[c-nsp] BGP peer groups / performance question

[Bill Blackford]

Recently, I had to break up my iBGP group because with its presence, I couldn't from an export policy (prefix-list out). Therefore, I was forced to duplicate every line to every peer in my AS. Luckily for me, this is a small list of peers.

-b

-----Original Message-----
From: cisco-nsp-bounces at puck.nether.net [mailto:cisco-nsp-bounces at puck.nether.net] On Behalf Of Richard A Steenbergen
Sent: Monday, April 19, 2010 8:26 PM
To: Paul Stewart
Cc: cisco-nsp at puck.nether.net
Subject: Re: [c-nsp] BGP peer groups / performance question

On Mon, Apr 19, 2010 at 10:02:34PM -0400, Paul Stewart wrote:
> Hi folks..
> 
> Has anyone studied how much faster BGP peer groups are on very loaded
> routers?  I have a customer that I'm doing work with currently that
> has about 300 BGP peers on a Sup720/7613 box.  Downstream customers
> take about 10 minutes to receive a full routing table. 
> 
> I'm used to seeing this in under 2 minutes normally - a Cisco doc
> recommends moving to BGP peer groups which is quite possible to do but
> before starting to migrate these sessions into peer groups I wanted to
> find out if there's a major change I should be looking for or
> something quite minor...?

Peer groups are not necessary, modern IOS will do update grouping on its
own when the export policies and necessary bgp parameters match. If you
want to see what it does to your performance without update grouping,
configure "capability orf" (which breaks it) then stand back. I'll give
you a hint, it's not pretty. :)

-- 
Richard A Steenbergen <ras at e-gerbil.net>       http://www.e-gerbil.net/ras
GPG Key ID: 0xF8B12CBC (7535 7F59 8204 ED1F CC1C 53AF 4C41 5ECA F8B1 2CBC)
_______________________________________________
cisco-nsp mailing list  cisco-nsp at puck.nether.net
https://puck.nether.net/mailman/listinfo/cisco-nsp
archive at http://puck.nether.net/pipermail/cisco-nsp/