How far away is every datacentre, from here?
58 datacentres across 7 providers, measured from this browser against each provider’s own published test file — 12 of them in Asia, which is the part that decides anything from Vietnam.
Your connection
Through the tunnel measures browser → Cloudflare edge → context.vn → datacentre, and reports exact bytes — it is what a site hosted behind this tunnel actually feels like. Direct measures your browser straight to the datacentre. Only Vultr allows a browser to read its test file, so direct mode reports latency only; everything else would be a guess. Compare both runs the two in the same pass and shows the difference per datacentre — a negative Δ means the tunnel wins. Throughput always comes from the tunnel, for the same CORS reason.
Latency by datacentre
| Datacentre | Edge → datacentre | ms | |
|---|---|---|---|
| SingaporeSGVultr | Vultr | ||
| TokyoJPVultr | Vultr | ||
| BangaloreINVultr | Vultr | ||
| FrankfurtDEVultr | Vultr | ||
| AmsterdamNLVultr | Vultr | ||
| LondonGBVultr | Vultr | ||
| ParisFRVultr | Vultr | ||
| MadridESVultr | Vultr | ||
| StockholmSEVultr | Vultr | ||
| WarsawPLVultr | Vultr | ||
| New JerseyUSVultr | Vultr | ||
| Silicon ValleyUSVultr | Vultr | ||
| TorontoCAVultr | Vultr | ||
| Mexico CityMXVultr | Vultr | ||
| São PauloBRVultr | Vultr | ||
| SydneyAUVultr | Vultr | ||
| MelbourneAUVultr | Vultr | ||
| JohannesburgZAVultr | Vultr | ||
| Tel AvivILVultr | Vultr | ||
| SingaporeSGAkamai (Linode) | Akamai (Linode) | ||
| TokyoJPAkamai (Linode) | Akamai (Linode) | ||
| OsakaJPAkamai (Linode) | Akamai (Linode) | ||
| JakartaIDAkamai (Linode) | Akamai (Linode) | ||
| ChennaiINAkamai (Linode) | Akamai (Linode) | ||
| FrankfurtDEAkamai (Linode) | Akamai (Linode) | ||
| LondonGBAkamai (Linode) | Akamai (Linode) | ||
| ParisFRAkamai (Linode) | Akamai (Linode) | ||
| MilanITAkamai (Linode) | Akamai (Linode) | ||
| StockholmSEAkamai (Linode) | Akamai (Linode) | ||
| NewarkUSAkamai (Linode) | Akamai (Linode) | ||
| DallasUSAkamai (Linode) | Akamai (Linode) | ||
| AtlantaUSAkamai (Linode) | Akamai (Linode) | ||
| MiamiUSAkamai (Linode) | Akamai (Linode) | ||
| SeattleUSAkamai (Linode) | Akamai (Linode) | ||
| FremontUSAkamai (Linode) | Akamai (Linode) | ||
| Washington DCUSAkamai (Linode) | Akamai (Linode) | ||
| São PauloBRAkamai (Linode) | Akamai (Linode) | ||
| SydneyAUAkamai (Linode) | Akamai (Linode) | ||
| SingaporeSGHetzner | Hetzner | ||
| FalkensteinDEHetzner | Hetzner | ||
| NurembergDEHetzner | Hetzner | ||
| HelsinkiFIHetzner | Hetzner | ||
| AshburnUSHetzner | Hetzner | ||
| HillsboroUSHetzner | Hetzner | ||
| SingaporeSGOVHcloud | OVHcloud | ||
| RoubaixFROVHcloud | OVHcloud | ||
| GravelinesFROVHcloud | OVHcloud | ||
| StrasbourgFROVHcloud | OVHcloud | ||
| Erith (London)GBOVHcloud | OVHcloud | ||
| SydneyAUOVHcloud | OVHcloud | ||
| NurembergDEnetcup | netcup | ||
| SingaporeSGContabo | Contabo | ||
| TokyoJPContabo | Contabo | ||
| SeattleUSContabo | Contabo | ||
| St. LouisUSContabo | Contabo | ||
| PortsmouthGBContabo | Contabo | ||
| SydneyAUContabo | Contabo | ||
| ParisFRScaleway (Online.net) | Scaleway (Online.net) |
What Δ includes: the tunnel figure is browser → edge → this origin → datacentre, so it measures this deployment rather than Cloudflare in the abstract — the browser → origin leg is the figure at the top of the page, and a Δ is only as meaningful as that leg is real. Direct is a full browser request including its TLS handshake. Contabo publishes a reachable speed-test host per location but no file behind it, so those rows measure latency only and are skipped by the download test rather than reported as 0 Mbps. Each result is the median of several samples; the hairline behind the bar is the fastest and slowest of them. A wide spread means an unstable route, which matters more than a good average. Files are each provider’s own published test file, requested by byte range so no more is transferred than the test needs — never more than 50 MB per request. Upload is measured to context.vn only.