Measure first. Then decide.
32 speed-test nodes across 10 datacentres and 4 backhaul classes. Every test IP is copy-ready — measure real latency and packet loss before you commit to a line.
The four backhaul classes
The line class determines latency, packet loss and price — the same machine on different lines can differ by up to 4×. That is why the catalogue treats line class as a first-class filter.
Global reach at the lowest price. The default choice for overseas users and workloads that do not target mainland China.
Covers: Tokyo, Taipei, Singapore, Frankfurt, Los Angeles, Silicon Valley, Seattle, Seoul, Hong Kong, MalaysiaReturn path optimised for mainland China; better latency and loss than Global BGP at a mid-range price.
Covers: Tokyo, Taipei, Singapore, Frankfurt, Los Angeles, Silicon Valley, Seattle, Seoul, Hong Kong, MalaysiaChina Telecom CN2 enterprise routing — the best return-path quality across all three mainland carriers, and the highest price.
Covers: Tokyo, Singapore, Frankfurt, Los Angeles, Silicon Valley, Seattle, Seoul, Hong Kong, MalaysiaA higher grade of China-optimised return path, with better stability during peak hours.
Covers: Los Angeles, Silicon Valley, SeattleAll nodes
Click an IP to copy it. Full measured data (mainland return-path latency, loss, transit and peering) lives on the speed-test node page。
| # | Datacentres | Line | Test IP | Transit |
|---|---|---|---|---|
| 1 | Tokyo NRT | Global BGP | 6 transit networks | |
| 2 | Seoul ICN | Global BGP | 5 transit networks | |
| 3 | Hong Kong HKG | Global BGP | 4 transit networks | |
| 4 | Singapore SIN | Global BGP | 3 transit networks | |
| 5 | Malaysia KUL | Global BGP | 3 transit networks | |
| 6 | Taipei TPE | Global BGP | 4 transit networks | |
| 7 | Los Angeles LAX | Global BGP | 5 transit networks | |
| 8 | Frankfurt FRA | Global BGP | 2 transit networks | |
| 9 | Seattle SEA | Global BGP | 2 transit networks | |
| 10 | Silicon Valley SJC | Global BGP | 2 transit networks | |
| 11 | Seattle SEA | China Optimized VIP | 5 transit networks | |
| 12 | Singapore SIN | China Optimized | 5 transit networks | |
| 13 | Malaysia KUL | China Optimized | 7 transit networks | |
| 14 | Frankfurt FRA | China Optimized | 6 transit networks | |
| 15 | Los Angeles LAX | China Optimized VIP | 5 transit networks | |
| 16 | Silicon Valley SJC | China Optimized VIP | 4 transit networks | |
| 17 | Hong Kong HKG | China Optimized | 5 transit networks | |
| 18 | Malaysia KUL | Enterprise CN2 | 6 transit networks | |
| 19 | Taipei TPE | China Optimized | 5 transit networks | |
| 20 | Seattle SEA | Enterprise CN2 | 7 transit networks | |
| 21 | Tokyo NRT | China Optimized | 4 transit networks | |
| 22 | Seoul ICN | China Optimized | 4 transit networks | |
| 23 | Los Angeles LAX | Enterprise CN2 | 6 transit networks | |
| 24 | Silicon Valley SJC | Enterprise CN2 | 6 transit networks | |
| 25 | Los Angeles LAX | China Optimized | 4 transit networks | |
| 26 | Seattle SEA | China Optimized | 4 transit networks | |
| 27 | Silicon Valley SJC | China Optimized | 4 transit networks | |
| 28 | Frankfurt FRA | Enterprise CN2 | 3 transit networks | |
| 29 | Singapore SIN | Enterprise CN2 | 4 transit networks | |
| 30 | Hong Kong HKG | Enterprise CN2 | 1 transit networks | |
| 31 | Tokyo NRT | Enterprise CN2 | 4 transit networks | |
| 32 | Seoul ICN | Enterprise CN2 | 4 transit networks |
Diagnostics
MTR / traceroute
Generate MTR command templates for the three mainland carriers using the test IPs above, and see hop by hop where latency and loss appear. Use it to tell a local egress problem from a submarine-cable or transit-peering one.
Reading the measurements
How to interpret results across all 32 nodes: which metrics matter, what counts as normal, and how to compare peak against off-peak.
Hit a problem? Open a support requestwith MTR output attached shortens diagnosis time substantially.