On Sat, 12 Jun 2004, Randy Bush wrote:

...

It appears that AboveNet is having major worldwide backbone issues at the moment. We were seeing high latency from the US to Europe, and now some European routes are no longer being advertised to the US.

it might be interesting to know how you determined this and what are "major worldwide backbone issues" in the sense of how they are defined and measured.

Maybe they told him. :) They don't say exactly what's broken, but Above.net did send out a notice Date: Sat, 12 Jun 2004 10:11:27 -0700 (PDT) Subject: Network Issues US & Europe ~12:03 EDT June 12, 2004 I think someone was thinking faster than they were typing though. At approximately 12:03 EDT widespread networking issues. This is causing networking issues through out our network. We are now diagnosing the problem. We do not know what caused the failure at this time. Apparently networking issues are causing networking issues on their network. I hate it when that happens. ---------------------------------------------------------------------- Jon Lewis | I route Senior Network Engineer | therefore you are Atlantic Net | _________ http://www.lewis.org/~jlewis/pgp for PGP public key_________

On Sat, 12 Jun 2004, Randy Bush wrote:

...

...

it might be interesting to know how you determined this and what are "major worldwide backbone issues" in the sense of how they are defined and measured. Maybe they told him. :)

damn. and i really meant my question. a lot of researchers are investing a lot of effort into recognizing and sizing major network problems from general/external evidence, e.g. route-views, traces, ippm measurements, ...

So, would RIPE's RIS project or some of the other route monitoring projects have noticed this as well? What is a 'major backbone outage' versus a peering link bounce from their perspective? Could they/should they monitor and report to some 'central' place when these larger events happen? What's the cutoff from 'minor' to 'major' event? -Chris

On Sat, 12 Jun 2004, Randy Bush wrote:

...

It appears that AboveNet is having major worldwide backbone issues at the moment. We were seeing high latency from the US to Europe, and now some European routes are no longer being advertised to the US.

it might be interesting to know how you determined this and what are "major worldwide backbone issues" in the sense of how they are defined and measured.

agreed, as a datapoint though, I noticed some things I monitor on above.net went unreachable several times while I was attempting to sleep in :) -Chris