[c-nsp] ISP / MPLS "POP" design

[Mark Tinka]

On Wednesday, October 30, 2013 01:54:03 AM Phil Bedard 
wrote:

> If I was designing it from scratch there isn't much need
> for the P routers to speak BGP if there are labeled
> paths to the remote PEs.   Why are they VPNv4 RRs?  You
> do not want a full mesh between PEs or what?   How many
> pops are you talking about?

Native IPv6?

No chance of LDPv6 or RSVPv6 seeing the light of day. The 
only way we are likely to get native label switching for 
IPv6 traffic is in the form of Segment Routing.

> Another kind of popular thing to do these days is carry
> the Internet table in a VRF as well.  That way your
> entire infrastructure is not open to the internet or
> your customers by default.  If you have good edge
> infrastructure ACLs it is kind of moot.

Not sure if it's "popular". I'd never do it, but this list 
is full of folk that have.

Mark.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 836 bytes
Desc: This is a digitally signed message part.
URL: <https://puck.nether.net/pipermail/cisco-nsp/attachments/20131030/900be2a6/attachment.sig>