IETF 86 - STRAW (Sip Traversal Required for Applications to Work) FRIDAY, March 15, 2013 Friday Afternoon Session I (1120-1220 - Boca 1) Charter: http://datatracker.ietf.org/wg/straw/charter/ Chairs: Christer Holmberg and Victor Pascual Note takers: Keith Drage, Shida Schubert Jabber scribe: Lorenzo Miniero Jabber log: http://www.ietf.org/jabber/logs/straw/2013-03-15.html Meetecho Session Recording: http://ietf86.conf.meetecho.com/index.php/Recorded_Sessions#STRAW Audio Recording: http://www.ietf.org/audio/ietf86/ietf86-boca1-20130315-1120-am1.mp3 Agenda and WG update Presented by: Chairs Slides: http://tools.ietf.org/agenda/86/slides/slides-86-straw-0.pdf ----------------------------------------------- Role-types of B2BUA (http://tools.ietf.org/html/draft-ietf-straw-b2bua-taxonomy-01) Chair : Will extend the WGLC for another 2 weeks. hop-by-hop media based traceroute (http://tools.ietf.org/id/draft-kaplan-straw-sip-traceroute-00.txt) Chair : It has been suggested to remove the milestone Hadriel : Will explain why I suggested to remove the milestone. Loop detection/prevention Draft : draft-kaplan-straw-b2bua-loop-detection-01 Presented by Hadriel Kaplan Slides: http://tools.ietf.org/agenda/86/slides/slides-86-straw-1.pdf ----------------------------------------------- Andrew : Why not just decrement the Max-Forward. Hadriel : Many just reset for topology-hiding, security reason OR claiming B2BUA starting a complete new session thus should have fresh Max-forward value. Paul : What about when it across TDM? Robert : No way to solve the TDM case. Robert : Are you going to implement Max-Breadth? Hadriel : Good question.. Robert : Think this will help.. Great! Have seen some implementations that implement Max-Breadth. Hadriel : Some B2BUA is like having TDM inside, they will potentially benefit from this. Robert : It is the best we can do. Chair : Will ask on the list whether to adopt this draft as a WG item as a baseline for the milestone. Hadriel : Can we get a consensus in the room? Chair : Is anybody going to propose another draft addressing this milestone? - Silence Chair : Does anyone object to adopt this draft as a baseline? - No objection. Final call to list. Why reomve media traceroute (Wasn't in the original agenda) Draft : draft-kaplan-straw-b2bua-sip-traceroute-00 Presented by Hadriel Kaplan Slides: http://tools.ietf.org/agenda/86/slides/slides-86-straw-2.pdf ----------------------------------------------- Point about "Indistinguishable from loops" Paul : If the goal is to get a media loopback Sending an offer-less INVITE is not a way to do it. Hadriel : Can't use offerless invite There are vendors their default behavior is to send offerless INVITE. Paul : How do you do it? Hadriel : Needs to be a SIP header (b2bua header). Problem exists never the less. Paul : If there is a loop in the path then no need for media loopback. Hadriel : It's another toolkit to debug. Not testing loop but checking the media path. Hadriel : Can't use the same header if you try to use the header causing the problem (max-forward). Hadriel : SDP will have media-loopback but other than that it looks like the previous INVITE. Paul : If the new header is needed for detecting this then what happens if the entity on the path doesn't support this. Hadriel : There is no 100% coverage, no matter what we do. If use Max-Forward, the mocking of Max-Forward Hadriel : What should we do as a next step? Chair : Even if we use Max-Forward, it will be important to document the uses. Not here to define new header. Robert : If you use the Max-Forward as presented, it will run into serious semantics issue. If the entity is authoritative enough I guess it can work. Hadriel : As a B2BUA it is authoritative to answer Max-Forward = 0 as a UAS. Robert : Splitting the semantic here but.. It's probably going to be okay. Add text why it's okay. Go forward with the mechanism proposed. Martin : If use Max-Forward will not solve issues Hadriel : Based on the assumption that previous draft will be implemented. Understand that you do whatever you decide to do. Andrew : What about 3pcc? Hadriel : Media loopback call just doesn't work in that case. (3pcc holding onto the SDP in INVITE and send the offerless INVITE to the target). Paul : They can decide to send SDP forward. Chair : Still think that it should document even with the suggested approach. Hadriel : Happy with working on it. Just thought didn't think WG document was necessary. Robert : Still believe that it should be an IETF stream draft. Hadriel : Want to hear what approach people prefer. - Use max-forward? 11 - Use b2bua-hop? 0 Chair : Will confirm on the list. Chairs would like a revision of individual document to see agreed solution. When resubmitted with this chairs will ask for adoption on mailing list Next steps Presented by Chairs ----------------------------------------------- No debates.