Global Ping Test
Real ICMP pings to your host from machines on five continents.
The host to ping
How this check works
A browser cannot send a ping — ICMP needs raw sockets, which neither a web page nor an edge worker has. Sites that claim otherwise time an HTTPS request and call it ping, which measures something real but is not the thing named. This page does it properly by delegating: the Globalping network is a set of community-run machines around the world that accept measurement requests over a public API, run the actual ping command, and return the output. Your browser asks for six probes spread across continents, waits the few seconds real packets take, and lays the round-trip times side by side. What the spread shows is mostly physics — light in fibre covers roughly 100 kilometres per millisecond of round trip, so a server in Frankfurt answering Sydney in under 250 ms is doing about as well as the planet allows.
What the results mean
- Average round trip
- Time for a packet to reach the target and the reply to come back, averaged over several pings from that probe. Distance sets the floor; congestion and routing add the rest.
- Packet loss
- The percentage of pings that got no reply. Zero is normal on healthy paths; persistent loss from one region points at a problem on that route rather than at the target.
- No reply
- The probe's pings went unanswered. Many hosts and firewalls drop ICMP as policy, so this means 'does not answer ping', which is not the same as 'down'.
- Median
- The middle value across all probes — a fairer single number than the average, since one far-away probe cannot drag it.
Common problems and fixes
- Every probe says no reply, but the site loads fine
- The host filters ping, which is common and deliberate. Test the thing you actually care about instead — an HTTP check tells you whether the site answers, which no amount of ICMP filtering hides.
- One continent is much slower than the rest
- First subtract geography: distance alone puts Sydney→Frankfurt near 300 ms. If the number is far beyond what distance explains, and consistent across runs, the routing to that region is the suspect — worth raising with the host if those users matter to you.
- The numbers change every run
- Different runs may get different probes, and the internet's paths shift by the minute. Tens of milliseconds of variation is weather, not signal. Run it three times and read the pattern.
- I want to ping from my own connection
- That is the one thing a remote probe network cannot do. Open a terminal: ping example.com works out of the box on Windows, macOS and Linux, and measures your path rather than a stranger's.
Frequently asked questions
Is this a real ping or a simulated one?
Real. Each probe runs the actual ping command — ICMP echo requests from its own network — and reports the raw statistics. That is exactly what your own terminal would send, just from a machine in another country. The measurement is made by the Globalping network, a free, community-run service, and this page credits it accordingly.
What is a good ping time?
Relative to distance, not in absolute terms. Same city: a few milliseconds. Same continent: tens. Opposite side of the planet: 200–300 ms is physics, not failure. The number worth acting on is one far above what distance predicts, appearing consistently from the same region.
Why do results differ from ping on my own machine?
Because the paths differ. Your terminal measures your ISP's route to the target; these probes measure their networks' routes. Both are true. A game feeling laggy while a probe reports 20 ms means the problem lives on your path — which is precisely what comparing the two tells you.
Can I choose which countries to ping from?
This page spreads probes across the world automatically, which answers the common question — is my host reachable and reasonably fast globally. For hand-picked locations, region-by-region comparisons and more probes, the Globalping site offers the full interface, free.