[c-nsp] 6PE our only option?

Blake Dunlap ikiris at gmail.com
Thu Mar 29 21:28:30 EDT 2012


I too would like to see LDPv6 very much, for the same reasons, as I have
had the same issues and was forced to redo my true dual stack in to 6PE due
to conversion to MPLS.

-Blake

On Thu, Mar 29, 2012 at 19:25, Lobo <lobotiger at gmail.com> wrote:

> Apologies if this is a little long but looking for some friendly advice
> from you experts on rolling out IPv6 on our network.
>
> Our network follows a traditional model where it's
> edge----distribution----core--**---gwy for our internet customers.  The
> entire distribution, core and gwy routers have had MPLS enabled on them for
> a couple of years in order to offer EoMPLS like services.  The edge routers
> have not had MPLS enabled on them as there was no real need since they only
> provide internet access.  The IGP is OSPF and BGP is running on the
> distribution & gwy routers towards our route-reflectors.  The core is BGP
> less now so they are functioning as true P routers.
>
> We have managed to dual stack (v4 & v6) all of our distribution, core and
> gwy routers' interfaces as we believed that dual stack is always the
> preferred option.  We followed the same principles as our v4 implementation
> (loopbacks & PTPs only in the IGP and static/connected routes at the edge
> distributed via iBGP).  Upon installing a new edge router that would
> participate in IPv6 we discovered that the core could not route the packets
> to other IPv6 destinations because it only knows about LBs & PTPs.  Even if
> it made it to the gwy because of a default route (::/0), there were times
> when another gwy router had a better route and then we would have packets
> bouncing back and forth between a gwy and core router until the TTL expired.
>
> Now we're at the point of wondering if 6PE is our only option in order to
> forward the packets or if we go back to the old way of doing things by
> re-enabling BGP on the core (for ipv6 only) and having a partial set of
> ipv6 routes?  Personally I've been configuring 6PE in our lab for the
> entire week and it's really racking my head.  There are times when I have
> things working and then I make the smallest change in a route-map and
> suddenly things no longer work.  Do I configure the 6PE stuff on the edge
> router or can it start at the distribution router?  Traceroutes look really
> odd with the core not showing and I'm wondering if I'm introducing more
> problems for our NOC to troubleshoot (and learn) vs going with some other,
> simpler option.
>
> Right now I'm really wishing that LDPv6 was implemented.  :)
>
> Appreciate any comments, feedback or suggestions.  Please let me know I
> can provide any further information.
>
> Thanks!
>
> Jose
>
> ______________________________**_________________
> cisco-nsp mailing list  cisco-nsp at puck.nether.net
> https://puck.nether.net/**mailman/listinfo/cisco-nsp<https://puck.nether.net/mailman/listinfo/cisco-nsp>
> archive at http://puck.nether.net/**pipermail/cisco-nsp/<http://puck.nether.net/pipermail/cisco-nsp/>
>


More information about the cisco-nsp mailing list