A traceroute shows the sequence of routers between a source and a destination, along with the round-trip time to each one. Knowing how to interpret the output — including what asterisks mean and how to spot where latency enters a path — turns a screenful of numbers into a useful diagnostic.
Here is a typical traceroute output to use as a reference throughout this guide:
Each numbered row is one router on the forward path between the source machine and the destination. The number on the left is the hop count — how many routers deep into the network that row is.
Traceroute works by sending packets with incrementally increasing TTL (Time to Live) values. Each router that handles a packet decrements its TTL by one. When the TTL reaches zero, the router discards the packet and sends back an ICMP Time Exceeded message, revealing its IP address and the round-trip time. By starting with TTL=1 and increasing by one each round, traceroute maps the path one hop at a time.
The three time values on each row — in milliseconds — are the round-trip times for three separate probe packets sent to that hop. Running three probes per hop gives a more reliable picture than a single measurement, since individual probes can be delayed by transient queue depth or CPU scheduling on the router.
Some variation between the three values is entirely normal. A large spread (e.g. 10 ms, 10 ms, 85 ms) can indicate queue jitter at that hop or inconsistent ICMP prioritization on that router. An average of the three values gives a better estimate of the typical latency at that hop.
An asterisk (*) in place of a time value means the probe packet timed out — no ICMP reply arrived before the tool's timeout window expired.
Asterisks do not mean the path is broken. The most common reason for asterisks is that the router at that hop deprioritizes or rate-limits ICMP Time Exceeded replies. The router is still forwarding traffic normally — it just does not generate a reply to traceroute probes. Look at the hops after the asterisk: if they respond with normal RTTs, the path is intact and working.
A row of asterisks does indicate a genuine problem only if all subsequent hops also show asterisks and the destination never responds — meaning packets stop moving forward at that point.
Look for a hop where the RTT jumps up noticeably — and stays elevated across all subsequent hops. In the example above, hop 6 shows an increase from ~12 ms to ~85 ms, and hops 7 and 8 remain at ~86 ms. That sustained jump at hop 6 is where the added latency enters: a long-haul link (such as a cross-continental or undersea cable segment) begins there.
The pattern to distinguish:
Traceroute measures round-trip time but only shows the forward path — the sequence of routers from source to destination. The ICMP reply from each hop travels back to the source via the return path, which may traverse completely different routers, different carriers, and different continents.
This means a high RTT at a particular hop does not necessarily mean the forward link from that router is slow. The extra latency could be on the return path from that router back to you. Networks route traffic based on cost and policy, not symmetry, so asymmetric paths are common, especially at peering points and international boundaries.
When diagnosing latency, traceroute from both ends if possible, or use a tool that runs probes from a vantage point close to the destination.
A single traceroute from one location gives you one data point. Running the same traceroute from multiple geographic vantage points shows whether high latency is a global issue or specific to one region or one network path.
The Path tool runs traceroutes from multiple probe locations simultaneously, so you can compare paths from different continents in a single view. If only probes from one region show high latency while others are normal, the problem is localized to the network segment or peering arrangement serving that region — not the destination itself.
This is especially useful for diagnosing anycast deployments, where different regions resolve the same IP to different physical servers. High latency from one region may mean the anycast routing is sending that region's traffic to a distant PoP rather than the nearest one.