Present: Vesselin Novov Chris Smith Steven Newhouse Marty Humphrey Glenn Wasson Blair Dillaway Andrew Grimshaw Minutes 1. OGF IPR Notice Notified. 2. Agenda Bashing No additions. 3. Experience document Version mailed out yesterday. Incorporates previous feedback. Two testing matrices for username/password and X509 certificates. Once we have accurate information collapse these matrices down to show active testing. Action: All to provide updates on the testing they have done for X509 and username/password. Action: All check details of the project URL and details in the table. Action: Bill to STILL provide details on Altair activity. Action: Submission timeline - we want to submit before Christmas. All updates by COB Monday. Final call to the mailing list for a week. Submit to the editor after Christmas. Enter GFSG 15 day review before discussion on call mid-January. Action: Marty/Glenn to refactor section 4 issues into spec and hosting environment issues. 4. HPCP - What next? Discussion points: * File staging (already in process) Underway. Need to bring it to a conclusion. * Kerberos credentials (with Active Directory support) Common in DoE and DoD environments. How can they be used to authenticate to the service and within the file staging. Use the WS-Security Kerberos scheme to pass the credential. Still use server side SSL to pass the message? In a pure Kerberos environment may want to use that as the server authentication token. * Application inventory (a simple list of available applications, maybe leveraging the Glue Scheme work) Do we need a 'standardised' catalogue? How can we describe (in an abstract way) that an application is accessible through a BES factory so that a client can produce a JSDL document that will execute on the platform. Are the application and machine inventory linked? Consider looking at GLUE to see what they are doing. Use case 1: Query the BES to populate a drop down list on a portal for a user to select. Primary goal... provide sufficient information to build JSDL. * Machine inventory (a list of hardware: hosts, cpus, memory, etc, as well as what's in use and what's free; maybe also leveraging the Glue Scheme) Need to not only know how many CPUs on a machine... but need to also know sockets, cores, threads. * Resource usage data (for simple accounting and status reporting) Provide updates during a run and afterward. Add in a new operation or put in an extension to the ActivityDocument operation that returns a Usage Record detail. Other areas: * Application Deployment. Treat as a separate action so it can be cached between job executions. Action: Chris - Session at OGF22 on Kerberos token. Blair at OGF22 Monday/Tuesday. Action: Chris - Use of GLUE to describe applications/machines. Session at OGF22. Also at OGSA F2F? Get Ellen involved. Action: Steven - FileStaging at OGF22. Action: Andrew - Application Deployment at OGF22. 5. Any Other Business DONM - Thursday 20th December, 10.30 Pacific.