The MPLS WG Archive

[Francis Arts <francis.arts@alcatel.be>]

Hello Kireeti,

Please find included some comments to your <draft-ietf-tewg-mib-00>
draft.

Kind regards,

    Francis.

----------------------

1) p.13, 'tePathExclude' object.

The text states that 'for each link that this path goes through, the
link MUST have colors associated with it, and the intersection of the
link's colors and the 'exclude' set MUST be null.

Question:
* Why is it mandatory for a link to have colors in order to pass the
'tePathExclude' test?

Referring to section 4.7.2 in <draft-ietf-mpls-rsvp-lsp-tunnel-08>
draft, the text does not impose this restriction. Therefore, according
to my understanding of the <draft-ietf-mpls-rsvp-lsp-tunnel-08> draft,
if all bits in the 'exclude-any' field of the SESSION ATTRIBUTE object
are set to 0, the test passes automatically.


2) tePathExclude.

Question: Could we rename the tePathExclude' object to
'tePathExcludeAny' in order to follow a similar convention as is done
for the 'tePathIncludeAny'
object and in order to align with the terminology used in section 4.7.2
of the <draft-ietf-mpls-rsvp-lsp-tunnel-08> draft?


3) Include vs exclude.

For the include case 2 variants are indicated ('include-any' and
'include-all'). However, for the exclude case only 1 variant is defined
('exclude-any').
Why is the other case ('exclude-all') not defined, just as was done for
the 'include-all'? [Of course, this would also require similar
adaptations in the SESSION ATTRIBUTE object of the RSVP signalling
protocol.]