Help Interpreting traceroute output

[Mark J. Dulcey]

dsr at tao.merseine.nu wrote:

> Remember that traceroute is not a performance measurement tool.
> It's a connectivity checker. 
> 
> Your measurements will differ from reality because:
> 
> - ICMP is not handled the same way as UDP which is not the same
>   as TCP 
> 
> - your packet size will be different
> 
> - your routing is not guaranteed
> 
> routers may handle traceroutes and/or pings preferentially, may
> artificially restrict them, may route them differently, store
> them in different buffers, all sorts of things.
> 
> Performance testing is best done between the points in question
> with the protocol in question using a payload as similar to
> "typical" as possible.

Nonetheless, traceroute can be a useful tool for tracking down the 
source of performance problems. If the ping time takes a sudden big jump 
(larger than can be accounted for by distance) between (say) hops 4 and 
5, it probably means that there is some sort of bottleneck at node 5 of 
your path, or in the link between 4 and 5.

Although routing is not guaranteed between source and destination, it is 
common for it to stay the same. This is especially true near the end 
points, where there is frequently only one way to get there from here.