Subject: Meeting minutes Participants: Jun Tatemura, Hiro Kishimoto, Chris Dabrowski, Kojo Tadashi, Jem Treadwell, Stuart Schaefer, Steve Loughran, Andrea Westerinen, and Dejan Milojicic - Andrea's discussion on DMTF liaising with CDDLM Put together a summary of the relationship with DMTF applications, does not think that there are any conflicts, based on the SF, XML schema or a quick scan of foundational document these two WG complement each other. + Short plan of actions. Need to update the work register and will call the collaboration with. Just wait with the utility computing WG, it gathers steam, but it aligns more generally with OGSA to manage general services and provide the details behind the underlying services. When it comes to config and deployment, we are more closely aligned with applications WG. More of a wait and see. Utility will work more on the underlying resources (NW, computing, storage,...) that service and apps need. Any app is service in DMTF. For the time being rely on DMTF App group to interact with UC. Hiro: very interested in the utility computing WG. They are going to define the model of resources and they are very close to the WSDM or CMM WG. Andrea agrees. Andrea asked UC WG if they can help OGSA not only in terms of services, but also in managing services. WSDM was always model-agnostic, defines operations behind the distributed mgmt, but semantics are open. DMTF has semantics in CIM, we need to make sure that the two work together, so that semantics works for them. Dejan: DMTF takes care of resources, OASIS about schemas and GGF about implementations. Andrea, DMTF about mgmt semantics, GGF about Grid service functionality and higher level services and implementations and OASIS about schemas and mgmt infrastructure. E.g. file systems, DMTF will not tell you how to use a file system, GGF will address it as services and OASIS will offer mgmt perspective. Steve and Andrea follow-up, the three standard bodies overlap in mgmt. Mgmt most important after function. The other issue is scalability. Multi-vendor. Profiles for achieving interoperability. Steve and Andrea will follow-up by email, we had to switch to the provisioning topic. - Jem & Hiro provisioning A few comments on Jem's proposal. One person from the OGSA said we add the resource in front of every functionality, Steve said we need to add undeployment as well. Hiro: we need to add much more functionality which is broader. Hiro sent the rough draft from the OGSA spec. Jem agrees to expand the definition. Chris: Is discovery part of the provisioning. Hiro: yes and know. Some services keep resources under the system, but discovery can be used by provisioning. Jem, this is all part of the glossary, and it needs to be sorted out. - Update on the submissions of the Foundation spec and SF-based language spec Dejan: gave a lot of thought to whether to submit the document or not. The summary was we go with it. Dejan will add two figures, get some input from Chris Dabrowski an put the new document not eh Web and announce a week for the last call on the mailing list. Steve: proposed 3 things. Minor changes to language spec to follow up on the comments at the GGF meeting. Wrote Grid-specific example, added about security and took out language related specifics that could be looked as common across representation and implementation. "Here are the attributes that are common". This should be a normative document on what stuff is required to be supported. Dejan: what to do with the run-time binding spec, does language spec depend on it? There are also implementation specific, which will come later. Schemas are different, we need to do a core minimum. Steve: give Patrick a chance to review pool out run-time binding. Will have the "final" spec available for the group by the end of the next week. Jun and Stuart will work with Steve on the run-time binding spec. We will have this document submitted after we have submitted the XML based spec. - Update on the progress of the basic services and component model specs Kojo: not much progress last week, but doing the liaison with PE, second week of May. In Chicago will be F2F meeting. The meeting will be for adjusting the I/F outside of the deployment. Kojo feels comfortable to have the document at the stage that Jun had XML-based spec at GGF10. Dejan: Can Kojo propose to Andrew that we do a joint demo in September. Stuart: working on 2 activities: a) relationship to CMM & WSDM, went through the documentation, will be finished this week and b) From the component model specification, he will send the draft next week, tries to sync with new CDL spec and WSRF. - Update of the XML-based language spec Jun is preparing the language example. This will be proposed by the next meeting. Dejan: The time is short, we need the final document in mid-May for the submission to GGF11. Jun will submit the final document either in end of April or beginning of May. Instead of submitting the intermediate versions, he will post the discussion on various topics. Steve would like to get the copy of the schema. - Update on the plans for the workshop at Brussels (we need to create a handout for the advertisement and plans for the demos well ahead of time) Dejan: form a small program committee that can work offline and make some agenda. Kojo and Dejan will lead this. Chris: NIST can possibly contribute some of the demo, will check. - Update from liaisons to related working and research groups. - Various (a possibility to have DMTF Applications co-chair participate- still working on this). Other meeting on Thursday 4/22, again at 6am.