Re: [mpls] Comments on draft-nitinb-lsp-ping-over-mpls-tunnel-01

["Nitin Bahadur" <nitinb at juniper.net>]

Hi Ben,

> During the ITU-T SG15 ad hoc meeting on T-MPLS in February, the ITU
model
> (G.8110) of MPLS was discussed.  G.8110 models a SWAP as a POP
followed by
> a
> PUSH and at least one of the IETF representatives at the meeting
expressed
> the view that modelling a SWAP as a POP followed by a PUSH did not
> sufficiently describe the behaviour as per the MPLS architecture
> (RFC3031).
> 
> You appear to be doing the same here, so is modelling a SWAP as a POP
> followed by a PUSH sufficient in this context?

Very good observation. In the forwarding plane, a POP/PUSH does not tell
you much. In the control plane, a POP/PUSH can be tailored to provide
sufficient detail so that the semantics are much clear. In this draft,
the POP/PUSH operations have FEC and peer-address associated with them.
That should help clarify the intended behavior. If there is still a
chance of ambiguity, we can modify the control information for clarity.


> Remote peer address
> "If the operation type is POP, the remote peer address MAY not be set
to
> Unspecified."
 
> I find the use of the term "MAY not" confusing.  If you really mean
MAY
> then
> as it is optional, IMO the not is superfluous.

Yes, that's correct. The intention is "MAY" and not "MAY not". I shall
have that corrected.

I've taken care of the editorials. Thanks for combing through the
document in detail.

Nitin 
_______________________________________________
mpls mailing list
mpls at ietf.org
https://www.ietf.org/mailman/listinfo/mpls