[c-nsp] L3VPN VPNv4 NLRI - Route Reflector Scaling

[Oliver Boehmer (oboehmer)]

Mark Tinka <> wrote on Monday, March 24, 2008 2:48 AM:

> Hello all.
> 
> (posted to NANOG too; please excuse the length of the
> message)
> 
> Considering the scaling techniques currently available for
> VPNv4/L3VPN deployments as regards MP-BGP route reflectors,
> what do folk think is currently the most elegant way to
> deploy this that provides an even compromise on
> manageability, cost and scale (see RFC 1925, section 2,
> part 7, :-))?
> 
[...]
> 
> How are folk handling these issues today?

Well, most of the L3VPN deployments I'm aware of (which includes some
very large SPs) still use a single iBGP mesh of dedicated VPNv4 RRs,
some flat, some using hierarchical RR structure. RR partioning via
rr-group or using other means is rarely done as the scalability
requirements are still able to be handled by the simpler design. I guess
once you reach 500.000 vpnv4 prefixes or more, RR partitioning comes
into play, with the caveats you've mentioned. What are your
requirements?

	oli