Minutes of the two SIPPING WG Sessions at IETF 69 ================================================= Minutes edited by Mary Barnes Based on detailed notes by Peter Blatherwicke, Eric Burger, John Elwell and Dale Worley Jabber: http://www3.ietf.org/meetings/ietf-logs/sipping/ Jabber scribe: Bruce Lowekamp Meeting chaired by Mary Barnes and Gonzalo Camarillo Slides presented included in the proceedings FIRST Session Tuesday, July 24th, 1740-1950 Grand Ballroom -------------------------------------------- Topic: Agenda Bash Discussions led by: Chairs The chairs presented the current status of the WG and the next planned charter updates, reflecting that the WG is doing a good job of meeting milestones overall. An important document dependency between the Configuration Framework and the Session Policy (one document in the SIP WG and two in the SIPPING WG) in the form of an individual draft was highlighted. In addition, a few minutes was allocated for Ravi to highlight his individual draft on "Perceptual MOS". Topic: Offer/Answer: o-line Usage Discussion led by: Paul Kyzivat Relevant documents: draft-ietf-sipping-offer-answer-02.txt The document is deemed to be nearly complete. The issue of the SIP session versus SDP session was discussed, with the proposal being to point out that the behavior is non- compliant. Conclusion: Once the document is updated resolving that issue and various editorial changes, it will undergo a one month WGLC to provide sufficient time to elicit feedback from implementers. Topic: Offer/Answer: Race Conditions Discussion led by: Paul Kyzivat Relevant documents: draft-ietf-sipping-race-examples-02.txt This document has completed WGLC, with a summary of changes post-WGLC presented. This document and is deemed to be ready. Conclusion: Chairs to post a note to mailing list to send any additional comments. The document will then be forwarded to the AD for progression. Topic: Service Identification Discussion led by: Jonathan Rosenberg Relevant documents: draft-rosenberg-sipping-service-identification-03.txt draft-rosenberg-sip-app-media-tag-00 draft-drage-sipping-service-identification-00 Note: these 3 drafts were updated/produced as a result of the compromise solution proposed and discussed at IETF-68. Discussion on the first document, which is the proposed long term solution for this general problem, centered on what to do with offerless INVITEs. General support for this document was expressed. A concern was also raised (and agreed) that this problem is bigger than SIP and needs to be evaluated by the APPS folks. Conclusion: Draft agreed as the baseline for a WG document, for the previously (IETF-68) agreed Service Identification charter item. Jonathan to submit -00 post IETF-69. Jonathan briefly discussed the media feature tag for application sub-types, which is being progressed as an individual AD-sponsored submission and is currently undergoing IETF LC. There are still some concerns on this solution approach, but 3GPP prime (Andrew Allen) highlighted that this approach has been agreed in 3GPP in terms of a compromise. Jonathan also discussed the P-header draft on behalf of Keith Drage. There are concerns over the usefulness of P-Asserted-Service and what some refer to as the "evilness" of P-Preferred-Service. Folks are encouraged to send comments to the IETF mailing list as the document (and the media-tag document) is undergoing IETF LC as an individual AD-sponsored submission. Topic: Configuration Framework Discussion led by: Sumanth Channabasappa Relevant documents: draft-ietf-sipping-config-framework-12.txt As chair, Mary presented background of the evolution of this document that has been a WG item for 5 years with a plea for the importance of completing this work. It was highlighted that the changes relative to the -11 are significant normative changes. Sumanth presented the technical and document content changes, highlighting that any changes were a result of WG feedback. There was a concern raised about the relation of this document to Outbound (in terms of locating a server that might be behind a NAT). It was suggested that this was a broader issue than config FW and that the solution to such might not belong in this document. Also, some folks considered this to be non-issue. Markus (who raised the initial issue) is to follow-up with an email identifying the concern. It was suggested that this document has suffered from over-editting and it was suggested to have a single editor to do the final clean-up. The chair attempted to do a hum vote on the way forward and a question was raised on the relation of the "Simple Application Configuration Protocol" and this document (i.e., is it a competing solution?) The chair's response was that the other document is complementary as it addresses a pure HTTP based solution that had been part of this document, but was removed as the solution was incomplete. It was decided that a hum vote was not necessary. conclusion: Any additional editorial changes are the sole discretion of the editor, Sumanth, with folks to raise any technical concerns that were not raised as part of WGLC ASAP, so we can complete this document. Topic: Simple Application Configuration Protocol Discussion led by: Simo Veikkolainen Relevant documents: draft-veikkolainen-sipping-app-config-00.txt Simo summarized the motivation for this document in terms of addressing non-SIP environments and very simply applications that didn't need a Notify mechanism, with a deliberately restricted scope. It was questioned as to whether the SIPPING WG was the right place to do this work and whether there was a list of requirements that aren't covered by the SIP Config FW. Simo noted that this document had been presented and discussed also in the APPS area this week. It was re-iterated that SIPPING WG needs enumerated requirements to consider the work. Conclusion: Discussion on this topic to continue on the list. SECOND Session Thursday, July 26th, 2007, 1300-1500 Red Lacquer ------------------------------------------------ Topic: Agenda Bash Discussions led by: Chairs The chairs also pointed out a proposal to expedite the processing of new requirements that impact SIP by not progressing those as separate documents in SIPPING, but rather once the requirements are agreed, the work can transition (in the form of an individual solution draft) to the SIP WG. The chairs also highlighted draft-houri-sipping-presence-scaling-requirements for working group feedback. Jonathan Rosenberg provided a brief update on the Overload Design team activity. The results from two different simulators yielded results that were quite different than the currently drafted mechanism. The team is in the process of coding up two new proposals for feeding into the simulator. Topic: An Extension to SIP Events for Pausing and Resuming Notifications Discussion led by: Mohammad Vakil Relevant documents: draft-vakil-sipping-notify-pause-01.txt There was some general discussion on how large the costs of a establishing a subscription really are. It was mentioned that without numbers, it was not possible to assess the gains of this proposed mechanism (i.e., would the overhead be worth the potential savings). Rosenberg agreed with the author that there is some overhead reduction possible, particularly in an IMS environment. Conclusion: Hum vote indicated that there was slightly more interest in pursuing this problem than not and discussion should continue on the mailing list. Topic: SIP File Directory Discussion led by: Miguel A. Garcia-Martin Relevant documents: draft-garcia-sipping-file-sharing-framework-00.txt draft-garcia-sipping-file-event-package-00.txt draft-garcia-sipping-file-desc-pidf-00.txt Note: the name change from "-resource-" to "-file-" and reduced scope from IETF-68 discussions on this topic. The chair prefaced the discussion with the importance of focusing on the question of "Why use SIP for file subscriptions?" The author highlighted the related document in MMUSIC, draft-ietf-mmusic-file-transfer-mech that isn't sufficient for all cases. It was mentioned that while doing the event package in SIPPING might be okay, the problem is broader than SIP and might be more appropriately done in the APPS area, although Cullen (RAI AD) thinks it's okay for this work to happen in SIPPING. Folks expressed concern over whether this functionality is really useful (i.e., keeping a subscription for cases where a file might be available). Robert Sparks highlighted that this topic brings us back to the general question as to where the line is wrt event packages. Conclusion: Needs more discussion on the mailing list. Topic: Transparent B2BUA Discussion led by: Xavier Marjou Relevant document: draft-marjou-sipping-b2bua-01.txt The chair prefaced the discussion with the need to understand the purpose of the draft at this point, rather than delve into mechanisms. The author explained why he believed the draft was needed in terms of the things broken by B2BUAs, that have led to adoption of other solutions. Folks expressed concern over the term "transparent". It was also suggested that this draft might in a sense be "BEHAVE for SBCs". Hum vote: Would this document be useful or harmful? Conclusion: More folks believe pursuing this work would be more harmful than useful. Topic: Replace Instant Message in SIP Discussion led by: Da Qi Ren Relevant document: draft-ren-sipping-replace-instant-message-00.txt The discussion centered on whether message recall/replacement was possible, desireable or would be demanded by users. It was suggested that this would work only in a closed environment. It was also suggested that perhaps the recall (versus replace) might be something to look at or even a "never mind". Hum vote: Is this deemed useful (or useless) work? Conclusion: This work was not deemed useful. Topic: Response Code for Indication of Terminated Dialog Discussion led by: Christer Holmberg Relevant document: draft-holmberg-sipping-199-02.txt The chair prefaced that the focus of the discussion should be whether this is a problem we want to solve. There was discussion as to whether this problem is HERFP (Heterogeneous Error Forking Problem). The author said that it was not, whereas others felt that it was, although the problems being solved were lesser magnitude than in past HERFP solution proposals. There were divided views on the usefulness of this mechanism with some wanting to solve at least part of the HERFP problem with this mechanism, but others concerned that it could be harmful if it caused intermediaries to keep more state than they would otherwise. Hum vote: Is there interest in working on this problem? Conclusion: Hum vote in support of working on this problem. As to whether it should expand to cover the entire HERFP problem, discussion should continue on the mailing list. Topic: Media Description for IKE in SDP Discussion led by: Makoto Saito Relevant document: draft-saito-mmusic-sdp-ike-01.txt This draft had been discussed earlier in the MMUSIC WG and it had been suggested to get feedback from the SIP community. The chairs state that the intention is for the SIPPING WG to review the proposal and decide whether it's a sensible use of SIP. Since no one in the room had read the draft, the chairs asked for volunteers to review. Rohan Mahy, Richard Barnes and Rémi Denis-Courmont volunteered. Conclusion: SIPPING WG members to review the draft within 2 weeks and provide feedback to the author and mailing lists.