OGSA Teleconference - 7 November 2005 ===================================== * Participants Pete Ziu (Northorp Grumman) Jay Unger (IBM) Jem Treadwell (HP) Frank Siebenlist (ANL) Andreas Savva (Fujitsu) Darren Pulsipher (Xango) Takuya Mori (NEC) Tom Maguire (EMC) Chris Jordan (SDSC) Hiro Kishimoto (Fujitsu) Mike Behrens (R2AD, LLC) Apologies: Ellen Stokes, Fred Maciel Minutes: Andreas Savva * November 2 minutes approved with no changes * Profile Definition 1.0 update - Public comment has finished. There are a small number of comments but nothing major. - The GGF Editor has not officially returned the pen to OGSA-WG. The group will start preparing a new version. - Jem will take the pen and make an initial draft. - Tom will take the pen after Jem. * EMS, CDDLM, GRAAP joint call preparation ** CDDLM discussion preparation Reviewed Dejan's meeting notes from October 12 in preparation for the next call. - No issue with communication between CDDLM and ACS via data services (option 2) but eventually details of this interaction should be filled in. - ACS does not place restrictions on the application data it stores. It is up to EMS(CSG) to decide what is where. Action: Hiro will talk with the RSS-WG (owners of the CSG spec) and ask them to outline in a memo what information they expect to have and where it may be coming from (ACS, Information Services, etc.) - [And match that against what the ACS spec already defines.] - Estimate of deployment time - Agreed that sophisticated time estimation is not a high-priority item at the moment. - Defining the interface and not detailing how the estimation is carried out is a reasonable first step. (Estimation could be history based and also involve some prediction when deployment requests are received for the first time.) - Should this be a simple interface receiving a CDL fragment or one that unwinds a bigger CDL tree? For example, ask for estimates for deploying a CDL tree made of (A,B,C,D) (and considering deployments may be done in parallel) or ask separately for each A,B,C,D part? - Also in the more complex form what should happen if there is history for deploying B,C,D but not A. - Agreed that initially this interface may be very simple: either returns an estimate for whatever CDL (simple or complex) it receives; or "don't know" - Probably need a new sequence diagram to show the relative positioning of this service. [Need to decide when in the EMS Roadmap it appears.] - Hiro will follow up on other issues with Jay ** WS-Agreement discussion preparation - Not sure what the state of this work is as there has been no update from the volunteer team. * Glossary 1.5 review - Client/Service provider/Service requestor - Not sure if these are really needed. The definitions look ok so include them for now. - ACID : leave as is - CSG: leave as is - Component: Jem and Andreas to discuss offline - Context: leave as is - IPC: leave as is - Legacy program: leave as is - Manageability: Refer to WSDM and WS-Manageability specs for consistency (Jem) - Management: Revised definition Action: Jem will update the working glossary copy and the trackers. Action: Andreas to check the use of the term 'use case' in the Architecture document. * Naming proposal status update - Recap of key points from last call - Tom will have the written proposal (agreed on during the previous call) ready within the day. * Basic Security Profile - Draft not ready - postponed for next week * Teleconference schedule ** Next call - EMS discussion first; followed by Naming proposal review - UML tool review (given time) - Mike will send out his evaluation to the list before the call * Next week - SC'05 is next week. - Next Monday: OGSA-AuthZ joint call. (Deliverables, charter, collaboration topics, etc.) Action: Takuya to invite the OGSA-AuthZ team to the call.