The MPLS WG Archive

["Adrian Farrel" <adrian@olddog.co.uk>]

Hi Stefaan,

PMFJI



> > 5) Question on the tunnel index
> >
> >  mplsFrrConstTunnelIndex OBJECT-TYPE
> >        SYNTAX        MplsTunnelIndex
> >        MAX-ACCESS    not-accessible
> >        STATUS        current
> >        DESCRIPTION
> >          "Uniquely identifies a tunnel for which fast reroute is
> >           requested."
> >        ::= { mplsFrrConstEntry 2 }
> >
> >     mplsFrrConstTunnelInstance OBJECT-TYPE
> >        SYNTAX        MplsTunnelInstanceIndex
> >        MAX-ACCESS    not-accessible
> >        STATUS        current
> >        DESCRIPTION
> >          "Uniquely identifies an instance of this tunnel for which fast
> >           reroute is requested.
> >
> >          - lower 16 bits : protected tunnel instance
> >          - higher 16 bits: must be all zeros"
> >        ::= { mplsFrrConstEntry 3 }
> >
> > mplsFrrConstTunnelIndex OBJECT-TYPE - what would be the required value here?
> >
>
> Not really clear what you mean: the TunnelIndex is the tunnel index
> of the tunnel being protected...

I haven't read this enough, but:

- If mplsFrrConstTunnelIndex contains enough to uniquely identify
   the TunnelIndex of the protected tunnel, why don't you add a reference
   to the field in the TE MIB?
- Is a tunnel index enough to uniquely identify a protected tunnel?
   Don't you also need destination and extended tunnel ID?
   Perhaps you have these and they're not in the snippet shown here.

> > 9) Questions on One2one backup
> >
> > We don't have a variable that specifies the bw on the detour. Don't we
> > need one?
> >
> > There are no notifications for the 1-to-1 backup. Wouldn't we want to
> > track when an lsp starts using the detour and when it stops using it?
>
> I checked the TE MIB and there are up and down notifications defined
> at the tunnel level. There are no up/down notifications at the
> instance level and since also notifications for the one2one case
> would be at the instance level I believe it is better not to have
> notifications.
>
> I have not checked with the co-authors but I would like to hear
> their opinion as well, as well as other people's opinion of course.

I seem to recall someone (Nurit Sprecher?) requesting this in the very late days of the TE
MIB.
The idea was deferred for inclusion in this MIB.

Thanks,
Adrian