[mpls] FW: RE: draft-smack-mpls-rfc4379bis

[David Allan I <[e-mail]>]

Cc'd the MPLS list as quested
-------------------------------------------------
HI Loa:

I've read it though and will state up front, that as a next step in the standards process (bis version), I'm applying a tougher standard to the document and looking at this as more of a "last call". Even more so as it is OAM and therefore IMO requires a high degree of interoperability to be useful. As pretty much all of this is from the original document and outside the scope of what has been proposed (errata incorporation
etc.) and likely just complicates life, you can take it with "a grain of salt".

My chief concern with the document is that the language is very non specific and vague, and what I would expect to say "this is..." 
typically says "this might...".  I appreciate that much of this originates with the original RFC 4379, but for a BIS I would expect it to be cleaned up.

Some examples:
"An MPLS echo request with 1 (Do not reply) in the Reply Mode field  may be used for one-way connectivity tests;". IMO it is not a "may be", it "is" used for one way connectivity tests. If there is no reply, what else are you going to do.

"Sub-TLVs have  independent types and MUST also be 4-octet aligned. 
Two examples follow."  The examples would appear to be a normative definition of the ONLY sub-TLVs that are defined in this I-D. Other RFCs (e.g. VCCV) may define more, but these are the ones defined here, they are NOT examples.


And finally....
I'm hoping I'm reading this incorrectly, but assuming my read IS correct, It is obviously far far too late to change it but the FEC stack would have been easier to use if it listed the FECs from BOS to TOS, and not TOS to BOS as it is described now.  An intermediate LSR would need to count from BOS backwards given the depth of the current stack would be arbitrary at any intermediate LSR. Should have caught that years ago....

All of this being said, I have no idea what leeway in an IETF bis that one has to fix editorial and style issues. Meanwhile I do consider correcting the references and applying all errata to be a useful step.

Cheers
Dave