OGSA Teleconference - 13 October 2004 ===================================== * Participants Michael Behrens (R2AD, LLC) Dave Berry (NeSC) Bill Horn (IBM) Hiro Kishimoto (Fujitsu) Mark Morgan (UVa) Andreas Savva (Fujitsu) Latha Srinivasan (HP) Jem Treadwell (HP) Jay Unger (IBM) Jeffrin Von Reich (HP) Peter Ziu (Northrop Grumman) Apologies: Chris Smith Minutes: Andreas Savva, Jem Treadwell * OGSA F2F #9 update - Dates and location fixed: Dec 8-10, Washington D.C. http://www-unix.gridforum.org/mail_archive/ogsa-wg/2004/10/msg00017.html - Let Hiro know if you are not on the list and you are planning to attend. - Jay already booked IBM facilities: room, AV equipment. He is confirming the network arrangements, etc. - Location is about 28 miles from the national airport, 38 miles from Dulles. - Dial-in facilities are also planned. - Jay & Hiro will be sending more information, including exact location and hotel information. - Jay is also planning to ask Heather Kreger (WSDM) to attend and continue the OGSA-WSDM discussion: naming, modelling, etc. - Bill has a conflict and probably cannot make this F2F. * Teleconference minutes approval - Minutes for Sep 15, Sep 29 & Oct 6 approved with no changes * OGSA Use Case tier1 document review [Background: OGSA Use Case tier-1 came out of public comment some time ago. The public comments were discussed in a previous teleconferences (July 14 & 28) and Jeffrin did the revisions agreed on. This review is to check the revisions and how ready the document is for re-submission to the GGF Editor.] References: Public comment tracker: http://forge.gridforum.org/forum/forum.php?forum_id=411 - Takuya was mistakenly added as the author of the Mutual Authorization use case. - Changed to the correct contributor: Shawn Mullen - "Functional requirements for OGSA platform" comment: structure and relation of use cases to OGSA spec is unclear: - Added explanation on use case structure - Referring to appropriate OGSA spec version (014) for specific terminology - Changed reference to version 014 to "Version 1.0 Draft 14." - Changed "advanced reservation" to "advance reservation." Added explanation. - General requirements for workflow engine - Rejected; It is the UC author's decision to describe other aspects. - Change "enterprise" to "in-house": Rejected - "IT system integrator": rejected - Abstract rewritten to address clarity comment - Editorial: unify document formatting/reference/etc - Jeffrin did some work but focusing on content at the moment - Need to fix everything we notice; worried that there is quite a bit of work to do (and document seems to be using lots of different formatting styles). - Also possible unresolved references: - global update and check for errors. - Minor english corrections - Service descriptions: added comment on interface definition - Comment on concept of 'discovery' - Workflow as grid service etc: It is already in the doc. - Compatible with semantic web is out of scope for the use case description. - VO explanation : add text - Not sure if one more review is needed - Proceed by email and review if Jeffrin/Jem/Hiro think it is necessary. - Hiro should put in a status check at the next call to check on progress - Longer term: - EGR-RG to take over use cases. There is no need for OGSA-WG to find a new editor for the Use Case document. Action: Jeffrin do first pass and then send to Jem. Hiro will have it after that. Action: Jeffrin to reply to comments on public comment trackers. (Gridforge will automatically email the person who submitted the tracker; no need for a separate email.) Action: Andreas to send Jeffrin/Jem/Hiro the url to new GGF template. >Done. * V1 comments review https://forge.gridforum.org/projects/ogsa-wg/documen/ogsa-v1-public-comments/en/2 - Reviewed Data section comments. Numbers correspond to rows in the spreadsheet. 73: "Description of something existing..." - Not clear if the comment refers to just the tense or something more? Andreas to confirm. - Layered architecture not a functional requirement - Remove from section title and revise body text slightly (it is only mentioned in passing) - 'Virtualization' vs 'resource specific' interfaces - Add example and push rest of the work to v2 74: Re-phrase or rewrite: schema is provided not by resource itself but by virtual source. If it is a needed term then put in glossary. 75: Add example 76: Formatting: ok 77: Already agreed that this is v2 work. 78: Compare with EMS revision when it is ready. Meanwhile DaveB to look adding some explanatory text to tie the figure better with the surrounding text. 79: Rephrase to make clearer why it is a functional requirement (tackle together with 73). 80: Transactions: - Need to rephrase text to make clearer. It does not refer to specific services but rather to a requirement. (There are 2 aspects: transaction used or provided by.) - Agreed that 'transactions' are not just a data requirement and we are not covering them sufficiently in v1. We are not at a stage we can make definitive statements. - Agreed to fix just 3.5.5.1 & 3.5.3.7 for this version and postpone more detailed work for v2. - Proposal to keep track of these topics: e.g., a what we do not do section. 81: Reject: orthogonal. - Andreas to talk to Philipp to make sure we understood this comment correctly. 82: Rephrase 83: Add text 84: Add text (Data services provide information or stimulus (trigger) but they don't do configuration). * Glossary review - Mike checked the document for usage of choreography/orchestration/workflow. - Choreography: 1 instance - Orchestration: many instances - Workflow: many instances - All seem to talk about the same thing; with varying degrees of nuance. Might want to restrict the terms we use to just one of these. - Summarized the definitions in a document; sent to Jem. Action: Jem to send out email on this topic. [Restart comments review from Jem's doc):p2, c4)] - Choreography comment - Agreed that there is no need to define related terms such as "orchestrator" or "orchestratee." Reject. - DoS: Consensus that the current definition of DoS is adequate, and doesn't require change. Reject. - Comment on definition of "entity": Agreed to add example. - Jem will add the sentence "For example, in OGSA an entity might be a resource or a service." - Event consumer/producer: The comment makes the assumption that we regard an event as a service. However, the subjects of these definitions are the producer and the consumer, which are defined here as being services. (While the Event definition does not say that event is a service.) We saw no reason to make changes. Reject. - Interface comment: Jem will preface the definition with "In a service-oriented architecture..." to restrict its scope. We didn't understand the issue with the second part of the comment. - Port type: Jem will change port type to portType throughout to make the connection to the WSDL term more immediate. - Manageability: Agreed that the current definition could be improved, but we need to do it in line with CMM and with WSDM. Bill will make some suggestions. * AOB - Agreed to do Glossary glossary review first, or at least as early in the call as possible, from next week to accomodate Jem's schedule.