-1

I ran the traceroute command from Server A to Server B, two of the hops display a latency between 250ms to 480ms.

However when running ping from Server A to Server B, the response is always within 1ms.

My network admin team is of the opinion that there are no issues since ping always responds back with 1ms and the latency observed via traceroute should't be considered.

How do I rule out network latency issues?

Oliver
  • 13
  • 2
  • Where are the servers located in relation to each other? – ewwhite Oct 15 '16 at 08:00
  • Both servers are within my organisation, however the physical locations differ. I don't have the exact details, all I tell you is that traceroute takes 9 hops to reach server B from server A. – Oliver Oct 15 '16 at 08:03
  • 2
    I'm voting to close this question as off-topic because I'm not sure you've provided enough information or detail to allow us to help you. – ewwhite Oct 15 '16 at 08:15
  • @ewwhite Since `traceroute` shows a latency and `ping` doesn't, I was wondering which one of the two is a better indication of network issues. Can the latency by `traceroute` be ignored since `ping` seems to be returning within 1ms? – Oliver Oct 15 '16 at 08:20
  • `Should traceroute be used to diagnose network latency issues?` - No it shouldn't. – joeqwerty Oct 15 '16 at 14:00

2 Answers2

1

Traceroute is not a good measure of end to end latency because many routers in the forwarding path purposefully rate-limit ICMP traffic. If you are consitently getting a <1ms response between your two servers, it sounds like you have already ruled out a network latency issue.

Mark Riddell
  • 1,103
  • 1
  • 6
  • 10
0

Try the mtr command, as it can show you the latency between source and destination.

ewwhite
  • 194,921
  • 91
  • 434
  • 799