If you're only seeing this problem on this one target, then it's pretty important that you can trace to that target. You have packet loss and latency variations at earlier hops (as discussed), but 2% loss isn't going to cause chronic timeout problems. You might try using PingPlotter Standard (an evaluation version is available so you can test it without having to buy a license) and use the TCP packet type to try and get a picture of the response times of the final target (see here for some notes on doing that: http://www.nessoft.com/kb/51).

Hop 3 is probably inside the provider - although 7ms of latency for a DSL line is very uncommon (that's why I assumed cable). Most DSL connections have 35ms of latency or more to get to the provider. Hop 1 and hop 2 are probably pretty close to you - within the same building or complex, but I'm speculating again here.

In any case, any problem that only shows up to a single target strongly implicates that target or that target's immediate internet connection as a problem point. PingPlotter can't tell you much about that problem if it's not getting a response back in that point. Normally, problems are close to home (e.g., your DSL / Cable connection), but if everyplace else is working fine, then you have to rule that out as the source of the problem you're seeing, and look further downstream.

- Pete