Gate46
My IP All tools Learn
Address line
Names line
Routes line
Checks line
  1. All tools
  2. Routes line

Traceroute

See every router between our probe and a destination, and how long each one takes to answer. It shows where a delay begins.

Protocol
Enter a hostname or IP address to trace the route to it.

You will see each router on the way from our probe to the destination, the network that owns it, and three round-trip times per hop. Try one of these:

How traceroute works

Every packet carries a hop counter called TTL, short for time to live. Each router that forwards the packet subtracts one. The router that brings the counter to zero throws the packet away and sends back a short "time exceeded" message.

Traceroute uses that rule on purpose. It sends probes with a TTL of 1, which expire at the first router, then 2, then 3, and so on. Each expired probe makes one more router reveal its address, until a probe reaches the destination itself. Three probes go to every hop, which is why each row has three times. IPv6 works the same way and calls the counter the hop limit.

Reading the result

  • Hop is the position on the path. Hop 1 is the first router after our probe, and the last row is the destination.
  • Host is the router's reverse DNS name, when it has one, above its address. Names often carry an interface such as ae-3 and a city code such as fra for Frankfurt.
  • Network is the number of the autonomous system (AS) that announces the address. When it changes between two rows, traffic has passed from one network to the next.
  • Probe 1 to 3 are round-trip times from our probe to that router and back. The bar shows the best of the three.

In a healthy trace the times rise gradually, with one or two clear steps where the path covers a long distance, such as an ocean cable. A step that stays for every later hop is distance, not a fault.

A slow hop in the middle is usually not a problem

Routers forward traffic in dedicated hardware, but they answer traceroute probes with a general-purpose processor that has more important work to do. Many also limit how many of these replies they send each second. So one hop can show 80 ms while the hops after it show 20 ms. If the delay does not carry through to the following hops and the destination, real traffic is not affected.

Rows that show * * * are routers that did not answer at all. That is normal, and the trace carries on past them. What matters is latency or loss that starts at one hop and continues on every hop to the end. That pattern points at the hop where it starts, or the link leading into it.

The return path is invisible

Every time in the table includes the reply's trip back to our probe, and the reply may take a different route than the probe did. Each network chooses its own way out, so paths on the internet are often asymmetric. Traceroute only shows the forward path. A delay that appears at hop 7 may really sit on the way back from hop 7. When you can, a trace from the other end fills in the missing half.

What to send your provider

This trace starts at our probe, so it shows the route from us to the destination, not from your connection. For a fault on your own line, your provider needs a trace from your device. Run tracert example.com in Command Prompt on Windows, or traceroute example.com in Terminal on macOS and Linux. Then send:

  • The full output as text, not a screenshot of one line.
  • The date, the time with your time zone, and your public IP address at that moment.
  • A ping to the same destination that shows the loss or the high latency.
  • A trace taken while things work, for comparison, if you have one.
  • Whether the problem also happens on a wired connection, which rules out Wi-Fi.

Common questions

Why does the trace end in stars without reaching the destination?

The destination, or a firewall in front of it, is dropping the probes. Many servers do this on purpose. If the site or service works normally, nothing is wrong. Confirm with the port checker on the port the service uses.

Why is this route different from the one my computer shows?

The two traces start in different places. Ours begins at the probe server, yours begins at your router and crosses your provider first. Only the last few hops, inside the destination's network, are likely to match.

Do more hops mean a slower connection?

Not really. A modern router adds a fraction of a millisecond. Distance and congestion decide latency, so a path with 9 hops across an ocean is slower than one with 15 hops inside one country.

Why do some hops show private addresses such as 10.x.x.x?

Providers often number routers inside their own network from private ranges, or from the carrier-grade NAT range 100.64.0.0/10. Those routers still forward your traffic and still answer probes. The addresses just cannot be reached from outside that network.