Transparent Interconnection of Lots of Links (TRILL) Operations, Administration, and Maintenance (OAM) Framework
draft-ietf-trill-oam-framework-04

I have no objection to the publication of this document. Here are a few Comments that might improve the document.

---

Please note that, after some soul-searching, the RFC Editor has "OAM" 
expanded as "Operations, Administration, and Maintenance" per RFC 6291
that you reference.

---

      ECMP - Equal Cost Multi-Pathing

Isn't this more normally "Equal Cost Multipath"?

---

The MPLS working group had a lot of discussion with the ITU-T around 
terminology for OAM. This resulted in draft-ietf-mpls-tp-rosetta-stone
which has been approved for publication as an RFC. You may find it
helpful to allign your trminology (such as MEP and MIP) with that 
document. You might consider that your variations are not significantly
different, and I might find it hard to explain how they are different,
but the folk from the ITU-T considered precision very important and
claimed they were aligning with Y.1731 amongst other documents.

---

Of course, Section 1.2 could contain an endless list of prior art. I
should have liked it had you referenced RFCs 4377/8 and 5860/6371 just
because of the amount of time and effort that went in to them and the
consensus between two SDOs that the second pair embody. But I would
completely understand if you decided you didn't even want to start to
think about this at this stage.

---

Figure 1

I think there is an alignment problem with the BFD line.

---                            

Section 3.1

   The TRILL OAM packet format proposed below provides

s/proposed/defined/

Francis Dupont made some editorial comments in his Gen-ART review. And the authors responded. Hopefully there's a new version somewhere or instructions to the RFC Editor so that these comments are not lost. At least the one on the expansion of ECMP term seemed like a fix that should be done, not sure if the others are more RFC Editor style issues.

This was very well-written and clear to me. Thank you for that.

> BFD, which is typically used for fast convergence
> BFD provides fast convergence characteristics to TRILL networks.

BFD only provides fast failure detection. How convergence happens is up to ISIS.

> The recommended default value is the character string "DEFAULT".

Presumably you need to identify a character set to send this in.