyo-gru-rtt): 230.5 ms * [Taipei to São Paulo latency and RTT](/docs/network/latency/pairs/tpe-gru-rtt): 273 ms * [Hong Kong to São Paulo latency and RTT](/docs/network/latency/pairs/hkg-gru-rtt): 287.6 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/sin-gru.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Singapore → São Paulo round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/sin-gru-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Singapore → Hong Kong RTT 🇸🇬 **Singapore (SIN)** → 🇭🇰 **Hong Kong (HKG)** Measurement round: `2026-08-16T04:07:28Z`. Singapore to Hong Kong averaged 30.8 ms across 50 replies. The observed range was 30.29 to 33.37 ms; these figures describe one probe round, not a long-term service target. This is number 1 of 19 outgoing routes by average RTT. That ranking compares different destinations and is not adjusted for distance. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------------------- | | RTT | **30.8 ms** | | Jitter | **0.4 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **25.24 ms** | | Fiber Efficiency | **81.9%** | | Latency Tier | Excellent | | Source Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | | Destination Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Singapore](https://www.geonames.org/1880252) (1.28967, 103.85007) and [Hong Kong](https://www.geonames.org/1819729) (22.27832, 114.17469). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **2,577.2 km** | | Vacuum RTT floor | **17.19 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **25.24 ms** | | Low-latency fiber material floor | **25.14 ms** | | Engineering floor (5% path allowance) | **26.5 ms** | | Research 1.33× mapped-fiber reference | **33.57 ms** | | Estimated unamplified path loss | **541.2 dB** | | Transparent optical spans / inline amplifiers | **34 / 33** | | Published RTT inflation over fiber floor | **1.22×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------ | | Minimum RTT | **30.29 ms** | | Average RTT | **30.8 ms** | | Maximum RTT | **33.37 ms** | | Standard deviation | **0.58 ms** | | Stdev / average | **1.9%** | ## Route Context * Departure PoP: [Singapore (SIN)](/docs/network/latency/sin-singapore) * Destination PoP: [Hong Kong (HKG)](/docs/network/latency/hkg-hong-kong) * Region overview: [Asia Pacific](/docs/network/latency/regions/asia-pacific) → [Asia Pacific](/docs/network/latency/regions/asia-pacific) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Hong Kong to Singapore latency and RTT](/docs/network/latency/pairs/hkg-sin-rtt): 31.5 ms **Fastest routes departing Singapore (SIN)** * [Singapore to Taipei latency and RTT](/docs/network/latency/pairs/sin-tpe-rtt): 45.8 ms * [Singapore to Tokyo latency and RTT](/docs/network/latency/pairs/sin-tyo-rtt): 68.9 ms * [Singapore to Melbourne latency and RTT](/docs/network/latency/pairs/sin-mel-rtt): 88.6 ms * [Singapore to Sydney latency and RTT](/docs/network/latency/pairs/sin-syd-rtt): 94.5 ms * [Singapore to Marseille latency and RTT](/docs/network/latency/pairs/sin-mrs-rtt): 140.6 ms **Fastest routes arriving at Hong Kong (HKG)** * [Taipei to Hong Kong latency and RTT](/docs/network/latency/pairs/tpe-hkg-rtt): 15.5 ms * [Tokyo to Hong Kong latency and RTT](/docs/network/latency/pairs/tyo-hkg-rtt): 44.8 ms * [Moscow to Hong Kong latency and RTT](/docs/network/latency/pairs/mow-hkg-rtt): 118.6 ms * [Seattle to Hong Kong latency and RTT](/docs/network/latency/pairs/sea-hkg-rtt): 130.4 ms * [Sydney to Hong Kong latency and RTT](/docs/network/latency/pairs/syd-hkg-rtt): 135.7 ms **Same corridor (Asia Pacific → Asia Pacific)** * [Melbourne to Sydney latency and RTT](/docs/network/latency/pairs/mel-syd-rtt): 9.8 ms * [Sydney to Melbourne latency and RTT](/docs/network/latency/pairs/syd-mel-rtt): 9.8 ms * [Hong Kong to Taipei latency and RTT](/docs/network/latency/pairs/hkg-tpe-rtt): 14.7 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/sin-hkg.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Singapore → Hong Kong round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/sin-hkg-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Singapore → Ashburn, USA RTT 🇸🇬 **Singapore (SIN)** → 🇺🇸 **Ashburn, USA (IAD)** Measurement round: `2026-08-16T04:07:28Z`. Singapore to Ashburn averaged 210 ms across 50 replies. The observed range was 198.98 to 248.94 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 9.22 ms, or 4.4% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **210 ms** | | Jitter | **8.9 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **152.14 ms** | | Fiber Efficiency | **72.4%** | | Latency Tier | Fair | | Source Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | | Destination Region | [North America](/docs/network/latency/regions/north-america) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Singapore](https://www.geonames.org/1880252) (1.28967, 103.85007) and [Ashburn](https://www.geonames.org/4744870) (39.04372, -77.48749). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | --------------- | | WGS-84 geodesic distance | **15,535.9 km** | | Vacuum RTT floor | **103.64 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **152.14 ms** | | Low-latency fiber material floor | **151.53 ms** | | Engineering floor (5% path allowance) | **159.77 ms** | | Research 1.33× mapped-fiber reference | **202.35 ms** | | Estimated unamplified path loss | **3262.5 dB** | | Transparent optical spans / inline amplifiers | **204 / 203** | | Published RTT inflation over fiber floor | **1.38×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **198.98 ms** | | Average RTT | **210 ms** | | Maximum RTT | **248.94 ms** | | Standard deviation | **9.22 ms** | | Stdev / average | **4.4%** | ## Route Context * Departure PoP: [Singapore (SIN)](/docs/network/latency/sin-singapore) * Destination PoP: [Ashburn (IAD)](/docs/network/latency/iad-ashburn) * Region overview: [Asia Pacific](/docs/network/latency/regions/asia-pacific) → [North America](/docs/network/latency/regions/north-america) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Ashburn to Singapore latency and RTT](/docs/network/latency/pairs/iad-sin-rtt): 211.5 ms **Fastest routes departing Singapore (SIN)** * [Singapore to Hong Kong latency and RTT](/docs/network/latency/pairs/sin-hkg-rtt): 30.8 ms * [Singapore to Taipei latency and RTT](/docs/network/latency/pairs/sin-tpe-rtt): 45.8 ms * [Singapore to Tokyo latency and RTT](/docs/network/latency/pairs/sin-tyo-rtt): 68.9 ms * [Singapore to Melbourne latency and RTT](/docs/network/latency/pairs/sin-mel-rtt): 88.6 ms * [Singapore to Sydney latency and RTT](/docs/network/latency/pairs/sin-syd-rtt): 94.5 ms **Fastest routes arriving at Ashburn (IAD)** * [New York to Ashburn latency and RTT](/docs/network/latency/pairs/nyc-iad-rtt): 6 ms * [Miami to Ashburn latency and RTT](/docs/network/latency/pairs/mia-iad-rtt): 27.3 ms * [Los Angeles to Ashburn latency and RTT](/docs/network/latency/pairs/lax-iad-rtt): 60.4 ms * [Seattle to Ashburn latency and RTT](/docs/network/latency/pairs/sea-iad-rtt): 61.5 ms * [London to Ashburn latency and RTT](/docs/network/latency/pairs/lon-iad-rtt): 71 ms **Same corridor (Asia Pacific → North America)** * [Tokyo to Seattle latency and RTT](/docs/network/latency/pairs/tyo-sea-rtt): 84.8 ms * [Tokyo to Los Angeles latency and RTT](/docs/network/latency/pairs/tyo-lax-rtt): 101.2 ms * [Taipei to Seattle latency and RTT](/docs/network/latency/pairs/tpe-sea-rtt): 115.8 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/sin-iad.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Singapore → Ashburn round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/sin-iad-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Singapore → Johannesburg, South Africa RTT 🇸🇬 **Singapore (SIN)** → 🇿🇦 **Johannesburg, South Africa (JNB)** Measurement round: `2026-08-16T04:07:28Z`. Singapore to Johannesburg averaged 312.6 ms, compared with a 84.85 ms city-centre fiber reference. The difference cannot be assigned to a particular cable or carrier from RTT alone. Standard deviation was 7.16 ms, or 2.3% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------------------- | | RTT | **312.6 ms** | | Jitter | **7.32 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **84.85 ms** | | Fiber Efficiency | **27.1%** | | Latency Tier | High | | Source Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | | Destination Region | [Africa](/docs/network/latency/regions/africa) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Singapore](https://www.geonames.org/1880252) (1.28967, 103.85007) and [Johannesburg](https://www.geonames.org/993800) (-26.20227, 28.04363). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **8,664.2 km** | | Vacuum RTT floor | **57.8 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **84.85 ms** | | Low-latency fiber material floor | **84.51 ms** | | Engineering floor (5% path allowance) | **89.1 ms** | | Research 1.33× mapped-fiber reference | **112.85 ms** | | Estimated unamplified path loss | **1819.5 dB** | | Transparent optical spans / inline amplifiers | **114 / 113** | | Published RTT inflation over fiber floor | **3.68×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **303.47 ms** | | Average RTT | **312.6 ms** | | Maximum RTT | **332.3 ms** | | Standard deviation | **7.16 ms** | | Stdev / average | **2.3%** | ## Route Context * Departure PoP: [Singapore (SIN)](/docs/network/latency/sin-singapore) * Destination PoP: [Johannesburg (JNB)](/docs/network/latency/jnb-johannesburg) * Region overview: [Asia Pacific](/docs/network/latency/regions/asia-pacific) → [Africa](/docs/network/latency/regions/africa) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Johannesburg to Singapore latency and RTT](/docs/network/latency/pairs/jnb-sin-rtt): 311.6 ms **Fastest routes departing Singapore (SIN)** * [Singapore to Hong Kong latency and RTT](/docs/network/latency/pairs/sin-hkg-rtt): 30.8 ms * [Singapore to Taipei latency and RTT](/docs/network/latency/pairs/sin-tpe-rtt): 45.8 ms * [Singapore to Tokyo latency and RTT](/docs/network/latency/pairs/sin-tyo-rtt): 68.9 ms * [Singapore to Melbourne latency and RTT](/docs/network/latency/pairs/sin-mel-rtt): 88.6 ms * [Singapore to Sydney latency and RTT](/docs/network/latency/pairs/sin-syd-rtt): 94.5 ms **Fastest routes arriving at Johannesburg (JNB)** * [Amsterdam to Johannesburg latency and RTT](/docs/network/latency/pairs/ams-jnb-rtt): 171.9 ms * [Berlin to Johannesburg latency and RTT](/docs/network/latency/pairs/ber-jnb-rtt): 172.1 ms * [London to Johannesburg latency and RTT](/docs/network/latency/pairs/lon-jnb-rtt): 179.6 ms * [Paris to Johannesburg latency and RTT](/docs/network/latency/pairs/par-jnb-rtt): 180.9 ms * [Frankfurt to Johannesburg latency and RTT](/docs/network/latency/pairs/fra-jnb-rtt): 188.4 ms **Same corridor (Asia Pacific → Africa)** * [Hong Kong to Johannesburg latency and RTT](/docs/network/latency/pairs/hkg-jnb-rtt): 316.7 ms * [Taipei to Johannesburg latency and RTT](/docs/network/latency/pairs/tpe-jnb-rtt): 331 ms * [Tokyo to Johannesburg latency and RTT](/docs/network/latency/pairs/tyo-jnb-rtt): 359.2 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/sin-jnb.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Singapore → Johannesburg round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/sin-jnb-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Singapore → Los Angeles, USA RTT 🇸🇬 **Singapore (SIN)** → 🇺🇸 **Los Angeles, USA (LAX)** Measurement round: `2026-08-16T04:07:28Z`. Singapore to Los Angeles averaged 168.2 ms across 50 replies. The observed range was 160.69 to 194.08 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 6.01 ms, or 3.6% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **168.2 ms** | | Jitter | **4.76 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **138.43 ms** | | Fiber Efficiency | **82.3%** | | Latency Tier | Fair | | Source Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | | Destination Region | [North America](/docs/network/latency/regions/north-america) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Singapore](https://www.geonames.org/1880252) (1.28967, 103.85007) and [Los Angeles](https://www.geonames.org/5368361) (34.05223, -118.24368). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | --------------- | | WGS-84 geodesic distance | **14,135.8 km** | | Vacuum RTT floor | **94.3 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **138.43 ms** | | Low-latency fiber material floor | **137.87 ms** | | Engineering floor (5% path allowance) | **145.37 ms** | | Research 1.33× mapped-fiber reference | **184.11 ms** | | Estimated unamplified path loss | **2968.5 dB** | | Transparent optical spans / inline amplifiers | **186 / 185** | | Published RTT inflation over fiber floor | **1.22×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **160.69 ms** | | Average RTT | **168.2 ms** | | Maximum RTT | **194.08 ms** | | Standard deviation | **6.01 ms** | | Stdev / average | **3.6%** | ## Route Context * Departure PoP: [Singapore (SIN)](/docs/network/latency/sin-singapore) * Destination PoP: [Los Angeles (LAX)](/docs/network/latency/lax-los-angeles) * Region overview: [Asia Pacific](/docs/network/latency/regions/asia-pacific) → [North America](/docs/network/latency/regions/north-america) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Los Angeles to Singapore latency and RTT](/docs/network/latency/pairs/lax-sin-rtt): 167.7 ms **Fastest routes departing Singapore (SIN)** * [Singapore to Hong Kong latency and RTT](/docs/network/latency/pairs/sin-hkg-rtt): 30.8 ms * [Singapore to Taipei latency and RTT](/docs/network/latency/pairs/sin-tpe-rtt): 45.8 ms * [Singapore to Tokyo latency and RTT](/docs/network/latency/pairs/sin-tyo-rtt): 68.9 ms * [Singapore to Melbourne latency and RTT](/docs/network/latency/pairs/sin-mel-rtt): 88.6 ms * [Singapore to Sydney latency and RTT](/docs/network/latency/pairs/sin-syd-rtt): 94.5 ms **Fastest routes arriving at Los Angeles (LAX)** * [Seattle to Los Angeles latency and RTT](/docs/network/latency/pairs/sea-lax-rtt): 25.9 ms * [Miami to Los Angeles latency and RTT](/docs/network/latency/pairs/mia-lax-rtt): 56.8 ms * [New York to Los Angeles latency and RTT](/docs/network/latency/pairs/nyc-lax-rtt): 58.9 ms * [Ashburn to Los Angeles latency and RTT](/docs/network/latency/pairs/iad-lax-rtt): 60.7 ms * [Tokyo to Los Angeles latency and RTT](/docs/network/latency/pairs/tyo-lax-rtt): 101.2 ms **Same corridor (Asia Pacific → North America)** * [Tokyo to Seattle latency and RTT](/docs/network/latency/pairs/tyo-sea-rtt): 84.8 ms * [Taipei to Seattle latency and RTT](/docs/network/latency/pairs/tpe-sea-rtt): 115.8 ms * [Hong Kong to Seattle latency and RTT](/docs/network/latency/pairs/hkg-sea-rtt): 131.6 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/sin-lax.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Singapore → Los Angeles round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/sin-lax-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Singapore → London, UK RTT 🇸🇬 **Singapore (SIN)** → 🇬🇧 **London, UK (LON)** Measurement round: `2026-08-16T04:07:28Z`. Singapore to London averaged 155.4 ms across 50 replies. The observed range was 150.37 to 180.35 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 5.16 ms, or 3.3% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------------------- | | RTT | **155.4 ms** | | Jitter | **3.28 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **106.35 ms** | | Fiber Efficiency | **68.4%** | | Latency Tier | Fair | | Source Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | | Destination Region | [Europe](/docs/network/latency/regions/europe) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Singapore](https://www.geonames.org/1880252) (1.28967, 103.85007) and [London](https://www.geonames.org/2643743) (51.50853, -0.12574). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | --------------- | | WGS-84 geodesic distance | **10,859.7 km** | | Vacuum RTT floor | **72.45 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **106.35 ms** | | Low-latency fiber material floor | **105.92 ms** | | Engineering floor (5% path allowance) | **111.68 ms** | | Research 1.33× mapped-fiber reference | **141.44 ms** | | Estimated unamplified path loss | **2280.5 dB** | | Transparent optical spans / inline amplifiers | **143 / 142** | | Published RTT inflation over fiber floor | **1.46×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **150.37 ms** | | Average RTT | **155.4 ms** | | Maximum RTT | **180.35 ms** | | Standard deviation | **5.16 ms** | | Stdev / average | **3.3%** | ## Route Context * Departure PoP: [Singapore (SIN)](/docs/network/latency/sin-singapore) * Destination PoP: [London (LON)](/docs/network/latency/lon-london) * Region overview: [Asia Pacific](/docs/network/latency/regions/asia-pacific) → [Europe](/docs/network/latency/regions/europe) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [London to Singapore latency and RTT](/docs/network/latency/pairs/lon-sin-rtt): 156.5 ms **Fastest routes departing Singapore (SIN)** * [Singapore to Hong Kong latency and RTT](/docs/network/latency/pairs/sin-hkg-rtt): 30.8 ms * [Singapore to Taipei latency and RTT](/docs/network/latency/pairs/sin-tpe-rtt): 45.8 ms * [Singapore to Tokyo latency and RTT](/docs/network/latency/pairs/sin-tyo-rtt): 68.9 ms * [Singapore to Melbourne latency and RTT](/docs/network/latency/pairs/sin-mel-rtt): 88.6 ms * [Singapore to Sydney latency and RTT](/docs/network/latency/pairs/sin-syd-rtt): 94.5 ms **Fastest routes arriving at London (LON)** * [Amsterdam to London latency and RTT](/docs/network/latency/pairs/ams-lon-rtt): 5.2 ms * [Paris to London latency and RTT](/docs/network/latency/pairs/par-lon-rtt): 6.4 ms * [Frankfurt to London latency and RTT](/docs/network/latency/pairs/fra-lon-rtt): 12.9 ms * [Berlin to London latency and RTT](/docs/network/latency/pairs/ber-lon-rtt): 15.2 ms * [Marseille to London latency and RTT](/docs/network/latency/pairs/mrs-lon-rtt): 18.9 ms **Same corridor (Asia Pacific → Europe)** * [Hong Kong to Moscow latency and RTT](/docs/network/latency/pairs/hkg-mow-rtt): 118.2 ms * [Taipei to Moscow latency and RTT](/docs/network/latency/pairs/tpe-mow-rtt): 132.7 ms * [Singapore to Marseille latency and RTT](/docs/network/latency/pairs/sin-mrs-rtt): 140.6 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/sin-lon.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Singapore → London round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/sin-lon-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Singapore → Melbourne, Australia RTT 🇸🇬 **Singapore (SIN)** → 🇦🇺 **Melbourne, Australia (MEL)** Measurement round: `2026-08-16T04:07:28Z`. Singapore to Melbourne averaged 88.6 ms across 50 replies. The observed range was 86.6 to 94.44 ms; these figures describe one probe round, not a long-term service target. This is number 4 of 19 outgoing routes by average RTT. That ranking compares different destinations and is not adjusted for distance. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------------------- | | RTT | **88.6 ms** | | Jitter | **1.59 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **59.22 ms** | | Fiber Efficiency | **66.8%** | | Latency Tier | Good | | Source Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | | Destination Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Singapore](https://www.geonames.org/1880252) (1.28967, 103.85007) and [Melbourne](https://www.geonames.org/2158177) (-37.81400, 144.96332). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **6,046.9 km** | | Vacuum RTT floor | **40.34 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **59.22 ms** | | Low-latency fiber material floor | **58.98 ms** | | Engineering floor (5% path allowance) | **62.18 ms** | | Research 1.33× mapped-fiber reference | **78.76 ms** | | Estimated unamplified path loss | **1269.9 dB** | | Transparent optical spans / inline amplifiers | **80 / 79** | | Published RTT inflation over fiber floor | **1.5×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------ | | Minimum RTT | **86.6 ms** | | Average RTT | **88.6 ms** | | Maximum RTT | **94.44 ms** | | Standard deviation | **1.9 ms** | | Stdev / average | **2.1%** | ## Route Context * Departure PoP: [Singapore (SIN)](/docs/network/latency/sin-singapore) * Destination PoP: [Melbourne (MEL)](/docs/network/latency/mel-melbourne) * Region overview: [Asia Pacific](/docs/network/latency/regions/asia-pacific) → [Asia Pacific](/docs/network/latency/regions/asia-pacific) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Melbourne to Singapore latency and RTT](/docs/network/latency/pairs/mel-sin-rtt): 88.4 ms **Fastest routes departing Singapore (SIN)** * [Singapore to Hong Kong latency and RTT](/docs/network/latency/pairs/sin-hkg-rtt): 30.8 ms * [Singapore to Taipei latency and RTT](/docs/network/latency/pairs/sin-tpe-rtt): 45.8 ms * [Singapore to Tokyo latency and RTT](/docs/network/latency/pairs/sin-tyo-rtt): 68.9 ms * [Singapore to Sydney latency and RTT](/docs/network/latency/pairs/sin-syd-rtt): 94.5 ms * [Singapore to Marseille latency and RTT](/docs/network/latency/pairs/sin-mrs-rtt): 140.6 ms **Fastest routes arriving at Melbourne (MEL)** * [Sydney to Melbourne latency and RTT](/docs/network/latency/pairs/syd-mel-rtt): 9.8 ms * [Tokyo to Melbourne latency and RTT](/docs/network/latency/pairs/tyo-mel-rtt): 113.9 ms * [Hong Kong to Melbourne latency and RTT](/docs/network/latency/pairs/hkg-mel-rtt): 138.4 ms * [Taipei to Melbourne latency and RTT](/docs/network/latency/pairs/tpe-mel-rtt): 142 ms * [Los Angeles to Melbourne latency and RTT](/docs/network/latency/pairs/lax-mel-rtt): 148.2 ms **Same corridor (Asia Pacific → Asia Pacific)** * [Melbourne to Sydney latency and RTT](/docs/network/latency/pairs/mel-syd-rtt): 9.8 ms * [Hong Kong to Taipei latency and RTT](/docs/network/latency/pairs/hkg-tpe-rtt): 14.7 ms * [Taipei to Hong Kong latency and RTT](/docs/network/latency/pairs/tpe-hkg-rtt): 15.5 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/sin-mel.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Singapore → Melbourne round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/sin-mel-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Singapore → Miami, USA RTT 🇸🇬 **Singapore (SIN)** → 🇺🇸 **Miami, USA (MIA)** Measurement round: `2026-08-16T04:07:28Z`. Singapore to Miami averaged 221.7 ms across 50 replies. The observed range was 212 to 252.74 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 8.4 ms, or 3.8% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **221.7 ms** | | Jitter | **6.57 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **166.27 ms** | | Fiber Efficiency | **75%** | | Latency Tier | Fair | | Source Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | | Destination Region | [North America](/docs/network/latency/regions/north-america) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Singapore](https://www.geonames.org/1880252) (1.28967, 103.85007) and [Miami](https://www.geonames.org/4164138) (25.77427, -80.19366). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | --------------- | | WGS-84 geodesic distance | **16,978.7 km** | | Vacuum RTT floor | **113.27 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **166.27 ms** | | Low-latency fiber material floor | **165.6 ms** | | Engineering floor (5% path allowance) | **174.6 ms** | | Research 1.33× mapped-fiber reference | **221.14 ms** | | Estimated unamplified path loss | **3565.5 dB** | | Transparent optical spans / inline amplifiers | **223 / 222** | | Published RTT inflation over fiber floor | **1.33×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **212 ms** | | Average RTT | **221.7 ms** | | Maximum RTT | **252.74 ms** | | Standard deviation | **8.4 ms** | | Stdev / average | **3.8%** | ## Route Context * Departure PoP: [Singapore (SIN)](/docs/network/latency/sin-singapore) * Destination PoP: [Miami (MIA)](/docs/network/latency/mia-miami) * Region overview: [Asia Pacific](/docs/network/latency/regions/asia-pacific) → [North America](/docs/network/latency/regions/north-america) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Miami to Singapore latency and RTT](/docs/network/latency/pairs/mia-sin-rtt): 223.1 ms **Fastest routes departing Singapore (SIN)** * [Singapore to Hong Kong latency and RTT](/docs/network/latency/pairs/sin-hkg-rtt): 30.8 ms * [Singapore to Taipei latency and RTT](/docs/network/latency/pairs/sin-tpe-rtt): 45.8 ms * [Singapore to Tokyo latency and RTT](/docs/network/latency/pairs/sin-tyo-rtt): 68.9 ms * [Singapore to Melbourne latency and RTT](/docs/network/latency/pairs/sin-mel-rtt): 88.6 ms * [Singapore to Sydney latency and RTT](/docs/network/latency/pairs/sin-syd-rtt): 94.5 ms **Fastest routes arriving at Miami (MIA)** * [Ashburn to Miami latency and RTT](/docs/network/latency/pairs/iad-mia-rtt): 27.4 ms * [New York to Miami latency and RTT](/docs/network/latency/pairs/nyc-mia-rtt): 32.9 ms * [Los Angeles to Miami latency and RTT](/docs/network/latency/pairs/lax-mia-rtt): 57.3 ms * [São Paulo to Miami latency and RTT](/docs/network/latency/pairs/gru-mia-rtt): 74 ms * [Seattle to Miami latency and RTT](/docs/network/latency/pairs/sea-mia-rtt): 81.7 ms **Same corridor (Asia Pacific → North America)** * [Tokyo to Seattle latency and RTT](/docs/network/latency/pairs/tyo-sea-rtt): 84.8 ms * [Tokyo to Los Angeles latency and RTT](/docs/network/latency/pairs/tyo-lax-rtt): 101.2 ms * [Taipei to Seattle latency and RTT](/docs/network/latency/pairs/tpe-sea-rtt): 115.8 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/sin-mia.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Singapore → Miami round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/sin-mia-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Singapore → Moscow, Russia RTT 🇸🇬 **Singapore (SIN)** → 🇷🇺 **Moscow, Russia (MOW)** Measurement round: `2026-08-16T04:07:28Z`. Singapore to Moscow averaged 147.7 ms across 50 replies. The observed range was 144.38 to 154.89 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 2.64 ms, or 1.8% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------------------- | | RTT | **147.7 ms** | | Jitter | **2.33 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **82.49 ms** | | Fiber Efficiency | **55.8%** | | Latency Tier | Good | | Source Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | | Destination Region | [Europe](/docs/network/latency/regions/europe) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Singapore](https://www.geonames.org/1880252) (1.28967, 103.85007) and [Moscow](https://www.geonames.org/524901) (55.75204, 37.61781). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **8,423.2 km** | | Vacuum RTT floor | **56.19 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **82.49 ms** | | Low-latency fiber material floor | **82.15 ms** | | Engineering floor (5% path allowance) | **86.62 ms** | | Research 1.33× mapped-fiber reference | **109.71 ms** | | Estimated unamplified path loss | **1768.9 dB** | | Transparent optical spans / inline amplifiers | **111 / 110** | | Published RTT inflation over fiber floor | **1.79×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **144.38 ms** | | Average RTT | **147.7 ms** | | Maximum RTT | **154.89 ms** | | Standard deviation | **2.64 ms** | | Stdev / average | **1.8%** | ## Route Context * Departure PoP: [Singapore (SIN)](/docs/network/latency/sin-singapore) * Destination PoP: [Moscow (MOW)](/docs/network/latency/mow-moscow) * Region overview: [Asia Pacific](/docs/network/latency/regions/asia-pacific) → [Europe](/docs/network/latency/regions/europe) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Moscow to Singapore latency and RTT](/docs/network/latency/pairs/mow-sin-rtt): 147.3 ms **Fastest routes departing Singapore (SIN)** * [Singapore to Hong Kong latency and RTT](/docs/network/latency/pairs/sin-hkg-rtt): 30.8 ms * [Singapore to Taipei latency and RTT](/docs/network/latency/pairs/sin-tpe-rtt): 45.8 ms * [Singapore to Tokyo latency and RTT](/docs/network/latency/pairs/sin-tyo-rtt): 68.9 ms * [Singapore to Melbourne latency and RTT](/docs/network/latency/pairs/sin-mel-rtt): 88.6 ms * [Singapore to Sydney latency and RTT](/docs/network/latency/pairs/sin-syd-rtt): 94.5 ms **Fastest routes arriving at Moscow (MOW)** * [Berlin to Moscow latency and RTT](/docs/network/latency/pairs/ber-mow-rtt): 28.6 ms * [Frankfurt to Moscow latency and RTT](/docs/network/latency/pairs/fra-mow-rtt): 34.7 ms * [Amsterdam to Moscow latency and RTT](/docs/network/latency/pairs/ams-mow-rtt): 38.1 ms * [London to Moscow latency and RTT](/docs/network/latency/pairs/lon-mow-rtt): 43.5 ms * [Paris to Moscow latency and RTT](/docs/network/latency/pairs/par-mow-rtt): 44.7 ms **Same corridor (Asia Pacific → Europe)** * [Hong Kong to Moscow latency and RTT](/docs/network/latency/pairs/hkg-mow-rtt): 118.2 ms * [Taipei to Moscow latency and RTT](/docs/network/latency/pairs/tpe-mow-rtt): 132.7 ms * [Singapore to Marseille latency and RTT](/docs/network/latency/pairs/sin-mrs-rtt): 140.6 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/sin-mow.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Singapore → Moscow round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/sin-mow-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Singapore → Marseille, France RTT 🇸🇬 **Singapore (SIN)** → 🇫🇷 **Marseille, France (MRS)** Measurement round: `2026-08-16T04:07:28Z`. Singapore to Marseille averaged 140.6 ms across 50 replies. The observed range was 138.21 to 150.15 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 2.29 ms, or 1.6% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------------------- | | RTT | **140.6 ms** | | Jitter | **1.84 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **103.79 ms** | | Fiber Efficiency | **73.8%** | | Latency Tier | Good | | Source Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | | Destination Region | [Europe](/docs/network/latency/regions/europe) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Singapore](https://www.geonames.org/1880252) (1.28967, 103.85007) and [Marseille](https://www.geonames.org/2995469) (43.29695, 5.38107). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | --------------- | | WGS-84 geodesic distance | **10,598.5 km** | | Vacuum RTT floor | **70.71 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **103.79 ms** | | Low-latency fiber material floor | **103.37 ms** | | Engineering floor (5% path allowance) | **108.99 ms** | | Research 1.33× mapped-fiber reference | **138.04 ms** | | Estimated unamplified path loss | **2225.7 dB** | | Transparent optical spans / inline amplifiers | **140 / 139** | | Published RTT inflation over fiber floor | **1.35×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **138.21 ms** | | Average RTT | **140.6 ms** | | Maximum RTT | **150.15 ms** | | Standard deviation | **2.29 ms** | | Stdev / average | **1.6%** | ## Route Context * Departure PoP: [Singapore (SIN)](/docs/network/latency/sin-singapore) * Destination PoP: [Marseille (MRS)](/docs/network/latency/mrs-marseille) * Region overview: [Asia Pacific](/docs/network/latency/regions/asia-pacific) → [Europe](/docs/network/latency/regions/europe) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Marseille to Singapore latency and RTT](/docs/network/latency/pairs/mrs-sin-rtt): 139.6 ms **Fastest routes departing Singapore (SIN)** * [Singapore to Hong Kong latency and RTT](/docs/network/latency/pairs/sin-hkg-rtt): 30.8 ms * [Singapore to Taipei latency and RTT](/docs/network/latency/pairs/sin-tpe-rtt): 45.8 ms * [Singapore to Tokyo latency and RTT](/docs/network/latency/pairs/sin-tyo-rtt): 68.9 ms * [Singapore to Melbourne latency and RTT](/docs/network/latency/pairs/sin-mel-rtt): 88.6 ms * [Singapore to Sydney latency and RTT](/docs/network/latency/pairs/sin-syd-rtt): 94.5 ms **Fastest routes arriving at Marseille (MRS)** * [Paris to Marseille latency and RTT](/docs/network/latency/pairs/par-mrs-rtt): 8.9 ms * [Frankfurt to Marseille latency and RTT](/docs/network/latency/pairs/fra-mrs-rtt): 16.4 ms * [London to Marseille latency and RTT](/docs/network/latency/pairs/lon-mrs-rtt): 17.9 ms * [Amsterdam to Marseille latency and RTT](/docs/network/latency/pairs/ams-mrs-rtt): 19.6 ms * [Berlin to Marseille latency and RTT](/docs/network/latency/pairs/ber-mrs-rtt): 20.2 ms **Same corridor (Asia Pacific → Europe)** * [Hong Kong to Moscow latency and RTT](/docs/network/latency/pairs/hkg-mow-rtt): 118.2 ms * [Taipei to Moscow latency and RTT](/docs/network/latency/pairs/tpe-mow-rtt): 132.7 ms * [Hong Kong to Berlin latency and RTT](/docs/network/latency/pairs/hkg-ber-rtt): 145.1 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/sin-mrs.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Singapore → Marseille round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/sin-mrs-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Singapore → New York, USA RTT 🇸🇬 **Singapore (SIN)** → 🇺🇸 **New York, USA (NYC)** Measurement round: `2026-08-16T04:07:28Z`. Singapore to New York averaged 209.3 ms across 50 replies. The observed range was 199.54 to 235.91 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 8.69 ms, or 4.2% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **209.3 ms** | | Jitter | **7.95 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **150.29 ms** | | Fiber Efficiency | **71.8%** | | Latency Tier | Fair | | Source Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | | Destination Region | [North America](/docs/network/latency/regions/north-america) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Singapore](https://www.geonames.org/1880252) (1.28967, 103.85007) and [New York City](https://www.geonames.org/5128581) (40.71427, -74.00597). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | --------------- | | WGS-84 geodesic distance | **15,347.5 km** | | Vacuum RTT floor | **102.39 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **150.29 ms** | | Low-latency fiber material floor | **149.69 ms** | | Engineering floor (5% path allowance) | **157.83 ms** | | Research 1.33× mapped-fiber reference | **199.89 ms** | | Estimated unamplified path loss | **3223 dB** | | Transparent optical spans / inline amplifiers | **202 / 201** | | Published RTT inflation over fiber floor | **1.39×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **199.54 ms** | | Average RTT | **209.3 ms** | | Maximum RTT | **235.91 ms** | | Standard deviation | **8.69 ms** | | Stdev / average | **4.2%** | ## Route Context * Departure PoP: [Singapore (SIN)](/docs/network/latency/sin-singapore) * Destination PoP: [New York (NYC)](/docs/network/latency/nyc-new-york) * Region overview: [Asia Pacific](/docs/network/latency/regions/asia-pacific) → [North America](/docs/network/latency/regions/north-america) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [New York to Singapore latency and RTT](/docs/network/latency/pairs/nyc-sin-rtt): 208.7 ms **Fastest routes departing Singapore (SIN)** * [Singapore to Hong Kong latency and RTT](/docs/network/latency/pairs/sin-hkg-rtt): 30.8 ms * [Singapore to Taipei latency and RTT](/docs/network/latency/pairs/sin-tpe-rtt): 45.8 ms * [Singapore to Tokyo latency and RTT](/docs/network/latency/pairs/sin-tyo-rtt): 68.9 ms * [Singapore to Melbourne latency and RTT](/docs/network/latency/pairs/sin-mel-rtt): 88.6 ms * [Singapore to Sydney latency and RTT](/docs/network/latency/pairs/sin-syd-rtt): 94.5 ms **Fastest routes arriving at New York (NYC)** * [Ashburn to New York latency and RTT](/docs/network/latency/pairs/iad-nyc-rtt): 6 ms * [Miami to New York latency and RTT](/docs/network/latency/pairs/mia-nyc-rtt): 33.3 ms * [Los Angeles to New York latency and RTT](/docs/network/latency/pairs/lax-nyc-rtt): 57.9 ms * [Seattle to New York latency and RTT](/docs/network/latency/pairs/sea-nyc-rtt): 59.5 ms * [London to New York latency and RTT](/docs/network/latency/pairs/lon-nyc-rtt): 63.8 ms **Same corridor (Asia Pacific → North America)** * [Tokyo to Seattle latency and RTT](/docs/network/latency/pairs/tyo-sea-rtt): 84.8 ms * [Tokyo to Los Angeles latency and RTT](/docs/network/latency/pairs/tyo-lax-rtt): 101.2 ms * [Taipei to Seattle latency and RTT](/docs/network/latency/pairs/tpe-sea-rtt): 115.8 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/sin-nyc.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Singapore → New York round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/sin-nyc-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Singapore → Paris, France RTT 🇸🇬 **Singapore (SIN)** → 🇫🇷 **Paris, France (PAR)** Measurement round: `2026-08-16T04:07:28Z`. Singapore to Paris averaged 151.4 ms across 50 replies. The observed range was 142.1 to 172.35 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 6.94 ms, or 4.6% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------------------- | | RTT | **151.4 ms** | | Jitter | **6.44 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **105.19 ms** | | Fiber Efficiency | **69.5%** | | Latency Tier | Fair | | Source Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | | Destination Region | [Europe](/docs/network/latency/regions/europe) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Singapore](https://www.geonames.org/1880252) (1.28967, 103.85007) and [Paris](https://www.geonames.org/2988507) (48.85341, 2.34880). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | --------------- | | WGS-84 geodesic distance | **10,741.6 km** | | Vacuum RTT floor | **71.66 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **105.19 ms** | | Low-latency fiber material floor | **104.77 ms** | | Engineering floor (5% path allowance) | **110.46 ms** | | Research 1.33× mapped-fiber reference | **139.9 ms** | | Estimated unamplified path loss | **2255.7 dB** | | Transparent optical spans / inline amplifiers | **141 / 140** | | Published RTT inflation over fiber floor | **1.44×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **142.1 ms** | | Average RTT | **151.4 ms** | | Maximum RTT | **172.35 ms** | | Standard deviation | **6.94 ms** | | Stdev / average | **4.6%** | ## Route Context * Departure PoP: [Singapore (SIN)](/docs/network/latency/sin-singapore) * Destination PoP: [Paris (PAR)](/docs/network/latency/par-paris) * Region overview: [Asia Pacific](/docs/network/latency/regions/asia-pacific) → [Europe](/docs/network/latency/regions/europe) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Paris to Singapore latency and RTT](/docs/network/latency/pairs/par-sin-rtt): 150.3 ms **Fastest routes departing Singapore (SIN)** * [Singapore to Hong Kong latency and RTT](/docs/network/latency/pairs/sin-hkg-rtt): 30.8 ms * [Singapore to Taipei latency and RTT](/docs/network/latency/pairs/sin-tpe-rtt): 45.8 ms * [Singapore to Tokyo latency and RTT](/docs/network/latency/pairs/sin-tyo-rtt): 68.9 ms * [Singapore to Melbourne latency and RTT](/docs/network/latency/pairs/sin-mel-rtt): 88.6 ms * [Singapore to Sydney latency and RTT](/docs/network/latency/pairs/sin-syd-rtt): 94.5 ms **Fastest routes arriving at Paris (PAR)** * [London to Paris latency and RTT](/docs/network/latency/pairs/lon-par-rtt): 6.4 ms * [Amsterdam to Paris latency and RTT](/docs/network/latency/pairs/ams-par-rtt): 7 ms * [Frankfurt to Paris latency and RTT](/docs/network/latency/pairs/fra-par-rtt): 7.6 ms * [Marseille to Paris latency and RTT](/docs/network/latency/pairs/mrs-par-rtt): 8.9 ms * [Berlin to Paris latency and RTT](/docs/network/latency/pairs/ber-par-rtt): 15.6 ms **Same corridor (Asia Pacific → Europe)** * [Hong Kong to Moscow latency and RTT](/docs/network/latency/pairs/hkg-mow-rtt): 118.2 ms * [Taipei to Moscow latency and RTT](/docs/network/latency/pairs/tpe-mow-rtt): 132.7 ms * [Singapore to Marseille latency and RTT](/docs/network/latency/pairs/sin-mrs-rtt): 140.6 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/sin-par.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Singapore → Paris round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/sin-par-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Singapore → Seattle, USA RTT 🇸🇬 **Singapore (SIN)** → 🇺🇸 **Seattle, USA (SEA)** Measurement round: `2026-08-16T04:07:28Z`. Singapore to Seattle averaged 151.5 ms across 50 replies. The observed range was 141.74 to 170.28 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 7.34 ms, or 4.8% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **151.5 ms** | | Jitter | **7.22 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **127.22 ms** | | Fiber Efficiency | **84%** | | Latency Tier | Fair | | Source Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | | Destination Region | [North America](/docs/network/latency/regions/north-america) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Singapore](https://www.geonames.org/1880252) (1.28967, 103.85007) and [Seattle](https://www.geonames.org/5809844) (47.60621, -122.33207). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | --------------- | | WGS-84 geodesic distance | **12,991.1 km** | | Vacuum RTT floor | **86.67 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **127.22 ms** | | Low-latency fiber material floor | **126.71 ms** | | Engineering floor (5% path allowance) | **133.6 ms** | | Research 1.33× mapped-fiber reference | **169.2 ms** | | Estimated unamplified path loss | **2728.1 dB** | | Transparent optical spans / inline amplifiers | **171 / 170** | | Published RTT inflation over fiber floor | **1.19×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **141.74 ms** | | Average RTT | **151.5 ms** | | Maximum RTT | **170.28 ms** | | Standard deviation | **7.34 ms** | | Stdev / average | **4.8%** | ## Route Context * Departure PoP: [Singapore (SIN)](/docs/network/latency/sin-singapore) * Destination PoP: [Seattle (SEA)](/docs/network/latency/sea-seattle) * Region overview: [Asia Pacific](/docs/network/latency/regions/asia-pacific) → [North America](/docs/network/latency/regions/north-america) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Seattle to Singapore latency and RTT](/docs/network/latency/pairs/sea-sin-rtt): 151.1 ms **Fastest routes departing Singapore (SIN)** * [Singapore to Hong Kong latency and RTT](/docs/network/latency/pairs/sin-hkg-rtt): 30.8 ms * [Singapore to Taipei latency and RTT](/docs/network/latency/pairs/sin-tpe-rtt): 45.8 ms * [Singapore to Tokyo latency and RTT](/docs/network/latency/pairs/sin-tyo-rtt): 68.9 ms * [Singapore to Melbourne latency and RTT](/docs/network/latency/pairs/sin-mel-rtt): 88.6 ms * [Singapore to Sydney latency and RTT](/docs/network/latency/pairs/sin-syd-rtt): 94.5 ms **Fastest routes arriving at Seattle (SEA)** * [Los Angeles to Seattle latency and RTT](/docs/network/latency/pairs/lax-sea-rtt): 26.9 ms * [New York to Seattle latency and RTT](/docs/network/latency/pairs/nyc-sea-rtt): 58.7 ms * [Ashburn to Seattle latency and RTT](/docs/network/latency/pairs/iad-sea-rtt): 62.3 ms * [Miami to Seattle latency and RTT](/docs/network/latency/pairs/mia-sea-rtt): 81.9 ms * [Tokyo to Seattle latency and RTT](/docs/network/latency/pairs/tyo-sea-rtt): 84.8 ms **Same corridor (Asia Pacific → North America)** * [Tokyo to Los Angeles latency and RTT](/docs/network/latency/pairs/tyo-lax-rtt): 101.2 ms * [Taipei to Seattle latency and RTT](/docs/network/latency/pairs/tpe-sea-rtt): 115.8 ms * [Hong Kong to Seattle latency and RTT](/docs/network/latency/pairs/hkg-sea-rtt): 131.6 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/sin-sea.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Singapore → Seattle round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/sin-sea-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Singapore → Sydney, Australia RTT 🇸🇬 **Singapore (SIN)** → 🇦🇺 **Sydney, Australia (SYD)** Measurement round: `2026-08-16T04:07:28Z`. Singapore to Sydney averaged 94.5 ms across 50 replies. The observed range was 92.37 to 99.4 ms; these figures describe one probe round, not a long-term service target. This is number 5 of 19 outgoing routes by average RTT. That ranking compares different destinations and is not adjusted for distance. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------------------- | | RTT | **94.5 ms** | | Jitter | **1.47 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **61.63 ms** | | Fiber Efficiency | **65.2%** | | Latency Tier | Good | | Source Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | | Destination Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Singapore](https://www.geonames.org/1880252) (1.28967, 103.85007) and [Sydney](https://www.geonames.org/2147714) (-33.86785, 151.20732). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **6,293.5 km** | | Vacuum RTT floor | **41.99 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **61.63 ms** | | Low-latency fiber material floor | **61.38 ms** | | Engineering floor (5% path allowance) | **64.72 ms** | | Research 1.33× mapped-fiber reference | **81.97 ms** | | Estimated unamplified path loss | **1321.6 dB** | | Transparent optical spans / inline amplifiers | **83 / 82** | | Published RTT inflation over fiber floor | **1.53×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------ | | Minimum RTT | **92.37 ms** | | Average RTT | **94.5 ms** | | Maximum RTT | **99.4 ms** | | Standard deviation | **1.76 ms** | | Stdev / average | **1.9%** | ## Route Context * Departure PoP: [Singapore (SIN)](/docs/network/latency/sin-singapore) * Destination PoP: [Sydney (SYD)](/docs/network/latency/syd-sydney) * Region overview: [Asia Pacific](/docs/network/latency/regions/asia-pacific) → [Asia Pacific](/docs/network/latency/regions/asia-pacific) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Sydney to Singapore latency and RTT](/docs/network/latency/pairs/syd-sin-rtt): 94.5 ms **Fastest routes departing Singapore (SIN)** * [Singapore to Hong Kong latency and RTT](/docs/network/latency/pairs/sin-hkg-rtt): 30.8 ms * [Singapore to Taipei latency and RTT](/docs/network/latency/pairs/sin-tpe-rtt): 45.8 ms * [Singapore to Tokyo latency and RTT](/docs/network/latency/pairs/sin-tyo-rtt): 68.9 ms * [Singapore to Melbourne latency and RTT](/docs/network/latency/pairs/sin-mel-rtt): 88.6 ms * [Singapore to Marseille latency and RTT](/docs/network/latency/pairs/sin-mrs-rtt): 140.6 ms **Fastest routes arriving at Sydney (SYD)** * [Melbourne to Sydney latency and RTT](/docs/network/latency/pairs/mel-syd-rtt): 9.8 ms * [Tokyo to Sydney latency and RTT](/docs/network/latency/pairs/tyo-syd-rtt): 101.9 ms * [Taipei to Sydney latency and RTT](/docs/network/latency/pairs/tpe-syd-rtt): 133.2 ms * [Hong Kong to Sydney latency and RTT](/docs/network/latency/pairs/hkg-syd-rtt): 137.2 ms * [Los Angeles to Sydney latency and RTT](/docs/network/latency/pairs/lax-syd-rtt): 137.6 ms **Same corridor (Asia Pacific → Asia Pacific)** * [Sydney to Melbourne latency and RTT](/docs/network/latency/pairs/syd-mel-rtt): 9.8 ms * [Hong Kong to Taipei latency and RTT](/docs/network/latency/pairs/hkg-tpe-rtt): 14.7 ms * [Taipei to Hong Kong latency and RTT](/docs/network/latency/pairs/tpe-hkg-rtt): 15.5 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/sin-syd.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Singapore → Sydney round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/sin-syd-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Singapore → Taipei, Taiwan RTT 🇸🇬 **Singapore (SIN)** → 🇹🇼 **Taipei, Taiwan (TPE)** Measurement round: `2026-08-16T04:07:28Z`. Singapore to Taipei averaged 45.8 ms across 50 replies. The observed range was 43.12 to 50.4 ms; these figures describe one probe round, not a long-term service target. This is number 2 of 19 outgoing routes by average RTT. That ranking compares different destinations and is not adjusted for distance. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------------------- | | RTT | **45.8 ms** | | Jitter | **1.86 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **31.77 ms** | | Fiber Efficiency | **69.4%** | | Latency Tier | Excellent | | Source Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | | Destination Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Singapore](https://www.geonames.org/1880252) (1.28967, 103.85007) and [Taipei](https://www.geonames.org/1668341) (25.05306, 121.52639). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **3,243.9 km** | | Vacuum RTT floor | **21.64 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **31.77 ms** | | Low-latency fiber material floor | **31.64 ms** | | Engineering floor (5% path allowance) | **33.36 ms** | | Research 1.33× mapped-fiber reference | **42.25 ms** | | Estimated unamplified path loss | **681.2 dB** | | Transparent optical spans / inline amplifiers | **43 / 42** | | Published RTT inflation over fiber floor | **1.44×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------ | | Minimum RTT | **43.12 ms** | | Average RTT | **45.8 ms** | | Maximum RTT | **50.4 ms** | | Standard deviation | **1.72 ms** | | Stdev / average | **3.8%** | ## Route Context * Departure PoP: [Singapore (SIN)](/docs/network/latency/sin-singapore) * Destination PoP: [Taipei (TPE)](/docs/network/latency/tpe-taipei) * Region overview: [Asia Pacific](/docs/network/latency/regions/asia-pacific) → [Asia Pacific](/docs/network/latency/regions/asia-pacific) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Taipei to Singapore latency and RTT](/docs/network/latency/pairs/tpe-sin-rtt): 44.3 ms **Fastest routes departing Singapore (SIN)** * [Singapore to Hong Kong latency and RTT](/docs/network/latency/pairs/sin-hkg-rtt): 30.8 ms * [Singapore to Tokyo latency and RTT](/docs/network/latency/pairs/sin-tyo-rtt): 68.9 ms * [Singapore to Melbourne latency and RTT](/docs/network/latency/pairs/sin-mel-rtt): 88.6 ms * [Singapore to Sydney latency and RTT](/docs/network/latency/pairs/sin-syd-rtt): 94.5 ms * [Singapore to Marseille latency and RTT](/docs/network/latency/pairs/sin-mrs-rtt): 140.6 ms **Fastest routes arriving at Taipei (TPE)** * [Hong Kong to Taipei latency and RTT](/docs/network/latency/pairs/hkg-tpe-rtt): 14.7 ms * [Tokyo to Taipei latency and RTT](/docs/network/latency/pairs/tyo-tpe-rtt): 32.3 ms * [Seattle to Taipei latency and RTT](/docs/network/latency/pairs/sea-tpe-rtt): 114.7 ms * [Sydney to Taipei latency and RTT](/docs/network/latency/pairs/syd-tpe-rtt): 131.7 ms * [Moscow to Taipei latency and RTT](/docs/network/latency/pairs/mow-tpe-rtt): 132.1 ms **Same corridor (Asia Pacific → Asia Pacific)** * [Melbourne to Sydney latency and RTT](/docs/network/latency/pairs/mel-syd-rtt): 9.8 ms * [Sydney to Melbourne latency and RTT](/docs/network/latency/pairs/syd-mel-rtt): 9.8 ms * [Taipei to Hong Kong latency and RTT](/docs/network/latency/pairs/tpe-hkg-rtt): 15.5 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/sin-tpe.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Singapore → Taipei round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/sin-tpe-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Singapore → Tokyo, Japan RTT 🇸🇬 **Singapore (SIN)** → 🇯🇵 **Tokyo, Japan (TYO)** Measurement round: `2026-08-16T04:07:28Z`. Singapore to Tokyo averaged 68.9 ms across 50 replies. The observed range was 65.47 to 78.16 ms; these figures describe one probe round, not a long-term service target. This is number 3 of 19 outgoing routes by average RTT. That ranking compares different destinations and is not adjusted for distance. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------------------- | | RTT | **68.9 ms** | | Jitter | **1.99 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **52 ms** | | Fiber Efficiency | **75.5%** | | Latency Tier | Excellent | | Source Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | | Destination Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Singapore](https://www.geonames.org/1880252) (1.28967, 103.85007) and [Tokyo](https://www.geonames.org/1850147) (35.68950, 139.69171). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **5,310.2 km** | | Vacuum RTT floor | **35.43 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **52 ms** | | Low-latency fiber material floor | **51.79 ms** | | Engineering floor (5% path allowance) | **54.61 ms** | | Research 1.33× mapped-fiber reference | **69.16 ms** | | Estimated unamplified path loss | **1115.1 dB** | | Transparent optical spans / inline amplifiers | **70 / 69** | | Published RTT inflation over fiber floor | **1.32×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------ | | Minimum RTT | **65.47 ms** | | Average RTT | **68.9 ms** | | Maximum RTT | **78.16 ms** | | Standard deviation | **2.72 ms** | | Stdev / average | **3.9%** | ## Route Context * Departure PoP: [Singapore (SIN)](/docs/network/latency/sin-singapore) * Destination PoP: [Tokyo (TYO)](/docs/network/latency/tyo-tokyo) * Region overview: [Asia Pacific](/docs/network/latency/regions/asia-pacific) → [Asia Pacific](/docs/network/latency/regions/asia-pacific) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Tokyo to Singapore latency and RTT](/docs/network/latency/pairs/tyo-sin-rtt): 69.4 ms **Fastest routes departing Singapore (SIN)** * [Singapore to Hong Kong latency and RTT](/docs/network/latency/pairs/sin-hkg-rtt): 30.8 ms * [Singapore to Taipei latency and RTT](/docs/network/latency/pairs/sin-tpe-rtt): 45.8 ms * [Singapore to Melbourne latency and RTT](/docs/network/latency/pairs/sin-mel-rtt): 88.6 ms * [Singapore to Sydney latency and RTT](/docs/network/latency/pairs/sin-syd-rtt): 94.5 ms * [Singapore to Marseille latency and RTT](/docs/network/latency/pairs/sin-mrs-rtt): 140.6 ms **Fastest routes arriving at Tokyo (TYO)** * [Taipei to Tokyo latency and RTT](/docs/network/latency/pairs/tpe-tyo-rtt): 31.9 ms * [Hong Kong to Tokyo latency and RTT](/docs/network/latency/pairs/hkg-tyo-rtt): 44.9 ms * [Seattle to Tokyo latency and RTT](/docs/network/latency/pairs/sea-tyo-rtt): 85 ms * [Los Angeles to Tokyo latency and RTT](/docs/network/latency/pairs/lax-tyo-rtt): 101.2 ms * [Sydney to Tokyo latency and RTT](/docs/network/latency/pairs/syd-tyo-rtt): 101.3 ms **Same corridor (Asia Pacific → Asia Pacific)** * [Melbourne to Sydney latency and RTT](/docs/network/latency/pairs/mel-syd-rtt): 9.8 ms * [Sydney to Melbourne latency and RTT](/docs/network/latency/pairs/syd-mel-rtt): 9.8 ms * [Hong Kong to Taipei latency and RTT](/docs/network/latency/pairs/hkg-tpe-rtt): 14.7 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/sin-tyo.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Singapore → Tokyo round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/sin-tyo-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Sydney, Australia → Amsterdam, Netherlands RTT 🇦🇺 **Sydney, Australia (SYD)** → 🇳🇱 **Amsterdam, Netherlands (AMS)** Measurement round: `2026-08-16T04:07:28Z`. Sydney to Amsterdam averaged 258.9 ms across 50 replies. The observed range was 249.45 to 283.53 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 7.86 ms, or 3.0% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------------------- | | RTT | **258.9 ms** | | Jitter | **6.38 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **162.93 ms** | | Fiber Efficiency | **62.9%** | | Latency Tier | High | | Source Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | | Destination Region | [Europe](/docs/network/latency/regions/europe) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Sydney](https://www.geonames.org/2147714) (-33.86785, 151.20732) and [Amsterdam](https://www.geonames.org/2759794) (52.37403, 4.88969). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | --------------- | | WGS-84 geodesic distance | **16,638.1 km** | | Vacuum RTT floor | **111 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **162.93 ms** | | Low-latency fiber material floor | **162.28 ms** | | Engineering floor (5% path allowance) | **171.1 ms** | | Research 1.33× mapped-fiber reference | **216.7 ms** | | Estimated unamplified path loss | **3494 dB** | | Transparent optical spans / inline amplifiers | **219 / 218** | | Published RTT inflation over fiber floor | **1.59×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **249.45 ms** | | Average RTT | **258.9 ms** | | Maximum RTT | **283.53 ms** | | Standard deviation | **7.86 ms** | | Stdev / average | **3.0%** | ## Route Context * Departure PoP: [Sydney (SYD)](/docs/network/latency/syd-sydney) * Destination PoP: [Amsterdam (AMS)](/docs/network/latency/ams-amsterdam) * Region overview: [Asia Pacific](/docs/network/latency/regions/asia-pacific) → [Europe](/docs/network/latency/regions/europe) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Amsterdam to Sydney latency and RTT](/docs/network/latency/pairs/ams-syd-rtt): 258 ms **Fastest routes departing Sydney (SYD)** * [Sydney to Melbourne latency and RTT](/docs/network/latency/pairs/syd-mel-rtt): 9.8 ms * [Sydney to Singapore latency and RTT](/docs/network/latency/pairs/syd-sin-rtt): 94.5 ms * [Sydney to Tokyo latency and RTT](/docs/network/latency/pairs/syd-tyo-rtt): 101.3 ms * [Sydney to Taipei latency and RTT](/docs/network/latency/pairs/syd-tpe-rtt): 131.7 ms * [Sydney to Hong Kong latency and RTT](/docs/network/latency/pairs/syd-hkg-rtt): 135.7 ms **Fastest routes arriving at Amsterdam (AMS)** * [London to Amsterdam latency and RTT](/docs/network/latency/pairs/lon-ams-rtt): 5.2 ms * [Frankfurt to Amsterdam latency and RTT](/docs/network/latency/pairs/fra-ams-rtt): 5.9 ms * [Paris to Amsterdam latency and RTT](/docs/network/latency/pairs/par-ams-rtt): 7 ms * [Berlin to Amsterdam latency and RTT](/docs/network/latency/pairs/ber-ams-rtt): 7.2 ms * [Marseille to Amsterdam latency and RTT](/docs/network/latency/pairs/mrs-ams-rtt): 20.3 ms **Same corridor (Asia Pacific → Europe)** * [Hong Kong to Moscow latency and RTT](/docs/network/latency/pairs/hkg-mow-rtt): 118.2 ms * [Taipei to Moscow latency and RTT](/docs/network/latency/pairs/tpe-mow-rtt): 132.7 ms * [Singapore to Marseille latency and RTT](/docs/network/latency/pairs/sin-mrs-rtt): 140.6 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/syd-ams.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Sydney → Amsterdam round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/syd-ams-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Sydney, Australia → Berlin, Germany RTT 🇦🇺 **Sydney, Australia (SYD)** → 🇩🇪 **Berlin, Germany (BER)** Measurement round: `2026-08-16T04:07:28Z`. Sydney to Berlin averaged 260.5 ms across 50 replies. The observed range was 249.51 to 293.46 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 9.56 ms, or 3.7% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------------------- | | RTT | **260.5 ms** | | Jitter | **8.36 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **157.54 ms** | | Fiber Efficiency | **60.5%** | | Latency Tier | High | | Source Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | | Destination Region | [Europe](/docs/network/latency/regions/europe) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Sydney](https://www.geonames.org/2147714) (-33.86785, 151.20732) and [Berlin](https://www.geonames.org/2950159) (52.52437, 13.41053). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | --------------- | | WGS-84 geodesic distance | **16,087.7 km** | | Vacuum RTT floor | **107.33 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **157.54 ms** | | Low-latency fiber material floor | **156.91 ms** | | Engineering floor (5% path allowance) | **165.44 ms** | | Research 1.33× mapped-fiber reference | **209.53 ms** | | Estimated unamplified path loss | **3378.4 dB** | | Transparent optical spans / inline amplifiers | **212 / 211** | | Published RTT inflation over fiber floor | **1.65×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **249.51 ms** | | Average RTT | **260.5 ms** | | Maximum RTT | **293.46 ms** | | Standard deviation | **9.56 ms** | | Stdev / average | **3.7%** | ## Route Context * Departure PoP: [Sydney (SYD)](/docs/network/latency/syd-sydney) * Destination PoP: [Berlin (BER)](/docs/network/latency/ber-berlin) * Region overview: [Asia Pacific](/docs/network/latency/regions/asia-pacific) → [Europe](/docs/network/latency/regions/europe) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Berlin to Sydney latency and RTT](/docs/network/latency/pairs/ber-syd-rtt): 262.1 ms **Fastest routes departing Sydney (SYD)** * [Sydney to Melbourne latency and RTT](/docs/network/latency/pairs/syd-mel-rtt): 9.8 ms * [Sydney to Singapore latency and RTT](/docs/network/latency/pairs/syd-sin-rtt): 94.5 ms * [Sydney to Tokyo latency and RTT](/docs/network/latency/pairs/syd-tyo-rtt): 101.3 ms * [Sydney to Taipei latency and RTT](/docs/network/latency/pairs/syd-tpe-rtt): 131.7 ms * [Sydney to Hong Kong latency and RTT](/docs/network/latency/pairs/syd-hkg-rtt): 135.7 ms **Fastest routes arriving at Berlin (BER)** * [Frankfurt to Berlin latency and RTT](/docs/network/latency/pairs/fra-ber-rtt): 6.1 ms * [Amsterdam to Berlin latency and RTT](/docs/network/latency/pairs/ams-ber-rtt): 7.2 ms * [Paris to Berlin latency and RTT](/docs/network/latency/pairs/par-ber-rtt): 15.5 ms * [London to Berlin latency and RTT](/docs/network/latency/pairs/lon-ber-rtt): 16.2 ms * [Marseille to Berlin latency and RTT](/docs/network/latency/pairs/mrs-ber-rtt): 20.6 ms **Same corridor (Asia Pacific → Europe)** * [Hong Kong to Moscow latency and RTT](/docs/network/latency/pairs/hkg-mow-rtt): 118.2 ms * [Taipei to Moscow latency and RTT](/docs/network/latency/pairs/tpe-mow-rtt): 132.7 ms * [Singapore to Marseille latency and RTT](/docs/network/latency/pairs/sin-mrs-rtt): 140.6 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/syd-ber.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Sydney → Berlin round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/syd-ber-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Sydney, Australia → Frankfurt, Germany RTT 🇦🇺 **Sydney, Australia (SYD)** → 🇩🇪 **Frankfurt, Germany (FRA)** Measurement round: `2026-08-16T04:07:28Z`. Sydney to Frankfurt averaged 254.4 ms across 50 replies. The observed range was 239.27 to 292.11 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 11.5 ms, or 4.5% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------------------- | | RTT | **254.4 ms** | | Jitter | **9.95 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **161.37 ms** | | Fiber Efficiency | **63.4%** | | Latency Tier | High | | Source Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | | Destination Region | [Europe](/docs/network/latency/regions/europe) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Sydney](https://www.geonames.org/2147714) (-33.86785, 151.20732) and [Frankfurt am Main](https://www.geonames.org/2925533) (50.11552, 8.68417). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | --------------- | | WGS-84 geodesic distance | **16,478.2 km** | | Vacuum RTT floor | **109.93 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **161.37 ms** | | Low-latency fiber material floor | **160.72 ms** | | Engineering floor (5% path allowance) | **169.46 ms** | | Research 1.33× mapped-fiber reference | **214.62 ms** | | Estimated unamplified path loss | **3460.4 dB** | | Transparent optical spans / inline amplifiers | **217 / 216** | | Published RTT inflation over fiber floor | **1.58×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **239.27 ms** | | Average RTT | **254.4 ms** | | Maximum RTT | **292.11 ms** | | Standard deviation | **11.5 ms** | | Stdev / average | **4.5%** | ## Route Context * Departure PoP: [Sydney (SYD)](/docs/network/latency/syd-sydney) * Destination PoP: [Frankfurt (FRA)](/docs/network/latency/fra-frankfurt) * Region overview: [Asia Pacific](/docs/network/latency/regions/asia-pacific) → [Europe](/docs/network/latency/regions/europe) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Frankfurt to Sydney latency and RTT](/docs/network/latency/pairs/fra-syd-rtt): 256 ms **Fastest routes departing Sydney (SYD)** * [Sydney to Melbourne latency and RTT](/docs/network/latency/pairs/syd-mel-rtt): 9.8 ms * [Sydney to Singapore latency and RTT](/docs/network/latency/pairs/syd-sin-rtt): 94.5 ms * [Sydney to Tokyo latency and RTT](/docs/network/latency/pairs/syd-tyo-rtt): 101.3 ms * [Sydney to Taipei latency and RTT](/docs/network/latency/pairs/syd-tpe-rtt): 131.7 ms * [Sydney to Hong Kong latency and RTT](/docs/network/latency/pairs/syd-hkg-rtt): 135.7 ms **Fastest routes arriving at Frankfurt (FRA)** * [Amsterdam to Frankfurt latency and RTT](/docs/network/latency/pairs/ams-fra-rtt): 5.9 ms * [Berlin to Frankfurt latency and RTT](/docs/network/latency/pairs/ber-fra-rtt): 6.1 ms * [Paris to Frankfurt latency and RTT](/docs/network/latency/pairs/par-fra-rtt): 7.6 ms * [London to Frankfurt latency and RTT](/docs/network/latency/pairs/lon-fra-rtt): 13.6 ms * [Marseille to Frankfurt latency and RTT](/docs/network/latency/pairs/mrs-fra-rtt): 16 ms **Same corridor (Asia Pacific → Europe)** * [Hong Kong to Moscow latency and RTT](/docs/network/latency/pairs/hkg-mow-rtt): 118.2 ms * [Taipei to Moscow latency and RTT](/docs/network/latency/pairs/tpe-mow-rtt): 132.7 ms * [Singapore to Marseille latency and RTT](/docs/network/latency/pairs/sin-mrs-rtt): 140.6 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/syd-fra.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Sydney → Frankfurt round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/syd-fra-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Sydney, Australia → São Paulo, Brazil RTT 🇦🇺 **Sydney, Australia (SYD)** → 🇧🇷 **São Paulo, Brazil (GRU)** Measurement round: `2026-08-16T04:07:28Z`. Sydney to São Paulo averaged 296.7 ms, compared with a 131 ms city-centre fiber reference. The difference cannot be assigned to a particular cable or carrier from RTT alone. Standard deviation was 13.03 ms, or 4.4% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **296.7 ms** | | Jitter | **10.12 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **131 ms** | | Fiber Efficiency | **44.2%** | | Latency Tier | High | | Source Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | | Destination Region | [South America](/docs/network/latency/regions/south-america) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Sydney](https://www.geonames.org/2147714) (-33.86785, 151.20732) and [São Paulo](https://www.geonames.org/3448439) (-23.54750, -46.63611). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | ------------- | | WGS-84 geodesic distance | **13,377 km** | | Vacuum RTT floor | **89.24 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **131 ms** | | Low-latency fiber material floor | **130.47 ms** | | Engineering floor (5% path allowance) | **137.57 ms** | | Research 1.33× mapped-fiber reference | **174.23 ms** | | Estimated unamplified path loss | **2809.2 dB** | | Transparent optical spans / inline amplifiers | **176 / 175** | | Published RTT inflation over fiber floor | **2.26×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **278.26 ms** | | Average RTT | **296.7 ms** | | Maximum RTT | **337.27 ms** | | Standard deviation | **13.03 ms** | | Stdev / average | **4.4%** | ## Route Context * Departure PoP: [Sydney (SYD)](/docs/network/latency/syd-sydney) * Destination PoP: [São Paulo (GRU)](/docs/network/latency/gru-sao-paulo) * Region overview: [Asia Pacific](/docs/network/latency/regions/asia-pacific) → [South America](/docs/network/latency/regions/south-america) * Corridor overview: [Asia Pacific to South America](/docs/network/latency/regions/asia-pacific-to-south-america) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [São Paulo to Sydney latency and RTT](/docs/network/latency/pairs/gru-syd-rtt): 265.5 ms **Fastest routes departing Sydney (SYD)** * [Sydney to Melbourne latency and RTT](/docs/network/latency/pairs/syd-mel-rtt): 9.8 ms * [Sydney to Singapore latency and RTT](/docs/network/latency/pairs/syd-sin-rtt): 94.5 ms * [Sydney to Tokyo latency and RTT](/docs/network/latency/pairs/syd-tyo-rtt): 101.3 ms * [Sydney to Taipei latency and RTT](/docs/network/latency/pairs/syd-tpe-rtt): 131.7 ms * [Sydney to Hong Kong latency and RTT](/docs/network/latency/pairs/syd-hkg-rtt): 135.7 ms **Fastest routes arriving at São Paulo (GRU)** * [Ashburn to São Paulo latency and RTT](/docs/network/latency/pairs/iad-gru-rtt): 101.4 ms * [New York to São Paulo latency and RTT](/docs/network/latency/pairs/nyc-gru-rtt): 106.9 ms * [Miami to São Paulo latency and RTT](/docs/network/latency/pairs/mia-gru-rtt): 128.7 ms * [Los Angeles to São Paulo latency and RTT](/docs/network/latency/pairs/lax-gru-rtt): 131.3 ms * [Seattle to São Paulo latency and RTT](/docs/network/latency/pairs/sea-gru-rtt): 155.7 ms **Same corridor (Asia Pacific → South America)** * [Tokyo to São Paulo latency and RTT](/docs/network/latency/pairs/tyo-gru-rtt): 230.5 ms * [Taipei to São Paulo latency and RTT](/docs/network/latency/pairs/tpe-gru-rtt): 273 ms * [Hong Kong to São Paulo latency and RTT](/docs/network/latency/pairs/hkg-gru-rtt): 287.6 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/syd-gru.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Sydney → São Paulo round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/syd-gru-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Sydney, Australia → Hong Kong RTT 🇦🇺 **Sydney, Australia (SYD)** → 🇭🇰 **Hong Kong (HKG)** Measurement round: `2026-08-16T04:07:28Z`. Sydney to Hong Kong averaged 135.7 ms across 50 replies. The observed range was 124.18 to 155.08 ms; these figures describe one probe round, not a long-term service target. This is number 5 of 19 outgoing routes by average RTT. That ranking compares different destinations and is not adjusted for distance. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------------------- | | RTT | **135.7 ms** | | Jitter | **6.06 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **71.97 ms** | | Fiber Efficiency | **53%** | | Latency Tier | Good | | Source Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | | Destination Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Sydney](https://www.geonames.org/2147714) (-33.86785, 151.20732) and [Hong Kong](https://www.geonames.org/1819729) (22.27832, 114.17469). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **7,349.1 km** | | Vacuum RTT floor | **49.03 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **71.97 ms** | | Low-latency fiber material floor | **71.68 ms** | | Engineering floor (5% path allowance) | **75.58 ms** | | Research 1.33× mapped-fiber reference | **95.72 ms** | | Estimated unamplified path loss | **1543.3 dB** | | Transparent optical spans / inline amplifiers | **97 / 96** | | Published RTT inflation over fiber floor | **1.89×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **124.18 ms** | | Average RTT | **135.7 ms** | | Maximum RTT | **155.08 ms** | | Standard deviation | **6.42 ms** | | Stdev / average | **4.7%** | ## Route Context * Departure PoP: [Sydney (SYD)](/docs/network/latency/syd-sydney) * Destination PoP: [Hong Kong (HKG)](/docs/network/latency/hkg-hong-kong) * Region overview: [Asia Pacific](/docs/network/latency/regions/asia-pacific) → [Asia Pacific](/docs/network/latency/regions/asia-pacific) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Hong Kong to Sydney latency and RTT](/docs/network/latency/pairs/hkg-syd-rtt): 137.2 ms **Fastest routes departing Sydney (SYD)** * [Sydney to Melbourne latency and RTT](/docs/network/latency/pairs/syd-mel-rtt): 9.8 ms * [Sydney to Singapore latency and RTT](/docs/network/latency/pairs/syd-sin-rtt): 94.5 ms * [Sydney to Tokyo latency and RTT](/docs/network/latency/pairs/syd-tyo-rtt): 101.3 ms * [Sydney to Taipei latency and RTT](/docs/network/latency/pairs/syd-tpe-rtt): 131.7 ms * [Sydney to Los Angeles latency and RTT](/docs/network/latency/pairs/syd-lax-rtt): 136.5 ms **Fastest routes arriving at Hong Kong (HKG)** * [Taipei to Hong Kong latency and RTT](/docs/network/latency/pairs/tpe-hkg-rtt): 15.5 ms * [Singapore to Hong Kong latency and RTT](/docs/network/latency/pairs/sin-hkg-rtt): 30.8 ms * [Tokyo to Hong Kong latency and RTT](/docs/network/latency/pairs/tyo-hkg-rtt): 44.8 ms * [Moscow to Hong Kong latency and RTT](/docs/network/latency/pairs/mow-hkg-rtt): 118.6 ms * [Seattle to Hong Kong latency and RTT](/docs/network/latency/pairs/sea-hkg-rtt): 130.4 ms **Same corridor (Asia Pacific → Asia Pacific)** * [Melbourne to Sydney latency and RTT](/docs/network/latency/pairs/mel-syd-rtt): 9.8 ms * [Hong Kong to Taipei latency and RTT](/docs/network/latency/pairs/hkg-tpe-rtt): 14.7 ms * [Hong Kong to Singapore latency and RTT](/docs/network/latency/pairs/hkg-sin-rtt): 31.5 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/syd-hkg.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Sydney → Hong Kong round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/syd-hkg-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Sydney, Australia → Ashburn, USA RTT 🇦🇺 **Sydney, Australia (SYD)** → 🇺🇸 **Ashburn, USA (IAD)** Measurement round: `2026-08-16T04:07:28Z`. Sydney to Ashburn averaged 195.3 ms across 50 replies. The observed range was 186.39 to 216.22 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 7.74 ms, or 4.0% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **195.3 ms** | | Jitter | **6.41 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **153.48 ms** | | Fiber Efficiency | **78.6%** | | Latency Tier | Fair | | Source Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | | Destination Region | [North America](/docs/network/latency/regions/north-america) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Sydney](https://www.geonames.org/2147714) (-33.86785, 151.20732) and [Ashburn](https://www.geonames.org/4744870) (39.04372, -77.48749). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | --------------- | | WGS-84 geodesic distance | **15,673.1 km** | | Vacuum RTT floor | **104.56 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **153.48 ms** | | Low-latency fiber material floor | **152.87 ms** | | Engineering floor (5% path allowance) | **161.18 ms** | | Research 1.33× mapped-fiber reference | **204.13 ms** | | Estimated unamplified path loss | **3291.4 dB** | | Transparent optical spans / inline amplifiers | **206 / 205** | | Published RTT inflation over fiber floor | **1.27×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **186.39 ms** | | Average RTT | **195.3 ms** | | Maximum RTT | **216.22 ms** | | Standard deviation | **7.74 ms** | | Stdev / average | **4.0%** | ## Route Context * Departure PoP: [Sydney (SYD)](/docs/network/latency/syd-sydney) * Destination PoP: [Ashburn (IAD)](/docs/network/latency/iad-ashburn) * Region overview: [Asia Pacific](/docs/network/latency/regions/asia-pacific) → [North America](/docs/network/latency/regions/north-america) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Ashburn to Sydney latency and RTT](/docs/network/latency/pairs/iad-syd-rtt): 193.6 ms **Fastest routes departing Sydney (SYD)** * [Sydney to Melbourne latency and RTT](/docs/network/latency/pairs/syd-mel-rtt): 9.8 ms * [Sydney to Singapore latency and RTT](/docs/network/latency/pairs/syd-sin-rtt): 94.5 ms * [Sydney to Tokyo latency and RTT](/docs/network/latency/pairs/syd-tyo-rtt): 101.3 ms * [Sydney to Taipei latency and RTT](/docs/network/latency/pairs/syd-tpe-rtt): 131.7 ms * [Sydney to Hong Kong latency and RTT](/docs/network/latency/pairs/syd-hkg-rtt): 135.7 ms **Fastest routes arriving at Ashburn (IAD)** * [New York to Ashburn latency and RTT](/docs/network/latency/pairs/nyc-iad-rtt): 6 ms * [Miami to Ashburn latency and RTT](/docs/network/latency/pairs/mia-iad-rtt): 27.3 ms * [Los Angeles to Ashburn latency and RTT](/docs/network/latency/pairs/lax-iad-rtt): 60.4 ms * [Seattle to Ashburn latency and RTT](/docs/network/latency/pairs/sea-iad-rtt): 61.5 ms * [London to Ashburn latency and RTT](/docs/network/latency/pairs/lon-iad-rtt): 71 ms **Same corridor (Asia Pacific → North America)** * [Tokyo to Seattle latency and RTT](/docs/network/latency/pairs/tyo-sea-rtt): 84.8 ms * [Tokyo to Los Angeles latency and RTT](/docs/network/latency/pairs/tyo-lax-rtt): 101.2 ms * [Taipei to Seattle latency and RTT](/docs/network/latency/pairs/tpe-sea-rtt): 115.8 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/syd-iad.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Sydney → Ashburn round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/syd-iad-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Sydney, Australia → Johannesburg, South Africa RTT 🇦🇺 **Sydney, Australia (SYD)** → 🇿🇦 **Johannesburg, South Africa (JNB)** Measurement round: `2026-08-16T04:07:28Z`. Sydney to Johannesburg averaged 412.5 ms, compared with a 108.32 ms city-centre fiber reference. The difference cannot be assigned to a particular cable or carrier from RTT alone. Standard deviation was 12.82 ms, or 3.1% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------------------- | | RTT | **412.5 ms** | | Jitter | **9.3 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **108.32 ms** | | Fiber Efficiency | **26.3%** | | Latency Tier | High | | Source Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | | Destination Region | [Africa](/docs/network/latency/regions/africa) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Sydney](https://www.geonames.org/2147714) (-33.86785, 151.20732) and [Johannesburg](https://www.geonames.org/993800) (-26.20227, 28.04363). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | --------------- | | WGS-84 geodesic distance | **11,060.7 km** | | Vacuum RTT floor | **73.79 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **108.32 ms** | | Low-latency fiber material floor | **107.88 ms** | | Engineering floor (5% path allowance) | **113.75 ms** | | Research 1.33× mapped-fiber reference | **144.06 ms** | | Estimated unamplified path loss | **2322.8 dB** | | Transparent optical spans / inline amplifiers | **146 / 145** | | Published RTT inflation over fiber floor | **3.81×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **399.32 ms** | | Average RTT | **412.5 ms** | | Maximum RTT | **451.86 ms** | | Standard deviation | **12.82 ms** | | Stdev / average | **3.1%** | ## Route Context * Departure PoP: [Sydney (SYD)](/docs/network/latency/syd-sydney) * Destination PoP: [Johannesburg (JNB)](/docs/network/latency/jnb-johannesburg) * Region overview: [Asia Pacific](/docs/network/latency/regions/asia-pacific) → [Africa](/docs/network/latency/regions/africa) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Johannesburg to Sydney latency and RTT](/docs/network/latency/pairs/jnb-syd-rtt): 412.6 ms **Fastest routes departing Sydney (SYD)** * [Sydney to Melbourne latency and RTT](/docs/network/latency/pairs/syd-mel-rtt): 9.8 ms * [Sydney to Singapore latency and RTT](/docs/network/latency/pairs/syd-sin-rtt): 94.5 ms * [Sydney to Tokyo latency and RTT](/docs/network/latency/pairs/syd-tyo-rtt): 101.3 ms * [Sydney to Taipei latency and RTT](/docs/network/latency/pairs/syd-tpe-rtt): 131.7 ms * [Sydney to Hong Kong latency and RTT](/docs/network/latency/pairs/syd-hkg-rtt): 135.7 ms **Fastest routes arriving at Johannesburg (JNB)** * [Amsterdam to Johannesburg latency and RTT](/docs/network/latency/pairs/ams-jnb-rtt): 171.9 ms * [Berlin to Johannesburg latency and RTT](/docs/network/latency/pairs/ber-jnb-rtt): 172.1 ms * [London to Johannesburg latency and RTT](/docs/network/latency/pairs/lon-jnb-rtt): 179.6 ms * [Paris to Johannesburg latency and RTT](/docs/network/latency/pairs/par-jnb-rtt): 180.9 ms * [Frankfurt to Johannesburg latency and RTT](/docs/network/latency/pairs/fra-jnb-rtt): 188.4 ms **Same corridor (Asia Pacific → Africa)** * [Singapore to Johannesburg latency and RTT](/docs/network/latency/pairs/sin-jnb-rtt): 312.6 ms * [Hong Kong to Johannesburg latency and RTT](/docs/network/latency/pairs/hkg-jnb-rtt): 316.7 ms * [Taipei to Johannesburg latency and RTT](/docs/network/latency/pairs/tpe-jnb-rtt): 331 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/syd-jnb.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Sydney → Johannesburg round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/syd-jnb-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Sydney, Australia → Los Angeles, USA RTT 🇦🇺 **Sydney, Australia (SYD)** → 🇺🇸 **Los Angeles, USA (LAX)** Measurement round: `2026-08-16T04:07:28Z`. Sydney to Los Angeles averaged 136.5 ms across 50 replies. The observed range was 132.98 to 143.81 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 2.85 ms, or 2.1% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **136.5 ms** | | Jitter | **2.5 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **118.13 ms** | | Fiber Efficiency | **86.5%** | | Latency Tier | Good | | Source Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | | Destination Region | [North America](/docs/network/latency/regions/north-america) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Sydney](https://www.geonames.org/2147714) (-33.86785, 151.20732) and [Los Angeles](https://www.geonames.org/5368361) (34.05223, -118.24368). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | --------------- | | WGS-84 geodesic distance | **12,063.2 km** | | Vacuum RTT floor | **80.48 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **118.13 ms** | | Low-latency fiber material floor | **117.66 ms** | | Engineering floor (5% path allowance) | **124.05 ms** | | Research 1.33× mapped-fiber reference | **157.12 ms** | | Estimated unamplified path loss | **2533.3 dB** | | Transparent optical spans / inline amplifiers | **159 / 158** | | Published RTT inflation over fiber floor | **1.16×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **132.98 ms** | | Average RTT | **136.5 ms** | | Maximum RTT | **143.81 ms** | | Standard deviation | **2.85 ms** | | Stdev / average | **2.1%** | ## Route Context * Departure PoP: [Sydney (SYD)](/docs/network/latency/syd-sydney) * Destination PoP: [Los Angeles (LAX)](/docs/network/latency/lax-los-angeles) * Region overview: [Asia Pacific](/docs/network/latency/regions/asia-pacific) → [North America](/docs/network/latency/regions/north-america) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Los Angeles to Sydney latency and RTT](/docs/network/latency/pairs/lax-syd-rtt): 137.6 ms **Fastest routes departing Sydney (SYD)** * [Sydney to Melbourne latency and RTT](/docs/network/latency/pairs/syd-mel-rtt): 9.8 ms * [Sydney to Singapore latency and RTT](/docs/network/latency/pairs/syd-sin-rtt): 94.5 ms * [Sydney to Tokyo latency and RTT](/docs/network/latency/pairs/syd-tyo-rtt): 101.3 ms * [Sydney to Taipei latency and RTT](/docs/network/latency/pairs/syd-tpe-rtt): 131.7 ms * [Sydney to Hong Kong latency and RTT](/docs/network/latency/pairs/syd-hkg-rtt): 135.7 ms **Fastest routes arriving at Los Angeles (LAX)** * [Seattle to Los Angeles latency and RTT](/docs/network/latency/pairs/sea-lax-rtt): 25.9 ms * [Miami to Los Angeles latency and RTT](/docs/network/latency/pairs/mia-lax-rtt): 56.8 ms * [New York to Los Angeles latency and RTT](/docs/network/latency/pairs/nyc-lax-rtt): 58.9 ms * [Ashburn to Los Angeles latency and RTT](/docs/network/latency/pairs/iad-lax-rtt): 60.7 ms * [Tokyo to Los Angeles latency and RTT](/docs/network/latency/pairs/tyo-lax-rtt): 101.2 ms **Same corridor (Asia Pacific → North America)** * [Tokyo to Seattle latency and RTT](/docs/network/latency/pairs/tyo-sea-rtt): 84.8 ms * [Taipei to Seattle latency and RTT](/docs/network/latency/pairs/tpe-sea-rtt): 115.8 ms * [Hong Kong to Seattle latency and RTT](/docs/network/latency/pairs/hkg-sea-rtt): 131.6 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/syd-lax.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Sydney → Los Angeles round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/syd-lax-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Sydney, Australia → London, UK RTT 🇦🇺 **Sydney, Australia (SYD)** → 🇬🇧 **London, UK (LON)** Measurement round: `2026-08-16T04:07:28Z`. Sydney to London averaged 256.5 ms across 50 replies. The observed range was 243.5 to 279.39 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 9.98 ms, or 3.9% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------------------- | | RTT | **256.5 ms** | | Jitter | **10.31 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **166.37 ms** | | Fiber Efficiency | **64.9%** | | Latency Tier | High | | Source Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | | Destination Region | [Europe](/docs/network/latency/regions/europe) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Sydney](https://www.geonames.org/2147714) (-33.86785, 151.20732) and [London](https://www.geonames.org/2643743) (51.50853, -0.12574). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | --------------- | | WGS-84 geodesic distance | **16,988.9 km** | | Vacuum RTT floor | **113.34 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **166.37 ms** | | Low-latency fiber material floor | **165.7 ms** | | Engineering floor (5% path allowance) | **174.71 ms** | | Research 1.33× mapped-fiber reference | **221.27 ms** | | Estimated unamplified path loss | **3567.7 dB** | | Transparent optical spans / inline amplifiers | **223 / 222** | | Published RTT inflation over fiber floor | **1.54×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **243.5 ms** | | Average RTT | **256.5 ms** | | Maximum RTT | **279.39 ms** | | Standard deviation | **9.98 ms** | | Stdev / average | **3.9%** | ## Route Context * Departure PoP: [Sydney (SYD)](/docs/network/latency/syd-sydney) * Destination PoP: [London (LON)](/docs/network/latency/lon-london) * Region overview: [Asia Pacific](/docs/network/latency/regions/asia-pacific) → [Europe](/docs/network/latency/regions/europe) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [London to Sydney latency and RTT](/docs/network/latency/pairs/lon-syd-rtt): 258.1 ms **Fastest routes departing Sydney (SYD)** * [Sydney to Melbourne latency and RTT](/docs/network/latency/pairs/syd-mel-rtt): 9.8 ms * [Sydney to Singapore latency and RTT](/docs/network/latency/pairs/syd-sin-rtt): 94.5 ms * [Sydney to Tokyo latency and RTT](/docs/network/latency/pairs/syd-tyo-rtt): 101.3 ms * [Sydney to Taipei latency and RTT](/docs/network/latency/pairs/syd-tpe-rtt): 131.7 ms * [Sydney to Hong Kong latency and RTT](/docs/network/latency/pairs/syd-hkg-rtt): 135.7 ms **Fastest routes arriving at London (LON)** * [Amsterdam to London latency and RTT](/docs/network/latency/pairs/ams-lon-rtt): 5.2 ms * [Paris to London latency and RTT](/docs/network/latency/pairs/par-lon-rtt): 6.4 ms * [Frankfurt to London latency and RTT](/docs/network/latency/pairs/fra-lon-rtt): 12.9 ms * [Berlin to London latency and RTT](/docs/network/latency/pairs/ber-lon-rtt): 15.2 ms * [Marseille to London latency and RTT](/docs/network/latency/pairs/mrs-lon-rtt): 18.9 ms **Same corridor (Asia Pacific → Europe)** * [Hong Kong to Moscow latency and RTT](/docs/network/latency/pairs/hkg-mow-rtt): 118.2 ms * [Taipei to Moscow latency and RTT](/docs/network/latency/pairs/tpe-mow-rtt): 132.7 ms * [Singapore to Marseille latency and RTT](/docs/network/latency/pairs/sin-mrs-rtt): 140.6 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/syd-lon.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Sydney → London round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/syd-lon-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Sydney, Australia → Melbourne, Australia RTT 🇦🇺 **Sydney, Australia (SYD)** → 🇦🇺 **Melbourne, Australia (MEL)** Measurement round: `2026-08-16T04:07:28Z`. Sydney to Melbourne averaged 9.8 ms across 50 replies. The observed range was 9.38 to 11.84 ms; these figures describe one probe round, not a long-term service target. This is number 1 of 19 outgoing routes by average RTT. That ranking compares different destinations and is not adjusted for distance. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------------------- | | RTT | **9.8 ms** | | Jitter | **0.41 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **6.99 ms** | | Fiber Efficiency | **71.3%** | | Latency Tier | Ultra-Low | | Source Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | | Destination Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Sydney](https://www.geonames.org/2147714) (-33.86785, 151.20732) and [Melbourne](https://www.geonames.org/2158177) (-37.81400, 144.96332). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | ------------ | | WGS-84 geodesic distance | **713.8 km** | | Vacuum RTT floor | **4.76 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **6.99 ms** | | Low-latency fiber material floor | **6.96 ms** | | Engineering floor (5% path allowance) | **7.34 ms** | | Research 1.33× mapped-fiber reference | **9.3 ms** | | Estimated unamplified path loss | **149.9 dB** | | Transparent optical spans / inline amplifiers | **10 / 9** | | Published RTT inflation over fiber floor | **1.4×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------ | | Minimum RTT | **9.38 ms** | | Average RTT | **9.8 ms** | | Maximum RTT | **11.84 ms** | | Standard deviation | **0.46 ms** | | Stdev / average | **4.7%** | ## Route Context * Departure PoP: [Sydney (SYD)](/docs/network/latency/syd-sydney) * Destination PoP: [Melbourne (MEL)](/docs/network/latency/mel-melbourne) * Region overview: [Asia Pacific](/docs/network/latency/regions/asia-pacific) → [Asia Pacific](/docs/network/latency/regions/asia-pacific) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Melbourne to Sydney latency and RTT](/docs/network/latency/pairs/mel-syd-rtt): 9.8 ms **Fastest routes departing Sydney (SYD)** * [Sydney to Singapore latency and RTT](/docs/network/latency/pairs/syd-sin-rtt): 94.5 ms * [Sydney to Tokyo latency and RTT](/docs/network/latency/pairs/syd-tyo-rtt): 101.3 ms * [Sydney to Taipei latency and RTT](/docs/network/latency/pairs/syd-tpe-rtt): 131.7 ms * [Sydney to Hong Kong latency and RTT](/docs/network/latency/pairs/syd-hkg-rtt): 135.7 ms * [Sydney to Los Angeles latency and RTT](/docs/network/latency/pairs/syd-lax-rtt): 136.5 ms **Fastest routes arriving at Melbourne (MEL)** * [Singapore to Melbourne latency and RTT](/docs/network/latency/pairs/sin-mel-rtt): 88.6 ms * [Tokyo to Melbourne latency and RTT](/docs/network/latency/pairs/tyo-mel-rtt): 113.9 ms * [Hong Kong to Melbourne latency and RTT](/docs/network/latency/pairs/hkg-mel-rtt): 138.4 ms * [Taipei to Melbourne latency and RTT](/docs/network/latency/pairs/tpe-mel-rtt): 142 ms * [Los Angeles to Melbourne latency and RTT](/docs/network/latency/pairs/lax-mel-rtt): 148.2 ms **Same corridor (Asia Pacific → Asia Pacific)** * [Hong Kong to Taipei latency and RTT](/docs/network/latency/pairs/hkg-tpe-rtt): 14.7 ms * [Taipei to Hong Kong latency and RTT](/docs/network/latency/pairs/tpe-hkg-rtt): 15.5 ms * [Singapore to Hong Kong latency and RTT](/docs/network/latency/pairs/sin-hkg-rtt): 30.8 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/syd-mel.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Sydney → Melbourne round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/syd-mel-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Sydney, Australia → Miami, USA RTT 🇦🇺 **Sydney, Australia (SYD)** → 🇺🇸 **Miami, USA (MIA)** Measurement round: `2026-08-16T04:07:28Z`. Sydney to Miami averaged 191.5 ms across 50 replies. The observed range was 181.12 to 215.13 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 7.49 ms, or 3.9% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **191.5 ms** | | Jitter | **6.69 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **147.18 ms** | | Fiber Efficiency | **76.9%** | | Latency Tier | Fair | | Source Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | | Destination Region | [North America](/docs/network/latency/regions/north-america) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Sydney](https://www.geonames.org/2147714) (-33.86785, 151.20732) and [Miami](https://www.geonames.org/4164138) (25.77427, -80.19366). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | --------------- | | WGS-84 geodesic distance | **15,029.6 km** | | Vacuum RTT floor | **100.27 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **147.18 ms** | | Low-latency fiber material floor | **146.59 ms** | | Engineering floor (5% path allowance) | **154.56 ms** | | Research 1.33× mapped-fiber reference | **195.75 ms** | | Estimated unamplified path loss | **3156.2 dB** | | Transparent optical spans / inline amplifiers | **198 / 197** | | Published RTT inflation over fiber floor | **1.3×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **181.12 ms** | | Average RTT | **191.5 ms** | | Maximum RTT | **215.13 ms** | | Standard deviation | **7.49 ms** | | Stdev / average | **3.9%** | ## Route Context * Departure PoP: [Sydney (SYD)](/docs/network/latency/syd-sydney) * Destination PoP: [Miami (MIA)](/docs/network/latency/mia-miami) * Region overview: [Asia Pacific](/docs/network/latency/regions/asia-pacific) → [North America](/docs/network/latency/regions/north-america) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Miami to Sydney latency and RTT](/docs/network/latency/pairs/mia-syd-rtt): 191.5 ms **Fastest routes departing Sydney (SYD)** * [Sydney to Melbourne latency and RTT](/docs/network/latency/pairs/syd-mel-rtt): 9.8 ms * [Sydney to Singapore latency and RTT](/docs/network/latency/pairs/syd-sin-rtt): 94.5 ms * [Sydney to Tokyo latency and RTT](/docs/network/latency/pairs/syd-tyo-rtt): 101.3 ms * [Sydney to Taipei latency and RTT](/docs/network/latency/pairs/syd-tpe-rtt): 131.7 ms * [Sydney to Hong Kong latency and RTT](/docs/network/latency/pairs/syd-hkg-rtt): 135.7 ms **Fastest routes arriving at Miami (MIA)** * [Ashburn to Miami latency and RTT](/docs/network/latency/pairs/iad-mia-rtt): 27.4 ms * [New York to Miami latency and RTT](/docs/network/latency/pairs/nyc-mia-rtt): 32.9 ms * [Los Angeles to Miami latency and RTT](/docs/network/latency/pairs/lax-mia-rtt): 57.3 ms * [São Paulo to Miami latency and RTT](/docs/network/latency/pairs/gru-mia-rtt): 74 ms * [Seattle to Miami latency and RTT](/docs/network/latency/pairs/sea-mia-rtt): 81.7 ms **Same corridor (Asia Pacific → North America)** * [Tokyo to Seattle latency and RTT](/docs/network/latency/pairs/tyo-sea-rtt): 84.8 ms * [Tokyo to Los Angeles latency and RTT](/docs/network/latency/pairs/tyo-lax-rtt): 101.2 ms * [Taipei to Seattle latency and RTT](/docs/network/latency/pairs/tpe-sea-rtt): 115.8 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/syd-mia.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Sydney → Miami round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/syd-mia-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Sydney, Australia → Moscow, Russia RTT 🇦🇺 **Sydney, Australia (SYD)** → 🇷🇺 **Moscow, Russia (MOW)** Measurement round: `2026-08-16T04:07:28Z`. Sydney to Moscow averaged 242.2 ms across 50 replies. The observed range was 230.07 to 270.93 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 9.04 ms, or 3.7% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------------------- | | RTT | **242.2 ms** | | Jitter | **7.98 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **141.84 ms** | | Fiber Efficiency | **58.6%** | | Latency Tier | Fair | | Source Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | | Destination Region | [Europe](/docs/network/latency/regions/europe) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Sydney](https://www.geonames.org/2147714) (-33.86785, 151.20732) and [Moscow](https://www.geonames.org/524901) (55.75204, 37.61781). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | --------------- | | WGS-84 geodesic distance | **14,483.9 km** | | Vacuum RTT floor | **96.63 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **141.84 ms** | | Low-latency fiber material floor | **141.27 ms** | | Engineering floor (5% path allowance) | **148.95 ms** | | Research 1.33× mapped-fiber reference | **188.64 ms** | | Estimated unamplified path loss | **3041.6 dB** | | Transparent optical spans / inline amplifiers | **191 / 190** | | Published RTT inflation over fiber floor | **1.71×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **230.07 ms** | | Average RTT | **242.2 ms** | | Maximum RTT | **270.93 ms** | | Standard deviation | **9.04 ms** | | Stdev / average | **3.7%** | ## Route Context * Departure PoP: [Sydney (SYD)](/docs/network/latency/syd-sydney) * Destination PoP: [Moscow (MOW)](/docs/network/latency/mow-moscow) * Region overview: [Asia Pacific](/docs/network/latency/regions/asia-pacific) → [Europe](/docs/network/latency/regions/europe) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Moscow to Sydney latency and RTT](/docs/network/latency/pairs/mow-syd-rtt): 241.8 ms **Fastest routes departing Sydney (SYD)** * [Sydney to Melbourne latency and RTT](/docs/network/latency/pairs/syd-mel-rtt): 9.8 ms * [Sydney to Singapore latency and RTT](/docs/network/latency/pairs/syd-sin-rtt): 94.5 ms * [Sydney to Tokyo latency and RTT](/docs/network/latency/pairs/syd-tyo-rtt): 101.3 ms * [Sydney to Taipei latency and RTT](/docs/network/latency/pairs/syd-tpe-rtt): 131.7 ms * [Sydney to Hong Kong latency and RTT](/docs/network/latency/pairs/syd-hkg-rtt): 135.7 ms **Fastest routes arriving at Moscow (MOW)** * [Berlin to Moscow latency and RTT](/docs/network/latency/pairs/ber-mow-rtt): 28.6 ms * [Frankfurt to Moscow latency and RTT](/docs/network/latency/pairs/fra-mow-rtt): 34.7 ms * [Amsterdam to Moscow latency and RTT](/docs/network/latency/pairs/ams-mow-rtt): 38.1 ms * [London to Moscow latency and RTT](/docs/network/latency/pairs/lon-mow-rtt): 43.5 ms * [Paris to Moscow latency and RTT](/docs/network/latency/pairs/par-mow-rtt): 44.7 ms **Same corridor (Asia Pacific → Europe)** * [Hong Kong to Moscow latency and RTT](/docs/network/latency/pairs/hkg-mow-rtt): 118.2 ms * [Taipei to Moscow latency and RTT](/docs/network/latency/pairs/tpe-mow-rtt): 132.7 ms * [Singapore to Marseille latency and RTT](/docs/network/latency/pairs/sin-mrs-rtt): 140.6 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/syd-mow.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Sydney → Moscow round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/syd-mow-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Sydney, Australia → Marseille, France RTT 🇦🇺 **Sydney, Australia (SYD)** → 🇫🇷 **Marseille, France (MRS)** Measurement round: `2026-08-16T04:07:28Z`. Sydney to Marseille averaged 240.5 ms across 50 replies. The observed range was 231.2 to 264.13 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 7.87 ms, or 3.3% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------------------- | | RTT | **240.5 ms** | | Jitter | **6.37 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **165.41 ms** | | Fiber Efficiency | **68.8%** | | Latency Tier | Fair | | Source Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | | Destination Region | [Europe](/docs/network/latency/regions/europe) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Sydney](https://www.geonames.org/2147714) (-33.86785, 151.20732) and [Marseille](https://www.geonames.org/2995469) (43.29695, 5.38107). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | ------------- | | WGS-84 geodesic distance | **16,891 km** | | Vacuum RTT floor | **112.68 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **165.41 ms** | | Low-latency fiber material floor | **164.74 ms** | | Engineering floor (5% path allowance) | **173.7 ms** | | Research 1.33× mapped-fiber reference | **219.99 ms** | | Estimated unamplified path loss | **3547.1 dB** | | Transparent optical spans / inline amplifiers | **222 / 221** | | Published RTT inflation over fiber floor | **1.45×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **231.2 ms** | | Average RTT | **240.5 ms** | | Maximum RTT | **264.13 ms** | | Standard deviation | **7.87 ms** | | Stdev / average | **3.3%** | ## Route Context * Departure PoP: [Sydney (SYD)](/docs/network/latency/syd-sydney) * Destination PoP: [Marseille (MRS)](/docs/network/latency/mrs-marseille) * Region overview: [Asia Pacific](/docs/network/latency/regions/asia-pacific) → [Europe](/docs/network/latency/regions/europe) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Marseille to Sydney latency and RTT](/docs/network/latency/pairs/mrs-syd-rtt): 240.6 ms **Fastest routes departing Sydney (SYD)** * [Sydney to Melbourne latency and RTT](/docs/network/latency/pairs/syd-mel-rtt): 9.8 ms * [Sydney to Singapore latency and RTT](/docs/network/latency/pairs/syd-sin-rtt): 94.5 ms * [Sydney to Tokyo latency and RTT](/docs/network/latency/pairs/syd-tyo-rtt): 101.3 ms * [Sydney to Taipei latency and RTT](/docs/network/latency/pairs/syd-tpe-rtt): 131.7 ms * [Sydney to Hong Kong latency and RTT](/docs/network/latency/pairs/syd-hkg-rtt): 135.7 ms **Fastest routes arriving at Marseille (MRS)** * [Paris to Marseille latency and RTT](/docs/network/latency/pairs/par-mrs-rtt): 8.9 ms * [Frankfurt to Marseille latency and RTT](/docs/network/latency/pairs/fra-mrs-rtt): 16.4 ms * [London to Marseille latency and RTT](/docs/network/latency/pairs/lon-mrs-rtt): 17.9 ms * [Amsterdam to Marseille latency and RTT](/docs/network/latency/pairs/ams-mrs-rtt): 19.6 ms * [Berlin to Marseille latency and RTT](/docs/network/latency/pairs/ber-mrs-rtt): 20.2 ms **Same corridor (Asia Pacific → Europe)** * [Hong Kong to Moscow latency and RTT](/docs/network/latency/pairs/hkg-mow-rtt): 118.2 ms * [Taipei to Moscow latency and RTT](/docs/network/latency/pairs/tpe-mow-rtt): 132.7 ms * [Singapore to Marseille latency and RTT](/docs/network/latency/pairs/sin-mrs-rtt): 140.6 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/syd-mrs.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Sydney → Marseille round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/syd-mrs-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Sydney, Australia → New York, USA RTT 🇦🇺 **Sydney, Australia (SYD)** → 🇺🇸 **New York, USA (NYC)** Measurement round: `2026-08-16T04:07:28Z`. Sydney to New York averaged 191.9 ms across 50 replies. The observed range was 183.18 to 221.95 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 7.82 ms, or 4.1% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **191.9 ms** | | Jitter | **4.83 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **156.57 ms** | | Fiber Efficiency | **81.6%** | | Latency Tier | Fair | | Source Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | | Destination Region | [North America](/docs/network/latency/regions/north-america) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Sydney](https://www.geonames.org/2147714) (-33.86785, 151.20732) and [New York City](https://www.geonames.org/5128581) (40.71427, -74.00597). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | --------------- | | WGS-84 geodesic distance | **15,988.1 km** | | Vacuum RTT floor | **106.66 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **156.57 ms** | | Low-latency fiber material floor | **155.94 ms** | | Engineering floor (5% path allowance) | **164.42 ms** | | Research 1.33× mapped-fiber reference | **208.24 ms** | | Estimated unamplified path loss | **3357.5 dB** | | Transparent optical spans / inline amplifiers | **210 / 209** | | Published RTT inflation over fiber floor | **1.23×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **183.18 ms** | | Average RTT | **191.9 ms** | | Maximum RTT | **221.95 ms** | | Standard deviation | **7.82 ms** | | Stdev / average | **4.1%** | ## Route Context * Departure PoP: [Sydney (SYD)](/docs/network/latency/syd-sydney) * Destination PoP: [New York (NYC)](/docs/network/latency/nyc-new-york) * Region overview: [Asia Pacific](/docs/network/latency/regions/asia-pacific) → [North America](/docs/network/latency/regions/north-america) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [New York to Sydney latency and RTT](/docs/network/latency/pairs/nyc-syd-rtt): 192.6 ms **Fastest routes departing Sydney (SYD)** * [Sydney to Melbourne latency and RTT](/docs/network/latency/pairs/syd-mel-rtt): 9.8 ms * [Sydney to Singapore latency and RTT](/docs/network/latency/pairs/syd-sin-rtt): 94.5 ms * [Sydney to Tokyo latency and RTT](/docs/network/latency/pairs/syd-tyo-rtt): 101.3 ms * [Sydney to Taipei latency and RTT](/docs/network/latency/pairs/syd-tpe-rtt): 131.7 ms * [Sydney to Hong Kong latency and RTT](/docs/network/latency/pairs/syd-hkg-rtt): 135.7 ms **Fastest routes arriving at New York (NYC)** * [Ashburn to New York latency and RTT](/docs/network/latency/pairs/iad-nyc-rtt): 6 ms * [Miami to New York latency and RTT](/docs/network/latency/pairs/mia-nyc-rtt): 33.3 ms * [Los Angeles to New York latency and RTT](/docs/network/latency/pairs/lax-nyc-rtt): 57.9 ms * [Seattle to New York latency and RTT](/docs/network/latency/pairs/sea-nyc-rtt): 59.5 ms * [London to New York latency and RTT](/docs/network/latency/pairs/lon-nyc-rtt): 63.8 ms **Same corridor (Asia Pacific → North America)** * [Tokyo to Seattle latency and RTT](/docs/network/latency/pairs/tyo-sea-rtt): 84.8 ms * [Tokyo to Los Angeles latency and RTT](/docs/network/latency/pairs/tyo-lax-rtt): 101.2 ms * [Taipei to Seattle latency and RTT](/docs/network/latency/pairs/tpe-sea-rtt): 115.8 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/syd-nyc.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Sydney → New York round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/syd-nyc-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Sydney, Australia → Paris, France RTT 🇦🇺 **Sydney, Australia (SYD)** → 🇫🇷 **Paris, France (PAR)** Measurement round: `2026-08-16T04:07:28Z`. Sydney to Paris averaged 250.3 ms across 50 replies. The observed range was 243.57 to 268.33 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 5.4 ms, or 2.2% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------------------- | | RTT | **250.3 ms** | | Jitter | **3.75 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **166.06 ms** | | Fiber Efficiency | **66.3%** | | Latency Tier | High | | Source Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | | Destination Region | [Europe](/docs/network/latency/regions/europe) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Sydney](https://www.geonames.org/2147714) (-33.86785, 151.20732) and [Paris](https://www.geonames.org/2988507) (48.85341, 2.34880). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | --------------- | | WGS-84 geodesic distance | **16,957.7 km** | | Vacuum RTT floor | **113.13 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **166.06 ms** | | Low-latency fiber material floor | **165.4 ms** | | Engineering floor (5% path allowance) | **174.39 ms** | | Research 1.33× mapped-fiber reference | **220.86 ms** | | Estimated unamplified path loss | **3561.1 dB** | | Transparent optical spans / inline amplifiers | **223 / 222** | | Published RTT inflation over fiber floor | **1.51×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **243.57 ms** | | Average RTT | **250.3 ms** | | Maximum RTT | **268.33 ms** | | Standard deviation | **5.4 ms** | | Stdev / average | **2.2%** | ## Route Context * Departure PoP: [Sydney (SYD)](/docs/network/latency/syd-sydney) * Destination PoP: [Paris (PAR)](/docs/network/latency/par-paris) * Region overview: [Asia Pacific](/docs/network/latency/regions/asia-pacific) → [Europe](/docs/network/latency/regions/europe) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Paris to Sydney latency and RTT](/docs/network/latency/pairs/par-syd-rtt): 249.7 ms **Fastest routes departing Sydney (SYD)** * [Sydney to Melbourne latency and RTT](/docs/network/latency/pairs/syd-mel-rtt): 9.8 ms * [Sydney to Singapore latency and RTT](/docs/network/latency/pairs/syd-sin-rtt): 94.5 ms * [Sydney to Tokyo latency and RTT](/docs/network/latency/pairs/syd-tyo-rtt): 101.3 ms * [Sydney to Taipei latency and RTT](/docs/network/latency/pairs/syd-tpe-rtt): 131.7 ms * [Sydney to Hong Kong latency and RTT](/docs/network/latency/pairs/syd-hkg-rtt): 135.7 ms **Fastest routes arriving at Paris (PAR)** * [London to Paris latency and RTT](/docs/network/latency/pairs/lon-par-rtt): 6.4 ms * [Amsterdam to Paris latency and RTT](/docs/network/latency/pairs/ams-par-rtt): 7 ms * [Frankfurt to Paris latency and RTT](/docs/network/latency/pairs/fra-par-rtt): 7.6 ms * [Marseille to Paris latency and RTT](/docs/network/latency/pairs/mrs-par-rtt): 8.9 ms * [Berlin to Paris latency and RTT](/docs/network/latency/pairs/ber-par-rtt): 15.6 ms **Same corridor (Asia Pacific → Europe)** * [Hong Kong to Moscow latency and RTT](/docs/network/latency/pairs/hkg-mow-rtt): 118.2 ms * [Taipei to Moscow latency and RTT](/docs/network/latency/pairs/tpe-mow-rtt): 132.7 ms * [Singapore to Marseille latency and RTT](/docs/network/latency/pairs/sin-mrs-rtt): 140.6 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/syd-par.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Sydney → Paris round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/syd-par-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Sydney, Australia → Seattle, USA RTT 🇦🇺 **Sydney, Australia (SYD)** → 🇺🇸 **Seattle, USA (SEA)** Measurement round: `2026-08-16T04:07:28Z`. Sydney to Seattle averaged 161.9 ms across 50 replies. The observed range was 153.84 to 176.63 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 5.72 ms, or 3.5% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **161.9 ms** | | Jitter | **3.81 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **121.96 ms** | | Fiber Efficiency | **75.3%** | | Latency Tier | Fair | | Source Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | | Destination Region | [North America](/docs/network/latency/regions/north-america) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Sydney](https://www.geonames.org/2147714) (-33.86785, 151.20732) and [Seattle](https://www.geonames.org/5809844) (47.60621, -122.33207). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | --------------- | | WGS-84 geodesic distance | **12,453.7 km** | | Vacuum RTT floor | **83.08 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **121.96 ms** | | Low-latency fiber material floor | **121.47 ms** | | Engineering floor (5% path allowance) | **128.07 ms** | | Research 1.33× mapped-fiber reference | **162.2 ms** | | Estimated unamplified path loss | **2615.3 dB** | | Transparent optical spans / inline amplifiers | **164 / 163** | | Published RTT inflation over fiber floor | **1.33×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **153.84 ms** | | Average RTT | **161.9 ms** | | Maximum RTT | **176.63 ms** | | Standard deviation | **5.72 ms** | | Stdev / average | **3.5%** | ## Route Context * Departure PoP: [Sydney (SYD)](/docs/network/latency/syd-sydney) * Destination PoP: [Seattle (SEA)](/docs/network/latency/sea-seattle) * Region overview: [Asia Pacific](/docs/network/latency/regions/asia-pacific) → [North America](/docs/network/latency/regions/north-america) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Seattle to Sydney latency and RTT](/docs/network/latency/pairs/sea-syd-rtt): 161.3 ms **Fastest routes departing Sydney (SYD)** * [Sydney to Melbourne latency and RTT](/docs/network/latency/pairs/syd-mel-rtt): 9.8 ms * [Sydney to Singapore latency and RTT](/docs/network/latency/pairs/syd-sin-rtt): 94.5 ms * [Sydney to Tokyo latency and RTT](/docs/network/latency/pairs/syd-tyo-rtt): 101.3 ms * [Sydney to Taipei latency and RTT](/docs/network/latency/pairs/syd-tpe-rtt): 131.7 ms * [Sydney to Hong Kong latency and RTT](/docs/network/latency/pairs/syd-hkg-rtt): 135.7 ms **Fastest routes arriving at Seattle (SEA)** * [Los Angeles to Seattle latency and RTT](/docs/network/latency/pairs/lax-sea-rtt): 26.9 ms * [New York to Seattle latency and RTT](/docs/network/latency/pairs/nyc-sea-rtt): 58.7 ms * [Ashburn to Seattle latency and RTT](/docs/network/latency/pairs/iad-sea-rtt): 62.3 ms * [Miami to Seattle latency and RTT](/docs/network/latency/pairs/mia-sea-rtt): 81.9 ms * [Tokyo to Seattle latency and RTT](/docs/network/latency/pairs/tyo-sea-rtt): 84.8 ms **Same corridor (Asia Pacific → North America)** * [Tokyo to Los Angeles latency and RTT](/docs/network/latency/pairs/tyo-lax-rtt): 101.2 ms * [Taipei to Seattle latency and RTT](/docs/network/latency/pairs/tpe-sea-rtt): 115.8 ms * [Hong Kong to Seattle latency and RTT](/docs/network/latency/pairs/hkg-sea-rtt): 131.6 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/syd-sea.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Sydney → Seattle round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/syd-sea-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Sydney, Australia → Singapore RTT 🇦🇺 **Sydney, Australia (SYD)** → 🇸🇬 **Singapore (SIN)** Measurement round: `2026-08-16T04:07:28Z`. Sydney to Singapore averaged 94.5 ms across 50 replies. The observed range was 89.04 to 106.31 ms; these figures describe one probe round, not a long-term service target. This is number 2 of 19 outgoing routes by average RTT. That ranking compares different destinations and is not adjusted for distance. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------------------- | | RTT | **94.5 ms** | | Jitter | **3.48 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **61.63 ms** | | Fiber Efficiency | **65.2%** | | Latency Tier | Good | | Source Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | | Destination Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Sydney](https://www.geonames.org/2147714) (-33.86785, 151.20732) and [Singapore](https://www.geonames.org/1880252) (1.28967, 103.85007). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **6,293.5 km** | | Vacuum RTT floor | **41.99 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **61.63 ms** | | Low-latency fiber material floor | **61.38 ms** | | Engineering floor (5% path allowance) | **64.72 ms** | | Research 1.33× mapped-fiber reference | **81.97 ms** | | Estimated unamplified path loss | **1321.6 dB** | | Transparent optical spans / inline amplifiers | **83 / 82** | | Published RTT inflation over fiber floor | **1.53×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **89.04 ms** | | Average RTT | **94.5 ms** | | Maximum RTT | **106.31 ms** | | Standard deviation | **3.56 ms** | | Stdev / average | **3.8%** | ## Route Context * Departure PoP: [Sydney (SYD)](/docs/network/latency/syd-sydney) * Destination PoP: [Singapore (SIN)](/docs/network/latency/sin-singapore) * Region overview: [Asia Pacific](/docs/network/latency/regions/asia-pacific) → [Asia Pacific](/docs/network/latency/regions/asia-pacific) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Singapore to Sydney latency and RTT](/docs/network/latency/pairs/sin-syd-rtt): 94.5 ms **Fastest routes departing Sydney (SYD)** * [Sydney to Melbourne latency and RTT](/docs/network/latency/pairs/syd-mel-rtt): 9.8 ms * [Sydney to Tokyo latency and RTT](/docs/network/latency/pairs/syd-tyo-rtt): 101.3 ms * [Sydney to Taipei latency and RTT](/docs/network/latency/pairs/syd-tpe-rtt): 131.7 ms * [Sydney to Hong Kong latency and RTT](/docs/network/latency/pairs/syd-hkg-rtt): 135.7 ms * [Sydney to Los Angeles latency and RTT](/docs/network/latency/pairs/syd-lax-rtt): 136.5 ms **Fastest routes arriving at Singapore (SIN)** * [Hong Kong to Singapore latency and RTT](/docs/network/latency/pairs/hkg-sin-rtt): 31.5 ms * [Taipei to Singapore latency and RTT](/docs/network/latency/pairs/tpe-sin-rtt): 44.3 ms * [Tokyo to Singapore latency and RTT](/docs/network/latency/pairs/tyo-sin-rtt): 69.4 ms * [Melbourne to Singapore latency and RTT](/docs/network/latency/pairs/mel-sin-rtt): 88.4 ms * [Marseille to Singapore latency and RTT](/docs/network/latency/pairs/mrs-sin-rtt): 139.6 ms **Same corridor (Asia Pacific → Asia Pacific)** * [Melbourne to Sydney latency and RTT](/docs/network/latency/pairs/mel-syd-rtt): 9.8 ms * [Hong Kong to Taipei latency and RTT](/docs/network/latency/pairs/hkg-tpe-rtt): 14.7 ms * [Taipei to Hong Kong latency and RTT](/docs/network/latency/pairs/tpe-hkg-rtt): 15.5 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/syd-sin.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Sydney → Singapore round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/syd-sin-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Sydney, Australia → Taipei, Taiwan RTT 🇦🇺 **Sydney, Australia (SYD)** → 🇹🇼 **Taipei, Taiwan (TPE)** Measurement round: `2026-08-16T04:07:28Z`. Sydney to Taipei averaged 131.7 ms across 50 replies. The observed range was 126.36 to 151.94 ms; these figures describe one probe round, not a long-term service target. This is number 4 of 19 outgoing routes by average RTT. That ranking compares different destinations and is not adjusted for distance. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------------------- | | RTT | **131.7 ms** | | Jitter | **3.58 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **70.88 ms** | | Fiber Efficiency | **53.8%** | | Latency Tier | Good | | Source Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | | Destination Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Sydney](https://www.geonames.org/2147714) (-33.86785, 151.20732) and [Taipei](https://www.geonames.org/1668341) (25.05306, 121.52639). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **7,237.6 km** | | Vacuum RTT floor | **48.28 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **70.88 ms** | | Low-latency fiber material floor | **70.59 ms** | | Engineering floor (5% path allowance) | **74.43 ms** | | Research 1.33× mapped-fiber reference | **94.26 ms** | | Estimated unamplified path loss | **1519.9 dB** | | Transparent optical spans / inline amplifiers | **95 / 94** | | Published RTT inflation over fiber floor | **1.86×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **126.36 ms** | | Average RTT | **131.7 ms** | | Maximum RTT | **151.94 ms** | | Standard deviation | **4.46 ms** | | Stdev / average | **3.4%** | ## Route Context * Departure PoP: [Sydney (SYD)](/docs/network/latency/syd-sydney) * Destination PoP: [Taipei (TPE)](/docs/network/latency/tpe-taipei) * Region overview: [Asia Pacific](/docs/network/latency/regions/asia-pacific) → [Asia Pacific](/docs/network/latency/regions/asia-pacific) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Taipei to Sydney latency and RTT](/docs/network/latency/pairs/tpe-syd-rtt): 133.2 ms **Fastest routes departing Sydney (SYD)** * [Sydney to Melbourne latency and RTT](/docs/network/latency/pairs/syd-mel-rtt): 9.8 ms * [Sydney to Singapore latency and RTT](/docs/network/latency/pairs/syd-sin-rtt): 94.5 ms * [Sydney to Tokyo latency and RTT](/docs/network/latency/pairs/syd-tyo-rtt): 101.3 ms * [Sydney to Hong Kong latency and RTT](/docs/network/latency/pairs/syd-hkg-rtt): 135.7 ms * [Sydney to Los Angeles latency and RTT](/docs/network/latency/pairs/syd-lax-rtt): 136.5 ms **Fastest routes arriving at Taipei (TPE)** * [Hong Kong to Taipei latency and RTT](/docs/network/latency/pairs/hkg-tpe-rtt): 14.7 ms * [Tokyo to Taipei latency and RTT](/docs/network/latency/pairs/tyo-tpe-rtt): 32.3 ms * [Singapore to Taipei latency and RTT](/docs/network/latency/pairs/sin-tpe-rtt): 45.8 ms * [Seattle to Taipei latency and RTT](/docs/network/latency/pairs/sea-tpe-rtt): 114.7 ms * [Moscow to Taipei latency and RTT](/docs/network/latency/pairs/mow-tpe-rtt): 132.1 ms **Same corridor (Asia Pacific → Asia Pacific)** * [Melbourne to Sydney latency and RTT](/docs/network/latency/pairs/mel-syd-rtt): 9.8 ms * [Taipei to Hong Kong latency and RTT](/docs/network/latency/pairs/tpe-hkg-rtt): 15.5 ms * [Singapore to Hong Kong latency and RTT](/docs/network/latency/pairs/sin-hkg-rtt): 30.8 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/syd-tpe.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Sydney → Taipei round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/syd-tpe-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Sydney, Australia → Tokyo, Japan RTT 🇦🇺 **Sydney, Australia (SYD)** → 🇯🇵 **Tokyo, Japan (TYO)** Measurement round: `2026-08-16T04:07:28Z`. Sydney to Tokyo averaged 101.3 ms across 50 replies. The observed range was 99.25 to 108.62 ms; these figures describe one probe round, not a long-term service target. This is number 3 of 19 outgoing routes by average RTT. That ranking compares different destinations and is not adjusted for distance. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------------------- | | RTT | **101.3 ms** | | Jitter | **1.46 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **76.31 ms** | | Fiber Efficiency | **75.3%** | | Latency Tier | Good | | Source Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | | Destination Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Sydney](https://www.geonames.org/2147714) (-33.86785, 151.20732) and [Tokyo](https://www.geonames.org/1850147) (35.68950, 139.69171). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **7,792.8 km** | | Vacuum RTT floor | **51.99 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **76.31 ms** | | Low-latency fiber material floor | **76.01 ms** | | Engineering floor (5% path allowance) | **80.14 ms** | | Research 1.33× mapped-fiber reference | **101.5 ms** | | Estimated unamplified path loss | **1636.5 dB** | | Transparent optical spans / inline amplifiers | **103 / 102** | | Published RTT inflation over fiber floor | **1.33×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **99.25 ms** | | Average RTT | **101.3 ms** | | Maximum RTT | **108.62 ms** | | Standard deviation | **2.05 ms** | | Stdev / average | **2.0%** | ## Route Context * Departure PoP: [Sydney (SYD)](/docs/network/latency/syd-sydney) * Destination PoP: [Tokyo (TYO)](/docs/network/latency/tyo-tokyo) * Region overview: [Asia Pacific](/docs/network/latency/regions/asia-pacific) → [Asia Pacific](/docs/network/latency/regions/asia-pacific) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Tokyo to Sydney latency and RTT](/docs/network/latency/pairs/tyo-syd-rtt): 101.9 ms **Fastest routes departing Sydney (SYD)** * [Sydney to Melbourne latency and RTT](/docs/network/latency/pairs/syd-mel-rtt): 9.8 ms * [Sydney to Singapore latency and RTT](/docs/network/latency/pairs/syd-sin-rtt): 94.5 ms * [Sydney to Taipei latency and RTT](/docs/network/latency/pairs/syd-tpe-rtt): 131.7 ms * [Sydney to Hong Kong latency and RTT](/docs/network/latency/pairs/syd-hkg-rtt): 135.7 ms * [Sydney to Los Angeles latency and RTT](/docs/network/latency/pairs/syd-lax-rtt): 136.5 ms **Fastest routes arriving at Tokyo (TYO)** * [Taipei to Tokyo latency and RTT](/docs/network/latency/pairs/tpe-tyo-rtt): 31.9 ms * [Hong Kong to Tokyo latency and RTT](/docs/network/latency/pairs/hkg-tyo-rtt): 44.9 ms * [Singapore to Tokyo latency and RTT](/docs/network/latency/pairs/sin-tyo-rtt): 68.9 ms * [Seattle to Tokyo latency and RTT](/docs/network/latency/pairs/sea-tyo-rtt): 85 ms * [Los Angeles to Tokyo latency and RTT](/docs/network/latency/pairs/lax-tyo-rtt): 101.2 ms **Same corridor (Asia Pacific → Asia Pacific)** * [Melbourne to Sydney latency and RTT](/docs/network/latency/pairs/mel-syd-rtt): 9.8 ms * [Hong Kong to Taipei latency and RTT](/docs/network/latency/pairs/hkg-tpe-rtt): 14.7 ms * [Taipei to Hong Kong latency and RTT](/docs/network/latency/pairs/tpe-hkg-rtt): 15.5 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/syd-tyo.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Sydney → Tokyo round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/syd-tyo-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Taipei, Taiwan → Amsterdam, Netherlands RTT 🇹🇼 **Taipei, Taiwan (TPE)** → 🇳🇱 **Amsterdam, Netherlands (AMS)** Measurement round: `2026-08-16T04:07:28Z`. Taipei to Amsterdam averaged 166.5 ms across 50 replies. The observed range was 161.36 to 179.84 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 3.92 ms, or 2.4% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------------------- | | RTT | **166.5 ms** | | Jitter | **3.56 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **92.72 ms** | | Fiber Efficiency | **55.7%** | | Latency Tier | Fair | | Source Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | | Destination Region | [Europe](/docs/network/latency/regions/europe) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Taipei](https://www.geonames.org/1668341) (25.05306, 121.52639) and [Amsterdam](https://www.geonames.org/2759794) (52.37403, 4.88969). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **9,467.8 km** | | Vacuum RTT floor | **63.16 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **92.72 ms** | | Low-latency fiber material floor | **92.34 ms** | | Engineering floor (5% path allowance) | **97.36 ms** | | Research 1.33× mapped-fiber reference | **123.31 ms** | | Estimated unamplified path loss | **1988.2 dB** | | Transparent optical spans / inline amplifiers | **125 / 124** | | Published RTT inflation over fiber floor | **1.8×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **161.36 ms** | | Average RTT | **166.5 ms** | | Maximum RTT | **179.84 ms** | | Standard deviation | **3.92 ms** | | Stdev / average | **2.4%** | ## Route Context * Departure PoP: [Taipei (TPE)](/docs/network/latency/tpe-taipei) * Destination PoP: [Amsterdam (AMS)](/docs/network/latency/ams-amsterdam) * Region overview: [Asia Pacific](/docs/network/latency/regions/asia-pacific) → [Europe](/docs/network/latency/regions/europe) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Amsterdam to Taipei latency and RTT](/docs/network/latency/pairs/ams-tpe-rtt): 168.7 ms **Fastest routes departing Taipei (TPE)** * [Taipei to Hong Kong latency and RTT](/docs/network/latency/pairs/tpe-hkg-rtt): 15.5 ms * [Taipei to Tokyo latency and RTT](/docs/network/latency/pairs/tpe-tyo-rtt): 31.9 ms * [Taipei to Singapore latency and RTT](/docs/network/latency/pairs/tpe-sin-rtt): 44.3 ms * [Taipei to Seattle latency and RTT](/docs/network/latency/pairs/tpe-sea-rtt): 115.8 ms * [Taipei to Los Angeles latency and RTT](/docs/network/latency/pairs/tpe-lax-rtt): 132 ms **Fastest routes arriving at Amsterdam (AMS)** * [London to Amsterdam latency and RTT](/docs/network/latency/pairs/lon-ams-rtt): 5.2 ms * [Frankfurt to Amsterdam latency and RTT](/docs/network/latency/pairs/fra-ams-rtt): 5.9 ms * [Paris to Amsterdam latency and RTT](/docs/network/latency/pairs/par-ams-rtt): 7 ms * [Berlin to Amsterdam latency and RTT](/docs/network/latency/pairs/ber-ams-rtt): 7.2 ms * [Marseille to Amsterdam latency and RTT](/docs/network/latency/pairs/mrs-ams-rtt): 20.3 ms **Same corridor (Asia Pacific → Europe)** * [Hong Kong to Moscow latency and RTT](/docs/network/latency/pairs/hkg-mow-rtt): 118.2 ms * [Taipei to Moscow latency and RTT](/docs/network/latency/pairs/tpe-mow-rtt): 132.7 ms * [Singapore to Marseille latency and RTT](/docs/network/latency/pairs/sin-mrs-rtt): 140.6 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/tpe-ams.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Taipei → Amsterdam round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/tpe-ams-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Taipei, Taiwan → Berlin, Germany RTT 🇹🇼 **Taipei, Taiwan (TPE)** → 🇩🇪 **Berlin, Germany (BER)** Measurement round: `2026-08-16T04:07:28Z`. Taipei to Berlin averaged 158.6 ms across 50 replies. The observed range was 154.12 to 170.16 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 3.77 ms, or 2.4% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------------------- | | RTT | **158.6 ms** | | Jitter | **3.48 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **87.84 ms** | | Fiber Efficiency | **55.4%** | | Latency Tier | Fair | | Source Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | | Destination Region | [Europe](/docs/network/latency/regions/europe) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Taipei](https://www.geonames.org/1668341) (25.05306, 121.52639) and [Berlin](https://www.geonames.org/2950159) (52.52437, 13.41053). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **8,970.4 km** | | Vacuum RTT floor | **59.84 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **87.84 ms** | | Low-latency fiber material floor | **87.49 ms** | | Engineering floor (5% path allowance) | **92.25 ms** | | Research 1.33× mapped-fiber reference | **116.83 ms** | | Estimated unamplified path loss | **1883.8 dB** | | Transparent optical spans / inline amplifiers | **118 / 117** | | Published RTT inflation over fiber floor | **1.81×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **154.12 ms** | | Average RTT | **158.6 ms** | | Maximum RTT | **170.16 ms** | | Standard deviation | **3.77 ms** | | Stdev / average | **2.4%** | ## Route Context * Departure PoP: [Taipei (TPE)](/docs/network/latency/tpe-taipei) * Destination PoP: [Berlin (BER)](/docs/network/latency/ber-berlin) * Region overview: [Asia Pacific](/docs/network/latency/regions/asia-pacific) → [Europe](/docs/network/latency/regions/europe) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Berlin to Taipei latency and RTT](/docs/network/latency/pairs/ber-tpe-rtt): 156.8 ms **Fastest routes departing Taipei (TPE)** * [Taipei to Hong Kong latency and RTT](/docs/network/latency/pairs/tpe-hkg-rtt): 15.5 ms * [Taipei to Tokyo latency and RTT](/docs/network/latency/pairs/tpe-tyo-rtt): 31.9 ms * [Taipei to Singapore latency and RTT](/docs/network/latency/pairs/tpe-sin-rtt): 44.3 ms * [Taipei to Seattle latency and RTT](/docs/network/latency/pairs/tpe-sea-rtt): 115.8 ms * [Taipei to Los Angeles latency and RTT](/docs/network/latency/pairs/tpe-lax-rtt): 132 ms **Fastest routes arriving at Berlin (BER)** * [Frankfurt to Berlin latency and RTT](/docs/network/latency/pairs/fra-ber-rtt): 6.1 ms * [Amsterdam to Berlin latency and RTT](/docs/network/latency/pairs/ams-ber-rtt): 7.2 ms * [Paris to Berlin latency and RTT](/docs/network/latency/pairs/par-ber-rtt): 15.5 ms * [London to Berlin latency and RTT](/docs/network/latency/pairs/lon-ber-rtt): 16.2 ms * [Marseille to Berlin latency and RTT](/docs/network/latency/pairs/mrs-ber-rtt): 20.6 ms **Same corridor (Asia Pacific → Europe)** * [Hong Kong to Moscow latency and RTT](/docs/network/latency/pairs/hkg-mow-rtt): 118.2 ms * [Taipei to Moscow latency and RTT](/docs/network/latency/pairs/tpe-mow-rtt): 132.7 ms * [Singapore to Marseille latency and RTT](/docs/network/latency/pairs/sin-mrs-rtt): 140.6 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/tpe-ber.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Taipei → Berlin round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/tpe-ber-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Taipei, Taiwan → Frankfurt, Germany RTT 🇹🇼 **Taipei, Taiwan (TPE)** → 🇩🇪 **Frankfurt, Germany (FRA)** Measurement round: `2026-08-16T04:07:28Z`. Taipei to Frankfurt averaged 166 ms across 50 replies. The observed range was 162.67 to 174.81 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 2.89 ms, or 1.7% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------------------- | | RTT | **166 ms** | | Jitter | **2.22 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **91.96 ms** | | Fiber Efficiency | **55.4%** | | Latency Tier | Fair | | Source Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | | Destination Region | [Europe](/docs/network/latency/regions/europe) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Taipei](https://www.geonames.org/1668341) (25.05306, 121.52639) and [Frankfurt am Main](https://www.geonames.org/2925533) (50.11552, 8.68417). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **9,390.8 km** | | Vacuum RTT floor | **62.65 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **91.96 ms** | | Low-latency fiber material floor | **91.59 ms** | | Engineering floor (5% path allowance) | **96.57 ms** | | Research 1.33× mapped-fiber reference | **122.31 ms** | | Estimated unamplified path loss | **1972.1 dB** | | Transparent optical spans / inline amplifiers | **124 / 123** | | Published RTT inflation over fiber floor | **1.81×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **162.67 ms** | | Average RTT | **166 ms** | | Maximum RTT | **174.81 ms** | | Standard deviation | **2.89 ms** | | Stdev / average | **1.7%** | ## Route Context * Departure PoP: [Taipei (TPE)](/docs/network/latency/tpe-taipei) * Destination PoP: [Frankfurt (FRA)](/docs/network/latency/fra-frankfurt) * Region overview: [Asia Pacific](/docs/network/latency/regions/asia-pacific) → [Europe](/docs/network/latency/regions/europe) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Frankfurt to Taipei latency and RTT](/docs/network/latency/pairs/fra-tpe-rtt): 163.8 ms **Fastest routes departing Taipei (TPE)** * [Taipei to Hong Kong latency and RTT](/docs/network/latency/pairs/tpe-hkg-rtt): 15.5 ms * [Taipei to Tokyo latency and RTT](/docs/network/latency/pairs/tpe-tyo-rtt): 31.9 ms * [Taipei to Singapore latency and RTT](/docs/network/latency/pairs/tpe-sin-rtt): 44.3 ms * [Taipei to Seattle latency and RTT](/docs/network/latency/pairs/tpe-sea-rtt): 115.8 ms * [Taipei to Los Angeles latency and RTT](/docs/network/latency/pairs/tpe-lax-rtt): 132 ms **Fastest routes arriving at Frankfurt (FRA)** * [Amsterdam to Frankfurt latency and RTT](/docs/network/latency/pairs/ams-fra-rtt): 5.9 ms * [Berlin to Frankfurt latency and RTT](/docs/network/latency/pairs/ber-fra-rtt): 6.1 ms * [Paris to Frankfurt latency and RTT](/docs/network/latency/pairs/par-fra-rtt): 7.6 ms * [London to Frankfurt latency and RTT](/docs/network/latency/pairs/lon-fra-rtt): 13.6 ms * [Marseille to Frankfurt latency and RTT](/docs/network/latency/pairs/mrs-fra-rtt): 16 ms **Same corridor (Asia Pacific → Europe)** * [Hong Kong to Moscow latency and RTT](/docs/network/latency/pairs/hkg-mow-rtt): 118.2 ms * [Taipei to Moscow latency and RTT](/docs/network/latency/pairs/tpe-mow-rtt): 132.7 ms * [Singapore to Marseille latency and RTT](/docs/network/latency/pairs/sin-mrs-rtt): 140.6 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/tpe-fra.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Taipei → Frankfurt round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/tpe-fra-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Taipei, Taiwan → São Paulo, Brazil RTT 🇹🇼 **Taipei, Taiwan (TPE)** → 🇧🇷 **São Paulo, Brazil (GRU)** Measurement round: `2026-08-16T04:07:28Z`. Taipei to São Paulo averaged 273 ms across 50 replies. The observed range was 261.55 to 292.28 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 7.67 ms, or 2.8% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **273 ms** | | Jitter | **8.08 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **184.29 ms** | | Fiber Efficiency | **67.5%** | | Latency Tier | High | | Source Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | | Destination Region | [South America](/docs/network/latency/regions/south-america) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Taipei](https://www.geonames.org/1668341) (25.05306, 121.52639) and [São Paulo](https://www.geonames.org/3448439) (-23.54750, -46.63611). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | --------------- | | WGS-84 geodesic distance | **18,818.5 km** | | Vacuum RTT floor | **125.54 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **184.29 ms** | | Low-latency fiber material floor | **183.54 ms** | | Engineering floor (5% path allowance) | **193.52 ms** | | Research 1.33× mapped-fiber reference | **245.1 ms** | | Estimated unamplified path loss | **3951.9 dB** | | Transparent optical spans / inline amplifiers | **247 / 246** | | Published RTT inflation over fiber floor | **1.48×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **261.55 ms** | | Average RTT | **273 ms** | | Maximum RTT | **292.28 ms** | | Standard deviation | **7.67 ms** | | Stdev / average | **2.8%** | ## Route Context * Departure PoP: [Taipei (TPE)](/docs/network/latency/tpe-taipei) * Destination PoP: [São Paulo (GRU)](/docs/network/latency/gru-sao-paulo) * Region overview: [Asia Pacific](/docs/network/latency/regions/asia-pacific) → [South America](/docs/network/latency/regions/south-america) * Corridor overview: [Asia Pacific to South America](/docs/network/latency/regions/asia-pacific-to-south-america) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [São Paulo to Taipei latency and RTT](/docs/network/latency/pairs/gru-tpe-rtt): 259.1 ms **Fastest routes departing Taipei (TPE)** * [Taipei to Hong Kong latency and RTT](/docs/network/latency/pairs/tpe-hkg-rtt): 15.5 ms * [Taipei to Tokyo latency and RTT](/docs/network/latency/pairs/tpe-tyo-rtt): 31.9 ms * [Taipei to Singapore latency and RTT](/docs/network/latency/pairs/tpe-sin-rtt): 44.3 ms * [Taipei to Seattle latency and RTT](/docs/network/latency/pairs/tpe-sea-rtt): 115.8 ms * [Taipei to Los Angeles latency and RTT](/docs/network/latency/pairs/tpe-lax-rtt): 132 ms **Fastest routes arriving at São Paulo (GRU)** * [Ashburn to São Paulo latency and RTT](/docs/network/latency/pairs/iad-gru-rtt): 101.4 ms * [New York to São Paulo latency and RTT](/docs/network/latency/pairs/nyc-gru-rtt): 106.9 ms * [Miami to São Paulo latency and RTT](/docs/network/latency/pairs/mia-gru-rtt): 128.7 ms * [Los Angeles to São Paulo latency and RTT](/docs/network/latency/pairs/lax-gru-rtt): 131.3 ms * [Seattle to São Paulo latency and RTT](/docs/network/latency/pairs/sea-gru-rtt): 155.7 ms **Same corridor (Asia Pacific → South America)** * [Tokyo to São Paulo latency and RTT](/docs/network/latency/pairs/tyo-gru-rtt): 230.5 ms * [Hong Kong to São Paulo latency and RTT](/docs/network/latency/pairs/hkg-gru-rtt): 287.6 ms * [Sydney to São Paulo latency and RTT](/docs/network/latency/pairs/syd-gru-rtt): 296.7 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/tpe-gru.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Taipei → São Paulo round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/tpe-gru-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Taipei, Taiwan → Hong Kong RTT 🇹🇼 **Taipei, Taiwan (TPE)** → 🇭🇰 **Hong Kong (HKG)** Measurement round: `2026-08-16T04:07:28Z`. Taipei to Hong Kong averaged 15.5 ms across 50 replies. The observed range was 15.09 to 17.58 ms; these figures describe one probe round, not a long-term service target. This is number 1 of 19 outgoing routes by average RTT. That ranking compares different destinations and is not adjusted for distance. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------------------- | | RTT | **15.5 ms** | | Jitter | **0.34 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **7.94 ms** | | Fiber Efficiency | **51.2%** | | Latency Tier | Ultra-Low | | Source Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | | Destination Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Taipei](https://www.geonames.org/1668341) (25.05306, 121.52639) and [Hong Kong](https://www.geonames.org/1819729) (22.27832, 114.17469). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | ------------ | | WGS-84 geodesic distance | **810.3 km** | | Vacuum RTT floor | **5.41 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **7.94 ms** | | Low-latency fiber material floor | **7.9 ms** | | Engineering floor (5% path allowance) | **8.33 ms** | | Research 1.33× mapped-fiber reference | **10.55 ms** | | Estimated unamplified path loss | **170.2 dB** | | Transparent optical spans / inline amplifiers | **11 / 10** | | Published RTT inflation over fiber floor | **1.95×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------ | | Minimum RTT | **15.09 ms** | | Average RTT | **15.5 ms** | | Maximum RTT | **17.58 ms** | | Standard deviation | **0.41 ms** | | Stdev / average | **2.6%** | ## Route Context * Departure PoP: [Taipei (TPE)](/docs/network/latency/tpe-taipei) * Destination PoP: [Hong Kong (HKG)](/docs/network/latency/hkg-hong-kong) * Region overview: [Asia Pacific](/docs/network/latency/regions/asia-pacific) → [Asia Pacific](/docs/network/latency/regions/asia-pacific) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Hong Kong to Taipei latency and RTT](/docs/network/latency/pairs/hkg-tpe-rtt): 14.7 ms **Fastest routes departing Taipei (TPE)** * [Taipei to Tokyo latency and RTT](/docs/network/latency/pairs/tpe-tyo-rtt): 31.9 ms * [Taipei to Singapore latency and RTT](/docs/network/latency/pairs/tpe-sin-rtt): 44.3 ms * [Taipei to Seattle latency and RTT](/docs/network/latency/pairs/tpe-sea-rtt): 115.8 ms * [Taipei to Los Angeles latency and RTT](/docs/network/latency/pairs/tpe-lax-rtt): 132 ms * [Taipei to Moscow latency and RTT](/docs/network/latency/pairs/tpe-mow-rtt): 132.7 ms **Fastest routes arriving at Hong Kong (HKG)** * [Singapore to Hong Kong latency and RTT](/docs/network/latency/pairs/sin-hkg-rtt): 30.8 ms * [Tokyo to Hong Kong latency and RTT](/docs/network/latency/pairs/tyo-hkg-rtt): 44.8 ms * [Moscow to Hong Kong latency and RTT](/docs/network/latency/pairs/mow-hkg-rtt): 118.6 ms * [Seattle to Hong Kong latency and RTT](/docs/network/latency/pairs/sea-hkg-rtt): 130.4 ms * [Sydney to Hong Kong latency and RTT](/docs/network/latency/pairs/syd-hkg-rtt): 135.7 ms **Same corridor (Asia Pacific → Asia Pacific)** * [Melbourne to Sydney latency and RTT](/docs/network/latency/pairs/mel-syd-rtt): 9.8 ms * [Sydney to Melbourne latency and RTT](/docs/network/latency/pairs/syd-mel-rtt): 9.8 ms * [Hong Kong to Singapore latency and RTT](/docs/network/latency/pairs/hkg-sin-rtt): 31.5 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/tpe-hkg.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Taipei → Hong Kong round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/tpe-hkg-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Taipei, Taiwan → Ashburn, USA RTT 🇹🇼 **Taipei, Taiwan (TPE)** → 🇺🇸 **Ashburn, USA (IAD)** Measurement round: `2026-08-16T04:07:28Z`. Taipei to Ashburn averaged 171.6 ms across 50 replies. The observed range was 163.88 to 199.15 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 7 ms, or 4.1% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **171.6 ms** | | Jitter | **4.89 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **123.78 ms** | | Fiber Efficiency | **72.1%** | | Latency Tier | Fair | | Source Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | | Destination Region | [North America](/docs/network/latency/regions/north-america) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Taipei](https://www.geonames.org/1668341) (25.05306, 121.52639) and [Ashburn](https://www.geonames.org/4744870) (39.04372, -77.48749). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | --------------- | | WGS-84 geodesic distance | **12,639.6 km** | | Vacuum RTT floor | **84.32 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **123.78 ms** | | Low-latency fiber material floor | **123.28 ms** | | Engineering floor (5% path allowance) | **129.98 ms** | | Research 1.33× mapped-fiber reference | **164.62 ms** | | Estimated unamplified path loss | **2654.3 dB** | | Transparent optical spans / inline amplifiers | **166 / 165** | | Published RTT inflation over fiber floor | **1.39×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **163.88 ms** | | Average RTT | **171.6 ms** | | Maximum RTT | **199.15 ms** | | Standard deviation | **7 ms** | | Stdev / average | **4.1%** | ## Route Context * Departure PoP: [Taipei (TPE)](/docs/network/latency/tpe-taipei) * Destination PoP: [Ashburn (IAD)](/docs/network/latency/iad-ashburn) * Region overview: [Asia Pacific](/docs/network/latency/regions/asia-pacific) → [North America](/docs/network/latency/regions/north-america) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Ashburn to Taipei latency and RTT](/docs/network/latency/pairs/iad-tpe-rtt): 172.3 ms **Fastest routes departing Taipei (TPE)** * [Taipei to Hong Kong latency and RTT](/docs/network/latency/pairs/tpe-hkg-rtt): 15.5 ms * [Taipei to Tokyo latency and RTT](/docs/network/latency/pairs/tpe-tyo-rtt): 31.9 ms * [Taipei to Singapore latency and RTT](/docs/network/latency/pairs/tpe-sin-rtt): 44.3 ms * [Taipei to Seattle latency and RTT](/docs/network/latency/pairs/tpe-sea-rtt): 115.8 ms * [Taipei to Los Angeles latency and RTT](/docs/network/latency/pairs/tpe-lax-rtt): 132 ms **Fastest routes arriving at Ashburn (IAD)** * [New York to Ashburn latency and RTT](/docs/network/latency/pairs/nyc-iad-rtt): 6 ms * [Miami to Ashburn latency and RTT](/docs/network/latency/pairs/mia-iad-rtt): 27.3 ms * [Los Angeles to Ashburn latency and RTT](/docs/network/latency/pairs/lax-iad-rtt): 60.4 ms * [Seattle to Ashburn latency and RTT](/docs/network/latency/pairs/sea-iad-rtt): 61.5 ms * [London to Ashburn latency and RTT](/docs/network/latency/pairs/lon-iad-rtt): 71 ms **Same corridor (Asia Pacific → North America)** * [Tokyo to Seattle latency and RTT](/docs/network/latency/pairs/tyo-sea-rtt): 84.8 ms * [Tokyo to Los Angeles latency and RTT](/docs/network/latency/pairs/tyo-lax-rtt): 101.2 ms * [Hong Kong to Seattle latency and RTT](/docs/network/latency/pairs/hkg-sea-rtt): 131.6 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/tpe-iad.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Taipei → Ashburn round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/tpe-iad-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Taipei, Taiwan → Johannesburg, South Africa RTT 🇹🇼 **Taipei, Taiwan (TPE)** → 🇿🇦 **Johannesburg, South Africa (JNB)** Measurement round: `2026-08-16T04:07:28Z`. Taipei to Johannesburg averaged 331 ms, compared with a 112.87 ms city-centre fiber reference. The difference cannot be assigned to a particular cable or carrier from RTT alone. Standard deviation was 6.31 ms, or 1.9% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------------------- | | RTT | **331 ms** | | Jitter | **5.35 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **112.87 ms** | | Fiber Efficiency | **34.1%** | | Latency Tier | High | | Source Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | | Destination Region | [Africa](/docs/network/latency/regions/africa) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Taipei](https://www.geonames.org/1668341) (25.05306, 121.52639) and [Johannesburg](https://www.geonames.org/993800) (-26.20227, 28.04363). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | --------------- | | WGS-84 geodesic distance | **11,525.4 km** | | Vacuum RTT floor | **76.89 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **112.87 ms** | | Low-latency fiber material floor | **112.41 ms** | | Engineering floor (5% path allowance) | **118.52 ms** | | Research 1.33× mapped-fiber reference | **150.11 ms** | | Estimated unamplified path loss | **2420.3 dB** | | Transparent optical spans / inline amplifiers | **152 / 151** | | Published RTT inflation over fiber floor | **2.93×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **322.36 ms** | | Average RTT | **331 ms** | | Maximum RTT | **356.12 ms** | | Standard deviation | **6.31 ms** | | Stdev / average | **1.9%** | ## Route Context * Departure PoP: [Taipei (TPE)](/docs/network/latency/tpe-taipei) * Destination PoP: [Johannesburg (JNB)](/docs/network/latency/jnb-johannesburg) * Region overview: [Asia Pacific](/docs/network/latency/regions/asia-pacific) → [Africa](/docs/network/latency/regions/africa) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Johannesburg to Taipei latency and RTT](/docs/network/latency/pairs/jnb-tpe-rtt): 329.8 ms **Fastest routes departing Taipei (TPE)** * [Taipei to Hong Kong latency and RTT](/docs/network/latency/pairs/tpe-hkg-rtt): 15.5 ms * [Taipei to Tokyo latency and RTT](/docs/network/latency/pairs/tpe-tyo-rtt): 31.9 ms * [Taipei to Singapore latency and RTT](/docs/network/latency/pairs/tpe-sin-rtt): 44.3 ms * [Taipei to Seattle latency and RTT](/docs/network/latency/pairs/tpe-sea-rtt): 115.8 ms * [Taipei to Los Angeles latency and RTT](/docs/network/latency/pairs/tpe-lax-rtt): 132 ms **Fastest routes arriving at Johannesburg (JNB)** * [Amsterdam to Johannesburg latency and RTT](/docs/network/latency/pairs/ams-jnb-rtt): 171.9 ms * [Berlin to Johannesburg latency and RTT](/docs/network/latency/pairs/ber-jnb-rtt): 172.1 ms * [London to Johannesburg latency and RTT](/docs/network/latency/pairs/lon-jnb-rtt): 179.6 ms * [Paris to Johannesburg latency and RTT](/docs/network/latency/pairs/par-jnb-rtt): 180.9 ms * [Frankfurt to Johannesburg latency and RTT](/docs/network/latency/pairs/fra-jnb-rtt): 188.4 ms **Same corridor (Asia Pacific → Africa)** * [Singapore to Johannesburg latency and RTT](/docs/network/latency/pairs/sin-jnb-rtt): 312.6 ms * [Hong Kong to Johannesburg latency and RTT](/docs/network/latency/pairs/hkg-jnb-rtt): 316.7 ms * [Tokyo to Johannesburg latency and RTT](/docs/network/latency/pairs/tyo-jnb-rtt): 359.2 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/tpe-jnb.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Taipei → Johannesburg round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/tpe-jnb-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Taipei, Taiwan → Los Angeles, USA RTT 🇹🇼 **Taipei, Taiwan (TPE)** → 🇺🇸 **Los Angeles, USA (LAX)** Measurement round: `2026-08-16T04:07:28Z`. Taipei to Los Angeles averaged 132 ms across 50 replies. The observed range was 129.33 to 144.07 ms; these figures describe one probe round, not a long-term service target. This is number 5 of 19 outgoing routes by average RTT. That ranking compares different destinations and is not adjusted for distance. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **132 ms** | | Jitter | **1.62 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **107 ms** | | Fiber Efficiency | **81.1%** | | Latency Tier | Good | | Source Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | | Destination Region | [North America](/docs/network/latency/regions/north-america) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Taipei](https://www.geonames.org/1668341) (25.05306, 121.52639) and [Los Angeles](https://www.geonames.org/5368361) (34.05223, -118.24368). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | --------------- | | WGS-84 geodesic distance | **10,926.8 km** | | Vacuum RTT floor | **72.9 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **107 ms** | | Low-latency fiber material floor | **106.57 ms** | | Engineering floor (5% path allowance) | **112.37 ms** | | Research 1.33× mapped-fiber reference | **142.32 ms** | | Estimated unamplified path loss | **2294.6 dB** | | Transparent optical spans / inline amplifiers | **144 / 143** | | Published RTT inflation over fiber floor | **1.23×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **129.33 ms** | | Average RTT | **132 ms** | | Maximum RTT | **144.07 ms** | | Standard deviation | **2.56 ms** | | Stdev / average | **1.9%** | ## Route Context * Departure PoP: [Taipei (TPE)](/docs/network/latency/tpe-taipei) * Destination PoP: [Los Angeles (LAX)](/docs/network/latency/lax-los-angeles) * Region overview: [Asia Pacific](/docs/network/latency/regions/asia-pacific) → [North America](/docs/network/latency/regions/north-america) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Los Angeles to Taipei latency and RTT](/docs/network/latency/pairs/lax-tpe-rtt): 132.9 ms **Fastest routes departing Taipei (TPE)** * [Taipei to Hong Kong latency and RTT](/docs/network/latency/pairs/tpe-hkg-rtt): 15.5 ms * [Taipei to Tokyo latency and RTT](/docs/network/latency/pairs/tpe-tyo-rtt): 31.9 ms * [Taipei to Singapore latency and RTT](/docs/network/latency/pairs/tpe-sin-rtt): 44.3 ms * [Taipei to Seattle latency and RTT](/docs/network/latency/pairs/tpe-sea-rtt): 115.8 ms * [Taipei to Moscow latency and RTT](/docs/network/latency/pairs/tpe-mow-rtt): 132.7 ms **Fastest routes arriving at Los Angeles (LAX)** * [Seattle to Los Angeles latency and RTT](/docs/network/latency/pairs/sea-lax-rtt): 25.9 ms * [Miami to Los Angeles latency and RTT](/docs/network/latency/pairs/mia-lax-rtt): 56.8 ms * [New York to Los Angeles latency and RTT](/docs/network/latency/pairs/nyc-lax-rtt): 58.9 ms * [Ashburn to Los Angeles latency and RTT](/docs/network/latency/pairs/iad-lax-rtt): 60.7 ms * [Tokyo to Los Angeles latency and RTT](/docs/network/latency/pairs/tyo-lax-rtt): 101.2 ms **Same corridor (Asia Pacific → North America)** * [Tokyo to Seattle latency and RTT](/docs/network/latency/pairs/tyo-sea-rtt): 84.8 ms * [Hong Kong to Seattle latency and RTT](/docs/network/latency/pairs/hkg-sea-rtt): 131.6 ms * [Sydney to Los Angeles latency and RTT](/docs/network/latency/pairs/syd-lax-rtt): 136.5 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/tpe-lax.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Taipei → Los Angeles round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/tpe-lax-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Taipei, Taiwan → London, UK RTT 🇹🇼 **Taipei, Taiwan (TPE)** → 🇬🇧 **London, UK (LON)** Measurement round: `2026-08-16T04:07:28Z`. Taipei to London averaged 173 ms across 50 replies. The observed range was 165.18 to 197.84 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 6.02 ms, or 3.5% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------------------- | | RTT | **173 ms** | | Jitter | **4.56 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **95.97 ms** | | Fiber Efficiency | **55.5%** | | Latency Tier | Fair | | Source Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | | Destination Region | [Europe](/docs/network/latency/regions/europe) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Taipei](https://www.geonames.org/1668341) (25.05306, 121.52639) and [London](https://www.geonames.org/2643743) (51.50853, -0.12574). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **9,799.7 km** | | Vacuum RTT floor | **65.38 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **95.97 ms** | | Low-latency fiber material floor | **95.58 ms** | | Engineering floor (5% path allowance) | **100.78 ms** | | Research 1.33× mapped-fiber reference | **127.63 ms** | | Estimated unamplified path loss | **2057.9 dB** | | Transparent optical spans / inline amplifiers | **129 / 128** | | Published RTT inflation over fiber floor | **1.8×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **165.18 ms** | | Average RTT | **173 ms** | | Maximum RTT | **197.84 ms** | | Standard deviation | **6.02 ms** | | Stdev / average | **3.5%** | ## Route Context * Departure PoP: [Taipei (TPE)](/docs/network/latency/tpe-taipei) * Destination PoP: [London (LON)](/docs/network/latency/lon-london) * Region overview: [Asia Pacific](/docs/network/latency/regions/asia-pacific) → [Europe](/docs/network/latency/regions/europe) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [London to Taipei latency and RTT](/docs/network/latency/pairs/lon-tpe-rtt): 173.8 ms **Fastest routes departing Taipei (TPE)** * [Taipei to Hong Kong latency and RTT](/docs/network/latency/pairs/tpe-hkg-rtt): 15.5 ms * [Taipei to Tokyo latency and RTT](/docs/network/latency/pairs/tpe-tyo-rtt): 31.9 ms * [Taipei to Singapore latency and RTT](/docs/network/latency/pairs/tpe-sin-rtt): 44.3 ms * [Taipei to Seattle latency and RTT](/docs/network/latency/pairs/tpe-sea-rtt): 115.8 ms * [Taipei to Los Angeles latency and RTT](/docs/network/latency/pairs/tpe-lax-rtt): 132 ms **Fastest routes arriving at London (LON)** * [Amsterdam to London latency and RTT](/docs/network/latency/pairs/ams-lon-rtt): 5.2 ms * [Paris to London latency and RTT](/docs/network/latency/pairs/par-lon-rtt): 6.4 ms * [Frankfurt to London latency and RTT](/docs/network/latency/pairs/fra-lon-rtt): 12.9 ms * [Berlin to London latency and RTT](/docs/network/latency/pairs/ber-lon-rtt): 15.2 ms * [Marseille to London latency and RTT](/docs/network/latency/pairs/mrs-lon-rtt): 18.9 ms **Same corridor (Asia Pacific → Europe)** * [Hong Kong to Moscow latency and RTT](/docs/network/latency/pairs/hkg-mow-rtt): 118.2 ms * [Taipei to Moscow latency and RTT](/docs/network/latency/pairs/tpe-mow-rtt): 132.7 ms * [Singapore to Marseille latency and RTT](/docs/network/latency/pairs/sin-mrs-rtt): 140.6 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/tpe-lon.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Taipei → London round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/tpe-lon-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Taipei, Taiwan → Melbourne, Australia RTT 🇹🇼 **Taipei, Taiwan (TPE)** → 🇦🇺 **Melbourne, Australia (MEL)** Measurement round: `2026-08-16T04:07:28Z`. Taipei to Melbourne averaged 142 ms across 50 replies. The observed range was 135.86 to 164.91 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 6.17 ms, or 4.3% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------------------- | | RTT | **142 ms** | | Jitter | **5.1 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **72.25 ms** | | Fiber Efficiency | **50.9%** | | Latency Tier | Good | | Source Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | | Destination Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Taipei](https://www.geonames.org/1668341) (25.05306, 121.52639) and [Melbourne](https://www.geonames.org/2158177) (-37.81400, 144.96332). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **7,377.8 km** | | Vacuum RTT floor | **49.22 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **72.25 ms** | | Low-latency fiber material floor | **71.96 ms** | | Engineering floor (5% path allowance) | **75.87 ms** | | Research 1.33× mapped-fiber reference | **96.09 ms** | | Estimated unamplified path loss | **1549.3 dB** | | Transparent optical spans / inline amplifiers | **97 / 96** | | Published RTT inflation over fiber floor | **1.97×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **135.86 ms** | | Average RTT | **142 ms** | | Maximum RTT | **164.91 ms** | | Standard deviation | **6.17 ms** | | Stdev / average | **4.3%** | ## Route Context * Departure PoP: [Taipei (TPE)](/docs/network/latency/tpe-taipei) * Destination PoP: [Melbourne (MEL)](/docs/network/latency/mel-melbourne) * Region overview: [Asia Pacific](/docs/network/latency/regions/asia-pacific) → [Asia Pacific](/docs/network/latency/regions/asia-pacific) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Melbourne to Taipei latency and RTT](/docs/network/latency/pairs/mel-tpe-rtt): 143.2 ms **Fastest routes departing Taipei (TPE)** * [Taipei to Hong Kong latency and RTT](/docs/network/latency/pairs/tpe-hkg-rtt): 15.5 ms * [Taipei to Tokyo latency and RTT](/docs/network/latency/pairs/tpe-tyo-rtt): 31.9 ms * [Taipei to Singapore latency and RTT](/docs/network/latency/pairs/tpe-sin-rtt): 44.3 ms * [Taipei to Seattle latency and RTT](/docs/network/latency/pairs/tpe-sea-rtt): 115.8 ms * [Taipei to Los Angeles latency and RTT](/docs/network/latency/pairs/tpe-lax-rtt): 132 ms **Fastest routes arriving at Melbourne (MEL)** * [Sydney to Melbourne latency and RTT](/docs/network/latency/pairs/syd-mel-rtt): 9.8 ms * [Singapore to Melbourne latency and RTT](/docs/network/latency/pairs/sin-mel-rtt): 88.6 ms * [Tokyo to Melbourne latency and RTT](/docs/network/latency/pairs/tyo-mel-rtt): 113.9 ms * [Hong Kong to Melbourne latency and RTT](/docs/network/latency/pairs/hkg-mel-rtt): 138.4 ms * [Los Angeles to Melbourne latency and RTT](/docs/network/latency/pairs/lax-mel-rtt): 148.2 ms **Same corridor (Asia Pacific → Asia Pacific)** * [Melbourne to Sydney latency and RTT](/docs/network/latency/pairs/mel-syd-rtt): 9.8 ms * [Hong Kong to Taipei latency and RTT](/docs/network/latency/pairs/hkg-tpe-rtt): 14.7 ms * [Singapore to Hong Kong latency and RTT](/docs/network/latency/pairs/sin-hkg-rtt): 30.8 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/tpe-mel.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Taipei → Melbourne round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/tpe-mel-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Taipei, Taiwan → Miami, USA RTT 🇹🇼 **Taipei, Taiwan (TPE)** → 🇺🇸 **Miami, USA (MIA)** Measurement round: `2026-08-16T04:07:28Z`. Taipei to Miami averaged 185.7 ms across 50 replies. The observed range was 175.97 to 211.22 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 9.15 ms, or 4.9% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **185.7 ms** | | Jitter | **7.6 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **136.3 ms** | | Fiber Efficiency | **73.4%** | | Latency Tier | Fair | | Source Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | | Destination Region | [North America](/docs/network/latency/regions/north-america) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Taipei](https://www.geonames.org/1668341) (25.05306, 121.52639) and [Miami](https://www.geonames.org/4164138) (25.77427, -80.19366). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | --------------- | | WGS-84 geodesic distance | **13,918.7 km** | | Vacuum RTT floor | **92.86 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **136.3 ms** | | Low-latency fiber material floor | **135.76 ms** | | Engineering floor (5% path allowance) | **143.14 ms** | | Research 1.33× mapped-fiber reference | **181.28 ms** | | Estimated unamplified path loss | **2922.9 dB** | | Transparent optical spans / inline amplifiers | **183 / 182** | | Published RTT inflation over fiber floor | **1.36×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **175.97 ms** | | Average RTT | **185.7 ms** | | Maximum RTT | **211.22 ms** | | Standard deviation | **9.15 ms** | | Stdev / average | **4.9%** | ## Route Context * Departure PoP: [Taipei (TPE)](/docs/network/latency/tpe-taipei) * Destination PoP: [Miami (MIA)](/docs/network/latency/mia-miami) * Region overview: [Asia Pacific](/docs/network/latency/regions/asia-pacific) → [North America](/docs/network/latency/regions/north-america) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Miami to Taipei latency and RTT](/docs/network/latency/pairs/mia-tpe-rtt): 185.1 ms **Fastest routes departing Taipei (TPE)** * [Taipei to Hong Kong latency and RTT](/docs/network/latency/pairs/tpe-hkg-rtt): 15.5 ms * [Taipei to Tokyo latency and RTT](/docs/network/latency/pairs/tpe-tyo-rtt): 31.9 ms * [Taipei to Singapore latency and RTT](/docs/network/latency/pairs/tpe-sin-rtt): 44.3 ms * [Taipei to Seattle latency and RTT](/docs/network/latency/pairs/tpe-sea-rtt): 115.8 ms * [Taipei to Los Angeles latency and RTT](/docs/network/latency/pairs/tpe-lax-rtt): 132 ms **Fastest routes arriving at Miami (MIA)** * [Ashburn to Miami latency and RTT](/docs/network/latency/pairs/iad-mia-rtt): 27.4 ms * [New York to Miami latency and RTT](/docs/network/latency/pairs/nyc-mia-rtt): 32.9 ms * [Los Angeles to Miami latency and RTT](/docs/network/latency/pairs/lax-mia-rtt): 57.3 ms * [São Paulo to Miami latency and RTT](/docs/network/latency/pairs/gru-mia-rtt): 74 ms * [Seattle to Miami latency and RTT](/docs/network/latency/pairs/sea-mia-rtt): 81.7 ms **Same corridor (Asia Pacific → North America)** * [Tokyo to Seattle latency and RTT](/docs/network/latency/pairs/tyo-sea-rtt): 84.8 ms * [Tokyo to Los Angeles latency and RTT](/docs/network/latency/pairs/tyo-lax-rtt): 101.2 ms * [Hong Kong to Seattle latency and RTT](/docs/network/latency/pairs/hkg-sea-rtt): 131.6 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/tpe-mia.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Taipei → Miami round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/tpe-mia-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Taipei, Taiwan → Moscow, Russia RTT 🇹🇼 **Taipei, Taiwan (TPE)** → 🇷🇺 **Moscow, Russia (MOW)** Measurement round: `2026-08-16T04:07:28Z`. Taipei to Moscow averaged 132.7 ms across 50 replies. The observed range was 127.77 to 145.28 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 3.91 ms, or 2.9% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------------------- | | RTT | **132.7 ms** | | Jitter | **3.84 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **72.16 ms** | | Fiber Efficiency | **54.4%** | | Latency Tier | Good | | Source Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | | Destination Region | [Europe](/docs/network/latency/regions/europe) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Taipei](https://www.geonames.org/1668341) (25.05306, 121.52639) and [Moscow](https://www.geonames.org/524901) (55.75204, 37.61781). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | ------------- | | WGS-84 geodesic distance | **7,369 km** | | Vacuum RTT floor | **49.16 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **72.16 ms** | | Low-latency fiber material floor | **71.87 ms** | | Engineering floor (5% path allowance) | **75.78 ms** | | Research 1.33× mapped-fiber reference | **95.98 ms** | | Estimated unamplified path loss | **1547.5 dB** | | Transparent optical spans / inline amplifiers | **97 / 96** | | Published RTT inflation over fiber floor | **1.84×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **127.77 ms** | | Average RTT | **132.7 ms** | | Maximum RTT | **145.28 ms** | | Standard deviation | **3.91 ms** | | Stdev / average | **2.9%** | ## Route Context * Departure PoP: [Taipei (TPE)](/docs/network/latency/tpe-taipei) * Destination PoP: [Moscow (MOW)](/docs/network/latency/mow-moscow) * Region overview: [Asia Pacific](/docs/network/latency/regions/asia-pacific) → [Europe](/docs/network/latency/regions/europe) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Moscow to Taipei latency and RTT](/docs/network/latency/pairs/mow-tpe-rtt): 132.1 ms **Fastest routes departing Taipei (TPE)** * [Taipei to Hong Kong latency and RTT](/docs/network/latency/pairs/tpe-hkg-rtt): 15.5 ms * [Taipei to Tokyo latency and RTT](/docs/network/latency/pairs/tpe-tyo-rtt): 31.9 ms * [Taipei to Singapore latency and RTT](/docs/network/latency/pairs/tpe-sin-rtt): 44.3 ms * [Taipei to Seattle latency and RTT](/docs/network/latency/pairs/tpe-sea-rtt): 115.8 ms * [Taipei to Los Angeles latency and RTT](/docs/network/latency/pairs/tpe-lax-rtt): 132 ms **Fastest routes arriving at Moscow (MOW)** * [Berlin to Moscow latency and RTT](/docs/network/latency/pairs/ber-mow-rtt): 28.6 ms * [Frankfurt to Moscow latency and RTT](/docs/network/latency/pairs/fra-mow-rtt): 34.7 ms * [Amsterdam to Moscow latency and RTT](/docs/network/latency/pairs/ams-mow-rtt): 38.1 ms * [London to Moscow latency and RTT](/docs/network/latency/pairs/lon-mow-rtt): 43.5 ms * [Paris to Moscow latency and RTT](/docs/network/latency/pairs/par-mow-rtt): 44.7 ms **Same corridor (Asia Pacific → Europe)** * [Hong Kong to Moscow latency and RTT](/docs/network/latency/pairs/hkg-mow-rtt): 118.2 ms * [Singapore to Marseille latency and RTT](/docs/network/latency/pairs/sin-mrs-rtt): 140.6 ms * [Hong Kong to Berlin latency and RTT](/docs/network/latency/pairs/hkg-ber-rtt): 145.1 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/tpe-mow.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Taipei → Moscow round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/tpe-mow-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Taipei, Taiwan → Marseille, France RTT 🇹🇼 **Taipei, Taiwan (TPE)** → 🇫🇷 **Marseille, France (MRS)** Measurement round: `2026-08-16T04:07:28Z`. Taipei to Marseille averaged 179.5 ms across 50 replies. The observed range was 175.71 to 189.07 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 3.41 ms, or 1.9% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------------------- | | RTT | **179.5 ms** | | Jitter | **2.52 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **98.2 ms** | | Fiber Efficiency | **54.7%** | | Latency Tier | Fair | | Source Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | | Destination Region | [Europe](/docs/network/latency/regions/europe) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Taipei](https://www.geonames.org/1668341) (25.05306, 121.52639) and [Marseille](https://www.geonames.org/2995469) (43.29695, 5.38107). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | --------------- | | WGS-84 geodesic distance | **10,027.3 km** | | Vacuum RTT floor | **66.9 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **98.2 ms** | | Low-latency fiber material floor | **97.8 ms** | | Engineering floor (5% path allowance) | **103.12 ms** | | Research 1.33× mapped-fiber reference | **130.6 ms** | | Estimated unamplified path loss | **2105.7 dB** | | Transparent optical spans / inline amplifiers | **132 / 131** | | Published RTT inflation over fiber floor | **1.83×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **175.71 ms** | | Average RTT | **179.5 ms** | | Maximum RTT | **189.07 ms** | | Standard deviation | **3.41 ms** | | Stdev / average | **1.9%** | ## Route Context * Departure PoP: [Taipei (TPE)](/docs/network/latency/tpe-taipei) * Destination PoP: [Marseille (MRS)](/docs/network/latency/mrs-marseille) * Region overview: [Asia Pacific](/docs/network/latency/regions/asia-pacific) → [Europe](/docs/network/latency/regions/europe) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Marseille to Taipei latency and RTT](/docs/network/latency/pairs/mrs-tpe-rtt): 180.9 ms **Fastest routes departing Taipei (TPE)** * [Taipei to Hong Kong latency and RTT](/docs/network/latency/pairs/tpe-hkg-rtt): 15.5 ms * [Taipei to Tokyo latency and RTT](/docs/network/latency/pairs/tpe-tyo-rtt): 31.9 ms * [Taipei to Singapore latency and RTT](/docs/network/latency/pairs/tpe-sin-rtt): 44.3 ms * [Taipei to Seattle latency and RTT](/docs/network/latency/pairs/tpe-sea-rtt): 115.8 ms * [Taipei to Los Angeles latency and RTT](/docs/network/latency/pairs/tpe-lax-rtt): 132 ms **Fastest routes arriving at Marseille (MRS)** * [Paris to Marseille latency and RTT](/docs/network/latency/pairs/par-mrs-rtt): 8.9 ms * [Frankfurt to Marseille latency and RTT](/docs/network/latency/pairs/fra-mrs-rtt): 16.4 ms * [London to Marseille latency and RTT](/docs/network/latency/pairs/lon-mrs-rtt): 17.9 ms * [Amsterdam to Marseille latency and RTT](/docs/network/latency/pairs/ams-mrs-rtt): 19.6 ms * [Berlin to Marseille latency and RTT](/docs/network/latency/pairs/ber-mrs-rtt): 20.2 ms **Same corridor (Asia Pacific → Europe)** * [Hong Kong to Moscow latency and RTT](/docs/network/latency/pairs/hkg-mow-rtt): 118.2 ms * [Taipei to Moscow latency and RTT](/docs/network/latency/pairs/tpe-mow-rtt): 132.7 ms * [Singapore to Marseille latency and RTT](/docs/network/latency/pairs/sin-mrs-rtt): 140.6 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/tpe-mrs.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Taipei → Marseille round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/tpe-mrs-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Taipei, Taiwan → New York, USA RTT 🇹🇼 **Taipei, Taiwan (TPE)** → 🇺🇸 **New York, USA (NYC)** Measurement round: `2026-08-16T04:07:28Z`. Taipei to New York averaged 174.2 ms across 50 replies. The observed range was 165.46 to 214.55 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 8.46 ms, or 4.9% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **174.2 ms** | | Jitter | **6.73 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **122.89 ms** | | Fiber Efficiency | **70.5%** | | Latency Tier | Fair | | Source Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | | Destination Region | [North America](/docs/network/latency/regions/north-america) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Taipei](https://www.geonames.org/1668341) (25.05306, 121.52639) and [New York City](https://www.geonames.org/5128581) (40.71427, -74.00597). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | --------------- | | WGS-84 geodesic distance | **12,549.1 km** | | Vacuum RTT floor | **83.72 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **122.89 ms** | | Low-latency fiber material floor | **122.4 ms** | | Engineering floor (5% path allowance) | **129.05 ms** | | Research 1.33× mapped-fiber reference | **163.44 ms** | | Estimated unamplified path loss | **2635.3 dB** | | Transparent optical spans / inline amplifiers | **165 / 164** | | Published RTT inflation over fiber floor | **1.42×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **165.46 ms** | | Average RTT | **174.2 ms** | | Maximum RTT | **214.55 ms** | | Standard deviation | **8.46 ms** | | Stdev / average | **4.9%** | ## Route Context * Departure PoP: [Taipei (TPE)](/docs/network/latency/tpe-taipei) * Destination PoP: [New York (NYC)](/docs/network/latency/nyc-new-york) * Region overview: [Asia Pacific](/docs/network/latency/regions/asia-pacific) → [North America](/docs/network/latency/regions/north-america) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [New York to Taipei latency and RTT](/docs/network/latency/pairs/nyc-tpe-rtt): 173.4 ms **Fastest routes departing Taipei (TPE)** * [Taipei to Hong Kong latency and RTT](/docs/network/latency/pairs/tpe-hkg-rtt): 15.5 ms * [Taipei to Tokyo latency and RTT](/docs/network/latency/pairs/tpe-tyo-rtt): 31.9 ms * [Taipei to Singapore latency and RTT](/docs/network/latency/pairs/tpe-sin-rtt): 44.3 ms * [Taipei to Seattle latency and RTT](/docs/network/latency/pairs/tpe-sea-rtt): 115.8 ms * [Taipei to Los Angeles latency and RTT](/docs/network/latency/pairs/tpe-lax-rtt): 132 ms **Fastest routes arriving at New York (NYC)** * [Ashburn to New York latency and RTT](/docs/network/latency/pairs/iad-nyc-rtt): 6 ms * [Miami to New York latency and RTT](/docs/network/latency/pairs/mia-nyc-rtt): 33.3 ms * [Los Angeles to New York latency and RTT](/docs/network/latency/pairs/lax-nyc-rtt): 57.9 ms * [Seattle to New York latency and RTT](/docs/network/latency/pairs/sea-nyc-rtt): 59.5 ms * [London to New York latency and RTT](/docs/network/latency/pairs/lon-nyc-rtt): 63.8 ms **Same corridor (Asia Pacific → North America)** * [Tokyo to Seattle latency and RTT](/docs/network/latency/pairs/tyo-sea-rtt): 84.8 ms * [Tokyo to Los Angeles latency and RTT](/docs/network/latency/pairs/tyo-lax-rtt): 101.2 ms * [Hong Kong to Seattle latency and RTT](/docs/network/latency/pairs/hkg-sea-rtt): 131.6 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/tpe-nyc.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Taipei → New York round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/tpe-nyc-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Taipei, Taiwan → Paris, France RTT 🇹🇼 **Taipei, Taiwan (TPE)** → 🇫🇷 **Paris, France (PAR)** Measurement round: `2026-08-16T04:07:28Z`. Taipei to Paris averaged 173.8 ms across 50 replies. The observed range was 169.08 to 187.89 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 3.91 ms, or 2.2% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------------------- | | RTT | **173.8 ms** | | Jitter | **3.31 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **96.42 ms** | | Fiber Efficiency | **55.5%** | | Latency Tier | Fair | | Source Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | | Destination Region | [Europe](/docs/network/latency/regions/europe) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Taipei](https://www.geonames.org/1668341) (25.05306, 121.52639) and [Paris](https://www.geonames.org/2988507) (48.85341, 2.34880). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **9,846.2 km** | | Vacuum RTT floor | **65.69 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **96.42 ms** | | Low-latency fiber material floor | **96.03 ms** | | Engineering floor (5% path allowance) | **101.26 ms** | | Research 1.33× mapped-fiber reference | **128.24 ms** | | Estimated unamplified path loss | **2067.7 dB** | | Transparent optical spans / inline amplifiers | **130 / 129** | | Published RTT inflation over fiber floor | **1.8×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **169.08 ms** | | Average RTT | **173.8 ms** | | Maximum RTT | **187.89 ms** | | Standard deviation | **3.91 ms** | | Stdev / average | **2.2%** | ## Route Context * Departure PoP: [Taipei (TPE)](/docs/network/latency/tpe-taipei) * Destination PoP: [Paris (PAR)](/docs/network/latency/par-paris) * Region overview: [Asia Pacific](/docs/network/latency/regions/asia-pacific) → [Europe](/docs/network/latency/regions/europe) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Paris to Taipei latency and RTT](/docs/network/latency/pairs/par-tpe-rtt): 171.9 ms **Fastest routes departing Taipei (TPE)** * [Taipei to Hong Kong latency and RTT](/docs/network/latency/pairs/tpe-hkg-rtt): 15.5 ms * [Taipei to Tokyo latency and RTT](/docs/network/latency/pairs/tpe-tyo-rtt): 31.9 ms * [Taipei to Singapore latency and RTT](/docs/network/latency/pairs/tpe-sin-rtt): 44.3 ms * [Taipei to Seattle latency and RTT](/docs/network/latency/pairs/tpe-sea-rtt): 115.8 ms * [Taipei to Los Angeles latency and RTT](/docs/network/latency/pairs/tpe-lax-rtt): 132 ms **Fastest routes arriving at Paris (PAR)** * [London to Paris latency and RTT](/docs/network/latency/pairs/lon-par-rtt): 6.4 ms * [Amsterdam to Paris latency and RTT](/docs/network/latency/pairs/ams-par-rtt): 7 ms * [Frankfurt to Paris latency and RTT](/docs/network/latency/pairs/fra-par-rtt): 7.6 ms * [Marseille to Paris latency and RTT](/docs/network/latency/pairs/mrs-par-rtt): 8.9 ms * [Berlin to Paris latency and RTT](/docs/network/latency/pairs/ber-par-rtt): 15.6 ms **Same corridor (Asia Pacific → Europe)** * [Hong Kong to Moscow latency and RTT](/docs/network/latency/pairs/hkg-mow-rtt): 118.2 ms * [Taipei to Moscow latency and RTT](/docs/network/latency/pairs/tpe-mow-rtt): 132.7 ms * [Singapore to Marseille latency and RTT](/docs/network/latency/pairs/sin-mrs-rtt): 140.6 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/tpe-par.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Taipei → Paris round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/tpe-par-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Taipei, Taiwan → Seattle, USA RTT 🇹🇼 **Taipei, Taiwan (TPE)** → 🇺🇸 **Seattle, USA (SEA)** Measurement round: `2026-08-16T04:07:28Z`. Taipei to Seattle averaged 115.8 ms across 50 replies. The observed range was 111.75 to 128.63 ms; these figures describe one probe round, not a long-term service target. This is number 4 of 19 outgoing routes by average RTT. That ranking compares different destinations and is not adjusted for distance. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **115.8 ms** | | Jitter | **2.57 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **95.46 ms** | | Fiber Efficiency | **82.4%** | | Latency Tier | Good | | Source Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | | Destination Region | [North America](/docs/network/latency/regions/north-america) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Taipei](https://www.geonames.org/1668341) (25.05306, 121.52639) and [Seattle](https://www.geonames.org/5809844) (47.60621, -122.33207). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **9,748.1 km** | | Vacuum RTT floor | **65.03 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **95.46 ms** | | Low-latency fiber material floor | **95.08 ms** | | Engineering floor (5% path allowance) | **100.25 ms** | | Research 1.33× mapped-fiber reference | **126.96 ms** | | Estimated unamplified path loss | **2047.1 dB** | | Transparent optical spans / inline amplifiers | **128 / 127** | | Published RTT inflation over fiber floor | **1.21×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **111.75 ms** | | Average RTT | **115.8 ms** | | Maximum RTT | **128.63 ms** | | Standard deviation | **4.07 ms** | | Stdev / average | **3.5%** | ## Route Context * Departure PoP: [Taipei (TPE)](/docs/network/latency/tpe-taipei) * Destination PoP: [Seattle (SEA)](/docs/network/latency/sea-seattle) * Region overview: [Asia Pacific](/docs/network/latency/regions/asia-pacific) → [North America](/docs/network/latency/regions/north-america) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Seattle to Taipei latency and RTT](/docs/network/latency/pairs/sea-tpe-rtt): 114.7 ms **Fastest routes departing Taipei (TPE)** * [Taipei to Hong Kong latency and RTT](/docs/network/latency/pairs/tpe-hkg-rtt): 15.5 ms * [Taipei to Tokyo latency and RTT](/docs/network/latency/pairs/tpe-tyo-rtt): 31.9 ms * [Taipei to Singapore latency and RTT](/docs/network/latency/pairs/tpe-sin-rtt): 44.3 ms * [Taipei to Los Angeles latency and RTT](/docs/network/latency/pairs/tpe-lax-rtt): 132 ms * [Taipei to Moscow latency and RTT](/docs/network/latency/pairs/tpe-mow-rtt): 132.7 ms **Fastest routes arriving at Seattle (SEA)** * [Los Angeles to Seattle latency and RTT](/docs/network/latency/pairs/lax-sea-rtt): 26.9 ms * [New York to Seattle latency and RTT](/docs/network/latency/pairs/nyc-sea-rtt): 58.7 ms * [Ashburn to Seattle latency and RTT](/docs/network/latency/pairs/iad-sea-rtt): 62.3 ms * [Miami to Seattle latency and RTT](/docs/network/latency/pairs/mia-sea-rtt): 81.9 ms * [Tokyo to Seattle latency and RTT](/docs/network/latency/pairs/tyo-sea-rtt): 84.8 ms **Same corridor (Asia Pacific → North America)** * [Tokyo to Los Angeles latency and RTT](/docs/network/latency/pairs/tyo-lax-rtt): 101.2 ms * [Hong Kong to Seattle latency and RTT](/docs/network/latency/pairs/hkg-sea-rtt): 131.6 ms * [Sydney to Los Angeles latency and RTT](/docs/network/latency/pairs/syd-lax-rtt): 136.5 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/tpe-sea.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Taipei → Seattle round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/tpe-sea-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Taipei, Taiwan → Singapore RTT 🇹🇼 **Taipei, Taiwan (TPE)** → 🇸🇬 **Singapore (SIN)** Measurement round: `2026-08-16T04:07:28Z`. Taipei to Singapore averaged 44.3 ms across 50 replies. The observed range was 42.61 to 50.6 ms; these figures describe one probe round, not a long-term service target. This is number 3 of 19 outgoing routes by average RTT. That ranking compares different destinations and is not adjusted for distance. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------------------- | | RTT | **44.3 ms** | | Jitter | **1.36 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **31.77 ms** | | Fiber Efficiency | **71.7%** | | Latency Tier | Excellent | | Source Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | | Destination Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Taipei](https://www.geonames.org/1668341) (25.05306, 121.52639) and [Singapore](https://www.geonames.org/1880252) (1.28967, 103.85007). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **3,243.9 km** | | Vacuum RTT floor | **21.64 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **31.77 ms** | | Low-latency fiber material floor | **31.64 ms** | | Engineering floor (5% path allowance) | **33.36 ms** | | Research 1.33× mapped-fiber reference | **42.25 ms** | | Estimated unamplified path loss | **681.2 dB** | | Transparent optical spans / inline amplifiers | **43 / 42** | | Published RTT inflation over fiber floor | **1.39×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------ | | Minimum RTT | **42.61 ms** | | Average RTT | **44.3 ms** | | Maximum RTT | **50.6 ms** | | Standard deviation | **1.64 ms** | | Stdev / average | **3.7%** | ## Route Context * Departure PoP: [Taipei (TPE)](/docs/network/latency/tpe-taipei) * Destination PoP: [Singapore (SIN)](/docs/network/latency/sin-singapore) * Region overview: [Asia Pacific](/docs/network/latency/regions/asia-pacific) → [Asia Pacific](/docs/network/latency/regions/asia-pacific) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Singapore to Taipei latency and RTT](/docs/network/latency/pairs/sin-tpe-rtt): 45.8 ms **Fastest routes departing Taipei (TPE)** * [Taipei to Hong Kong latency and RTT](/docs/network/latency/pairs/tpe-hkg-rtt): 15.5 ms * [Taipei to Tokyo latency and RTT](/docs/network/latency/pairs/tpe-tyo-rtt): 31.9 ms * [Taipei to Seattle latency and RTT](/docs/network/latency/pairs/tpe-sea-rtt): 115.8 ms * [Taipei to Los Angeles latency and RTT](/docs/network/latency/pairs/tpe-lax-rtt): 132 ms * [Taipei to Moscow latency and RTT](/docs/network/latency/pairs/tpe-mow-rtt): 132.7 ms **Fastest routes arriving at Singapore (SIN)** * [Hong Kong to Singapore latency and RTT](/docs/network/latency/pairs/hkg-sin-rtt): 31.5 ms * [Tokyo to Singapore latency and RTT](/docs/network/latency/pairs/tyo-sin-rtt): 69.4 ms * [Melbourne to Singapore latency and RTT](/docs/network/latency/pairs/mel-sin-rtt): 88.4 ms * [Sydney to Singapore latency and RTT](/docs/network/latency/pairs/syd-sin-rtt): 94.5 ms * [Marseille to Singapore latency and RTT](/docs/network/latency/pairs/mrs-sin-rtt): 139.6 ms **Same corridor (Asia Pacific → Asia Pacific)** * [Melbourne to Sydney latency and RTT](/docs/network/latency/pairs/mel-syd-rtt): 9.8 ms * [Sydney to Melbourne latency and RTT](/docs/network/latency/pairs/syd-mel-rtt): 9.8 ms * [Hong Kong to Taipei latency and RTT](/docs/network/latency/pairs/hkg-tpe-rtt): 14.7 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/tpe-sin.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Taipei → Singapore round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/tpe-sin-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Taipei, Taiwan → Sydney, Australia RTT 🇹🇼 **Taipei, Taiwan (TPE)** → 🇦🇺 **Sydney, Australia (SYD)** Measurement round: `2026-08-16T04:07:28Z`. Taipei to Sydney averaged 133.2 ms across 50 replies. The observed range was 129.47 to 143.28 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 3.02 ms, or 2.3% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------------------- | | RTT | **133.2 ms** | | Jitter | **2.55 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **70.88 ms** | | Fiber Efficiency | **53.2%** | | Latency Tier | Good | | Source Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | | Destination Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Taipei](https://www.geonames.org/1668341) (25.05306, 121.52639) and [Sydney](https://www.geonames.org/2147714) (-33.86785, 151.20732). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **7,237.6 km** | | Vacuum RTT floor | **48.28 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **70.88 ms** | | Low-latency fiber material floor | **70.59 ms** | | Engineering floor (5% path allowance) | **74.43 ms** | | Research 1.33× mapped-fiber reference | **94.26 ms** | | Estimated unamplified path loss | **1519.9 dB** | | Transparent optical spans / inline amplifiers | **95 / 94** | | Published RTT inflation over fiber floor | **1.88×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **129.47 ms** | | Average RTT | **133.2 ms** | | Maximum RTT | **143.28 ms** | | Standard deviation | **3.02 ms** | | Stdev / average | **2.3%** | ## Route Context * Departure PoP: [Taipei (TPE)](/docs/network/latency/tpe-taipei) * Destination PoP: [Sydney (SYD)](/docs/network/latency/syd-sydney) * Region overview: [Asia Pacific](/docs/network/latency/regions/asia-pacific) → [Asia Pacific](/docs/network/latency/regions/asia-pacific) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Sydney to Taipei latency and RTT](/docs/network/latency/pairs/syd-tpe-rtt): 131.7 ms **Fastest routes departing Taipei (TPE)** * [Taipei to Hong Kong latency and RTT](/docs/network/latency/pairs/tpe-hkg-rtt): 15.5 ms * [Taipei to Tokyo latency and RTT](/docs/network/latency/pairs/tpe-tyo-rtt): 31.9 ms * [Taipei to Singapore latency and RTT](/docs/network/latency/pairs/tpe-sin-rtt): 44.3 ms * [Taipei to Seattle latency and RTT](/docs/network/latency/pairs/tpe-sea-rtt): 115.8 ms * [Taipei to Los Angeles latency and RTT](/docs/network/latency/pairs/tpe-lax-rtt): 132 ms **Fastest routes arriving at Sydney (SYD)** * [Melbourne to Sydney latency and RTT](/docs/network/latency/pairs/mel-syd-rtt): 9.8 ms * [Singapore to Sydney latency and RTT](/docs/network/latency/pairs/sin-syd-rtt): 94.5 ms * [Tokyo to Sydney latency and RTT](/docs/network/latency/pairs/tyo-syd-rtt): 101.9 ms * [Hong Kong to Sydney latency and RTT](/docs/network/latency/pairs/hkg-syd-rtt): 137.2 ms * [Los Angeles to Sydney latency and RTT](/docs/network/latency/pairs/lax-syd-rtt): 137.6 ms **Same corridor (Asia Pacific → Asia Pacific)** * [Sydney to Melbourne latency and RTT](/docs/network/latency/pairs/syd-mel-rtt): 9.8 ms * [Hong Kong to Taipei latency and RTT](/docs/network/latency/pairs/hkg-tpe-rtt): 14.7 ms * [Singapore to Hong Kong latency and RTT](/docs/network/latency/pairs/sin-hkg-rtt): 30.8 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/tpe-syd.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Taipei → Sydney round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/tpe-syd-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Taipei, Taiwan → Tokyo, Japan RTT 🇹🇼 **Taipei, Taiwan (TPE)** → 🇯🇵 **Tokyo, Japan (TYO)** Measurement round: `2026-08-16T04:07:28Z`. Taipei to Tokyo averaged 31.9 ms across 50 replies. The observed range was 30.57 to 35.67 ms; these figures describe one probe round, not a long-term service target. This is number 2 of 19 outgoing routes by average RTT. That ranking compares different destinations and is not adjusted for distance. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------------------- | | RTT | **31.9 ms** | | Jitter | **1.06 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **20.58 ms** | | Fiber Efficiency | **64.5%** | | Latency Tier | Excellent | | Source Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | | Destination Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Taipei](https://www.geonames.org/1668341) (25.05306, 121.52639) and [Tokyo](https://www.geonames.org/1850147) (35.68950, 139.69171). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **2,101.2 km** | | Vacuum RTT floor | **14.02 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **20.58 ms** | | Low-latency fiber material floor | **20.49 ms** | | Engineering floor (5% path allowance) | **21.61 ms** | | Research 1.33× mapped-fiber reference | **27.37 ms** | | Estimated unamplified path loss | **441.3 dB** | | Transparent optical spans / inline amplifiers | **28 / 27** | | Published RTT inflation over fiber floor | **1.55×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------ | | Minimum RTT | **30.57 ms** | | Average RTT | **31.9 ms** | | Maximum RTT | **35.67 ms** | | Standard deviation | **1.12 ms** | | Stdev / average | **3.5%** | ## Route Context * Departure PoP: [Taipei (TPE)](/docs/network/latency/tpe-taipei) * Destination PoP: [Tokyo (TYO)](/docs/network/latency/tyo-tokyo) * Region overview: [Asia Pacific](/docs/network/latency/regions/asia-pacific) → [Asia Pacific](/docs/network/latency/regions/asia-pacific) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Tokyo to Taipei latency and RTT](/docs/network/latency/pairs/tyo-tpe-rtt): 32.3 ms **Fastest routes departing Taipei (TPE)** * [Taipei to Hong Kong latency and RTT](/docs/network/latency/pairs/tpe-hkg-rtt): 15.5 ms * [Taipei to Singapore latency and RTT](/docs/network/latency/pairs/tpe-sin-rtt): 44.3 ms * [Taipei to Seattle latency and RTT](/docs/network/latency/pairs/tpe-sea-rtt): 115.8 ms * [Taipei to Los Angeles latency and RTT](/docs/network/latency/pairs/tpe-lax-rtt): 132 ms * [Taipei to Moscow latency and RTT](/docs/network/latency/pairs/tpe-mow-rtt): 132.7 ms **Fastest routes arriving at Tokyo (TYO)** * [Hong Kong to Tokyo latency and RTT](/docs/network/latency/pairs/hkg-tyo-rtt): 44.9 ms * [Singapore to Tokyo latency and RTT](/docs/network/latency/pairs/sin-tyo-rtt): 68.9 ms * [Seattle to Tokyo latency and RTT](/docs/network/latency/pairs/sea-tyo-rtt): 85 ms * [Los Angeles to Tokyo latency and RTT](/docs/network/latency/pairs/lax-tyo-rtt): 101.2 ms * [Sydney to Tokyo latency and RTT](/docs/network/latency/pairs/syd-tyo-rtt): 101.3 ms **Same corridor (Asia Pacific → Asia Pacific)** * [Melbourne to Sydney latency and RTT](/docs/network/latency/pairs/mel-syd-rtt): 9.8 ms * [Sydney to Melbourne latency and RTT](/docs/network/latency/pairs/syd-mel-rtt): 9.8 ms * [Hong Kong to Taipei latency and RTT](/docs/network/latency/pairs/hkg-tpe-rtt): 14.7 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/tpe-tyo.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Taipei → Tokyo round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/tpe-tyo-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Tokyo, Japan → Amsterdam, Netherlands RTT 🇯🇵 **Tokyo, Japan (TYO)** → 🇳🇱 **Amsterdam, Netherlands (AMS)** Measurement round: `2026-08-16T04:07:28Z`. Tokyo to Amsterdam averaged 197.4 ms, compared with a 91.18 ms city-centre fiber reference. The difference cannot be assigned to a particular cable or carrier from RTT alone. Standard deviation was 6.74 ms, or 3.4% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------------------- | | RTT | **197.4 ms** | | Jitter | **5.76 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **91.18 ms** | | Fiber Efficiency | **46.2%** | | Latency Tier | Fair | | Source Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | | Destination Region | [Europe](/docs/network/latency/regions/europe) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Tokyo](https://www.geonames.org/1850147) (35.68950, 139.69171) and [Amsterdam](https://www.geonames.org/2759794) (52.37403, 4.88969). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **9,311.4 km** | | Vacuum RTT floor | **62.12 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **91.18 ms** | | Low-latency fiber material floor | **90.82 ms** | | Engineering floor (5% path allowance) | **95.76 ms** | | Research 1.33× mapped-fiber reference | **121.28 ms** | | Estimated unamplified path loss | **1955.4 dB** | | Transparent optical spans / inline amplifiers | **123 / 122** | | Published RTT inflation over fiber floor | **2.16×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **190.34 ms** | | Average RTT | **197.4 ms** | | Maximum RTT | **220.5 ms** | | Standard deviation | **6.74 ms** | | Stdev / average | **3.4%** | ## Route Context * Departure PoP: [Tokyo (TYO)](/docs/network/latency/tyo-tokyo) * Destination PoP: [Amsterdam (AMS)](/docs/network/latency/ams-amsterdam) * Region overview: [Asia Pacific](/docs/network/latency/regions/asia-pacific) → [Europe](/docs/network/latency/regions/europe) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Amsterdam to Tokyo latency and RTT](/docs/network/latency/pairs/ams-tyo-rtt): 198.2 ms **Fastest routes departing Tokyo (TYO)** * [Tokyo to Taipei latency and RTT](/docs/network/latency/pairs/tyo-tpe-rtt): 32.3 ms * [Tokyo to Hong Kong latency and RTT](/docs/network/latency/pairs/tyo-hkg-rtt): 44.8 ms * [Tokyo to Singapore latency and RTT](/docs/network/latency/pairs/tyo-sin-rtt): 69.4 ms * [Tokyo to Seattle latency and RTT](/docs/network/latency/pairs/tyo-sea-rtt): 84.8 ms * [Tokyo to Los Angeles latency and RTT](/docs/network/latency/pairs/tyo-lax-rtt): 101.2 ms **Fastest routes arriving at Amsterdam (AMS)** * [London to Amsterdam latency and RTT](/docs/network/latency/pairs/lon-ams-rtt): 5.2 ms * [Frankfurt to Amsterdam latency and RTT](/docs/network/latency/pairs/fra-ams-rtt): 5.9 ms * [Paris to Amsterdam latency and RTT](/docs/network/latency/pairs/par-ams-rtt): 7 ms * [Berlin to Amsterdam latency and RTT](/docs/network/latency/pairs/ber-ams-rtt): 7.2 ms * [Marseille to Amsterdam latency and RTT](/docs/network/latency/pairs/mrs-ams-rtt): 20.3 ms **Same corridor (Asia Pacific → Europe)** * [Hong Kong to Moscow latency and RTT](/docs/network/latency/pairs/hkg-mow-rtt): 118.2 ms * [Taipei to Moscow latency and RTT](/docs/network/latency/pairs/tpe-mow-rtt): 132.7 ms * [Singapore to Marseille latency and RTT](/docs/network/latency/pairs/sin-mrs-rtt): 140.6 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/tyo-ams.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Tokyo → Amsterdam round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/tyo-ams-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Tokyo, Japan → Berlin, Germany RTT 🇯🇵 **Tokyo, Japan (TYO)** → 🇩🇪 **Berlin, Germany (BER)** Measurement round: `2026-08-16T04:07:28Z`. Tokyo to Berlin averaged 186.1 ms, compared with a 87.53 ms city-centre fiber reference. The difference cannot be assigned to a particular cable or carrier from RTT alone. Standard deviation was 9.12 ms, or 4.9% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------------------- | | RTT | **186.1 ms** | | Jitter | **7.88 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **87.53 ms** | | Fiber Efficiency | **47%** | | Latency Tier | Fair | | Source Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | | Destination Region | [Europe](/docs/network/latency/regions/europe) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Tokyo](https://www.geonames.org/1850147) (35.68950, 139.69171) and [Berlin](https://www.geonames.org/2950159) (52.52437, 13.41053). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **8,937.7 km** | | Vacuum RTT floor | **59.63 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **87.53 ms** | | Low-latency fiber material floor | **87.17 ms** | | Engineering floor (5% path allowance) | **91.91 ms** | | Research 1.33× mapped-fiber reference | **116.41 ms** | | Estimated unamplified path loss | **1876.9 dB** | | Transparent optical spans / inline amplifiers | **118 / 117** | | Published RTT inflation over fiber floor | **2.13×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **174.48 ms** | | Average RTT | **186.1 ms** | | Maximum RTT | **227.17 ms** | | Standard deviation | **9.12 ms** | | Stdev / average | **4.9%** | ## Route Context * Departure PoP: [Tokyo (TYO)](/docs/network/latency/tyo-tokyo) * Destination PoP: [Berlin (BER)](/docs/network/latency/ber-berlin) * Region overview: [Asia Pacific](/docs/network/latency/regions/asia-pacific) → [Europe](/docs/network/latency/regions/europe) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Berlin to Tokyo latency and RTT](/docs/network/latency/pairs/ber-tyo-rtt): 187.5 ms **Fastest routes departing Tokyo (TYO)** * [Tokyo to Taipei latency and RTT](/docs/network/latency/pairs/tyo-tpe-rtt): 32.3 ms * [Tokyo to Hong Kong latency and RTT](/docs/network/latency/pairs/tyo-hkg-rtt): 44.8 ms * [Tokyo to Singapore latency and RTT](/docs/network/latency/pairs/tyo-sin-rtt): 69.4 ms * [Tokyo to Seattle latency and RTT](/docs/network/latency/pairs/tyo-sea-rtt): 84.8 ms * [Tokyo to Los Angeles latency and RTT](/docs/network/latency/pairs/tyo-lax-rtt): 101.2 ms **Fastest routes arriving at Berlin (BER)** * [Frankfurt to Berlin latency and RTT](/docs/network/latency/pairs/fra-ber-rtt): 6.1 ms * [Amsterdam to Berlin latency and RTT](/docs/network/latency/pairs/ams-ber-rtt): 7.2 ms * [Paris to Berlin latency and RTT](/docs/network/latency/pairs/par-ber-rtt): 15.5 ms * [London to Berlin latency and RTT](/docs/network/latency/pairs/lon-ber-rtt): 16.2 ms * [Marseille to Berlin latency and RTT](/docs/network/latency/pairs/mrs-ber-rtt): 20.6 ms **Same corridor (Asia Pacific → Europe)** * [Hong Kong to Moscow latency and RTT](/docs/network/latency/pairs/hkg-mow-rtt): 118.2 ms * [Taipei to Moscow latency and RTT](/docs/network/latency/pairs/tpe-mow-rtt): 132.7 ms * [Singapore to Marseille latency and RTT](/docs/network/latency/pairs/sin-mrs-rtt): 140.6 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/tyo-ber.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Tokyo → Berlin round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/tyo-ber-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Tokyo, Japan → Frankfurt, Germany RTT 🇯🇵 **Tokyo, Japan (TYO)** → 🇩🇪 **Frankfurt, Germany (FRA)** Measurement round: `2026-08-16T04:07:28Z`. Tokyo to Frankfurt averaged 193.2 ms, compared with a 91.61 ms city-centre fiber reference. The difference cannot be assigned to a particular cable or carrier from RTT alone. Standard deviation was 3.37 ms, or 1.7% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------------------- | | RTT | **193.2 ms** | | Jitter | **2.63 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **91.61 ms** | | Fiber Efficiency | **47.4%** | | Latency Tier | Fair | | Source Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | | Destination Region | [Europe](/docs/network/latency/regions/europe) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Tokyo](https://www.geonames.org/1850147) (35.68950, 139.69171) and [Frankfurt am Main](https://www.geonames.org/2925533) (50.11552, 8.68417). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **9,354.4 km** | | Vacuum RTT floor | **62.41 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **91.61 ms** | | Low-latency fiber material floor | **91.24 ms** | | Engineering floor (5% path allowance) | **96.2 ms** | | Research 1.33× mapped-fiber reference | **121.84 ms** | | Estimated unamplified path loss | **1964.4 dB** | | Transparent optical spans / inline amplifiers | **123 / 122** | | Published RTT inflation over fiber floor | **2.11×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **189.92 ms** | | Average RTT | **193.2 ms** | | Maximum RTT | **203.54 ms** | | Standard deviation | **3.37 ms** | | Stdev / average | **1.7%** | ## Route Context * Departure PoP: [Tokyo (TYO)](/docs/network/latency/tyo-tokyo) * Destination PoP: [Frankfurt (FRA)](/docs/network/latency/fra-frankfurt) * Region overview: [Asia Pacific](/docs/network/latency/regions/asia-pacific) → [Europe](/docs/network/latency/regions/europe) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Frankfurt to Tokyo latency and RTT](/docs/network/latency/pairs/fra-tyo-rtt): 194.5 ms **Fastest routes departing Tokyo (TYO)** * [Tokyo to Taipei latency and RTT](/docs/network/latency/pairs/tyo-tpe-rtt): 32.3 ms * [Tokyo to Hong Kong latency and RTT](/docs/network/latency/pairs/tyo-hkg-rtt): 44.8 ms * [Tokyo to Singapore latency and RTT](/docs/network/latency/pairs/tyo-sin-rtt): 69.4 ms * [Tokyo to Seattle latency and RTT](/docs/network/latency/pairs/tyo-sea-rtt): 84.8 ms * [Tokyo to Los Angeles latency and RTT](/docs/network/latency/pairs/tyo-lax-rtt): 101.2 ms **Fastest routes arriving at Frankfurt (FRA)** * [Amsterdam to Frankfurt latency and RTT](/docs/network/latency/pairs/ams-fra-rtt): 5.9 ms * [Berlin to Frankfurt latency and RTT](/docs/network/latency/pairs/ber-fra-rtt): 6.1 ms * [Paris to Frankfurt latency and RTT](/docs/network/latency/pairs/par-fra-rtt): 7.6 ms * [London to Frankfurt latency and RTT](/docs/network/latency/pairs/lon-fra-rtt): 13.6 ms * [Marseille to Frankfurt latency and RTT](/docs/network/latency/pairs/mrs-fra-rtt): 16 ms **Same corridor (Asia Pacific → Europe)** * [Hong Kong to Moscow latency and RTT](/docs/network/latency/pairs/hkg-mow-rtt): 118.2 ms * [Taipei to Moscow latency and RTT](/docs/network/latency/pairs/tpe-mow-rtt): 132.7 ms * [Singapore to Marseille latency and RTT](/docs/network/latency/pairs/sin-mrs-rtt): 140.6 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/tyo-fra.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Tokyo → Frankfurt round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/tyo-fra-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Tokyo, Japan → São Paulo, Brazil RTT 🇯🇵 **Tokyo, Japan (TYO)** → 🇧🇷 **São Paulo, Brazil (GRU)** Measurement round: `2026-08-16T04:07:28Z`. Tokyo to São Paulo averaged 230.5 ms across 50 replies. The observed range was 223.65 to 246.27 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 5.46 ms, or 2.4% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **230.5 ms** | | Jitter | **4.64 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **181.46 ms** | | Fiber Efficiency | **78.7%** | | Latency Tier | Fair | | Source Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | | Destination Region | [South America](/docs/network/latency/regions/south-america) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Tokyo](https://www.geonames.org/1850147) (35.68950, 139.69171) and [São Paulo](https://www.geonames.org/3448439) (-23.54750, -46.63611). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | --------------- | | WGS-84 geodesic distance | **18,530.2 km** | | Vacuum RTT floor | **123.62 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **181.46 ms** | | Low-latency fiber material floor | **180.73 ms** | | Engineering floor (5% path allowance) | **190.56 ms** | | Research 1.33× mapped-fiber reference | **241.34 ms** | | Estimated unamplified path loss | **3891.3 dB** | | Transparent optical spans / inline amplifiers | **244 / 243** | | Published RTT inflation over fiber floor | **1.27×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **223.65 ms** | | Average RTT | **230.5 ms** | | Maximum RTT | **246.27 ms** | | Standard deviation | **5.46 ms** | | Stdev / average | **2.4%** | ## Route Context * Departure PoP: [Tokyo (TYO)](/docs/network/latency/tyo-tokyo) * Destination PoP: [São Paulo (GRU)](/docs/network/latency/gru-sao-paulo) * Region overview: [Asia Pacific](/docs/network/latency/regions/asia-pacific) → [South America](/docs/network/latency/regions/south-america) * Corridor overview: [Asia Pacific to South America](/docs/network/latency/regions/asia-pacific-to-south-america) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [São Paulo to Tokyo latency and RTT](/docs/network/latency/pairs/gru-tyo-rtt): 230.3 ms **Fastest routes departing Tokyo (TYO)** * [Tokyo to Taipei latency and RTT](/docs/network/latency/pairs/tyo-tpe-rtt): 32.3 ms * [Tokyo to Hong Kong latency and RTT](/docs/network/latency/pairs/tyo-hkg-rtt): 44.8 ms * [Tokyo to Singapore latency and RTT](/docs/network/latency/pairs/tyo-sin-rtt): 69.4 ms * [Tokyo to Seattle latency and RTT](/docs/network/latency/pairs/tyo-sea-rtt): 84.8 ms * [Tokyo to Los Angeles latency and RTT](/docs/network/latency/pairs/tyo-lax-rtt): 101.2 ms **Fastest routes arriving at São Paulo (GRU)** * [Ashburn to São Paulo latency and RTT](/docs/network/latency/pairs/iad-gru-rtt): 101.4 ms * [New York to São Paulo latency and RTT](/docs/network/latency/pairs/nyc-gru-rtt): 106.9 ms * [Miami to São Paulo latency and RTT](/docs/network/latency/pairs/mia-gru-rtt): 128.7 ms * [Los Angeles to São Paulo latency and RTT](/docs/network/latency/pairs/lax-gru-rtt): 131.3 ms * [Seattle to São Paulo latency and RTT](/docs/network/latency/pairs/sea-gru-rtt): 155.7 ms **Same corridor (Asia Pacific → South America)** * [Taipei to São Paulo latency and RTT](/docs/network/latency/pairs/tpe-gru-rtt): 273 ms * [Hong Kong to São Paulo latency and RTT](/docs/network/latency/pairs/hkg-gru-rtt): 287.6 ms * [Sydney to São Paulo latency and RTT](/docs/network/latency/pairs/syd-gru-rtt): 296.7 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/tyo-gru.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Tokyo → São Paulo round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/tyo-gru-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Tokyo, Japan → Hong Kong RTT 🇯🇵 **Tokyo, Japan (TYO)** → 🇭🇰 **Hong Kong (HKG)** Measurement round: `2026-08-16T04:07:28Z`. Tokyo to Hong Kong averaged 44.8 ms across 50 replies. The observed range was 43.35 to 48.43 ms; these figures describe one probe round, not a long-term service target. This is number 2 of 19 outgoing routes by average RTT. That ranking compares different destinations and is not adjusted for distance. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------------------- | | RTT | **44.8 ms** | | Jitter | **0.97 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **28.24 ms** | | Fiber Efficiency | **63%** | | Latency Tier | Excellent | | Source Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | | Destination Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Tokyo](https://www.geonames.org/1850147) (35.68950, 139.69171) and [Hong Kong](https://www.geonames.org/1819729) (22.27832, 114.17469). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **2,883.5 km** | | Vacuum RTT floor | **19.24 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **28.24 ms** | | Low-latency fiber material floor | **28.12 ms** | | Engineering floor (5% path allowance) | **29.65 ms** | | Research 1.33× mapped-fiber reference | **37.56 ms** | | Estimated unamplified path loss | **605.5 dB** | | Transparent optical spans / inline amplifiers | **38 / 37** | | Published RTT inflation over fiber floor | **1.59×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------ | | Minimum RTT | **43.35 ms** | | Average RTT | **44.8 ms** | | Maximum RTT | **48.43 ms** | | Standard deviation | **1.13 ms** | | Stdev / average | **2.5%** | ## Route Context * Departure PoP: [Tokyo (TYO)](/docs/network/latency/tyo-tokyo) * Destination PoP: [Hong Kong (HKG)](/docs/network/latency/hkg-hong-kong) * Region overview: [Asia Pacific](/docs/network/latency/regions/asia-pacific) → [Asia Pacific](/docs/network/latency/regions/asia-pacific) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Hong Kong to Tokyo latency and RTT](/docs/network/latency/pairs/hkg-tyo-rtt): 44.9 ms **Fastest routes departing Tokyo (TYO)** * [Tokyo to Taipei latency and RTT](/docs/network/latency/pairs/tyo-tpe-rtt): 32.3 ms * [Tokyo to Singapore latency and RTT](/docs/network/latency/pairs/tyo-sin-rtt): 69.4 ms * [Tokyo to Seattle latency and RTT](/docs/network/latency/pairs/tyo-sea-rtt): 84.8 ms * [Tokyo to Los Angeles latency and RTT](/docs/network/latency/pairs/tyo-lax-rtt): 101.2 ms * [Tokyo to Sydney latency and RTT](/docs/network/latency/pairs/tyo-syd-rtt): 101.9 ms **Fastest routes arriving at Hong Kong (HKG)** * [Taipei to Hong Kong latency and RTT](/docs/network/latency/pairs/tpe-hkg-rtt): 15.5 ms * [Singapore to Hong Kong latency and RTT](/docs/network/latency/pairs/sin-hkg-rtt): 30.8 ms * [Moscow to Hong Kong latency and RTT](/docs/network/latency/pairs/mow-hkg-rtt): 118.6 ms * [Seattle to Hong Kong latency and RTT](/docs/network/latency/pairs/sea-hkg-rtt): 130.4 ms * [Sydney to Hong Kong latency and RTT](/docs/network/latency/pairs/syd-hkg-rtt): 135.7 ms **Same corridor (Asia Pacific → Asia Pacific)** * [Melbourne to Sydney latency and RTT](/docs/network/latency/pairs/mel-syd-rtt): 9.8 ms * [Sydney to Melbourne latency and RTT](/docs/network/latency/pairs/syd-mel-rtt): 9.8 ms * [Hong Kong to Taipei latency and RTT](/docs/network/latency/pairs/hkg-tpe-rtt): 14.7 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/tyo-hkg.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Tokyo → Hong Kong round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/tyo-hkg-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Tokyo, Japan → Ashburn, USA RTT 🇯🇵 **Tokyo, Japan (TYO)** → 🇺🇸 **Ashburn, USA (IAD)** Measurement round: `2026-08-16T04:07:28Z`. Tokyo to Ashburn averaged 142.9 ms across 50 replies. The observed range was 137.64 to 155.57 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 4.84 ms, or 3.4% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **142.9 ms** | | Jitter | **3.66 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **106.69 ms** | | Fiber Efficiency | **74.7%** | | Latency Tier | Good | | Source Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | | Destination Region | [North America](/docs/network/latency/regions/north-america) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Tokyo](https://www.geonames.org/1850147) (35.68950, 139.69171) and [Ashburn](https://www.geonames.org/4744870) (39.04372, -77.48749). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | --------------- | | WGS-84 geodesic distance | **10,894.6 km** | | Vacuum RTT floor | **72.68 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **106.69 ms** | | Low-latency fiber material floor | **106.26 ms** | | Engineering floor (5% path allowance) | **112.04 ms** | | Research 1.33× mapped-fiber reference | **141.9 ms** | | Estimated unamplified path loss | **2287.9 dB** | | Transparent optical spans / inline amplifiers | **143 / 142** | | Published RTT inflation over fiber floor | **1.34×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **137.64 ms** | | Average RTT | **142.9 ms** | | Maximum RTT | **155.57 ms** | | Standard deviation | **4.84 ms** | | Stdev / average | **3.4%** | ## Route Context * Departure PoP: [Tokyo (TYO)](/docs/network/latency/tyo-tokyo) * Destination PoP: [Ashburn (IAD)](/docs/network/latency/iad-ashburn) * Region overview: [Asia Pacific](/docs/network/latency/regions/asia-pacific) → [North America](/docs/network/latency/regions/north-america) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Ashburn to Tokyo latency and RTT](/docs/network/latency/pairs/iad-tyo-rtt): 141.9 ms **Fastest routes departing Tokyo (TYO)** * [Tokyo to Taipei latency and RTT](/docs/network/latency/pairs/tyo-tpe-rtt): 32.3 ms * [Tokyo to Hong Kong latency and RTT](/docs/network/latency/pairs/tyo-hkg-rtt): 44.8 ms * [Tokyo to Singapore latency and RTT](/docs/network/latency/pairs/tyo-sin-rtt): 69.4 ms * [Tokyo to Seattle latency and RTT](/docs/network/latency/pairs/tyo-sea-rtt): 84.8 ms * [Tokyo to Los Angeles latency and RTT](/docs/network/latency/pairs/tyo-lax-rtt): 101.2 ms **Fastest routes arriving at Ashburn (IAD)** * [New York to Ashburn latency and RTT](/docs/network/latency/pairs/nyc-iad-rtt): 6 ms * [Miami to Ashburn latency and RTT](/docs/network/latency/pairs/mia-iad-rtt): 27.3 ms * [Los Angeles to Ashburn latency and RTT](/docs/network/latency/pairs/lax-iad-rtt): 60.4 ms * [Seattle to Ashburn latency and RTT](/docs/network/latency/pairs/sea-iad-rtt): 61.5 ms * [London to Ashburn latency and RTT](/docs/network/latency/pairs/lon-iad-rtt): 71 ms **Same corridor (Asia Pacific → North America)** * [Taipei to Seattle latency and RTT](/docs/network/latency/pairs/tpe-sea-rtt): 115.8 ms * [Hong Kong to Seattle latency and RTT](/docs/network/latency/pairs/hkg-sea-rtt): 131.6 ms * [Taipei to Los Angeles latency and RTT](/docs/network/latency/pairs/tpe-lax-rtt): 132 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/tyo-iad.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Tokyo → Ashburn round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/tyo-iad-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Tokyo, Japan → Johannesburg, South Africa RTT 🇯🇵 **Tokyo, Japan (TYO)** → 🇿🇦 **Johannesburg, South Africa (JNB)** Measurement round: `2026-08-16T04:07:28Z`. Tokyo to Johannesburg averaged 359.2 ms, compared with a 132.57 ms city-centre fiber reference. The difference cannot be assigned to a particular cable or carrier from RTT alone. Standard deviation was 15.79 ms, or 4.4% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------------------- | | RTT | **359.2 ms** | | Jitter | **13 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **132.57 ms** | | Fiber Efficiency | **36.9%** | | Latency Tier | High | | Source Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | | Destination Region | [Africa](/docs/network/latency/regions/africa) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Tokyo](https://www.geonames.org/1850147) (35.68950, 139.69171) and [Johannesburg](https://www.geonames.org/993800) (-26.20227, 28.04363). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | --------------- | | WGS-84 geodesic distance | **13,537.7 km** | | Vacuum RTT floor | **90.31 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **132.57 ms** | | Low-latency fiber material floor | **132.04 ms** | | Engineering floor (5% path allowance) | **139.22 ms** | | Research 1.33× mapped-fiber reference | **176.32 ms** | | Estimated unamplified path loss | **2842.9 dB** | | Transparent optical spans / inline amplifiers | **178 / 177** | | Published RTT inflation over fiber floor | **2.71×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **339.59 ms** | | Average RTT | **359.2 ms** | | Maximum RTT | **411.36 ms** | | Standard deviation | **15.79 ms** | | Stdev / average | **4.4%** | ## Route Context * Departure PoP: [Tokyo (TYO)](/docs/network/latency/tyo-tokyo) * Destination PoP: [Johannesburg (JNB)](/docs/network/latency/jnb-johannesburg) * Region overview: [Asia Pacific](/docs/network/latency/regions/asia-pacific) → [Africa](/docs/network/latency/regions/africa) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Johannesburg to Tokyo latency and RTT](/docs/network/latency/pairs/jnb-tyo-rtt): 360.5 ms **Fastest routes departing Tokyo (TYO)** * [Tokyo to Taipei latency and RTT](/docs/network/latency/pairs/tyo-tpe-rtt): 32.3 ms * [Tokyo to Hong Kong latency and RTT](/docs/network/latency/pairs/tyo-hkg-rtt): 44.8 ms * [Tokyo to Singapore latency and RTT](/docs/network/latency/pairs/tyo-sin-rtt): 69.4 ms * [Tokyo to Seattle latency and RTT](/docs/network/latency/pairs/tyo-sea-rtt): 84.8 ms * [Tokyo to Los Angeles latency and RTT](/docs/network/latency/pairs/tyo-lax-rtt): 101.2 ms **Fastest routes arriving at Johannesburg (JNB)** * [Amsterdam to Johannesburg latency and RTT](/docs/network/latency/pairs/ams-jnb-rtt): 171.9 ms * [Berlin to Johannesburg latency and RTT](/docs/network/latency/pairs/ber-jnb-rtt): 172.1 ms * [London to Johannesburg latency and RTT](/docs/network/latency/pairs/lon-jnb-rtt): 179.6 ms * [Paris to Johannesburg latency and RTT](/docs/network/latency/pairs/par-jnb-rtt): 180.9 ms * [Frankfurt to Johannesburg latency and RTT](/docs/network/latency/pairs/fra-jnb-rtt): 188.4 ms **Same corridor (Asia Pacific → Africa)** * [Singapore to Johannesburg latency and RTT](/docs/network/latency/pairs/sin-jnb-rtt): 312.6 ms * [Hong Kong to Johannesburg latency and RTT](/docs/network/latency/pairs/hkg-jnb-rtt): 316.7 ms * [Taipei to Johannesburg latency and RTT](/docs/network/latency/pairs/tpe-jnb-rtt): 331 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/tyo-jnb.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Tokyo → Johannesburg round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/tyo-jnb-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Tokyo, Japan → Los Angeles, USA RTT 🇯🇵 **Tokyo, Japan (TYO)** → 🇺🇸 **Los Angeles, USA (LAX)** Measurement round: `2026-08-16T04:07:28Z`. Tokyo to Los Angeles averaged 101.2 ms across 50 replies. The observed range was 98.89 to 105.47 ms; these figures describe one probe round, not a long-term service target. This is number 5 of 19 outgoing routes by average RTT. That ranking compares different destinations and is not adjusted for distance. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **101.2 ms** | | Jitter | **1.87 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **86.51 ms** | | Fiber Efficiency | **85.5%** | | Latency Tier | Good | | Source Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | | Destination Region | [North America](/docs/network/latency/regions/north-america) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Tokyo](https://www.geonames.org/1850147) (35.68950, 139.69171) and [Los Angeles](https://www.geonames.org/5368361) (34.05223, -118.24368). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **8,834.5 km** | | Vacuum RTT floor | **58.94 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **86.51 ms** | | Low-latency fiber material floor | **86.17 ms** | | Engineering floor (5% path allowance) | **90.85 ms** | | Research 1.33× mapped-fiber reference | **115.06 ms** | | Estimated unamplified path loss | **1855.3 dB** | | Transparent optical spans / inline amplifiers | **116 / 115** | | Published RTT inflation over fiber floor | **1.17×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **98.89 ms** | | Average RTT | **101.2 ms** | | Maximum RTT | **105.47 ms** | | Standard deviation | **1.62 ms** | | Stdev / average | **1.6%** | ## Route Context * Departure PoP: [Tokyo (TYO)](/docs/network/latency/tyo-tokyo) * Destination PoP: [Los Angeles (LAX)](/docs/network/latency/lax-los-angeles) * Region overview: [Asia Pacific](/docs/network/latency/regions/asia-pacific) → [North America](/docs/network/latency/regions/north-america) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Los Angeles to Tokyo latency and RTT](/docs/network/latency/pairs/lax-tyo-rtt): 101.2 ms **Fastest routes departing Tokyo (TYO)** * [Tokyo to Taipei latency and RTT](/docs/network/latency/pairs/tyo-tpe-rtt): 32.3 ms * [Tokyo to Hong Kong latency and RTT](/docs/network/latency/pairs/tyo-hkg-rtt): 44.8 ms * [Tokyo to Singapore latency and RTT](/docs/network/latency/pairs/tyo-sin-rtt): 69.4 ms * [Tokyo to Seattle latency and RTT](/docs/network/latency/pairs/tyo-sea-rtt): 84.8 ms * [Tokyo to Sydney latency and RTT](/docs/network/latency/pairs/tyo-syd-rtt): 101.9 ms **Fastest routes arriving at Los Angeles (LAX)** * [Seattle to Los Angeles latency and RTT](/docs/network/latency/pairs/sea-lax-rtt): 25.9 ms * [Miami to Los Angeles latency and RTT](/docs/network/latency/pairs/mia-lax-rtt): 56.8 ms * [New York to Los Angeles latency and RTT](/docs/network/latency/pairs/nyc-lax-rtt): 58.9 ms * [Ashburn to Los Angeles latency and RTT](/docs/network/latency/pairs/iad-lax-rtt): 60.7 ms * [London to Los Angeles latency and RTT](/docs/network/latency/pairs/lon-lax-rtt): 127.5 ms **Same corridor (Asia Pacific → North America)** * [Taipei to Seattle latency and RTT](/docs/network/latency/pairs/tpe-sea-rtt): 115.8 ms * [Hong Kong to Seattle latency and RTT](/docs/network/latency/pairs/hkg-sea-rtt): 131.6 ms * [Taipei to Los Angeles latency and RTT](/docs/network/latency/pairs/tpe-lax-rtt): 132 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/tyo-lax.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Tokyo → Los Angeles round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/tyo-lax-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Tokyo, Japan → London, UK RTT 🇯🇵 **Tokyo, Japan (TYO)** → 🇬🇧 **London, UK (LON)** Measurement round: `2026-08-16T04:07:28Z`. Tokyo to London averaged 202.7 ms, compared with a 93.84 ms city-centre fiber reference. The difference cannot be assigned to a particular cable or carrier from RTT alone. Standard deviation was 5.6 ms, or 2.8% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------------------- | | RTT | **202.7 ms** | | Jitter | **4.9 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **93.84 ms** | | Fiber Efficiency | **46.3%** | | Latency Tier | Fair | | Source Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | | Destination Region | [Europe](/docs/network/latency/regions/europe) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Tokyo](https://www.geonames.org/1850147) (35.68950, 139.69171) and [London](https://www.geonames.org/2643743) (51.50853, -0.12574). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **9,582.1 km** | | Vacuum RTT floor | **63.93 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **93.84 ms** | | Low-latency fiber material floor | **93.46 ms** | | Engineering floor (5% path allowance) | **98.54 ms** | | Research 1.33× mapped-fiber reference | **124.8 ms** | | Estimated unamplified path loss | **2012.2 dB** | | Transparent optical spans / inline amplifiers | **126 / 125** | | Published RTT inflation over fiber floor | **2.16×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **196.6 ms** | | Average RTT | **202.7 ms** | | Maximum RTT | **224.34 ms** | | Standard deviation | **5.6 ms** | | Stdev / average | **2.8%** | ## Route Context * Departure PoP: [Tokyo (TYO)](/docs/network/latency/tyo-tokyo) * Destination PoP: [London (LON)](/docs/network/latency/lon-london) * Region overview: [Asia Pacific](/docs/network/latency/regions/asia-pacific) → [Europe](/docs/network/latency/regions/europe) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [London to Tokyo latency and RTT](/docs/network/latency/pairs/lon-tyo-rtt): 202.9 ms **Fastest routes departing Tokyo (TYO)** * [Tokyo to Taipei latency and RTT](/docs/network/latency/pairs/tyo-tpe-rtt): 32.3 ms * [Tokyo to Hong Kong latency and RTT](/docs/network/latency/pairs/tyo-hkg-rtt): 44.8 ms * [Tokyo to Singapore latency and RTT](/docs/network/latency/pairs/tyo-sin-rtt): 69.4 ms * [Tokyo to Seattle latency and RTT](/docs/network/latency/pairs/tyo-sea-rtt): 84.8 ms * [Tokyo to Los Angeles latency and RTT](/docs/network/latency/pairs/tyo-lax-rtt): 101.2 ms **Fastest routes arriving at London (LON)** * [Amsterdam to London latency and RTT](/docs/network/latency/pairs/ams-lon-rtt): 5.2 ms * [Paris to London latency and RTT](/docs/network/latency/pairs/par-lon-rtt): 6.4 ms * [Frankfurt to London latency and RTT](/docs/network/latency/pairs/fra-lon-rtt): 12.9 ms * [Berlin to London latency and RTT](/docs/network/latency/pairs/ber-lon-rtt): 15.2 ms * [Marseille to London latency and RTT](/docs/network/latency/pairs/mrs-lon-rtt): 18.9 ms **Same corridor (Asia Pacific → Europe)** * [Hong Kong to Moscow latency and RTT](/docs/network/latency/pairs/hkg-mow-rtt): 118.2 ms * [Taipei to Moscow latency and RTT](/docs/network/latency/pairs/tpe-mow-rtt): 132.7 ms * [Singapore to Marseille latency and RTT](/docs/network/latency/pairs/sin-mrs-rtt): 140.6 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/tyo-lon.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Tokyo → London round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/tyo-lon-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Tokyo, Japan → Melbourne, Australia RTT 🇯🇵 **Tokyo, Japan (TYO)** → 🇦🇺 **Melbourne, Australia (MEL)** Measurement round: `2026-08-16T04:07:28Z`. Tokyo to Melbourne averaged 113.9 ms across 50 replies. The observed range was 107.58 to 128.46 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 4.6 ms, or 4.0% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------------------- | | RTT | **113.9 ms** | | Jitter | **4.76 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **79.87 ms** | | Fiber Efficiency | **70.1%** | | Latency Tier | Good | | Source Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | | Destination Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Tokyo](https://www.geonames.org/1850147) (35.68950, 139.69171) and [Melbourne](https://www.geonames.org/2158177) (-37.81400, 144.96332). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **8,156.1 km** | | Vacuum RTT floor | **54.41 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **79.87 ms** | | Low-latency fiber material floor | **79.55 ms** | | Engineering floor (5% path allowance) | **83.88 ms** | | Research 1.33× mapped-fiber reference | **106.23 ms** | | Estimated unamplified path loss | **1712.8 dB** | | Transparent optical spans / inline amplifiers | **108 / 107** | | Published RTT inflation over fiber floor | **1.43×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **107.58 ms** | | Average RTT | **113.9 ms** | | Maximum RTT | **128.46 ms** | | Standard deviation | **4.6 ms** | | Stdev / average | **4.0%** | ## Route Context * Departure PoP: [Tokyo (TYO)](/docs/network/latency/tyo-tokyo) * Destination PoP: [Melbourne (MEL)](/docs/network/latency/mel-melbourne) * Region overview: [Asia Pacific](/docs/network/latency/regions/asia-pacific) → [Asia Pacific](/docs/network/latency/regions/asia-pacific) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Melbourne to Tokyo latency and RTT](/docs/network/latency/pairs/mel-tyo-rtt): 112 ms **Fastest routes departing Tokyo (TYO)** * [Tokyo to Taipei latency and RTT](/docs/network/latency/pairs/tyo-tpe-rtt): 32.3 ms * [Tokyo to Hong Kong latency and RTT](/docs/network/latency/pairs/tyo-hkg-rtt): 44.8 ms * [Tokyo to Singapore latency and RTT](/docs/network/latency/pairs/tyo-sin-rtt): 69.4 ms * [Tokyo to Seattle latency and RTT](/docs/network/latency/pairs/tyo-sea-rtt): 84.8 ms * [Tokyo to Los Angeles latency and RTT](/docs/network/latency/pairs/tyo-lax-rtt): 101.2 ms **Fastest routes arriving at Melbourne (MEL)** * [Sydney to Melbourne latency and RTT](/docs/network/latency/pairs/syd-mel-rtt): 9.8 ms * [Singapore to Melbourne latency and RTT](/docs/network/latency/pairs/sin-mel-rtt): 88.6 ms * [Hong Kong to Melbourne latency and RTT](/docs/network/latency/pairs/hkg-mel-rtt): 138.4 ms * [Taipei to Melbourne latency and RTT](/docs/network/latency/pairs/tpe-mel-rtt): 142 ms * [Los Angeles to Melbourne latency and RTT](/docs/network/latency/pairs/lax-mel-rtt): 148.2 ms **Same corridor (Asia Pacific → Asia Pacific)** * [Melbourne to Sydney latency and RTT](/docs/network/latency/pairs/mel-syd-rtt): 9.8 ms * [Hong Kong to Taipei latency and RTT](/docs/network/latency/pairs/hkg-tpe-rtt): 14.7 ms * [Taipei to Hong Kong latency and RTT](/docs/network/latency/pairs/tpe-hkg-rtt): 15.5 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/tyo-mel.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Tokyo → Melbourne round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/tyo-mel-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Tokyo, Japan → Miami, USA RTT 🇯🇵 **Tokyo, Japan (TYO)** → 🇺🇸 **Miami, USA (MIA)** Measurement round: `2026-08-16T04:07:28Z`. Tokyo to Miami averaged 156.5 ms across 50 replies. The observed range was 154.18 to 166.87 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 2.45 ms, or 1.6% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **156.5 ms** | | Jitter | **1.73 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **117.71 ms** | | Fiber Efficiency | **75.2%** | | Latency Tier | Fair | | Source Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | | Destination Region | [North America](/docs/network/latency/regions/north-america) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Tokyo](https://www.geonames.org/1850147) (35.68950, 139.69171) and [Miami](https://www.geonames.org/4164138) (25.77427, -80.19366). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | --------------- | | WGS-84 geodesic distance | **12,019.9 km** | | Vacuum RTT floor | **80.19 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **117.71 ms** | | Low-latency fiber material floor | **117.24 ms** | | Engineering floor (5% path allowance) | **123.61 ms** | | Research 1.33× mapped-fiber reference | **156.55 ms** | | Estimated unamplified path loss | **2524.2 dB** | | Transparent optical spans / inline amplifiers | **158 / 157** | | Published RTT inflation over fiber floor | **1.33×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **154.18 ms** | | Average RTT | **156.5 ms** | | Maximum RTT | **166.87 ms** | | Standard deviation | **2.45 ms** | | Stdev / average | **1.6%** | ## Route Context * Departure PoP: [Tokyo (TYO)](/docs/network/latency/tyo-tokyo) * Destination PoP: [Miami (MIA)](/docs/network/latency/mia-miami) * Region overview: [Asia Pacific](/docs/network/latency/regions/asia-pacific) → [North America](/docs/network/latency/regions/north-america) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Miami to Tokyo latency and RTT](/docs/network/latency/pairs/mia-tyo-rtt): 156.3 ms **Fastest routes departing Tokyo (TYO)** * [Tokyo to Taipei latency and RTT](/docs/network/latency/pairs/tyo-tpe-rtt): 32.3 ms * [Tokyo to Hong Kong latency and RTT](/docs/network/latency/pairs/tyo-hkg-rtt): 44.8 ms * [Tokyo to Singapore latency and RTT](/docs/network/latency/pairs/tyo-sin-rtt): 69.4 ms * [Tokyo to Seattle latency and RTT](/docs/network/latency/pairs/tyo-sea-rtt): 84.8 ms * [Tokyo to Los Angeles latency and RTT](/docs/network/latency/pairs/tyo-lax-rtt): 101.2 ms **Fastest routes arriving at Miami (MIA)** * [Ashburn to Miami latency and RTT](/docs/network/latency/pairs/iad-mia-rtt): 27.4 ms * [New York to Miami latency and RTT](/docs/network/latency/pairs/nyc-mia-rtt): 32.9 ms * [Los Angeles to Miami latency and RTT](/docs/network/latency/pairs/lax-mia-rtt): 57.3 ms * [São Paulo to Miami latency and RTT](/docs/network/latency/pairs/gru-mia-rtt): 74 ms * [Seattle to Miami latency and RTT](/docs/network/latency/pairs/sea-mia-rtt): 81.7 ms **Same corridor (Asia Pacific → North America)** * [Taipei to Seattle latency and RTT](/docs/network/latency/pairs/tpe-sea-rtt): 115.8 ms * [Hong Kong to Seattle latency and RTT](/docs/network/latency/pairs/hkg-sea-rtt): 131.6 ms * [Taipei to Los Angeles latency and RTT](/docs/network/latency/pairs/tpe-lax-rtt): 132 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/tyo-mia.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Tokyo → Miami round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/tyo-mia-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Tokyo, Japan → Moscow, Russia RTT 🇯🇵 **Tokyo, Japan (TYO)** → 🇷🇺 **Moscow, Russia (MOW)** Measurement round: `2026-08-16T04:07:28Z`. Tokyo to Moscow averaged 161.7 ms, compared with a 73.42 ms city-centre fiber reference. The difference cannot be assigned to a particular cable or carrier from RTT alone. Standard deviation was 4.83 ms, or 3.0% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------------------- | | RTT | **161.7 ms** | | Jitter | **4.87 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **73.42 ms** | | Fiber Efficiency | **45.4%** | | Latency Tier | Fair | | Source Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | | Destination Region | [Europe](/docs/network/latency/regions/europe) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Tokyo](https://www.geonames.org/1850147) (35.68950, 139.69171) and [Moscow](https://www.geonames.org/524901) (55.75204, 37.61781). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | ------------- | | WGS-84 geodesic distance | **7,497 km** | | Vacuum RTT floor | **50.01 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **73.42 ms** | | Low-latency fiber material floor | **73.12 ms** | | Engineering floor (5% path allowance) | **77.1 ms** | | Research 1.33× mapped-fiber reference | **97.64 ms** | | Estimated unamplified path loss | **1574.4 dB** | | Transparent optical spans / inline amplifiers | **99 / 98** | | Published RTT inflation over fiber floor | **2.2×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **155.32 ms** | | Average RTT | **161.7 ms** | | Maximum RTT | **174.84 ms** | | Standard deviation | **4.83 ms** | | Stdev / average | **3.0%** | ## Route Context * Departure PoP: [Tokyo (TYO)](/docs/network/latency/tyo-tokyo) * Destination PoP: [Moscow (MOW)](/docs/network/latency/mow-moscow) * Region overview: [Asia Pacific](/docs/network/latency/regions/asia-pacific) → [Europe](/docs/network/latency/regions/europe) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Moscow to Tokyo latency and RTT](/docs/network/latency/pairs/mow-tyo-rtt): 162.2 ms **Fastest routes departing Tokyo (TYO)** * [Tokyo to Taipei latency and RTT](/docs/network/latency/pairs/tyo-tpe-rtt): 32.3 ms * [Tokyo to Hong Kong latency and RTT](/docs/network/latency/pairs/tyo-hkg-rtt): 44.8 ms * [Tokyo to Singapore latency and RTT](/docs/network/latency/pairs/tyo-sin-rtt): 69.4 ms * [Tokyo to Seattle latency and RTT](/docs/network/latency/pairs/tyo-sea-rtt): 84.8 ms * [Tokyo to Los Angeles latency and RTT](/docs/network/latency/pairs/tyo-lax-rtt): 101.2 ms **Fastest routes arriving at Moscow (MOW)** * [Berlin to Moscow latency and RTT](/docs/network/latency/pairs/ber-mow-rtt): 28.6 ms * [Frankfurt to Moscow latency and RTT](/docs/network/latency/pairs/fra-mow-rtt): 34.7 ms * [Amsterdam to Moscow latency and RTT](/docs/network/latency/pairs/ams-mow-rtt): 38.1 ms * [London to Moscow latency and RTT](/docs/network/latency/pairs/lon-mow-rtt): 43.5 ms * [Paris to Moscow latency and RTT](/docs/network/latency/pairs/par-mow-rtt): 44.7 ms **Same corridor (Asia Pacific → Europe)** * [Hong Kong to Moscow latency and RTT](/docs/network/latency/pairs/hkg-mow-rtt): 118.2 ms * [Taipei to Moscow latency and RTT](/docs/network/latency/pairs/tpe-mow-rtt): 132.7 ms * [Singapore to Marseille latency and RTT](/docs/network/latency/pairs/sin-mrs-rtt): 140.6 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/tyo-mow.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Tokyo → Moscow round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/tyo-mow-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Tokyo, Japan → Marseille, France RTT 🇯🇵 **Tokyo, Japan (TYO)** → 🇫🇷 **Marseille, France (MRS)** Measurement round: `2026-08-16T04:07:28Z`. Tokyo to Marseille averaged 205.6 ms, compared with a 99.03 ms city-centre fiber reference. The difference cannot be assigned to a particular cable or carrier from RTT alone. Standard deviation was 6.33 ms, or 3.1% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------------------- | | RTT | **205.6 ms** | | Jitter | **4.57 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **99.03 ms** | | Fiber Efficiency | **48.2%** | | Latency Tier | Fair | | Source Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | | Destination Region | [Europe](/docs/network/latency/regions/europe) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Tokyo](https://www.geonames.org/1850147) (35.68950, 139.69171) and [Marseille](https://www.geonames.org/2995469) (43.29695, 5.38107). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | --------------- | | WGS-84 geodesic distance | **10,112.8 km** | | Vacuum RTT floor | **67.47 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **99.03 ms** | | Low-latency fiber material floor | **98.63 ms** | | Engineering floor (5% path allowance) | **104 ms** | | Research 1.33× mapped-fiber reference | **131.71 ms** | | Estimated unamplified path loss | **2123.7 dB** | | Transparent optical spans / inline amplifiers | **133 / 132** | | Published RTT inflation over fiber floor | **2.08×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **198.77 ms** | | Average RTT | **205.6 ms** | | Maximum RTT | **227.7 ms** | | Standard deviation | **6.33 ms** | | Stdev / average | **3.1%** | ## Route Context * Departure PoP: [Tokyo (TYO)](/docs/network/latency/tyo-tokyo) * Destination PoP: [Marseille (MRS)](/docs/network/latency/mrs-marseille) * Region overview: [Asia Pacific](/docs/network/latency/regions/asia-pacific) → [Europe](/docs/network/latency/regions/europe) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Marseille to Tokyo latency and RTT](/docs/network/latency/pairs/mrs-tyo-rtt): 203.8 ms **Fastest routes departing Tokyo (TYO)** * [Tokyo to Taipei latency and RTT](/docs/network/latency/pairs/tyo-tpe-rtt): 32.3 ms * [Tokyo to Hong Kong latency and RTT](/docs/network/latency/pairs/tyo-hkg-rtt): 44.8 ms * [Tokyo to Singapore latency and RTT](/docs/network/latency/pairs/tyo-sin-rtt): 69.4 ms * [Tokyo to Seattle latency and RTT](/docs/network/latency/pairs/tyo-sea-rtt): 84.8 ms * [Tokyo to Los Angeles latency and RTT](/docs/network/latency/pairs/tyo-lax-rtt): 101.2 ms **Fastest routes arriving at Marseille (MRS)** * [Paris to Marseille latency and RTT](/docs/network/latency/pairs/par-mrs-rtt): 8.9 ms * [Frankfurt to Marseille latency and RTT](/docs/network/latency/pairs/fra-mrs-rtt): 16.4 ms * [London to Marseille latency and RTT](/docs/network/latency/pairs/lon-mrs-rtt): 17.9 ms * [Amsterdam to Marseille latency and RTT](/docs/network/latency/pairs/ams-mrs-rtt): 19.6 ms * [Berlin to Marseille latency and RTT](/docs/network/latency/pairs/ber-mrs-rtt): 20.2 ms **Same corridor (Asia Pacific → Europe)** * [Hong Kong to Moscow latency and RTT](/docs/network/latency/pairs/hkg-mow-rtt): 118.2 ms * [Taipei to Moscow latency and RTT](/docs/network/latency/pairs/tpe-mow-rtt): 132.7 ms * [Singapore to Marseille latency and RTT](/docs/network/latency/pairs/sin-mrs-rtt): 140.6 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/tyo-mrs.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Tokyo → Marseille round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/tyo-mrs-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Tokyo, Japan → New York, USA RTT 🇯🇵 **Tokyo, Japan (TYO)** → 🇺🇸 **New York, USA (NYC)** Measurement round: `2026-08-16T04:07:28Z`. Tokyo to New York averaged 144.6 ms across 50 replies. The observed range was 140.04 to 156.39 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 4.02 ms, or 2.8% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **144.6 ms** | | Jitter | **3.26 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **106.47 ms** | | Fiber Efficiency | **73.6%** | | Latency Tier | Good | | Source Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | | Destination Region | [North America](/docs/network/latency/regions/north-america) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Tokyo](https://www.geonames.org/1850147) (35.68950, 139.69171) and [New York City](https://www.geonames.org/5128581) (40.71427, -74.00597). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | --------------- | | WGS-84 geodesic distance | **10,872.7 km** | | Vacuum RTT floor | **72.53 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **106.47 ms** | | Low-latency fiber material floor | **106.05 ms** | | Engineering floor (5% path allowance) | **111.81 ms** | | Research 1.33× mapped-fiber reference | **141.61 ms** | | Estimated unamplified path loss | **2283.3 dB** | | Transparent optical spans / inline amplifiers | **143 / 142** | | Published RTT inflation over fiber floor | **1.36×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **140.04 ms** | | Average RTT | **144.6 ms** | | Maximum RTT | **156.39 ms** | | Standard deviation | **4.02 ms** | | Stdev / average | **2.8%** | ## Route Context * Departure PoP: [Tokyo (TYO)](/docs/network/latency/tyo-tokyo) * Destination PoP: [New York (NYC)](/docs/network/latency/nyc-new-york) * Region overview: [Asia Pacific](/docs/network/latency/regions/asia-pacific) → [North America](/docs/network/latency/regions/north-america) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [New York to Tokyo latency and RTT](/docs/network/latency/pairs/nyc-tyo-rtt): 143.1 ms **Fastest routes departing Tokyo (TYO)** * [Tokyo to Taipei latency and RTT](/docs/network/latency/pairs/tyo-tpe-rtt): 32.3 ms * [Tokyo to Hong Kong latency and RTT](/docs/network/latency/pairs/tyo-hkg-rtt): 44.8 ms * [Tokyo to Singapore latency and RTT](/docs/network/latency/pairs/tyo-sin-rtt): 69.4 ms * [Tokyo to Seattle latency and RTT](/docs/network/latency/pairs/tyo-sea-rtt): 84.8 ms * [Tokyo to Los Angeles latency and RTT](/docs/network/latency/pairs/tyo-lax-rtt): 101.2 ms **Fastest routes arriving at New York (NYC)** * [Ashburn to New York latency and RTT](/docs/network/latency/pairs/iad-nyc-rtt): 6 ms * [Miami to New York latency and RTT](/docs/network/latency/pairs/mia-nyc-rtt): 33.3 ms * [Los Angeles to New York latency and RTT](/docs/network/latency/pairs/lax-nyc-rtt): 57.9 ms * [Seattle to New York latency and RTT](/docs/network/latency/pairs/sea-nyc-rtt): 59.5 ms * [London to New York latency and RTT](/docs/network/latency/pairs/lon-nyc-rtt): 63.8 ms **Same corridor (Asia Pacific → North America)** * [Taipei to Seattle latency and RTT](/docs/network/latency/pairs/tpe-sea-rtt): 115.8 ms * [Hong Kong to Seattle latency and RTT](/docs/network/latency/pairs/hkg-sea-rtt): 131.6 ms * [Taipei to Los Angeles latency and RTT](/docs/network/latency/pairs/tpe-lax-rtt): 132 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/tyo-nyc.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Tokyo → New York round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/tyo-nyc-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Tokyo, Japan → Paris, France RTT 🇯🇵 **Tokyo, Japan (TYO)** → 🇫🇷 **Paris, France (PAR)** Measurement round: `2026-08-16T04:07:28Z`. Tokyo to Paris averaged 204.8 ms, compared with a 95.34 ms city-centre fiber reference. The difference cannot be assigned to a particular cable or carrier from RTT alone. Standard deviation was 6.58 ms, or 3.2% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------------------- | | RTT | **204.8 ms** | | Jitter | **5.58 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **95.34 ms** | | Fiber Efficiency | **46.6%** | | Latency Tier | Fair | | Source Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | | Destination Region | [Europe](/docs/network/latency/regions/europe) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Tokyo](https://www.geonames.org/1850147) (35.68950, 139.69171) and [Paris](https://www.geonames.org/2988507) (48.85341, 2.34880). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **9,736.1 km** | | Vacuum RTT floor | **64.95 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **95.34 ms** | | Low-latency fiber material floor | **94.96 ms** | | Engineering floor (5% path allowance) | **100.12 ms** | | Research 1.33× mapped-fiber reference | **126.81 ms** | | Estimated unamplified path loss | **2044.6 dB** | | Transparent optical spans / inline amplifiers | **128 / 127** | | Published RTT inflation over fiber floor | **2.15×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **196.94 ms** | | Average RTT | **204.8 ms** | | Maximum RTT | **227.87 ms** | | Standard deviation | **6.58 ms** | | Stdev / average | **3.2%** | ## Route Context * Departure PoP: [Tokyo (TYO)](/docs/network/latency/tyo-tokyo) * Destination PoP: [Paris (PAR)](/docs/network/latency/par-paris) * Region overview: [Asia Pacific](/docs/network/latency/regions/asia-pacific) → [Europe](/docs/network/latency/regions/europe) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Paris to Tokyo latency and RTT](/docs/network/latency/pairs/par-tyo-rtt): 204.7 ms **Fastest routes departing Tokyo (TYO)** * [Tokyo to Taipei latency and RTT](/docs/network/latency/pairs/tyo-tpe-rtt): 32.3 ms * [Tokyo to Hong Kong latency and RTT](/docs/network/latency/pairs/tyo-hkg-rtt): 44.8 ms * [Tokyo to Singapore latency and RTT](/docs/network/latency/pairs/tyo-sin-rtt): 69.4 ms * [Tokyo to Seattle latency and RTT](/docs/network/latency/pairs/tyo-sea-rtt): 84.8 ms * [Tokyo to Los Angeles latency and RTT](/docs/network/latency/pairs/tyo-lax-rtt): 101.2 ms **Fastest routes arriving at Paris (PAR)** * [London to Paris latency and RTT](/docs/network/latency/pairs/lon-par-rtt): 6.4 ms * [Amsterdam to Paris latency and RTT](/docs/network/latency/pairs/ams-par-rtt): 7 ms * [Frankfurt to Paris latency and RTT](/docs/network/latency/pairs/fra-par-rtt): 7.6 ms * [Marseille to Paris latency and RTT](/docs/network/latency/pairs/mrs-par-rtt): 8.9 ms * [Berlin to Paris latency and RTT](/docs/network/latency/pairs/ber-par-rtt): 15.6 ms **Same corridor (Asia Pacific → Europe)** * [Hong Kong to Moscow latency and RTT](/docs/network/latency/pairs/hkg-mow-rtt): 118.2 ms * [Taipei to Moscow latency and RTT](/docs/network/latency/pairs/tpe-mow-rtt): 132.7 ms * [Singapore to Marseille latency and RTT](/docs/network/latency/pairs/sin-mrs-rtt): 140.6 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/tyo-par.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Tokyo → Paris round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/tyo-par-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Tokyo, Japan → Seattle, USA RTT 🇯🇵 **Tokyo, Japan (TYO)** → 🇺🇸 **Seattle, USA (SEA)** Measurement round: `2026-08-16T04:07:28Z`. Tokyo to Seattle averaged 84.8 ms across 50 replies. The observed range was 81.06 to 94.31 ms; these figures describe one probe round, not a long-term service target. This is number 4 of 19 outgoing routes by average RTT. That ranking compares different destinations and is not adjusted for distance. ## Latency Summary | Metric | Value | | -------------------- | ------------------------------------------------------------ | | RTT | **84.8 ms** | | Jitter | **2.73 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **75.55 ms** | | Fiber Efficiency | **89.1%** | | Latency Tier | Good | | Source Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | | Destination Region | [North America](/docs/network/latency/regions/north-america) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Tokyo](https://www.geonames.org/1850147) (35.68950, 139.69171) and [Seattle](https://www.geonames.org/5809844) (47.60621, -122.33207). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **7,715.2 km** | | Vacuum RTT floor | **51.47 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **75.55 ms** | | Low-latency fiber material floor | **75.25 ms** | | Engineering floor (5% path allowance) | **79.34 ms** | | Research 1.33× mapped-fiber reference | **100.49 ms** | | Estimated unamplified path loss | **1620.2 dB** | | Transparent optical spans / inline amplifiers | **102 / 101** | | Published RTT inflation over fiber floor | **1.12×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------ | | Minimum RTT | **81.06 ms** | | Average RTT | **84.8 ms** | | Maximum RTT | **94.31 ms** | | Standard deviation | **2.93 ms** | | Stdev / average | **3.5%** | ## Route Context * Departure PoP: [Tokyo (TYO)](/docs/network/latency/tyo-tokyo) * Destination PoP: [Seattle (SEA)](/docs/network/latency/sea-seattle) * Region overview: [Asia Pacific](/docs/network/latency/regions/asia-pacific) → [North America](/docs/network/latency/regions/north-america) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Seattle to Tokyo latency and RTT](/docs/network/latency/pairs/sea-tyo-rtt): 85 ms **Fastest routes departing Tokyo (TYO)** * [Tokyo to Taipei latency and RTT](/docs/network/latency/pairs/tyo-tpe-rtt): 32.3 ms * [Tokyo to Hong Kong latency and RTT](/docs/network/latency/pairs/tyo-hkg-rtt): 44.8 ms * [Tokyo to Singapore latency and RTT](/docs/network/latency/pairs/tyo-sin-rtt): 69.4 ms * [Tokyo to Los Angeles latency and RTT](/docs/network/latency/pairs/tyo-lax-rtt): 101.2 ms * [Tokyo to Sydney latency and RTT](/docs/network/latency/pairs/tyo-syd-rtt): 101.9 ms **Fastest routes arriving at Seattle (SEA)** * [Los Angeles to Seattle latency and RTT](/docs/network/latency/pairs/lax-sea-rtt): 26.9 ms * [New York to Seattle latency and RTT](/docs/network/latency/pairs/nyc-sea-rtt): 58.7 ms * [Ashburn to Seattle latency and RTT](/docs/network/latency/pairs/iad-sea-rtt): 62.3 ms * [Miami to Seattle latency and RTT](/docs/network/latency/pairs/mia-sea-rtt): 81.9 ms * [Taipei to Seattle latency and RTT](/docs/network/latency/pairs/tpe-sea-rtt): 115.8 ms **Same corridor (Asia Pacific → North America)** * [Hong Kong to Seattle latency and RTT](/docs/network/latency/pairs/hkg-sea-rtt): 131.6 ms * [Taipei to Los Angeles latency and RTT](/docs/network/latency/pairs/tpe-lax-rtt): 132 ms * [Sydney to Los Angeles latency and RTT](/docs/network/latency/pairs/syd-lax-rtt): 136.5 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/tyo-sea.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Tokyo → Seattle round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/tyo-sea-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Tokyo, Japan → Singapore RTT 🇯🇵 **Tokyo, Japan (TYO)** → 🇸🇬 **Singapore (SIN)** Measurement round: `2026-08-16T04:07:28Z`. Tokyo to Singapore averaged 69.4 ms across 50 replies. The observed range was 66.32 to 80.7 ms; these figures describe one probe round, not a long-term service target. This is number 3 of 19 outgoing routes by average RTT. That ranking compares different destinations and is not adjusted for distance. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------------------- | | RTT | **69.4 ms** | | Jitter | **2.27 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **52 ms** | | Fiber Efficiency | **74.9%** | | Latency Tier | Excellent | | Source Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | | Destination Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Tokyo](https://www.geonames.org/1850147) (35.68950, 139.69171) and [Singapore](https://www.geonames.org/1880252) (1.28967, 103.85007). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **5,310.2 km** | | Vacuum RTT floor | **35.43 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **52 ms** | | Low-latency fiber material floor | **51.79 ms** | | Engineering floor (5% path allowance) | **54.61 ms** | | Research 1.33× mapped-fiber reference | **69.16 ms** | | Estimated unamplified path loss | **1115.1 dB** | | Transparent optical spans / inline amplifiers | **70 / 69** | | Published RTT inflation over fiber floor | **1.33×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------ | | Minimum RTT | **66.32 ms** | | Average RTT | **69.4 ms** | | Maximum RTT | **80.7 ms** | | Standard deviation | **3.27 ms** | | Stdev / average | **4.7%** | ## Route Context * Departure PoP: [Tokyo (TYO)](/docs/network/latency/tyo-tokyo) * Destination PoP: [Singapore (SIN)](/docs/network/latency/sin-singapore) * Region overview: [Asia Pacific](/docs/network/latency/regions/asia-pacific) → [Asia Pacific](/docs/network/latency/regions/asia-pacific) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Singapore to Tokyo latency and RTT](/docs/network/latency/pairs/sin-tyo-rtt): 68.9 ms **Fastest routes departing Tokyo (TYO)** * [Tokyo to Taipei latency and RTT](/docs/network/latency/pairs/tyo-tpe-rtt): 32.3 ms * [Tokyo to Hong Kong latency and RTT](/docs/network/latency/pairs/tyo-hkg-rtt): 44.8 ms * [Tokyo to Seattle latency and RTT](/docs/network/latency/pairs/tyo-sea-rtt): 84.8 ms * [Tokyo to Los Angeles latency and RTT](/docs/network/latency/pairs/tyo-lax-rtt): 101.2 ms * [Tokyo to Sydney latency and RTT](/docs/network/latency/pairs/tyo-syd-rtt): 101.9 ms **Fastest routes arriving at Singapore (SIN)** * [Hong Kong to Singapore latency and RTT](/docs/network/latency/pairs/hkg-sin-rtt): 31.5 ms * [Taipei to Singapore latency and RTT](/docs/network/latency/pairs/tpe-sin-rtt): 44.3 ms * [Melbourne to Singapore latency and RTT](/docs/network/latency/pairs/mel-sin-rtt): 88.4 ms * [Sydney to Singapore latency and RTT](/docs/network/latency/pairs/syd-sin-rtt): 94.5 ms * [Marseille to Singapore latency and RTT](/docs/network/latency/pairs/mrs-sin-rtt): 139.6 ms **Same corridor (Asia Pacific → Asia Pacific)** * [Melbourne to Sydney latency and RTT](/docs/network/latency/pairs/mel-syd-rtt): 9.8 ms * [Sydney to Melbourne latency and RTT](/docs/network/latency/pairs/syd-mel-rtt): 9.8 ms * [Hong Kong to Taipei latency and RTT](/docs/network/latency/pairs/hkg-tpe-rtt): 14.7 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/tyo-sin.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Tokyo → Singapore round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/tyo-sin-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Tokyo, Japan → Sydney, Australia RTT 🇯🇵 **Tokyo, Japan (TYO)** → 🇦🇺 **Sydney, Australia (SYD)** Measurement round: `2026-08-16T04:07:28Z`. Tokyo to Sydney averaged 101.9 ms across 50 replies. The observed range was 97.82 to 112.89 ms; these figures describe one probe round, not a long-term service target. Standard deviation was 3.32 ms, or 3.3% of the average. This describes variation within the sample without identifying its cause. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------------------- | | RTT | **101.9 ms** | | Jitter | **3.13 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **76.31 ms** | | Fiber Efficiency | **74.9%** | | Latency Tier | Good | | Source Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | | Destination Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Tokyo](https://www.geonames.org/1850147) (35.68950, 139.69171) and [Sydney](https://www.geonames.org/2147714) (-33.86785, 151.20732). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **7,792.8 km** | | Vacuum RTT floor | **51.99 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **76.31 ms** | | Low-latency fiber material floor | **76.01 ms** | | Engineering floor (5% path allowance) | **80.14 ms** | | Research 1.33× mapped-fiber reference | **101.5 ms** | | Estimated unamplified path loss | **1636.5 dB** | | Transparent optical spans / inline amplifiers | **103 / 102** | | Published RTT inflation over fiber floor | **1.34×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------- | | Minimum RTT | **97.82 ms** | | Average RTT | **101.9 ms** | | Maximum RTT | **112.89 ms** | | Standard deviation | **3.32 ms** | | Stdev / average | **3.3%** | ## Route Context * Departure PoP: [Tokyo (TYO)](/docs/network/latency/tyo-tokyo) * Destination PoP: [Sydney (SYD)](/docs/network/latency/syd-sydney) * Region overview: [Asia Pacific](/docs/network/latency/regions/asia-pacific) → [Asia Pacific](/docs/network/latency/regions/asia-pacific) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Sydney to Tokyo latency and RTT](/docs/network/latency/pairs/syd-tyo-rtt): 101.3 ms **Fastest routes departing Tokyo (TYO)** * [Tokyo to Taipei latency and RTT](/docs/network/latency/pairs/tyo-tpe-rtt): 32.3 ms * [Tokyo to Hong Kong latency and RTT](/docs/network/latency/pairs/tyo-hkg-rtt): 44.8 ms * [Tokyo to Singapore latency and RTT](/docs/network/latency/pairs/tyo-sin-rtt): 69.4 ms * [Tokyo to Seattle latency and RTT](/docs/network/latency/pairs/tyo-sea-rtt): 84.8 ms * [Tokyo to Los Angeles latency and RTT](/docs/network/latency/pairs/tyo-lax-rtt): 101.2 ms **Fastest routes arriving at Sydney (SYD)** * [Melbourne to Sydney latency and RTT](/docs/network/latency/pairs/mel-syd-rtt): 9.8 ms * [Singapore to Sydney latency and RTT](/docs/network/latency/pairs/sin-syd-rtt): 94.5 ms * [Taipei to Sydney latency and RTT](/docs/network/latency/pairs/tpe-syd-rtt): 133.2 ms * [Hong Kong to Sydney latency and RTT](/docs/network/latency/pairs/hkg-syd-rtt): 137.2 ms * [Los Angeles to Sydney latency and RTT](/docs/network/latency/pairs/lax-syd-rtt): 137.6 ms **Same corridor (Asia Pacific → Asia Pacific)** * [Sydney to Melbourne latency and RTT](/docs/network/latency/pairs/syd-mel-rtt): 9.8 ms * [Hong Kong to Taipei latency and RTT](/docs/network/latency/pairs/hkg-tpe-rtt): 14.7 ms * [Taipei to Hong Kong latency and RTT](/docs/network/latency/pairs/tpe-hkg-rtt): 15.5 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/tyo-syd.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Tokyo → Sydney round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/tyo-syd-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions. # Tokyo, Japan → Taipei, Taiwan RTT 🇯🇵 **Tokyo, Japan (TYO)** → 🇹🇼 **Taipei, Taiwan (TPE)** Measurement round: `2026-08-16T04:07:28Z`. Tokyo to Taipei averaged 32.3 ms across 50 replies. The observed range was 31.1 to 36.28 ms; these figures describe one probe round, not a long-term service target. This is number 1 of 19 outgoing routes by average RTT. That ranking compares different destinations and is not adjusted for distance. ## Latency Summary | Metric | Value | | -------------------- | ---------------------------------------------------------- | | RTT | **32.3 ms** | | Jitter | **0.75 ms** | | Packet Loss | **0%** | | Standard Fiber Floor | **20.58 ms** | | Fiber Efficiency | **63.7%** | | Latency Tier | Excellent | | Source Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | | Destination Region | [Asia Pacific](/docs/network/latency/regions/asia-pacific) | ## Ping Measurement Series 50 ICMP echo probes were recorded in this round. Individual replies and their distribution are shown below. ## Download This Route's Data Download the probe round as CSV, JSON or YAML under CC BY 4.0. [Citation details](#about-this-measurement) include the route and round timestamp. ## Theoretical Fiber Latency The propagation model uses public city-centre coordinates from GeoNames: [Tokyo](https://www.geonames.org/1850147) (35.68950, 139.69171) and [Taipei](https://www.geonames.org/1668341) (25.05306, 121.52639). These are not facility coordinates or a surveyed cable path. | Physical Reference | Value | | --------------------------------------------- | -------------- | | WGS-84 geodesic distance | **2,101.2 km** | | Vacuum RTT floor | **14.02 ms** | | Standard fiber RTT floor ($n_g=1.4679$) | **20.58 ms** | | Low-latency fiber material floor | **20.49 ms** | | Engineering floor (5% path allowance) | **21.61 ms** | | Research 1.33× mapped-fiber reference | **27.37 ms** | | Estimated unamplified path loss | **441.3 dB** | | Transparent optical spans / inline amplifiers | **28 / 27** | | Published RTT inflation over fiber floor | **1.57×** | The span count assumes 80 km optical spans. Attenuation is a power-loss term, not a propagation delay. See the [model assumptions and equations](/docs/network/latency/theoretical-fiber-latency). ## Stability & Trend Statistics from the same 50 replies. They describe this probe round, not continuous historical monitoring. | Round Statistic | Value | | ------------------ | ------------ | | Minimum RTT | **31.1 ms** | | Average RTT | **32.3 ms** | | Maximum RTT | **36.28 ms** | | Standard deviation | **0.96 ms** | | Stdev / average | **3.0%** | ## Route Context * Departure PoP: [Tokyo (TYO)](/docs/network/latency/tyo-tokyo) * Destination PoP: [Taipei (TPE)](/docs/network/latency/tpe-taipei) * Region overview: [Asia Pacific](/docs/network/latency/regions/asia-pacific) → [Asia Pacific](/docs/network/latency/regions/asia-pacific) * Full global matrix: [Backbone Latency Matrix](/docs/network/latency) * Physical methodology: [Theoretical Fiber Latency](/docs/network/latency/theoretical-fiber-latency) ## Related City-Pair Routes **Reverse direction** * [Taipei to Tokyo latency and RTT](/docs/network/latency/pairs/tpe-tyo-rtt): 31.9 ms **Fastest routes departing Tokyo (TYO)** * [Tokyo to Hong Kong latency and RTT](/docs/network/latency/pairs/tyo-hkg-rtt): 44.8 ms * [Tokyo to Singapore latency and RTT](/docs/network/latency/pairs/tyo-sin-rtt): 69.4 ms * [Tokyo to Seattle latency and RTT](/docs/network/latency/pairs/tyo-sea-rtt): 84.8 ms * [Tokyo to Los Angeles latency and RTT](/docs/network/latency/pairs/tyo-lax-rtt): 101.2 ms * [Tokyo to Sydney latency and RTT](/docs/network/latency/pairs/tyo-syd-rtt): 101.9 ms **Fastest routes arriving at Taipei (TPE)** * [Hong Kong to Taipei latency and RTT](/docs/network/latency/pairs/hkg-tpe-rtt): 14.7 ms * [Singapore to Taipei latency and RTT](/docs/network/latency/pairs/sin-tpe-rtt): 45.8 ms * [Seattle to Taipei latency and RTT](/docs/network/latency/pairs/sea-tpe-rtt): 114.7 ms * [Sydney to Taipei latency and RTT](/docs/network/latency/pairs/syd-tpe-rtt): 131.7 ms * [Moscow to Taipei latency and RTT](/docs/network/latency/pairs/mow-tpe-rtt): 132.1 ms **Same corridor (Asia Pacific → Asia Pacific)** * [Melbourne to Sydney latency and RTT](/docs/network/latency/pairs/mel-syd-rtt): 9.8 ms * [Sydney to Melbourne latency and RTT](/docs/network/latency/pairs/syd-mel-rtt): 9.8 ms * [Taipei to Hong Kong latency and RTT](/docs/network/latency/pairs/tpe-hkg-rtt): 15.5 ms ## About This Measurement These ICMP measurements cover the two backbone endpoints. Access networks, routing policy and load can change an end user's result. A probe round is not an SLA. The underlying replies are available under **CC BY 4.0** in the [Hats Network latency dataset](/opendata/latency/). Per-pair files are available at `/opendata/latency/latest/pairs/tyo-tpe.pings.{csv,json,yaml}`. Cite the route, round timestamp and dataset version. Example: "Hats Network latency dataset, Tokyo → Taipei round 2026-08-16T04:07:28Z, CC BY 4.0". *** *Round published August 16, 2026. See the [full interactive latency matrix](/docs/network/latency) or browse more [city-pair routes](/docs/network/latency/pairs).* --- ## License and attribution - **Canonical source:** [View the human-readable HTML page](https://hatsnet.io/docs/network/latency/pairs/tyo-tpe-rtt). - **Original documentation:** © Hats Network Inc., licensed under [CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/). - **Public latency data:** © Hats Network Inc., licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). - **Full terms:** [Content and Data License](https://hatsnet.io/docs/legal/content-and-data-license) — includes attribution requirements, third-party material, trademarks, source code and other exclusions.