Re: BGP advertisement logic?

Hi Keller,

I think this can be explained by the fact that BGP only advertises it's best route. For example, from your debug R3 advertises 205.90.31.0/24 over to R1. Next we see R1 advertise the same route back to R3 suggesting that R1 has an alternate path. Once R1 realises it's best path to 205.90.31.0/24 is through R3 it withdraws the route.

Cheers,
Ben.

From: Keller Giacomarro <keller.g_at_gmail.com<mailto:keller.g_at_gmail.com>>
Reply-To: Keller Giacomarro <keller.g_at_gmail.com<mailto:keller.g_at_gmail.com>>
Date: Sun, 20 Nov 2011 20:48:52 +1100
To: "ccielab_at_groupstudy.com<mailto:ccielab_at_groupstudy.com>" <ccielab_at_groupstudy.com<mailto:ccielab_at_groupstudy.com>>
Subject: BGP advertisement logic?

While doing some of the INE Vol1 BGP labs with 'debug ip bgp *neighbor' *turned
on, I noticed a strange behavior that I cannot explain.

I am doing a hard reset between two peers. R3 is in AS200, R1 is in AS100.
From the debug of reset on R3, what is happening is that R3 is advertising
all of its prefixes to R1, then withdrawing all the prefixes that contain
AS 100 in their AS_PATH.

Full debug logs @ http://pastebin.com/raw.php?i=cRxBGbwm

I understand the logic behind not wanting to advertise a route into an AS
that is originated from, but doesn't R1 handle that too (DENIED due to:
AS-PATH contains our own AS)? And R1 knows what AS R3 is in -- why is it
sending the updates in the first place? Is there some other mechanism that
is making R3 withdraw the routes it JUST sent?

As I study for the lab, I'm trying to ensure I don't just know commands,
but understand the guts of how the protocols work. Your insight is
appreciated!

Keller Giacomarro
keller.g_at_gmail.com<mailto:keller.g_at_gmail.com>

Blogs and organic groups at http://www.ccie.net
Received on Mon Nov 21 2011 - 09:57:00 ART