TSVWG IETF-76 (Hiroshmia) Thursday, November 12th, 2009 WG Chairs: James Polk Gorry Fairhurst TSVWG Meeting Chair: Magnus Westerlund Notetaker: Martin Stiemerling Magnus Westerlund is chairing, since neither chair could attend in person at this meeting. 1. Chair http://www.ietf.org/proceedings/09nov/slides/tsvwg-0.ppt NOTE WELL Document Status and Accomplishments Milestones Review With IESG: draft-ietf-tsvwg-rsvp-l3vpn draft-ietf-tsvwg-rsvp-proxy-approaches draft-ietf-tsvwg-rsvp-proxy-proto draft-ietf-tsvwg-admitted-realtime-dscp (waiting for author revision) Documents completed WGLC: draft-ietf-tsvwg-emergency-rsvp (with GF) draft-ietf-tsvwg-port-randomization (with JP) Documents in WGLC: draft-ietf-tsvwg-rsvp-security-groupkeying Reviews needed: draft-ietf-tsvwg-ecn-tunnel Janet Gunn (via jabber): When will we get the new version of Capacity Admitted Traffic? - The current I-D is over a year old. Magnus: We want to have new version, ask the authors or volunteer to help. 2. IPv6 UDP Checksum Considerations (Magnus Westerlund, as co-author) draft-fairhurst-tsvwg-6man-udpzero Magnus: This draft was discussed in the 6man working group (which owns the IPv6 specification). Magnus: Please watch 6man, mbone, lisp lists, and contribute when discussed there. Bob: When you add a UDP shim to a tunnel, should the IP header fields be copied to outside headers? Magnus: The draft talks about UDPv6, specifically checksum rules. This is not addressed. Joe: These issues are covered in the tunnel issues draft in the INT area (ingress and egress must reflect the signaling inside the tunnel), please read and contribute. We are taking edits to the document. 3. IANA Procedures for Management of Service and Transport Protocol Port Registry (Joe Touch) draft-ietf-tsvwg-iana-ports I-D has been updated and being taken to other IETF groups. There will be another document with advice to application designers. Dave Thaler: The I-D recommends constrains the syntax for the IANA registry. I think these changes do not apply to /etc/services, which was also mentioned in the presentation. Joe Touch: Correct. The format remains the same in /etc/services, it would add some entries. The document is to be updated (mid December), please review the next revision. 4. Integrated Services Extension to Allow Multiple TSPECs (Francois Le Faucheur, on behalf of James) draft-polk-tsvwg-intserv-multiple-tspec http://www.ietf.org/proceedings/09nov/slides Lars (as AD): The charter includes maintenance of RSVP. This I-D updates both RSVP and INTSERV. This is not a minor extension. Francois: This is an interesting comment, what is different in this case? Lars: The other I-Ds in TSVWG did not update RSVP and also update INTSERV. Francois: True. Lars: RSVP is a stable protocol. TSVWG is not currently charted to do big feature changes. If we do this, we need to recharter. Magnus: Agreed. How many others are interested in participating in such work? (nobody replied) Magnus: The overall interest in RSVP extensions is quite low, beyond the group of authors. It seems we need more energy. Lou Berger: RSVP related discussions should occur in IETF. Lars: RSVP use seems limited to specific use cases. I don't see the community for this work here. There seems to be one vendor, who wants to extend this, for a very specific customer. Francois: Does this apply in general to RSVP? Lars: The energy level in general and also for this. This I-D is not a small bugfix, it a pretty big extension. Lou: Are you trying say RSVP is Historic? Lars: I didn't say anything of this. There is a specific charter to do minor extensions. The protocol can be still used. Magnus: Are people interested in contributing to this? We need to see a community. Lars: Only a small number of people are contributing to the standard, driving this to be proprietary. is that really the way you wanted to go? Bob: Please don't discourage cisco from coming here. Cisco's customers want this to be standardized, so that they are not only relying on cisco. Joe: There are other ways to publish, e.g. Informational. Magnus: If nobody will review this, there is no meaning to do this in a WG. Lars: I do not want to push people to go proprietary. We hear you. However, if multiple vendors were to be involved discussing this, I would have a much better feeling. Magnus: The proxy work is on the border of the charter, but I am not going to stop WG work on proxy. Gorry (via Jabber): That would be my point, we need people to discuss and review, even if there is a small group of authors. Whether we do new things in the WG is an AD decision. Magnus: Try to get support for it. Lars: I have a problem to accept it as WG item. I am happy to discuss this if you come back with more interested people. Francois: Can we draw a line between small and big changes? Lars: Small extensions mentioned in charter. This I-D is not a small extension. It plans to update RSVP and INTSERV. Lou: if you're just looking for reviewers, contact me. Magnus: ... I encourage discussion on the list. 5. Multiple Preemption Priority Policy Element for RSVP (Francois Le Faucheur) draft-lefaucheur-tsvwg-rsvp-multiple-preemption The current draft is linked to the former. Francois thinks both would need to progress together. 6. RSVP Proxy - Approaches & Protocol (Francois Le Faucheur) draft-ietf-tsvwg-rsvp-proxy-approaches draft-ietf-tsvwg-rsvp-proxy-proto Magnus noted we need to clean up all the DISCUSS notes from the IESG based on the new revision. He will try to get the ADs that raised DISCUSSes to respond. 7. RSVP Extensions for Flexible Resource Sharing (Francois Le Faucheur) draft-narayanan-tsvwg-rsvp-resource-sharing There are common topics with another IETF WG. This work may be taken-up by CCAMP. 8. Layered Encapsulation of Congestion Notification (Bob Briscoe) draft-ietf-tsvwg-ecn-tunnel (Moved ahead in agenda at Bob's request) Gorry (via Jabber): I think this new revision is readable, and seems stable. I would love feedback about this document. Magnus: Please send comments to the list, appreciated. 11. Byte and Packet Congestion Notification (Bob Briscoe) draft-ietf-tsvwg-byte-pkt-congest Bob said there were comments on the list based on the recently revived I-D and a new revision is expected. 12. ECN for RTP/UDP/IP (Colin Perkins) draft-westerlund-avt-ecn-for-rtp This draft was presented in AVT this IETF meeting, but also includes ECN topics within the charter of TSVWG. A coordinated process would be used with WGLC in both groups. Bob: I thought that ECN-Nonce would break header compression. Magnus: No, it is just more inefficient to compress. Colin: My slides predated discussions with Bob. Lars: What is the recommended response to receiving an ECN CE marking? Colin: For each mark, you do the same as for a loss, we do not want to specify a new response. We will seek to do the same thing as 3168, if anyone spots issues, do let us know and we will wordsmith the correct text. The session ended 16:27