The MPLS WG Archive

["Thomas D. Nadeau" <tnadeau@cisco.com>]

At 11:52 AM 11/26/2002 -0800, Yuan Gu wrote:

> > Suppose I have three incoming Ethernet TLS flows on R1. Following
> > Martini's draft, I should encapsulate Ethernet frames with a
> > stack of two labels: the bottom label being the VC label, while
> > the top label being the tunnel label. In other words, R1 has to
> > push two labels on every incoming frame where the one at the top
> > is always the same.
> > I suppose each entry in mplsXCTable should point to a not null
> > mplsLabelStackTable containing the corresponding VC label. If the
> > tunnel label is set up by means of RSVP-TE and the VC label by
> > means of LDP, how does R1 decides that the VC label has to be put
> > in the mplsLabelStackTable and not in a new separate entry in
> > mplsXCTable? What information of the RSVP messages are used to
> > accomplish this?
> >
> > Regards Roberto
> >
>
>I don't think the LSR MIB take L2VPN into account at all. You may want to
>implement your propertory MIB for supporting L2VPN.

         The LSR MIB shows the LFIB and MPLS-enabled
interfaces of an LSR.  If you can represent labels (or label stacks)
in an LFIB, you can see it there.  If you want to attach application-
specific things to entries in the TFIB, you need application-specific
MIBs such as the PWE-MPLS MIB that joins pwe3 vcs of
any service type to MPLS LSPs (or TE tunnels).

         --Tom