Free network tool · 26 cities
Measure round-trip time from your browser to cities on six continents and find the region that answers fastest.
Test latency to 26 cities
Takes about 15 seconds. Each city is measured several times and the middle result is shown.
| City | Latency | Hosting |
|---|
Browsers can't send real ICMP ping packets, so this test times small HTTPS requests to public Amazon Web Services endpoints in each city. Figures run a little higher than a command-line ping, but the ranking between cities is a dependable guide.
More free network tools
IP address lookupYour public IP, ISP and locationSpeed testDownload, upload and jitterLatency is the time a small message takes to reach a server and come back. It's measured in milliseconds, and it's mostly a matter of distance. Light in optical fibre covers about 200 kilometres every millisecond, so each 1,000 km of cable adds roughly 10 ms to a round trip before any detours.
Detours are the other half of the story. Traffic follows cables and peering agreements, not straight lines. A city 800 km away can answer slower than one twice as far if your provider's route takes the long way round, and a test like this one is how you find out.
A rule of thumb for round-trip times. Your provider and time of day can shift these numbers either way.
| Round trip | Rating | What it usually means |
|---|---|---|
| Under 50 ms | Excellent | Usually the same country or a neighbour. Pages, SSH sessions and remote desktops feel instant. |
| 50–100 ms | Good | Typical within a continent. Websites feel quick; real-time games start to notice. |
| 100–200 ms | Fair | Usually across an ocean. Sites still work, but every extra request adds visible delay. |
| Over 200 ms | Slow | Far side of the world, or a congested route. Worth hosting closer to these users. |
Bandwidth decides how much data can flow. Latency decides how long every conversation takes to start.
Before the first byte arrives, a browser looks up the domain, opens a connection and agrees encryption with the server. That's at least three round trips, so 150 ms of latency costs close to half a second before anything appears.
Fonts, scripts, images and API calls each need their own trip. Some run in parallel, but a slow round trip still stretches the whole page.
SSH sessions, remote desktops and database admin tools send small messages back and forth constantly. Under 50 ms they feel local; over 150 ms every keystroke lags.
Small changes on your side can move the numbers by tens of milliseconds.
A busy connection queues the test's small requests behind everything else. Close video calls, game updates and cloud backups first.
Wi-Fi adds a few milliseconds and, more importantly, makes results jumpy. A wired connection gives steadier numbers.
A VPN sends everything through its own server first, so you'd be measuring the VPN's location, not yours.
The first run warms up DNS and connections. If two runs agree on the fastest region, you can trust the ranking.
A command-line ping uses the real ICMP protocol and is the standard way to check a single server. Replace example.com with any hostname or IP address.
To see where along the route the delay builds up, use tracert on Windows or traceroute on macOS and Linux. If you're comparing vpsNserver locations, our support team can give you a test IP for the data center you're considering.
# Windows: 10 pings
ping -n 10 example.com
# macOS / Linux: 10 pings
ping -c 10 example.com
# Trace the route
tracert example.com # Windows
traceroute example.com # macOS / LinuxThe command-line ping sends a tiny ICMP packet. A browser can only make HTTPS requests, which carry a little more overhead. Expect browser figures to sit a few milliseconds above a terminal ping to the same place.
The first request to each city also looks up the address and sets up an encrypted connection. The test sends a warm-up request first and reports the middle value of the samples after it.
For the people who use your site most, under 100 ms keeps pages feeling quick and under 50 ms feels instant. Measure from where your visitors are, not only from your office.
Usually, but not always. Routes follow cables and agreements between networks, so a slightly further city with a direct route can beat a nearer one. That's why it's worth testing rather than guessing.
No. The test measures public cloud endpoints in each city as a guide to network distance. vpsNserver has data centers in or near every listed country; ask support if you'd like a test IP for a specific location.
Host your site or app in the region that answered quickest and your users get the same short round trip.