Re: [CCAMP] BFD Directed Path discussion

["Nobo Akiya (nobo)" <[e-mail]>]

Hi Greg, Prasad, et al,

Great to see this being discussed. Adding CCAMP and authors of both documents.

Speaking as a BFD WG co-chair ...

> [1] draft-vgovindan-mpls-extended-bfd-disc-tlv
> [4] draft-mirsky-mpls-bfd-directed-00

Regarding above 2 documents, details should be discussed further but the needs of both look valid. However, we should not lead down the path of having more and more BFD bootstrapping TLVs, but have/aim-for one TLV that allows Sub-TLVs for extensions.

> [2] draft-ietf-mpls-lsp-ping-mpls-tp-oam-conf-06

Above document is dead.
IETF State: Dead WG Document

> [3] draft-ietf-ccamp-rsvp-te-mpls-tp-oam-ext-12

Above document is already LC'ed (long ago).
IETF State:  Submitted to IESG for Publication 

My best guess is that we do not want to disrupt/block [3] at this point.

Question to authors of [1] and [4].

Are you interested to change LSP Ping bootstrapping only? If so, then it seems the authors of [1] and [4] need to get together and come up with a single BFD bootstrapping TLV (given that [2] is dead).

Thanks!

-Nobo

> -----Original Message-----
> From: Vengada Prasad Govindan (venggovi)
> Sent: Thursday, July 24, 2014 6:35 AM
> To: Gregory Mirsky; [e-mail]
> Cc: [e-mail]; Nobo Akiya (nobo)
> Subject: RE: BFD Directed Path discussion
> 
> Hello Greg,
> 
> Thanks for your note, a couple of clarifications here:
> >how this work is related to Extended BFD Discriminator TLV work? Both
> >may be complementary, especially as parts of controlling BFD >session.
> >Would welcome authors of Extended BFD Discriminator TLV to discussion
> >and review draft-ietf-ccamp-rsvp-te-mpls-tp-oam-ext >and
> >draft-ietf-mpls-lsp-ping-mpls-tp-oam-conf drafts (authors are planning
> >to bring them over the finish line)
> a. Are you suggesting that we add sub-TLVs to convey the instance identifier
> [1] inside the BFD configuration sub-TLV of both MPLS LSP ping [2] and the
> RSVP TE OAM [3] messages?
> b. Are you considering the addition of BFD reverse-path TLV [4] as well to
> the LSP ping and RSVP-TE OAM messages?
> 
> Thanks
> Prasad
> [1] draft-vgovindan-mpls-extended-bfd-disc-tlv
> [2] draft-ietf-mpls-lsp-ping-mpls-tp-oam-conf-06
> [3] draft-ietf-ccamp-rsvp-te-mpls-tp-oam-ext-12
> [4] draft-mirsky-mpls-bfd-directed-00
> 
> 
> From: Rtg-bfd [mailto:[e-mail]] On Behalf Of Gregory
> Mirsky
> Sent: Wednesday, July 23, 2014 12:27 AM
> To: [e-mail]
> Cc: [e-mail]
> Subject: BFD Directed Path discussion
> 
> Dear All,
> on behalf of the authors of this work I thank you all who participated in the
> discussions Monday (BFD WG) and today (MPLS WG). Great comments and
> suggestions - very excited by your interest in this work.
> I've tried to capture summary of the comments and sketch possible ways to
> address these:
> . BFD Return Path TLV suitable for RFC 5880, what about S-BFD?
> In order to maintain stateless character of S-BFDReflector, S-BFD control
> packet may need to carry return path information. Need to discuss with
> authors of S-BFD Base and start work on S-BFD over MPLS LSP with IP and
> ACH encapsulations; . What would happen if Return Path conveyed to the
> far-end BFD peer is not available? Ingress may send another BFD Return
> Path in its TLV along with the same BFD Discriminator; . Can BFD Echo of RFC
> 5880 be used in conjunction with the BFD Return Path TLV over MPLS LSP? It
> may be interesting as BFD Echo already allows for a payload but work on
> proper handling and encapsulations of BFD Echo is needed. Besides, use of
> BFD Echo over MPLS LSP may be limited to single segment LSPs thus S-BFD
> may be more generic mechanism to have stateless far-end BFD peer; . how
> this work is related to Extended BFD Discriminator TLV work? Both may be
> complementary, especially as parts of controlling BFD session. Would
> welcome authors of Extended BFD Discriminator TLV to discussion and
> review draft-ietf-ccamp-rsvp-te-mpls-tp-oam-ext and draft-ietf-mpls-lsp-
> ping-mpls-tp-oam-conf drafts (authors are planning to bring them over the
> finish line)
> 
> Please excuse me if I've missed your question and please add it to this
> discussion.
> 
>                 Regards,
>                                 Greg