Since November 7th, we've been experiencing SSH connectivity issues with a ceratin server in Denver. After exhaustive troubleshooting, we did trace routes and found that we are losing a substantial amount of packets at teng-02-03-s2024.crsw02.den05.viawest.net . ViaWest Network Operations Center says this equipment belongs to Comcast/Time Warner. Here's a copy of our trace route (the % is % of packets LOST):
We beleive the 12% packet lost at den05.viawest.net is causing our connectivity issues to our final destination at login.donorware.net.
MTRs back from login.donorware.net are good. I beleive they take a different route outbound, not sure though as I do not have copies of their results. Can't test from teng-02-03-s2024.crsw02.den05.viawest.net of course. MTRs through other ISPs to login.donorware.net are good also.
Got more MTRs back from login.donorware.net - no lost packets or timeouts. I also just did an MTR through another Comcast account offsite using a different router, different account, etc. and am getting a slew of timeouts at the same hop ( teng-02-03-s2024.crsw02.den05.viawest.net). Contacted ViaWest.net and they say the problem is with Comcast, though provided no proof. MTRs from other locations in the US do not timeout.
Yeah, you will have a hard time getting somewhere without any kind of proof that it's a problem on Comcast's end.... TBH you would probably have a better chance getting someone from ViaWest.net to contact Comcast directly; they probably have higher-up contacts with people at Comcast's NOC who could deal with this sort of thing.
As it stands now, all signs point to an issue with ViaWest, just going by the traceroutes alone... just not enough evidence who's EXACTLY at fault yet.
I did contact ViaWest and they contacted Comcast NOC, as I believe Comcast owns the bad hop. Comcast NOC told ViaWest that I would have to contact Comcast. Around and 'round we go!