How trustworthy are existing BGP AS Paths?

Dale,

The reason for routes with different AS origins is not due to configuration
errors but routing arrangement between ASs.  When two ASs speak IGP between 
themselves but each BGPs with other ASs, such as the Michnet and UMnet
situation, this would occur.  Anyother problematic ASpath scnerarios?

						--jessica  

------- Forwarded Message

Return-Path: dsj at merit.edu
Received: from ncc.ripe.net (ncc.ripe.net [193.0.0.129]) by merit.edu (8.6.8.1/merit-1.0) with SMTP id NAA19465; Thu, 29 Sep 1994 13:09:30 -0400
Received: from merit.edu by ncc.ripe.net with SMTP
	id AA27687 (5.65a/NCC-2.8); Thu, 29 Sep 1994 18:09:21 +0100
Received: (dsj at localhost) by merit.edu (8.6.8.1/merit-1.0) id NAA19443 for rr-impl at ripe.net; Thu, 29 Sep 1994 13:09:19 -0400
Date: Thu, 29 Sep 1994 13:09:19 -0400
From: "Dale S. Johnson" <dsj at merit.edu>
Message-Id: <199409291709.NAA19443 at merit.edu>
To: rr-impl at ripe.net
Subject: How trustworthy are existing BGP AS Paths?

RIPE Folks,

  In discussions about how to generate GateD configs for the Route
Servers, we're coming up with choices between hard-configured net lists
(like the PRDB), lists of ASs, AS-Path expressions, and "interpreted"
AS lists (creating net lists by listing the Route objects which
reference the given ASs).  A couple of questions keep coming up in
these discussions:  How Trustworthy  are the AS-Paths actually seen in
current routing packets?  How secure are they?  (Can they be easily
misconfigured?)

  Last winter we noticed that there were a fair number of BGP Paths
that claimed by be complete, but which disagreed with each other about
what the origin ASs for certain nets were.  I believe Tony or Marten
also said that they had investigated determining Origin ASs from BGP
packets, and had found that the information currently being routed was
frequently wrong.  Is my memory correct on this?  Does anyone know if
there are still lots of misleading AS-paths being routed?  Further
vague memories:  Did someone say that some ASs were intentionally
fudging the AS-Paths, e.g., to make multiple ASs look like a single AS,
for policy reasons?   [Do I sound like Gordon Cook?]

  Also, AS-Path expressions are presumably not useful for expressing
policy about EPG speakers.  (?)

  All in all, if these things are true, do folks here have feelings
about how advisable it is to implement policy in terms of AS-path
expressions in 4Q94?  Guesses about 3Q95?

- --Dale

- ----------

PS:  After we get Dale's memory sorted out, it might be interesting to
take this same question to BGPD.

------- End of Forwarded Message




-------- Logged at  Thu Sep 29 18:42:52 MET 1994  ---------