Re: squid-cache.org connectivity

Adrian Chadd wrote:

> Jesus. 30 hops? You're lucky the traceroute even finished. :-)

Seems there is some peering problem/gap for traffic from Telia to Coop
causing the traffic to take a long turn over the US..

1-13 Telia (quite much, but not entirely unrealistic for a poor creature
far down in the broadband access network)
14-24 verio.net (a bit long, quite likely a problem)
25-26 pnap.net
27-28 coop.net
29 the measurement factory

In the other direction things look considerably better.

1 the measurement factory
2-3 coop.net
4-8 sprintlink.net
9 agis.net
10-22 Telia
23 Home

> Dunno. Those routers may be overloaded. THe trouble with traceroutes
> is that they are generally CPU generated, and so the router could be
> quite fine but the CPU could be trashed, resulting in slow routing
> updated, slow ping/traceroute returns, but fine packet throughput.

Which is true when you look at a single hop, but here all hops after a
certain point is performing very badly showing quite well where the problem
is, close to the peeringpoint Verio -> Internap/pnap.

--
Henrik