Re: On BGP and VPLS

Giles Heron <[email protected]>
Newsgroups gmane.ietf.ppvpn
Message-ID <1052743355.2873.22.camel@gizpad>
Hi Kireeti,

sorry for the late reply on this...

On Fri, 2003-05-09 at 19:18, Kireeti Kompella wrote:
> Hi Alex,
> 
> On Fri, 2 May 2003, Alex Zinin wrote:
[snip]
> >  2. Distribution of information
> >
> >    When used as an IP routing protocol, BGP distributes routes among
> >    all participating routers.
> 
> For 2547, you have the same issue: BGP distributes routes among *all*
> participating routers, whereas only a few are really interested in them.
> So why is this an issue for VPLS but not for 2547?  The distribution
> mechanisms are identical.

I'm not sure the issues are quite the same.

In 2547, as far as I understand it, a given labelled route announcement
from a PE must be sent to every other PE with a VRF matching the route
target for that announcement.

In VPLS a unique label is required by each other PE which is a member of
a given VPLS, so each label advertised is used by exactly one other PE.

So it's a question of degree I think?

Or to put it another way, with respect to the set of all PEs:

Distribution of IPv4 routes can be characterised as "broadcast".
Distribution of VPN-IPv4 labelled routes can be characterised as
"multicast".
Distribution of VPLS labels can be characterised as "unicast".

Why select a broadcast mechanism (BGP with RRs) to meet a unicast
requirement?

[rest snipped]

Giles

-- 
=================================================================
Giles Heron    Principal Network Architect    PacketExchange Ltd.
ph: +44 7880 506185              "if you build it they will yawn"
=================================================================
lmpx.com only provides a reader for public news (NNTP) servers. It is not affiliated with the servers or forums shown here and is not responsible for the content of articles, which is written by their respective authors.