Minutes of the OGSA-D WG telcon, November 30th 2005 1) Early discussion Roll call: Dave Berry, NeSC (note taker) Mike Behrens, R2AD Stephen Davey, NeSC Chris Jordan, SDSC Allen Luniewski, IBM Apologies: Peter Kunzst, CERN Mario Antonioletti, EPCC The minutes were approved. 2) Action report Completed actions - Peter & Stephen to add first draft of Storage Management scenario. - All to discuss plans for GGF16 OGSA Data meetings & F2F. New actions - Dave to mail references to documents for next meeting. - Chris to talk to GFS about telcon on 14th and a possible GGF16 session. - Dave to ask Ann whether she will be at Sunnyvale Ongoing actions - Dave to find out about status of WS-Agreement post GGF15. [OGSA WG have a plan to address this, but nothing has happened yet] - Ann to update replication section. - Everyone to comment on draft RNS specification and send comments to the list. [Stephen says he will mail some comments on Monday]. - Ann to check the Access section for suitability for file replication. - Mario to update the Access section and make text available for comment. - Section authors to add security notes to their sections. - Dave to revamp section 3 of the architecture document. - Everyone to talk with your contacts to get more participation in this WG. - Allen/Dave to contact David Martin & Hiro Kishimoto to get reviewers & expert participation (after internal reviews). 3) Progess & planning update Peter has updated the storage section. Stephen has updated the scenarios documents. Sunnyvale: Allen and Dave will be there. Peter and Stephen won't make it. Dave to ask Ann. GGF16: One session to present ideas. Chris to talk to GFS about telcon on 14th and a possible GGF16 session. 4) Review of Scenarios document Section 1: Add the GGF Boilerplate. Perhaps worth noting the role that this document plays: it is not a requirements document. Rather, it shows how the services defined in the services document can be combined to address particular scenarios/ use cases. Section 1.1: Data transfer covers more than just pipelining. Section 1.2: We agreed that this section helps to introduce people to the rest of the document. We may need to look at the definitions of service and resource being used here. Section 2: We do not intend to keep this in the final document. Section 3: We agreed with Allen's comments about databases. Also with his comment about replication vs. copying. The RealityGrid use case applies more to job scheduling; perhaps we need a use case that combines data staging and job migration? We agreed that the replica catalog does not need to be notified of changes to the data. We agreed that the description of how the data is copied needed more detail, e.g. the data transfer service does not allocate the storage for the copy. We discussed the question of how to tie the scenarios to particular interfaces. Should we use sequence diagrams with specific operations? There is a problem that being overly specific might lead the reader to think that specific operations have to be used - e.g. files vs. databases or a particular implementation of DAIS. Mike noted that OGSA WG are looking to use a standard UML package, particularly for sequence diagrams, so we would be consistent with other OGSA work if we followed suit. We decided that we didn't want to add sequence diagrams at this stage. Where diagrams have labelled steps, the document could expand the description to give examples of which interfaces might be used at that point. We will delete the old "interface specification" diagram. We discussed whether to clarify which parts of the examples are addressed by the use case. We agreed to add a clarifying paragraph at the end of each example, if necessary. The text at the beginning of section 3.4 is old and needs to be updated. We accepted Allen's comments about section 5.3. In section 5.4, step 8 will be kept as one arrow on the picture, and the description will be split into 3 substeps. We accepted Allen's comments in section 5.4.2. In 6.4, the job should be submitted to the simulation service. In 8, the reviewers didn't understand what was meant by an "echo node". We suggested just calling them "Discovery nodes". In 9.2.1 and Figure 1 in section 9 "Data Service" should probably be "Storage Management Service". We weren't sure how to fit naming into the scenario. 5) Wrap up Dave to mail references to documents for next meeting. Next meetings: Dec 7th: No meeting Dec 14th: Naming, RNS & GFS architecture Dec 21st: Full Data Architecture doc review Dec 28th: No meeting. Christmas hols.