*VidMid-VoD Conference Call*
October 19, 2001

*Attendees*

Mairead Martin (chair) - Tennessee
Steve Olshansky - Internet2
Lisa Hogeboom - Internet2
Bob Morgan - Washington
Samir Chatterjee - Claremont Graduate U.
Ted Hanss - Internet2
Grace Agnew - Georgia Tech
Phil Galanter - NYU
Renee Frost - Michigan/Internet2
Mary Trauner - Georgia Tech
Jill Gemmill - ViDe/Alabama at Birmingham
Ken Klingenstein - Colorado/Internet2
Ben Chinowsky (scribe) - Internet2

*Discussion*

The meeting opened with a discussion of relationships among NMI, VidMid-VoD, VidMid-VC, and Shibboleth. Ken noted that most of NMI's resources in its first year are going into directories, driven primarily by use scenarios from VidMid-VC and the Grid; he expects NMI to be less important to VidMid-VoD than to VidMid-VC. On the other hand, Ken pointed out, Shibboleth offers great opportunities to VidMid-VoD for work on access control; he strongly urged the group to orient to the Shibboleth architecture (http://middleware.internet2.edu/shibboleth/), as content providers interested in this architecture have already been identified, and "VoD is about providing content." Shibboleth pilots are planned for as early as January.

Mairead noted that the group is behind schedule in scenarios development, and asked for help. [AI] Mairead will collate and prioritize the existing VidMid-VoD scenarios. [AI] Steve will find out if any of the VidMid scenarios involve both videoconferencing and video on demand. [AI] Phil will send Mairead new VoD scenarios. Ken noted that a lesson learned from the development of Shibboleth is "engineer nothing out, but code very little in", and suggested that the group look at the scenarios in this light. He also noted that while the Shibboleth group correctly made many fine conceptual distinctions among the various components of the architecture, in the first implementation many of these components will be combined.

Mairead opened the projects-and-deliverables portion of the meeting by stressing the importance of digital rights metadata work, and pointed out that there will be fruitful overlap with other groups' work on this, in particular ViDe's. It has been suggested that VidMid-VoD develop a digital video portal; however, Ken pointed out that portals create difficult security problems -- a portal is a "man in the middle," which makes it hard for an infrastructure based on portals to defend against a man-in-the-middle attack. [See http://www.sans.org/infosecFAQ/threats/middle.htm for an explanation of this attack.] Ken suggested that VidMid-VoD focus instead on developing a policy decision point (PDP) that it can "slide into Shib". Phil noted that there is "a hunger for people to see something concrete" driving demand for a portal, and Ken replied that developing a Shibboleth-compatible rights-management PDP, and using it in a pilot with video resources, has great potential for satisfying this hunger. Bob agreed; [AI] Bob will either write a concise infrastructure model for a rights-management PDP or get someone else to do this. Ken listed a few questions that would need to be answered in the course of this effort. Do we need a PDP at the origin, e.g., to control congestion by doing access control based on the time of day? Where will the PDP get the information it needs to make its decisions? Will the PDP need to gather information from several places, and if so, how? Phil said that in pursuing this project, it would be more important to have a showcase implementation that faculty could use to encourage universities to put more resources into similar projects, than to have something that will itself satisfy all requirements for digital video rights management. There was general agreement that VidMid-VoD should work on infrastructure for digital video rights management.

Next the group discussed the many formats being developed to store and share digital rights metadata. Phil asked if video rights are being looked at as a "subset of other rights", and Grace replied "yes, absolutely", stressing the need for a unified digital rights metadata standard that will encompass text, images and audio as well as video -- "a Dublin Core for rights metadata." Grace noted that she is "more a metadata person than a video person" and that the ViDe metadata group she's working with, catalyzed by the recent Managing Digital Video Content workshop (http://www.vide.net/conferences/), will pursue a broad approach to rights metadata -- "for once video won't be the stepchild that gets attached to imaging." Grace noted that image vendors are just starting to realize that the proprietary rights management protocols they now use aren't going to work over the long haul; more generally, she observed that commercial entities tend to be focused on short term needs -- "pay the money or you don't get it" -- and not on collaboration. Grace noted that Real's metadata language is strictly for the Real video format and strictly commercial; ODRL and XRML also have a strong commercial focus. Bob suggested XACML as a possibly more hopeful alternative; it's being worked on in OASIS and is explicitly intended to build on SAML, and it appears to have attracted a lot of interest from digital rights management people. Grace observed that there is a need to survey the various protocols available, and that whichever protocol(s) VidMid-VoD adopts, it will need to map to those it has rejected. Ken suggested a division of labor between ViDe and VidMid-VoD: ViDe should "pick a winner among the protocols and develop mappings among them"; VidMid-VoD should focus on how to gather metadata and use it for authorization. Mairead expressed agreement with this formulation, while again stressing that there will need to be a lot of active cooperation between the two groups.

There was a short discussion of other things VidMid-VoD might want to work on further down the road, including multicast, QoS, and caching services. Samir noted that the September issue of IEEE Computer (http://www.computer.org/computer/co2001/r9toc.htm) focuses on VoD.

Ken asked if anyone knows of a VoD repository that has straightforward rights for higher education use and that VidMid-VoD might be able to use in a pilot deployment; ResearchChannel and the Variations project at Indiana University were suggested. Ted suggested using video from the Survivors of the Shoah Visual History Foundation (VHF); they have non-commercial reasons for needing digital rights management, e.g., some of their material is too horrific for small children, but faculty need to be able to access it. [AI] Ted will contact Sam Gustman about using VHF content in a digital rights management pilot project. [AI] Mairead will look for other video resources that VidMid-VoD could use as content in a digital rights management pilot project. There was a short discussion of the Ex Libris library management system; Bob noted the connection of Ex Libris with SFX and OpenURL technologies, and Grace expressed interest in looking into these systems in more detail.

[AI] Mairead and Steve will summarize VidMid-VoD's planned projects and get feedback from the list. VidMid-VoD agreed to a call schedule of every other Wednesday from 12:00 noon to 1:30 pm EST (1700-1830 UTC), starting November 7.

*Action Items*

[AI] 19-October-2001 - Mairead will collate and prioritize the existing VidMid-VoD scenarios.
[AI] 19-October-2001 - Steve will find out if any of the VidMid scenarios involve both videoconferencing and video on demand.
[AI] 19-October-2001 - Phil will send Mairead new VoD scenarios.
[AI] 19-October-2001 - Bob will either write a concise infrastructure model for a rights-management PDP or get someone else to do this.
[AI] 19-October-2001 - Ted will contact Sam Gustman about using VHF content in a digital rights management pilot project.
[AI] 19-October-2001 - Mairead will look for other video resources that VidMid-VoD could use as content in a digital rights management pilot project.
[AI] 19-October-2001 - Mairead and Steve will summarize VidMid-VoD's planned projects and get feedback from the list.