*VidMid Conference Call*
June 25, 2001

*Discussion*

Ken opened the call by soliciting comments on the progress of the group to date. The members were pleased with the amount that had been accomplished in a difficult set of topics with a great number of groups with conflicting interests. Ken mentioned a desire to begin to branch out technically with an eventual aim of pushing a product out the other end of the working group. It also seemed important to stay away from proprietary and corporate secrecy and reveal all information possible in the group's academic interest. The minutes from the last call featured a revision in that there was no agreement to boldly move forward with a focus on H.323 videoconferencing, as the importance of video on demand is no less. At the same time, a discussion of codecs, cards, and other elementary matters should be avoided as they are more "lower-layer" than the group's scope. It seems reasonable to contact Ted to ensure that the Internet2 Commons is a more appropriate place to discuss codecs to ensure that the specific hardware, encoding algorithms and cards are properly addressed at some point.

As the group works to build consensus and organize more support, it will be important to cull more participants from varied organizations and fields to help with development. Todd Needham has been asked repeatedly to suggest a possible Microsoft participant. Ken will also contact Philippe Galvez about joining the group. The University of Glasgow has developed a Dublin Core metadata schema which seems functional, and the European contingent may have better developed current implementations of video middleware than the Americans. It will be important to call them in.

The group also modified its format to "push forward" action items assigned but not completed in previous conference calls to ensure a better continuity and coverage of the group's functions. Ted has sent in his "why you need video out-takes" document, but Ken has yet to write anything about why video needs middleware. Ken will integrate Ted's inputs with his own to push out a one-page document to be placed on the web site explaining the group's purpose and necessity. With that reassigned item, the rest of the action items for the last week were closed off.

- video and metadirectory schemas -

There have been several proprietary systems designed to date to address the problem of culling usernames, passwords, and meta-information from an LDAP directory for both videoconferencing and VoD. Most are workarounds rather than true middleware implementations. The metadata and access controls are both of interest to the group, as they are things that could be solved using middleware. One corporate effort tells the UNIX server to reference the LDAP server to acquire a list of usernames and passwords rather than default to a UNIX login service. The need for a more comprehensive design will need to be demonstrated as the other problems looming in the video space are more pressing for many content providers. A middleware authentication and authorization infrastructure could potentially reach far beyond what could be done with this makeshift system.

Other institutions have begun to try to build schemas for content they host. These schemas need to be intelligible to both humans and machines, making for easy searches of, and references to, large repositories of video and audio information. Some current commercial schemas use over a thousand fields to allow for sufficiently detailed information about feeds to be retained. Ensuring interoperability and ease of use and configuration with this many fields would be difficult. The group would like to develop a set of metadata rich enough to satisfy end viewers, archivers and producers, while creating a standard mechanism for interoperability such that queries could someday cross institutional boundaries with ease.

- SIP, H.323 and commercial products -

The relationship between SIP, H.323, MGCP, and other standards can seem extremely murky to the uninitiated, and possibly even to the evangelists of the group. Karen wrote up a presentation to begin to discuss the inner workings of SIP, but there is currently nothing published that compares the standards well. A reference to help the uneducated begin to understand the distinctions between these protocols would be a welcome document. At the same time, such a document would better highlight the differences between the protocols that could potentially cause conflicts that would reduce the ability of these standards to interact. More important than merely highlighting the differences is to discover ways to make these different standards function seamlessly together and to ensure interoperable implementation in commercial products and designs. This provides another potential area in which to focus VidMid's efforts.

- scenarios -

Most working groups progress beyond initial stages by selecting a set of scenarios around which they will develop their recommendations and/or products. The VidMid group has reached a point where it will form these scenarios so that it can craft solutions that will work in all the necessary implementations. The three primary categories that may be considered for these scenarios are metadata, access controls, and codecs, though the last one is beyond the scope of this group and these scenarios. After representative scenarios have been developed to demonstrate the full realm of possibilities needed for each category, the group will proceed to develop software and implementations that will satisfy all potential needs.

Three scenarios have been proposed to date. The first would describe the use of proper middleware to implement appropriate access controls to video information. This would entail both fine- and course-grained definitions, allowing permissions to be issued across the spectrum from broad groups such as "all members of Internet2" to individuals. The second proposed scenario would describe the sorts of directory schema necessary to provide adequate archiving and access to information. The system would have to allow both humans and computers through traditional directory means to quickly determine the movies they are interested in retrieving. A third scenario would integrate real-time video with streaming video and a real-time capture of this information. An example given during the call was that of a security camera needing to both display real-time information and be able to quickly archive and recall specific events, often in time-critical applications.

Submissions of detailed versions of these scenarios as well as any others that might be useful for the group are welcome, between one and three paragraphs depending on the complexity of the situation. It is important to purify these scenarios into desired functionality so that the requirements can then be extracted from them. During the next phone call, the group will finalize a set of four scenarios to characterize the most urgent issues. The other major agenda item for the call is to look at call setup issues for SIP and H.323, as well as the interworkings between them.

*Action Items*

[AI] Ken will integrate Ted's inputs with his own and push out a document for why we need middleware, to be displayed on the website.
[AI] Ken will contact Philippe Galvez about joining the group.
[AI] Tyler will attempt to contact the appropriate vendors to have on the calls client-side, gatekeeper, and possibly open source representatives. He'll also talk with V-Con about their backend product.
[AI] Jim DeRoest will send out a URL describing his activities, while Philippe will send out a URL detailing the ongoings at the University of Glasgow. The group should contact colleagues in the UK to open communications.
[AI] Ken will ask Ted to ensure that the Internet2 Commons is a place where codecs can be discussed down to the hardware level.
[AI] Jim Whitlock will come up with a two-paragraph description of how he would use the capability to use real-time video alongside streaming in a manner not unlike TiVo for Internet video. Markus will be asked to write up a scenario using this capability for security cameras.
[AI] The group solicited access scenarios from various colleges, such as UW and early adopters in New York.
[AI] The group needs the URL for the Dublin Core metadata on video extensions.
[AI] Ken will track down what he did with the SIP slides.
[AI] Ken will add Dave Nash to the list.