The MPLS WG Archive

[Philip Matthews <pmatthews@hyperchip.com>]

Title: RE: Initiating the process to take LPD to draft standard

My assumption is that Loa and Bob are talking about RFC3036,
which does not mention any sort of LDP resiliency scheme.

The proposals for LDP Fault Tolerance and LDP Graceful Restart
are not part of RFC3036 and would advance separately. None of
these proposals are even at proposed standard status yet.

- Philip

> -----Original Message-----
> From: neil.2.harrison@bt.com [mailto:neil.2.harrison@bt.com]
> Sent: 20-Dec-01 09:25
> To: loa.andersson@utfors.se; mpls@UU.NET; pmatthews@hyperchip.com
> Cc: swallow@cisco.com; sob@harvard.edu; bwijnen@lucent.com;
> rhthomas@cisco.com
> Subject: RE: Initiating the process to take LPD to draft standard
>
>
> Loa,
>
> How does this square with the recent mail posting on LDP
> resiliency from
> Philip Matthews...see attached?
>
> I am thinking specifically about the interoperability combinations
> mentioned....seems like some more need adding.
>
> regards, Neil
>
> > -----Original Message-----
> > From: Loa Andersson [mailto:loa.andersson@utfors.se]
> > Sent: 20 December 2001 13:05
> > To: mpls@UU.NET
> > Cc: George Swallow; sob@harvard.edu; bwijnen@lucent.com; Bob
> > Thomas; Loa
> > Andersson
> > Subject: Initiating the process to take LPD to draft standard
> >
> >
> >
> > All,
> >
> > We have been notified by the AD's that it is time to advance LDP
> > (RFC3036) from proposed standard to draft standard.
> >
> > A formal requirement on an IETF protocol to advance to
> draft standard
> > is that there be at least two interoperable implementations of the
> > specification.
> >
> > This is a request for information about LDP implementations.  Please
> > send the information to:
> >
> >     Bob Thomas <rhthomas@cisco.com>
> >     Loa Andersson <loa.andersson@utfors.se>
> >
> > RFC3036 defines several modes of label distribution and allows an
> > implementation to support one or more of the modes. In addition,
> > RFC3036 specifies optional behavior. Therefore we need information
> > about the label distribution modes and optional behavior
> > supported by each implementation. 
> >
> > We ask that the information supplied include which of the following
> > label distribution modes are supported:
> >
> >        Advertisement             Retention      Control
> >
> >     1. Downstream Unsolicited    Ordered        Liberal
> >     2. Downstream Unsolicited    Ordered        Conservative
> >     3. Downstream Unsolicited    Independent    Liberal
> >     4. Downstream Unsolicited    Independent    Conservative
> >
> >     5. Downstream on Demand      Ordered        Liberal
> >     6. Downstream on Demand      Ordered        Conservative
> >     7. Downstream on Demand      Independent    Liberal
> >     8. Downstream on Demand      Independent    Conservative
> >
> > as well as optional features, such as loop detection, optional
> > TLVs, etc., supported by the implementation.  Also, please feel free
> > to include any additional material believed to be relevant.
> >
> > We also need information on interoperability testing and deployment
> > status in order to address the requirement of "two independent and
> > interoperable implementations".
> >
> >
> > Thanks,
> > Bob and Loa
> >
>
>