Oops - looks like I responded to that last one as "Anonymous". Sorry about that.<br><br>Here's an example of the situation I'm talking about.<br><br><br><br>Notice how hop 4 has 6% packet loss (for 8500 samples, too!), but hop 5 has 0%. Now hop 9 shows 10% packet loss - and the final destination shows 9%. The packet loss at hop 4 isn't contributing to the problem though - even though there's a bump in latency and packet loss at that hop - because it recovers at hop 5. Hop 9, however, adds packet loss that is *NOT* recovered from, so the packet loss at hop 9 is contributing to (and, it appears, causing all of) the packet loss at the final destination. Not completely on-topic, but another key indicator that hop 9 is a problem is that we change domains between hop 8 and hop 9 - from rr.com to alter.net. This points us to a problem in the link between those two routers.<br><br><br>