*VidMid-VC Conference Call*
April 22, 2002

*Attendees*

Egon Verharen - SURFnet (chair)
John McNair - U. Tennessee, Knoxville
William Rhodes - U. Tennessee, Knoxville
Karen Krivaa - RADVISION
Nadim El-Khoury - UNC-CH
Michael Gettes - Georgetown
Samir Chatterjee - Claremont Graduate University
Steve Olshansky - Internet2
Ann West - Internet2/EDUCAUSE
Art Vandenberg - Georgia State
Jill Gemmill - U. Alabama, Birmingham
Brent Zionic - U. Colorado, Boulder
Ken Klingenstein - Internet2
Tom Barton - U. Memphis
Doug Sicker - U. Colorado, Boulder

*Discussion*

Ken started off by offering his notes and comments on the Authentication / Authorization subgroup meeting on Tuesday, April 16th. At that time a couple of major questions arose regarding the differences in how to handle inter-realm vs. intra-realm authentication. Art clarified some of the information available on presence detection by saying that individuals may be detected just as devices are with SIP: "I am here on this device at one time, and there on another device at a different time." In this way, the registration process can pertain to an individual or to a device. Egon raised the question of whether the credentials for authentication would be the same for inter-realm as for intra-realm, and noted that the group's original focus was only on inter-realm authentication.

Ken, however, contended that there are two good reasons to discuss authentication in both contexts: 1) it provides ease of use to the user, and 2) it can reduce the exposure of the authentication mechanism to attack. Samir brought up the question of the trust model framework for authenticating users. The main point is that authentication happens through the home security domain, and authorization depends on the policy / business rules for the home domain that authenticates the user / device.

Michael Gettes offered some information on the use of Radius for local single sign-on. If an endpoint / user has been authenticated, and then wishes to utilize a service through a gatekeeper, then the gatekeeper is able to access the Radius server to check whether the endpoint / user has been validated.

Samir hypothesized that a slight modification to the existing Shibboleth design could utilize the Radius system, and Egon mentioned that they are implementing this same design for wireless LAN connectivity in Norway as well as for the SURFnet national dialin service and SURFspot software shop. They are pursuing this together with Uninett, a research network in Norway. However, the English translations of those documents were not written yet, though Egon promised to keep all of the interested parties up to date on their status.

Art added his commentary on the authentication issue as it relates to Resource Discovery by clarifying that authentication to the local domain happens first, and only then does registration and resource discovery occur. The registration involves linking the commObject to an entry in a directory, whether it is an enterprise directory or some other listing. The Access Grid people had talked about the question of encryption, but hadn't said much about authentication. Once authentication is handled, there are several ways to encrypt the datastream.

At the SURA/ViDe workshop Tyler, Egon and Larry will be presenting. Larry will be speaking on the video directory implementation, and Egon will give an update on VidMid-VC. During the upcoming Internet2 Spring Member meeting an open Authentication and Authorization workshop is organized on Monday afternoon. Furthermore, Egon will present on VidMid-VC and Mairead Martin on VidMid-VoD during the middleware track on Tuesday afternoon. VidMid will also be presented during the I2-DV update track on Tuesday morning. Egon will be sending out a URL with a new web interface for registration.

During a draft review call on Monday, April 15th, Tyler and Nadim captured some new feedback and plan to apply the necessary changes to the definitions paper and the white paper in order to have them ready for a final vetting call soon. With the mention of the vetting call, Ken wanted to make it clear where NMI stands: they will not be publishing any of these documents as standards. They are only making available these recommendations and findings to the academic and research communities. He proposed that we add "standards" as a future agenda item so that we can discuss how best to promote our work into the appropriate standards bodies.

*New Action Items*

1. [AI] 22-Apr-02 (Egon): Keep everyone posted on the status and availability of English translation documents detailing integration of local single sign-on and inter-realm authentication schemes for wireless LAN connectivity project in Norway.

2. [AI] 22-Apr-02 (Karen): Look into submitting the H.323 portion of our research to the ITU-T.

3. [AI] 22-Apr-02 (Egon): Send out a URL to the new web interface for registration.

4. [AI] 22-Apr-02 (Tyler and Nadim): Apply the final comments to the definition and white papers for final vetting call.

5. [AI] 22-Apr-02 (SteveO and Art): Send a copy of the Resource Discovery paper to the VC list for final comments.

*Old Action Items*

1. [AI] 22-Apr-02 Mary Fran will ask Terry Dixon and Bob Olson to provide an AccessGrid perspective on security, authentication and authorization.

2. [AI] 22-Apr-02 [TBD] will investigate potential benefit of leveraging Grid MDS v.2 (Meta-computing Directory Services) work on directory of directories (meta-directories).
In progress (est. 8 months ~ 1 year).

3. [AI] 22-Apr-02 Doug will develop scenario: interoperability between H.323 and SIP.
In progress.

4. [AI] 22-Apr-02 Steve and Tim Poe will follow up with vendors not already represented in VidMid-VC,identify and contact key individuals.
To be taken offline.

5. [AI] 22-Apr-02 Jill, Mike and Keith will collaborate on the VideoDoD/Globus investigations.
In progress - came up at the UNC meetings.

6. [AI] TBD Jill and Samir will conduct an Active Directory investigation.
In progress.