[55891] in North American Network Operators' Group
Re: "Selfish Routing"
daemon@ATHENA.MIT.EDU (Stephen Sprunk)
Sat Feb 15 01:40:38 2003
From: "Stephen Sprunk" <stephen@sprunk.org>
To: "Mikael Abrahamsson" <swmike@swm.pp.se>
Cc: <nanog@merit.edu>
Date: Sat, 15 Feb 2003 00:39:48 -0600
Errors-To: owner-nanog-outgoing@merit.edu
Thus spake "Mikael Abrahamsson" <swmike@swm.pp.se>
> When I read the article I didnt recognize anything in it that correlated
> with my understanding as to how BGP et all works. Am I the only
> one who after reading it thought the author needs to cut back on
> whatever drugs he is taking?
>
> I have never seen any router "send test packets and time them" nor take
> congestion into account when doing routing decisions?
There are indeed commercial and in-house products which use probes and
traffic measurement to influence the BGP decision process. Watch tarpit
logs for sites that ping every subnet on a regular schedule. I hope this
doesn't catch on or all the pings may drown out real traffic...
As a general description of Internet routing, I agree it's laughable -- but
the technology does exist.
S
