Minutes, SIPPING WG, IETF 57


Notes by Tom Taylor, AC Mahendran, and Hisham Khartabil
Minutes edited by Dean Willis
Chat room moderation by Brian Rosen
Meetings chaired by Gonzalo Camarillo, Rohan Mahy, Dean Willis

Session 1, July 14, 2003 1930-2200
==================================



Topic: Agenda
---------------------

Agenda accepted as previously posted.


Topic: Announcements and Status
----------------------------------

Chairs mentoin the SIP device reqts ad hoc 10-11 am Thur room J241/1, and that the QoS promotion scheme draft will be discussed in Friday's TSVWG meeting.

Slides presented reviewing status of working group documents.
          pres-chairs-SIPPING-WG-status-ietf-57.ppt


Topic: SIP to QSIG Mapping
------------------------------

Relevant document:
        draft-ietf-sipping-qsig2sip

Slides presented
         pres-elwell-qsig2sip-ietf-57.ppt

Discussion led by John Elwell

Issue: We need more qualified reviewers for this work. A hand-poll taken in the room indicated about a dozen people claiming to be familar with QSIG.  We discussed possibly working with external SDOs on this. Discussion indicated that ECMA is the likely SDO. John Elwell took the action item of providing a list of ECMA contacts to Allison Mankin.

No conclusion.

Actions: Francois Audet to formally review document.

Topic: Dialog Package
-----------------------

Relevant document:
         draft-ietf-sipping-dialog-package

Slides presented:
          pres-rosenberg-dialog-pkg-ietf-57.ppt

Discussion led by Jonathan Rosenberg

Status: Went through WGLC.  Minor comments incorporated.  Unable to restrict a particular XML type in XML -- has only the accompanying text to provide the constraint.

Issue: extensions proposed (Sean Olson).  Question: what to do with them?  Extend scope of main package, reject them, or document them separately?

Issue: "on hold".  Not clear what "on hold" means, since media can flow anyway. Suggestion that this isn't a dialog state.

Conclusion: Rough consensus that we are not ready to advance the dialog package at the present time.  Concerned parties instructed to discuss and report back to Thursday mtg.

Topic: Discussion of Configuration Issues
------------------------------------------

Relevant documents:

         draft-ietf-sipping-config-framework
         draft-ietf-sipping-ua-prof-framewk-reqs
         draft-ietf-simple-xcap-package

Slides presented:
          pres-petrie-config-fw-issues-ietf-57.ppt
          pres-mahy-config-issues-ietf-57.ppt


Some open issues w/ profiles.  No discussion. No conclusion on framrework or profiles.

Issue: potential conflict between XCAP package (SIMPLE work item) and sip-config. XCAP fits into framework: config retrieval, change notif, config upload. Differences in many details. Do they share the same requirements?  Do they share the same solution?  Should they be unified?

Proposal to unify led to comment that config is a much more general problem -- it may not be reasonable to use SIP for notifications in view of a more general model coming perhaps out of netconf.  This may reflect the views of a number of people in the IETF at large.  We will need to have a clear explanation of scope.

Key point: all of the entities involved have SIP addrs.  Fairly tightly scoped.

Proposal: merge or reorg the two drafts.  Need an author w/more cycles.  Poll indicated rough consensus in favour, with some hums opposed. Locus of work to be discussed w/ ADs, noting two different areas are involved.

Conclusion: Chairs to discuss with ADs.


Topic: Session Policy and Middle Box Issues
----------------------------------------------

Relevant Document:
         draft-ietf-sipping-session-policy-req-00.txt

Discussion led by Jonathan Rosenberg

Issue: Dynamic vs static policies
   -- Allison: not necessarily either/or
   -- Jonathan just concerned to eliminate unnecessary reqts
   -- CALEA an arg against dynamic policies

Rohan volunteers to work on mechanism if he has some help.

Questions on reqt for enforcement of policy.  Audience asked to review text and submit desired changes.

No conclusion.

Editors note: 3GPP is expecting some results here.

Topic: URI Leasing and GRUUS
----------------------------------

Relevant document:
         draft-rosenberg-sipping-lease-00.txt

An informal group met prior to this meeting. They noted that embedded route headers prevent proxy application of policy and user services, and that  leasing has a couple of other advantages.  This groups conclusion is that leasing seems to be the way to go.  There appear to be some open issues wrt GRUU grants. The author will add requirement that leasing be stateless to the requirements doc.

Conclusion: Agreed that requirements shall address stateless leasing. Consensus that work will continue with leasing model at this time.


Topic: Firewall and NAT Traversal
------------------------------------


Relevant Document:
         draft-rosenberg-sipping-ice-00.txt

Slides presented in:
          pres-rosenberg-sipping-nat-stuff-ietf-57.ppt

Status report

ice-01 addresses concerns raised in IETF 56
   -- backward compatibility
   --- case that breaks bkwd compatibility is multihomed host. alt framework vs. alt attribute being taken into off-line discussion

ice-01 has ~50 pages of example flows -- proposed replacement of sipping-nat-scenarios
   -- ice algorithm works out
   -- troublesome case is v4-v6
   -- need add call flow between ice and non-ice hosts

Question from chair: if Jonathan adds the required cases, is this a sufficient replacement for what we have now in the sipping-nat-scenarios document? Consensus is that it is adequate, provided that we add text to allow alternate approaches (such as ALG, MIDCOM) and add cases for MSRP relay. Francois Audet volunteered to send text relating to alternate approaches.

Process open issues:

1. Do we proceed, and if so, in which WG?
   -- main ice behaviour -- mmusic?
   -- SDP extensions -- mmusic
   -- preconditions -- if people care, would go to sip  (poss mmusic)
   -- usage scenarios -- repl sipping-nat-scenarios -- sipping

Conclusion: Hums indicated a consensus to do core ICE behavior and SDP changes in MMUSIC, usage scenarios in SIPPING, and preconditions work in SIP if needed.

Chairs are to discuss division of work with ADs.


Topic: Adding Realm Identifier for Private Addresses
------------------------------------------------------

Relevant Document:
         draft-audet-sipping-add-realm-00.txt

Slides presented in:
          pres-audet-sipping-add-realm-ietf-57.ppt

Notes: Adds explicit realm identifier for private (i.e. "local") addresses into SDP.  Useful in case both endpoints have a priori knowledge that they are in the same
realm -- can avoid ICE. Need to configure realm only if multiple media terminations behind the same NAT which will communicate with each other. Proposes SDP attribute.  Configuration mech out of scope of current draft.

Discussion on network toplogy and applicability indicated that some additional clarity is needed.  How does this relate to phone-context?

Conclusion: Requires further list dicussion.  Burden of proof is on Francois to say it is worth varying from general application of ICE.



Topic: Open Issues from Conferencing Design Team
-----------------------------------------------------


Relevant documents:
         draft-ietf-sipping-conferencing-framework
         draft-ietf-sipping-cc-conferencing  SIP
         draft-ietf-sipping-conferencing-requirements

Slides presented in:
          pres-levin-conf team-update-ietf-57.ppt

Discussion led by Orit Levin

Status: Conceptual issues resolved: basic conf by SIP, adv would use XCON (they hope).  Design team work is almost complete and they welcome general comments.  If XCON forms a working group, it may be possible to reduce the scope of draft-ietf-sipping-conferencing-requirements to "just the SIP requirements." 
Open issues with draft-ietf-sipping-conferencing-requirements: Focus discovery.

Open issues with draft-ietf-sipping-cc-conferncing: add details for selected msgs in presented call flows, or move into sip draft for mechanisms.

Open issues with draft-ietf-sipping-conference-package: Needs another rev e.g. re sidebar conversations Looking to future draft on implementation of sidebar



Topic: Open Issues from Transcoding/Deaf Design Team
---------------------------------------------------------


Relevant Document:
         draft-camarillo-sip-deaf-02.txt

Slides presented in:
          pres-camarillo-transcoding-emergency.ppt

Discussion led by Gonzalo Camarillo.

Status: Resuming design team meetings. Transcoding has dependencies which may be isolated in a separate draft.

No conclusion.


Topic: Report from Energency Calling Design Team
----------------------------------------------------


Slides presented in:
          pres-camarillo-transcoding-emergency.ppt

Discussion led by Gonzalo Camarillo

Status: Scenarios doc to appear within a couple of weeks.




Topic: Requirements for SIP Service Configuration
---------------------------------------------------


Note: This has some bearing on 3GPP's use of SIP.

Proposal: Extend config framework to include service info.

Open issue: Which server to contact for data manipulation (XCAP)?

Open issue: Download conference-factory URI for automatic conf creation

No conclusion.



Session 2, Thursday, July 17, 2003,  1300-1500
=====================================

Topic: Agenda and Status
----------------------------------

Review requested for new digest-AKA draft.



Topic: Open Issues from Application Interaction Design Team
---------------------------------------------------------------


Relevant Documents:
         draft-rosenberg-sipping-app-interaction-framework-00
         draft-jennings-sip-app-info-01
         draft-burger-sipping-kpml-02

Slides presented in:
          pres-rosenberg-sipping-day2-ietf-57.ppt

Discussion led by Jonathan Rosenberg

Issues:

KPML DTMF reporting problem
•    INFO ( Cannot work within call dialog)
•    NOTIFY w implicit subscription
•    Explicit subscription from application
•    HTTP
•    New method?
•    MESSAGE

Design team reluctantly chose MESSAGE!

Proposal:
•    Continue to hammer out the open issue
•    Adopt the framework and KPML as SIPPING items.

Conclusions:
a) Hum on “This is interesting work and that we should adopt the framework” – Accepted.
b) Hum on “All who believe KPML should be accepted as WI” – accepted.
c) Chairs will ask the AD to adopt framework and KPML as WG items.
d) The KPML reporting issue will be discussed more on the mailing list.

 

Topic: Early Media
-------------------

Relevant document:
            draft-camarillo-sipping-early-media-02.txt

Discussion led by Gonzalo Camarillo

Status: 
Changes since last rev include clarifying which features are specific to early media and which ones to SIP and other editorial changes.

To be done: Align application server model section with app design team

No conclusions.


Topic: Conveying Tones in SIP
-------------------------------


Relevant document:
          draft-mahy-sipping-tones-00.txt

Discussion led by Rohan Mahy

Issue:  How to provide tones?

Options discussed include
 
   1) In RTP
        a. Speech codec – poor choice
        b. Using audio/tone AVT payload
        c. Using audio/telephone-event AVT payload
 
    2) Referenced externally by URI with Alert-info or message/external

    3) In a SIP header (proposed back in Nov 2000)

    4) In the session description

    5) In SIP body
        a. Use Content-Disposition: Render

If in a body, What body types are available?
    1) Traditional
        a. Wav, au, mp3
    2) Audio/tone
    3) Audio/tone-info+xml
    4) Many more…
    5) XML based:  Audio/tone-info+xml

Discussions:
-Use of midi was suggested by Henning.

Conclusions: The issue on tones will be discussed more on the list.

Topic: Network Announcements
---------------------------------

Relevant document:
    draft-burger-sipping-netann-06

Discussion led by Eric Burger.

Open Issues:

    1) Early media:
        Punt. No definitions for early media
        Makes 487/409 problem go away.

    2) Media on hold
        Punt. Local matter

    3) Multiple media streams
        Punt. Netann is about objects not streams
        Only composite objects for multimedia

    4) VoiceXML keyword without value
        Generate 404 with explanation.

Poll taken for adoption as a working group item indicated little support. However, the work is not seen as being in opposition to chartered work.

Conclusions: The draft will be submitted for publication as an individual contribution.


Topic: Event Filtering and Throttling
------------------------------------


Relevant document:
        draft-niemi-sipping-event-throttle-reqs-01.txt

Slides presnted in:
       pres-niemi-event-throttle-ietf-57.ppt
       pres-niemi-sipping-ietf57-throttles-17072003.ppt

Discussion led by Aki Niemi
       
Changes:
    1) Updated model. A throttle defines minimum time period between two notifications
    2) Updated use cases
    3) Refined requirements
    4) Aligned language with model

Open issues:
    1)  Should use cases be more elaborate?
        Proposal: No
    2) Are Requirements are solid enough?
       Seem to be.
    3) Is scope for work well defined?
        Seem to be.

Discussions:
-    It was suggested that it would be worth noting in the draft about the type of buffering needed (like LIFO, FIFO etc)

Conclusion:  Chairs will recommended to the ADs to adopt this draft as WG item.


Topic: DPNSS to SIP Interworking
------------------------------------

Relevant document:
        draft-mukundan-sipping-dpnss-02.txt

Slides presented in:
          pres-mukundan-sipping-dpnss-ietf-57.ppt
      
Discussion led by Ranjith Mukundan

Difference between DPNSS MIME and QSIG/ISUP MIME
    1) Similar to RFC 3204 (MIME for ISUP/QSIG
    2) Mandates single binary coded octet message length field
    3) Specifies message buffering option
    4) Mandates single DPNSS call per SIP dialog

Discussions:
- Henning indicated that this doesn’t work with the MIME model.
- Some were skeptical about the usefulness of doing this work.
- There was a feeling that the group did not have enough expertise to take on this work.
- Gonzallo indicated that solving the MIME type is reasonable but the translation work is a tough thing to do.

Conclusions: This will NOT be taken as WG item and it will proceed as an individual contribution.


Topic: Discussion of End-to-Middle Security
---------------------------------------------

Relevant document:
        draft-ono-sipping-end-middle-security-00.txt

Slides presented in:
       pres-ono-end2middle-ietf-57.ppt

Discussion led by  Kumiko Ono

Issues: 
    1) End-to-end encryption may conflict with some features provided by intermediaries
    2 )Use cases:
        a. Logging services (IM logging, other logging)
        b. Hotspot service
        c. Connecting to home SIP server via partially trusted proxy
     3) Session-policy
     4) Transcoding

Proposed Mechanism:
•    Allows a UA to disclose data to selected intermediaries
•    End-to-middle encryption uses S/MIME CMS Enveloped data for intermediaries.

Discussions:
-    People felt that this is a very interesting WI.
-    Jonathan R said that we need to add more use cases to explain the problem of addressing data for intermediaries by user-agent.

Conclusions:
    1) There was consensus in the room that work has enough interest in the group.
    2) There was consensus that the requirements on end-to-middle & middle-to-end security should to be taken on as WG item.


Topic: Phone-related Status and Presence
------------------------------------------


Relevant document:
       draft-rosenberg-peterson-simple-pidf-phone

Slides presented in:
       pres-rosenberg-sipping-day2-ietf-57.ppt

Discussion led by Jonathan Rosenberg

Discussions:
-Difference between human presence vs device presence.
- Rohan said that states like dialing/ringing etc are not applicable to device or user; these are particular to a call/dialog. Some of the raw data is useful, but presence may not be right place for it.
- Henning: Information like line-state etc are not too useful as presence information.

Conclusions:
a) Hum on “Dealing with presence issues of phone” as useful work indicated a consensus..
b) Jonathan proposed to develop more use cases for the draft.
c) Is this a SIMPLE or SIPPING activity? Chairs to discuss with SIMPLE chairs.



Topic: Session Diagnostics in SIP
----------------------------------


Relevant document:
    draft-johnston-sipping-rtcp-summary-00.txt

Slides presented in:
       pres-johnston-sipping-rtcp-summary-ietf57.ppt

Discussion led by Alan Johnston
       
Summary:
    1) Delivery of RTCP summary reports to third parties
        a. Logging is main motivation
    2) Three alternatives
        a. Forking RTCP to multiple locations
        b. Carrying in SIP header in BYE
        c. Event package (preferred)

Discussions:

1)  Should RTCP be transferred to third parties?

2) What is the purpose of this? If this is for fault management or performance management, this is not needed. SNMP management tools should be sufficient.

3) How does this compare to RMON?

Conclusions: This will be deferred to this to the list. Alan is asked to investigate RMON work and provide some comparison here.


ACTION ITEMS
-------------------


1) John Elwell volunteered to provide a list of ECMA contacts for QSIG to SIP work o Allison Mankin.

2) Chairs to discuss netconf and XCAP reconciliation proposals with ADs.

3) Francois Audet volunteered to review the qsig2sip document and respons on-list.

4) Francois Audet volunteered to send text relating to alternate approaches such as ALG and MIDCOM for the ICE usage scenarios document.

5) Chairs to discuss moving some ICE work into MMUSIC and SIP with ADs.

6) Chairs to discuss  adding draft-rosenberg-sipping-app-interaction-framework-00 and  draft-burger-sipping-kpml-02 as WG items with ADs.

7) Chairs to work withs AD to add charter item(s) for end-to-middle and middle-to-end security requirements.

8) SIPPING chairs to discuss phone-presence with SIMPLE chairs and ADs to dtermine where thiswork belongs.

9) Alan Johnston was asked to consider how RMON work relates to proposed RTCP summary work and report to the mailing list.