Hong Kong → Taipei, Taiwan RTT
Hong Kong to Taipei, Taiwan ICMP RTT: 15.2 ms average latency, zero packet loss, and ultra-low jitter from the 2026-08-15 round.
🇭🇰 Hong Kong (HKG) → 🇹🇼 Taipei, Taiwan (TPE)
Measurement round: 2026-08-15T04:03:13Z.
In the 2026-08-15T04:03:13Z round, ICMP echo RTT from Hong Kong to Taipei averaged 15.2 ms, with a minimum of 14.92 ms and a maximum of 16.06 ms over 50 samples. There was no packet loss, and jitter was just 0.21 ms, positioning this route as the top-ranked outbound path among the 19 routes measured in this round.
The observed average is 1.92 times the 7.94 ms fiber floor for the 810 km distance, corresponding to 52.2% fiber efficiency. Despite the low absolute latency, the route still carries a modest overhead relative to the optical floor; that gap is consistent with the cable path being longer than the straight-line distance and with per-hop network forwarding.
With a standard deviation of 0.25 ms, the variability ratio is about 1.6%, well below the 3.08% median for the route set. This makes the Hong Kong-Taipei corridor an exceptionally consistent baseline, and the tight min-max range reinforces that impression.
Latency Summary
| Metric | Value |
|---|---|
| RTT | 15.2 ms |
| Jitter | 0.21 ms |
| Packet Loss | 0% |
| Standard Fiber Floor | 7.94 ms |
| Fiber Efficiency | 52.2% |
| Latency Tier | Ultra-Low |
| Source Region | Asia Pacific |
| Destination Region | Asia Pacific |
Ping Measurement Series
The chart below renders the 50-sample ICMP echo measurement round for this route; an accessible table of every sample is included with the chart.
Ping series: HKG → TPE
50 ICMP echo samples at 100 ms intervals · round 2026-08-15T04:03:13Z · avg 15.2 ± 0.04 ms (SE)
Sample distribution: box = interquartile range (15.0–15.3 ms), median 15.1 ms · whiskers = min–max · diamond = round average · each dot is one measured sample.
How to read this chart
The shaded band spans the measured minimum to maximum RTT of all samples collected up to that point, and the thin line traces the running average. The final position therefore matches the round statistics exactly — the band only widens when a new extreme sample arrives.
The dashed line is the round average, and the shaded band at the bottom is the theoretical fiber-optic floor (7.9 ms) for this route.
Full 50-sample dataset
| Seq | Offset (ms) | RTT (ms) |
|---|---|---|
| 1 | 0 | 14.92 |
| 2 | 100 | 15.29 |
| 3 | 200 | 15.04 |
| 4 | 300 | 15.18 |
| 5 | 400 | 15.31 |
| 6 | 500 | 15.06 |
| 7 | 600 | 15.27 |
| 8 | 700 | 15.12 |
| 9 | 800 | 15.00 |
| 10 | 900 | 15.01 |
| 11 | 1000 | 15.29 |
| 12 | 1100 | 15.06 |
| 13 | 1200 | 15.03 |
| 14 | 1300 | 14.96 |
| 15 | 1400 | 15.29 |
| 16 | 1500 | 15.11 |
| 17 | 1600 | 14.99 |
| 18 | 1700 | 14.94 |
| 19 | 1800 | 14.96 |
| 20 | 1900 | 15.26 |
| 21 | 2000 | 16.00 |
| 22 | 2100 | 15.34 |
| 23 | 2200 | 15.28 |
| 24 | 2300 | 15.18 |
| 25 | 2400 | 15.29 |
| 26 | 2500 | 15.10 |
| 27 | 2600 | 14.98 |
| 28 | 2700 | 14.99 |
| 29 | 2800 | 14.95 |
| 30 | 2900 | 14.97 |
| 31 | 3000 | 15.26 |
| 32 | 3100 | 15.24 |
| 33 | 3200 | 15.38 |
| 34 | 3300 | 15.94 |
| 35 | 3400 | 15.38 |
| 36 | 3500 | 15.08 |
| 37 | 3600 | 15.45 |
| 38 | 3700 | 15.10 |
| 39 | 3800 | 15.10 |
| 40 | 3900 | 15.03 |
| 41 | 4000 | 15.20 |
| 42 | 4100 | 15.26 |
| 43 | 4200 | 16.06 |
| 44 | 4300 | 15.55 |
| 45 | 4400 | 15.47 |
| 46 | 4500 | 15.21 |
| 47 | 4600 | 15.03 |
| 48 | 4700 | 15.15 |
| 49 | 4800 | 14.98 |
| 50 | 4900 | 14.96 |
Download This Route's Data
This route's measurement round is published as open data (CC BY 4.0) in CSV, JSON, and YAML. See About This Measurement for citation guidance.
Download this dataset
Per-route ICMP ping series, released under CC BY 4.0. Uncompressed plain text — no decompression step needed.
Theoretical Fiber Latency
The public physical reference uses the GeoNames city centres for Hong Kong (22.27832, 114.17469) and Taipei (25.05306, 121.52639). It does not disclose either facility address.
| Physical Reference | Value |
|---|---|
| WGS-84 geodesic distance | 810.3 km |
| Vacuum RTT floor | 5.41 ms |
| Standard fiber RTT floor () | 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 | 1.92× |
Attenuation is used to estimate the number of 80 km optical spans; it does not directly slow light. See the full theoretical fiber latency, GeoNames, attenuation, and amplifier methodology.
Stability & Trend
The current measurement round (2026-08-15T04:03:13Z) collected 50 ICMP echo samples on this route. Distribution statistics from that round:
| Round Statistic | Value |
|---|---|
| Minimum RTT | 14.92 ms |
| Average RTT | 15.2 ms |
| Maximum RTT | 16.06 ms |
| Standard deviation | 0.25 ms |
| Stdev / average | 1.6% |
RTT trend
Estimated route stability across the selected time window.
- Average
- 15.2 ms
- Minimum
- 15.1 ms
- Maximum
- 15.3 ms
- Average loss
- 0.01%
Route Context
- Departure PoP: Hong Kong (HKG)
- Destination PoP: Taipei (TPE)
- Region overview: Asia Pacific → Asia Pacific
- Full global matrix: Backbone Latency Matrix
- Physical methodology: Theoretical Fiber Latency
Related City-Pair Routes
Reverse direction
- Taipei to Hong Kong latency and RTT — 15.4 ms
Fastest routes departing Hong Kong (HKG)
- Hong Kong to Singapore latency and RTT — 31.2 ms
- Hong Kong to Tokyo latency and RTT — 45 ms
- Hong Kong to Moscow latency and RTT — 116.6 ms
- Hong Kong to Seattle latency and RTT — 129.9 ms
- Hong Kong to Sydney latency and RTT — 136.3 ms
Fastest routes arriving at Taipei (TPE)
- Tokyo to Taipei latency and RTT — 32.1 ms
- Singapore to Taipei latency and RTT — 44.3 ms
- Seattle to Taipei latency and RTT — 115.1 ms
- Moscow to Taipei latency and RTT — 130.9 ms
- Los Angeles to Taipei latency and RTT — 131.5 ms
Same corridor (Asia Pacific → Asia Pacific)
- Melbourne to Sydney latency and RTT — 9.8 ms
- Sydney to Melbourne latency and RTT — 9.8 ms
- Singapore to Hong Kong latency and RTT — 30.9 ms
About This Measurement
Published RTT reflects best-case backbone path behavior measured with ICMP echo probes between PoPs. End-user latency also depends on local access networks, congestion, routing policy, and traffic engineering. None of these values is an SLA.
The measurement round behind this page is published as open data under CC BY 4.0 in the Hats Network latency dataset. Per-pair files are available at /opendata/latency/latest/pairs/hkg-tpe.pings.{csv,json,yaml}. When citing, include the route, the measurement round timestamp, and the dataset version — for example: "Hats Network latency dataset, Hong Kong → Taipei round 2026-08-15T04:03:13Z, CC BY 4.0".
Data auto-generated on August 15, 2026. Explore the full interactive latency matrix or browse more city-pair routes.