[c-nsp] Cisco IOS 15.4(2)T unified mpls send-label

[Vitkovský Adam]

Hi Tim,

Labels allocated by BGP for loopback IP of PE in a remote area are not dependent on labels allocated by LDP for a next-hop IP(ABR loopback IP) on a path to that remote PE.
So if this remote PE’s loopback IP is actually an IPv6 address and BGP would allocate a label for it the forwarding would have worked up from ingress PE through local ABR until the packet hits the ABR in remote area where the BGP label would have been POP-ed and the packet is expected to be switched based on the remaining LDP label.
And since LDP does not allocate labels for IPv6 prefixes the packet would not be label-switched to the egress PE.

adam

From: Tim Durack [mailto:tdurack at gmail.com]
Sent: Tuesday, September 30, 2014 2:25 PM
To: Vitkovský Adam
Cc: cisco-nsp at puck.nether.net
Subject: Re: [c-nsp] Cisco IOS 15.4(2)T unified mpls send-label

Thanks for the response. The missing piece for me was understanding that send-label depends on labels allocated by ldp(v4).

On Tue, Sep 30, 2014 at 4:31 AM, Vitkovský Adam <adam.vitkovsky at swan.sk<mailto:adam.vitkovsky at swan.sk>> wrote:
> I would like to enable send-label on ipv6 bgp neighbors, and move vpnv6
> over to ipv6.

Even though you'd be able to advertise PE loopbacks along with their labels via ipv6 bgp transport sessions.
These PE loopbacks still needs to be IPv4 addresses so that LDP can allocate labels for them within the particular area.
So you would not be able to move VPNv6 to IPv6 anyways.

adam
> -----Original Message-----
> From: cisco-nsp [mailto:cisco-nsp-bounces at puck.nether.net<mailto:cisco-nsp-bounces at puck.nether.net>] On Behalf Of
> Tim Durack
> Sent: Friday, September 26, 2014 8:31 PM
> To: cisco-nsp at puck.nether.net<mailto:cisco-nsp at puck.nether.net>
> Subject: [c-nsp] Cisco IOS 15.4(2)T unified mpls send-label
>
> I am lab testing "unified mpls" using bgp send-label. This is working well for
> ipv4/vpnv4/vpnv6.
>
> I would like to enable send-label on ipv6 bgp neighbors, and move vpnv6
> over to ipv6.
>
> Doesn't look like send-label is a supported feature in 15.4(2)T1 for ipv6 bgp
> neighbors unless I am missing something.
>
> Anyone else run into this?
>
> --
> Tim:>
> _______________________________________________
> cisco-nsp mailing list  cisco-nsp at puck.nether.net<mailto:cisco-nsp at puck.nether.net>
> https://puck.nether.net/mailman/listinfo/cisco-nsp
> archive at http://puck.nether.net/pipermail/cisco-nsp/



--
Tim:>