Minutes for Meeting of SIP Working Group, IETF 72
Edited by Dean Willis based on notes by Brian Rosen, Alan Johnston,
Ben Campbell, and John Elwell.


Session 1, Tuesday 13:00 – 15:00 Convention 3 


Topic: Agenda and Status
Led by Chairs
Slides presented and included in proceedings

Agenda accepted as presented.


Topic: Identify requirements for test matrix to move SIP to Draft
Standard
Led by Robert Sparks
Slides presented and included in proceedings

Robert discussed the requirements for moving to draft standard. One
key issue is the decision of which RFCs to revise as a group, with the
two apparent alternatives being 1) 3261 and its dependencies including
TLS, SDP, etc. vs 2) The larger SIP Roadmap family, including Events,
Refer, 100Rel, etc.

Robert proposed three paths forward, including 1) move some set of
documents to draft, 2) revise the set of documents at "proposed", and
3) do nothing.

A poll was taken, and there was a strong consensus in favor of
recycling the core documents at the "proposed" level.

Robert discussed the concept of an implementation report, and a
majority in the room supported the position that such a report would
be useful. A number of participants volunteered to form a team to
pursue such a report. 


Topic: Delivery of Request URI and Parameters to UAS Through Proxy
Led by Jonathan Rosenberg 
Slides presented and included in proceedings

Jonathan reviewed the requirements and the previously proposed
solutions as the state of discussions on the topic. He proposed
solving the problem by extending the History-Info header field using a
new "Target" parameter. The room demonstrated consensus on this
general approach, although it was noted that the deletion of other H-I
fields as proposed might create a security issue.

Jonathan volunteered to write a draft documenting the proposal,
without including the deletion of other H-I fields as noted. 


Topic: INFO Issues
Led by Eric Burger 30 
Slides presented and included in proceedings

Eric reviewed the problem's history and the various proposals taht
have been made. 

The room showed a consensus that a user-to-user INFO like mechanism is
required, that said mechanism should be called INFO (rather than
something else), and that the SIP WG should adopt
draft-kaplan-sip-info-events as the basis for a solution. 

The chairs are to coordinate with the ADs on getting this work item
approved and setting a milestone for its completion.

Further discussion occurred on the need for documenting the guidelines
for signaling-path vs. media-path transport. There was a general
consensus (with some objections) for pursuing this work. SIPPING chair
Gonzalo Camarillo encouraged the discussion of requirements for this
work in the SIPPING working group.



Topic: Identity Issues

Led by John Elwell

Slides presented and included in proceedings

John discussed the problem background and discussed a significant use
case involving SBCs that perform media steering, with which RFC 4474
may not be effectively used.

Discussion followed as to whether this "failure" is actually a feature
and design goal of RFC 4474, with no consensus developing. This
discussion included a lengthy review of the rationale of RFC 4474 by
Jon Peterson. 

Further use cases, such as preservation of identity across multiple
service provider SBC boundaries, were raised.

Discussion on what to do next continued without final conclusions.



Session 2, Thursday 15:10 – 16:10 Convention 3 


Topic: Agenda bash and status

Noted that chairs will be assigning reviewers for
draft-ietf-sip-record-route-fix, as list review appears to be light.

Noted also that draft-ietf-sip-dtls-framework is in WGLC until 8/8/08,
and that there is a parallel doc in AVT also in WGLC. 

Agenda accepted as proposed.

 
Topic: Mechanisms for UA Initiated Privacy
Led by Mayumi Munakata
Slides presented and included in proceedings

Mayumi reviewed the status of the draft and its relationship to the
Privacy RFC (which it augments but does not replace). 

Issue: What should this draft say about sending privacy-id?

Noted that proxies that don't know better may insert
privacy-compromisng information, so requesting privacy is still
important. Also noted that location privacy (as per geopriv) is
outside the scope of this draft.

Noted that using GRUU will not anonymize the domain name. Author will
add text explaining use of third-party or anonymizing GRUU servers
servers to control this aspect of privacy. 

Noted that it would be useful to capture implementation info on this
draft at future SIPIts, and the SIPIt Coordinator is asked to do so.

The author expects to make one more revision of this draft based on
current feedback, then request WGLC.


Topic: Termination of early dialog prior to final response
Led by Christer Holmberg
Slides presented and included in proceedings
 
Issue: Allow reliable 199? If so, where does the resulting PRACK go

Noted that this might require updating the 100Rel RFC, which there
is general consensus to avoid. This issue will require further
discussion on-list.

Issue: Should a UAS be allowed to send a 199? 

Agreed that proxy-initiated 199 is a transitional measure, and that
the long-term solution requires UAS support. Robert Sparks volunteered
to contribute supporting text.

Issue: Should a 199 contain a sipfrag from the final response that
triggered it?

Discussion centered on the relationship between 199 and HERFP, with it
noted that solving HERFP was previously agreed explicitly out of scope
for 199. The conclusion was that the draft will remain silent on the
inclusion of sipfrag in the 199.

Issue: Do we need an option tag?

Discussion centered on the impact of not supporting 199, or of not
receiving a 199 when it is expected. The current consensus is that we
may not have all the use cases considered, but that the current use
cases do not require an option tag.


Topic: Keepalive Without Outbound
Led by  Christer Holmberg 
Slides presented and included in proceedings

Christer reviewed problem definition and use cases, without going into
the details of the proposed solution.

A poll on "who is interested in working on this problem" resulted in a
small but non-zero number of positive responses. A followup poll on
"Who thinks solving this problem is critical" received a larger number
of positive responses. 

Discussion was inconclusive, and the group resolved to continue the
discussion of use-cases on the mailing list.