The MPLS WG Archive

[riza.cetin@alcatel.be]

"Thomas D. Nadeau" wrote:

> At 09:38 AM 4/16/2003 +0200, riza.cetin@alcatel.be wrote:
>
> >"Thomas D. Nadeau" wrote:
> >
> > > >5. How can we represent label-stacking in LDP via this MIB. The
> > > >label-stacking could be either
> > > >- LDP LSPs over RSVP-TE tunnels.
> > > >- PWE-LDP (Martini) LSPs over core LDP/RSVP - LSPs.
> > > >
> > > >Suggestion - Each core LSP (LDP/RSVP) can be represented by an if-index
> > > >in the IF-MIB. If we include the if-index in the above tables then we
> > > >can find out whether the LSP goes via a physical-interface or a
> > > >'tunnel-interface'.
> > >
> > >          I don't understand how assigning an IfIndex to an LSP is going to
> > > work.  This sounds like something that is implementation-specific.
> >
> >MPLS-TE MIB (draft-ietf-mpls-te-mib-09.txt) allows for defining an MPLS tunnel
> >as an interface.
> >mplsTunnelIsIf object is set to true(1) for tunnels to be used as interface
> >and an entry is created in the ifTable with the ifIndex of the tunnel.
> >
> >Below is the text from the MPLS-TE-MIB.
> >
> >mplsTunnelEntry OBJECT-TYPE
> >    SYNTAX        MplsTunnelEntry
> >    MAX-ACCESS    not-accessible
> >    STATUS        current
> >    DESCRIPTION
> >         "An entry in this table represents an MPLS tunnel.
> >           An entry can be created by a network administrator
> >           or by an SNMP agent as instructed by an MPLS
> >           signalling protocol. Whenever a new entry is
> >           created with mplsTunnelIsIf set to true(1), then a
> >           corresponding entry is created in ifTable as well
> >           (see RFC 2863). The ifType of this entry is
> >           mplsTunnel(150)."
> >
> > >
> > > The label stack is represented either in the LSR MIB as it is
> > > defined today.  If you want to see the application-specific label,
> > > then look in the PWE3 MIBs as a short-hand. The PPVPN-MPLS-VPN MIB is
> > > also going to be modified to show the VPN label as well.  However,
> > > in either case, the ingress labels should be available in the LSR MIB
> > > as a point of reference.
> > >
> >
> >How would you represent the tunneling LDP into RSVP-TE at the ingress LSR of
> >the RSVP-TE tunnel?
>
>          You would show the LSP in the LSR MIB being switched to
> the TE label.
>
> >Where would you store the labels (LDP label and RSVP label) that need to be
> >pushed? In the label stack table of the LSR-MIB or higher layer.
>
>          In the LSR MIB's XC table.  The question above asked how to do LDP/TE.
> The forwarding/label stacking is different from the interface stacking
> used to accomplish this. The ifStack looks like:
>
>          Outgoing Interface:
>
>          TE
>          MPLS
>          underlying layer (enet for instance)
>
>          Incoming Interface:
>          MPLS
>          underlying layer (enet)
>
>          The cross-connection on a switch to do this might be:
>
>          Incoming LDP label, incoming interface -> NULL outgoing label
>
>          NULL incoming label -> outgoing TE tunnel label, outgoing TE interface
>                  XC defines a push with the incoming LDP label in the
>                  label stack table.
>

In this case, mplsLdpDownLabel object (defined in the mplsLdpDownLabelTable of
Neil's proposal) will be LDP label (inner label) while
mplsLdpDownLsrOutSegmentPointer object (of the same mplsLdpDownLabelTable entry)
points to an out-segment entry configured with the REVP-TE label. Would it not be
confusing?

>
>          So the resulting outgoing label stack would be LDP label/TE label.
> However, that would go out of the TE interface/MPLS interface.
>
>          --tom
>
> >Regards, Riza.
> >
> > >
> > >          --Tom
> > >
> > > http://www.elsevier-international.com/catalogue/title.cfm?ISBN=155860751X
>
> http://www.elsevier-international.com/catalogue/title.cfm?ISBN=155860751X