VidMid Conference Call, June 11, 2001

*VidMid Conference Call*
June 11, 2001

*Attendees*

Ken Klingenstein (chair) - Colorado/Internet2
Mary Trauner - Georgia Tech
Karen Krivaa - RADVISION
Tyler Johnson - UNC Chapel Hill
Renee Frost - Michigan/Internet2
Ellen Vaughan - Internet2
Lisa Hogeboom - Internet2
Ted Hanss - Internet2
Michael Gettes - Georgetown
Art Vandenberg - Georgia State
Jill Gemmill - ViDe/Alabama at Birmingham
Egon Verharen - SURFnet
Mary Fran Yafchak - SURA
John McNair - Tennessee at Knoxville
Bob Morgan - Washington
Amy Philipson - Washington
Ben Roy - Washington
Ben Chinowsky (scribe) - Internet2

*Discussion*

After approving the minutes, the group reviewed the action items from the previous call.

[Ken will negotiate a new conference call schedule via the VidMid list, planning on a maximum call length of 75 minutes.]
Done. VidMid calls will take place every other Monday at 1600 UTC, which is noon EDT and 9:00am PDT.

[Internet2 staff will review the current Internet2 glossary, and consider a separate glossary for VidMid.]
Ongoing. It was agreed that to the extent VidMid can meet its glossary needs by pointing to existing glossaries rather than creating new ones, it should do so. [AI] All who have URLs for glossaries of VidMid- related terms will send them to Ben for posting on the VidMid web site.

[Internet2 staff will look at options for audio capture of the VidMid calls.]
Recording of the calls will begin after Internet2 moves to a new conference-call provider sometime in the next few months.

[Over the next few months, Internet2 staff will survey the videoconferencing capabilities of VidMid participants, with a view to holding at least some of the VidMid meetings via videoconference.]
This will take place later this summer.

[Ted will try again to contact Todd Needham.]
Ted still hasn't heard from Todd; no one else has heard from Microsoft lately either. Karen noted that her colleagues at RADVISION have been finding lots of bugs in Windows XP.

[Ken will reply to Bob Dixon's request for VidMid promotional materials, suggesting that he use the VidMid web site, which will soon feature an expanded VidMid mission statement.]
[AI] Ken will tell Bob Dixon that the newly-reorganized VidMid site is ready for him to use in his pitch to Polycom.

[Ted will send Ken the why-you-need-video outtakes from his EDUCAUSE article, to form the basis of an expanded VidMid mission statement.]
Ted is working on his piece.

[Ken will add a why-video-needs-middleware section to Ted's material.]
Ken has completed a first draft of his piece.

[Ben will edit Ted and Ken's VidMid mission statement and send it to the list for review.]
Ben is waiting for Ted's and Ken's pieces.

[Ben will update the web site, a) creating separate Overview, Standards, and Technical sections, b) adding Ted's and Bob Dixon's EDUCAUSE articles to the Overview section, and c) structuring the Standards and Technical sections by organization.]
Done.

[Ted will produce an overview of the Internet2 Commons effort.]
Still to do.

[All will review Tyler's "Technical Directions" email, for in-depth discussion on the next call.]
See below.

[Ken will send Ben more SIP URLs for the site.]
Still to do.

[Egon will get David Chadwick's comments on eduPerson forwarded to the VidMid list.]
[Ken will forward David Chadwick's comments to MACE-Dir and others to whom they may be of interest.]
Chadwick's comments have been forwarded to both lists.

Ken noted that he is still looking for someone to take his place as chair of VidMid.

The group discussed what procedure to use for getting documents from the VidMid list to the web site. [AI] Renee and Lisa will find out if the VidMid list is being archived, and evaluate whether or not it should be. It was agreed that Ben should assume that any document sent to the list can be posted to the site, but that he will notify the document author(s) before doing so. [AI] In the draft documents section of the VidMid web site, Ben will a) add an explanation of the status of these documents, and b) list the date each was posted. There was a short discussion of the need to ensure that correct credit is given; Bob noted that he is working on revisions to the middleware document standards (see "Document Guidelines" on http://middleware.internet2.edu/) to address this and other issues.

The rest of the call was devoted to issues arising from Tyler's draft technical recommendations (see http://middleware.internet2.edu/video/). Tyler singled out two items in particular as being of "critical import" for interoperability: 1) data elements to put in eduPerson: most importantly the H.323 URL, and also an object that identifies an endpoint in a private dialing plan; and 2) getting this information into the gatekeepers (H.323 Annex O is of central importance here). Tyler stressed the quick-win potential of these two items taken together: if we put this information in eduPerson and partner with RADVISION to get it into the gatekeepers, "we really have something". Ken asked if putting this information in eduPerson would require that the data reside in the same enterprise directory that holds other eduPerson data; it was agreed it would not. Bob noted that how objects are described and where the data describing them resides are completely separate issues; there are multiple tradeoffs -- largely having to do with maintainability -- that will determine whether eduPerson data should be all stored in one big directory or parcelled out among application-specific directories.

Art asked if H.323 addresses were to be associated with people or with places, and this sparked a broader discussion of processes around directory usage. Comparisons were drawn between H.323 endpoints and telephones and email mailboxes. Bob stressed the importance of giving early attention to directory-process issues: "a really common problem with directory attributes is that people think about data without giving sufficient thought to process". There was a discussion of whether all calls, including point-to-point calls, should be under gatekeeper control; Tyler argued against putting much effort into non-gatekeeper scenarios, noting that "we're trying to stamp out" direct connections, in hopes of "bringing some order to the chaos". The current version of the H.323 standard recommends, but does not require, that calls go through gatekeepers. Ken and Tyler raised the issue of whether H.235 is likely to work well with existing authentication systems. Karen noted that H.235 requires implementation by the codecs, which isn't available yet; while it will be available in RADVISION's gatekeeper, Polycom's gatekeepers do things differently, so there's not likely to be much interest in promoting this for interoperability. Tyler argued for pursuing an authentication solution in parallel with H.323 identifier issues, stressing that lack of authentication is the principal obstacle to viable business models for videoconferencing, and thus the principal obstacle to deployment.

The group discussed the relative merits of H.323, SIP, and RTSP for call setup. Karen noted that it's not clear whether or not SIP will eventually overcome H.323's dominance in the videoconferencing world; she expects the picture to become clearer over the next six months. SIP is generally considered much more of a sure thing for VoIP. Tyler noted that while SIP is better call-setup technology than H.323, "we owe it to ourselves to make an H.323 environment work, even if it doesn't work well", and compared H.323 call setup to cable modems: it will get people used to having a capability, but also breed frustration with current technology, spurring deployment of improvements such as SIP. There was general approval of this "use H.323 and learn" approach. [AI] Karen will look through her presentations for comparisons of SIP and H.323 call setup; on the same subject, Ted recommended the book *Understanding the Session Initiation Protocol* by Alan B. Johnston. Egon identified RTSP as the standard for VoD call setup; see http://www.ietf.org/rfc/rfc2326.txt. [AI] Ken will make a discussion of H.323, SIP, and RTSP the main agenda item for the next call; the focus will be on understanding what information directories need to make available in order to support these protocols. [AI] All who have references to information on call setup processes for H.323, SIP, and/or RTSP, will send them to Ben, who will post them on the VidMid web site.

Ken summed up by saying that the group's immediate focus should be on the H.323-identifiers-and-authentication and VoD scenarios, and there was general agreement.

*Action Items*

[AI] All who have URLs for glossaries of VidMid- related terms will send them to Ben for posting on the VidMid web site.
[AI] Ken will tell Bob Dixon that the newly-reorganized VidMid site is ready for him to use in his pitch to Polycom.
[AI] Renee and Lisa will find out if the VidMid list is being archived, and evaluate whether or not it should be.
[AI] In the draft documents section of the VidMid web site, Ben will a) add an explanation of the status of these documents, and b) list the date each was posted.
[AI] Karen will look through her presentations for comparisons of SIP and H.323 call setup.
[AI] Ken will make a discussion of H.323, SIP, and RTSP the main agenda item for the next call; the focus will be on understanding what information directories need to make available in order to support these protocols.
[AI] All who have references to information on call setup processes for H.323, SIP, and/or RTSP, will send them to Ben, who will post them on the VidMid web site.