From: Gavin Lawson (GavinL@titan.net.au)
Date: Wed May 02 2007 - 03:22:14 ART
HI Narvin
No, a full mesh is not required between Confederation peers.
If a mesh of confederation peers is build the BGP process will decide if
a route is valid based on the AS Path.
i.e if R1 advertises a route to R2, R2 avertises the route to R3 and R3
advertises the route back to R1. Then R1 will see it's own AS# in the
path and ignore the advertisement. Just like normal eBGP.
GL
-----Original Message-----
From: nobody@groupstudy.com [mailto:nobody@groupstudy.com] On Behalf Of
Navin MS
Sent: Wednesday, 2 May 2007 2:09 PM
To: ccielab@groupstudy.com
Subject: Is full-mesh of BGP confed peers reqd ?
Hello all,
I have 3 routers as follows, each part of different AS.
R1-----R2-----R3
R1 = AS 65001
R2 = AS 65002
R3 = AS 65003
Question is .. "Rest of the BGP network should see the above routers to
be members of AS 200".
Fair enough, I configured "bgp confederation identifier 200" under each
routers BGP process. Looks like it is working as desired. Others ASes,
see them as belonging to AS 200.
But something else is troubling me very much.
(1) Should I have to do a full mesh of the 3 confed peers ? That is,
should all 3 sub-ASs belonging to AS 200 have full-mesh for BGP to work
correctly ?
(2) What are the implications of making it a full-mesh Versus
partial-mesh (like a Hub-n-spoke) ?
Plz share your experiences.
Thanks,
Naveen.
This archive was generated by hypermail 2.1.4 : Fri Jun 01 2007 - 06:55:19 ART