[c-nsp] BGP Signalled VPLS

Caillin Bathern caillinb at commtelns.com
Tue Apr 23 04:54:06 EDT 2013


My comment related to BGP VPLS with BGP signalling.  When you have two
VPLS PE interfaces with matching VE ID and VE block offset (in the same
VPLS) then you fall to the standard BGP tie-breakers for route selection
such as local-preference.  Hence you can have a single switched CE site
connect to two PEs and not have a loop between the core and access as
you do with LDP signalled VPLS.


-----Original Message-----
From: cisco-nsp [mailto:cisco-nsp-bounces at puck.nether.net] On Behalf Of
Saku Ytti
Sent: Tuesday, 23 April 2013 6:30 PM
To: cisco-nsp at puck.nether.net
Subject: Re: [c-nsp] BGP Signalled VPLS

On (2013-04-23 09:35 +0200), Adam Vitkovsky wrote:

> Anyways my question was regarding the "old school" VPLS as we all do 
> it right now and LDP vs BGP signaling in particular.
> I'd like to find out which one do you folks prefer and why. 

BGP, for all customer stuff, BGP. LDP only for IGP labels, and hopefully
not even that in 3 years time.

For VPLS if you're already doing discovery via BGP, running tLDP is just
additional complexity you don't need. LDP clearly scales poorly, BGP is
O(1) session to cater n remote PE, LDP is O(n)

If you're not doing discovery via BGP, configuration is awkward and
complex compared to L3 MPLS VPN.

cisco-nsp mailing list  cisco-nsp at puck.nether.net
archive at http://puck.nether.net/pipermail/cisco-nsp/
Message  protected by MailGuard: e-mail anti-virus, anti-spam and
content filtering.http://www.mailguard.com.au/mg

More information about the cisco-nsp mailing list