1) I've run a same time plot from another machine on the network and it looks completely different--any ideas there? We haven't seen any problems inside the building.
What does "completely different" mean? Different route? Different results? This is almost certainly worth looking in to - as differences in results can often lead to clues.
2) I've done extensive internal plotting and pinging and have seen no problems at all
That's what I would expect - since there are no lost packets on your internal router.
3) I've been in the middle of a email stall on a particular machine and able to check the internet connection and speed at the same time. Speed from CNET tests at all times is pretty consistent.
The packet loss starting at hop 8 would almost certainly not happen against *all* sites - it's probably specific to a certain subset of IP addresses from a specific area of the country. A trace back from the other mail server back to your network would help us understand this - especially if we could do this during a problem period. If you have control of that server (and can install a perl script), contact us at
support@pingplotter.com and we might be able to help you trace from that machine remotely with PingPlotter.
- Pete
Previous Topic
Index
