Minutes from Mediactrl Virtual Interim Meeting June 9, 2009 Working Group Home Page Chairs Eric Burger Spencer Dawkins Attendees Adnan Saleem Dan Romascanu Eric Berger Gary Munson Keith Drage Lorenzo Miniero Mary Barnes Spencer Dawkins Scott McGlashan Victor Paulsamy Pre-Agenda No changes to the proposed agenda… AD Review of Framework Current version is http://tools.ietf.org/html/draft-ietf-mediactrl-sip-control-framework-10. Robert Sparks (new Area Director) has provided AD Review comments (available at http://www.ietf.org/mail-archive/web/mediactrl/current/msg01255.html). These comments will require a new version of the draft, but Chris said he’s OK with the comments and can provide an update in a week or so. Update on MIXER Current version is http://tools.ietf.org/html/draft-ietf-mediactrl-mixer-control-package-07. Robert Sparks is actively working on the AD Review for IVR, so we’re going to wait until end-of-the-week to see those comments, before requesting MIXER publication, in case there are things common to both drafts that need to be addressed. MRB draft discussion Current version is http://tools.ietf.org/html/draft-ietf-mediactrl-mrb-00. MRB draft now reflects SFO discussion. A lot of mailing list discussion, we’re going topic-by-topic from recent e-mail threads. MRB-PUB-1 Active RTP sessions Want to distinguish between encode and decode directions. MRB wants to know how much MS capacity is left – number of sessions left? Percent left? Hard to do math, because load may be very dependent on specific sessions (hi-def vs. 3GPP speeds for same codec). Has the MRB got a way of calculating how full the MS is? No, but the MS should be the one figuring this out, anyway – is THAT even possible? Possibly establish a small numbers of thresholds? We’re not reserving resources in MRB – you’d need to do MIXER to figure out if you can add a 100-port conference or not. We’re thinking that the more information the MRB has, the better job it can do in avoiding out-of-resources situations that the AS encounters. Need a strawman for a profile. Codec names are strings in the schema, but there is a registry for them. MRB-PUB-2 Active Mixer Sessions Most of previous discussion about capacity applies to this topic as well. Gary asked about not reporting conference IDs due to privacy considerations. MRB may have this information already, but if it doesn’t, it shouldn’t learn the conference IDs through this interface? Is there a use case that requires this? Can audit through the MIXER package if you need the information. Would anonymizing the conference IDs work? Real conference IDs would be more useful for troubleshooting. Scott noted that the packages allow for trusted entities, but don’t require them. MRB-PUB-3 Non Active RTP Sessions Most of previous discussion about capacity applies to this topic as well. MSes will provide estimates of remaining capacity for each media type, MRBs may apply heuristics about how to mix-and-match this capacity information. MRB-PUB-4 MS Uptime Rather than uptime since entering service, isn’t it sufficient to just tell MRBs when MSes enter and leave service? MRB-PUB-5 Codecs Supported What about codecs tied to specific packages? Lorenzo thought capabilities would be for any package – licensing considerations are one reason why support would depend on the package being used. We would add package information to statements of codec support. MRB-PUB-6 Application Data This is pretty conceptual – authors need guidance from the group here. Would be good to include file formats (not just RTP-to-RTP endpoints). AMR, as one example. Gary gave two example use cases – operator services with attendants with different capabilities, or carrier-assigned categories. We can’t standardize these, but we can transport the unstandardized contents. Alternative is via backend/operations support system. Is the package capability auditing mechanism sufficient to meet these needs? Do package audit mechanisms overlap the publishing interface functionality? Do we think MRBs will have control channels to MSes? We need to have a clear rationale for why we have both mechanisms. MRB-PUB-7 File Formats Do we need both reading and writing? This is covered in detail in the IVR audit package – do we need any more? What happens if you say you support a file format, but don’t support IVR package? Need to be clear about the level of coupling here. MRB-PUB-8 Maximum Announcement Again, we need to be clear about the difference between the Publish interface and the package audit mechanism. Need to be clear about whether this means one time through the announcement, or including restarting it, etc. MRB-PUB-9 Variable Announcements Same discussion as previous – need to establish relationship with packages. MRB-PUB-10 DTMF Detection and Generation Support Provide some capabilities that the packages do not. If publishing interface mirrors package audits, need to address grammars for DTMF as well. MRB-PUB-11 Mixing Type Need to provide indication of panel size here. Summary Will do more on Publishing interface before we work on the Consumer interface – need continued discussion on the list. Gary asked that we continue the discovery process about the way we see the MRB working – Gary will suggest a time on the mailing list.