Hi team, Here is the minutes of the Aug F2F day 2. Please review it and point out where the record is not correct, if any. We already reviewed the minutes of day 1. I added the name of materials that are now on the GridForge. Thanks Keisuke! The minutes are also available at Gridforge: https://forge.gridforum.org/projects/acs-wg/document/minutes-20050801-f2f.tx... (day 1) https://forge.gridforum.org/projects/acs-wg/document/minutes-20050802-f2f.tx... (day 2) Sachiko ======================================================================= ACS F2F DAY 2 ----------------------------------------------------------------------- Date and Time: Aug 2 9:10-16:30 EDT 2005 Location: IBM, Durham RTP. Participants: Keisuke Fukui (Fujitsu) Peter Ziu (Northrop) Thomas Studwell (IBM) Michael Behrens (R2AD) Sachiko Wada (ASCADE) - minutes ----------------------------------------------------------------------- * 9:10 Role call, note taker, time keeper, agenda bashing. * 9:20 Review of the minutes Minutes approved. * 9:25 Review of the ACS requirements use case for SDD TC Material: cb_d3ACSreqsForSDDinput.doc With one refinement of the sentence from Tom, the requirements are approved. * 9:30 Review of the OGSA F2F slides Material: ACSoverview-OGSA-F2F-Aug05-draft.ppt We reviewed the OGSA presentation material which Mike posted on Aug 2nd. Action: Sachiko, to send the original file of the CDDLM sequence file to Mike. Done on Aug 3. Action: Mike, to send updated presentation material to the ml. Done on Aug 4. * 9:50 What is AAF? Agreed: AAF is a specification which defines the format of AA Document. AA Document logically consists of two elements: AAD and AC. AAD is always required but AC can be empty, in case the empty archive is to be created. AC consists of DD and other artifacts. * 10:50 break * 11:10 Asynchronous Messaging Agreed to allow asynchronous message exchange in case of the use case C that we discussed yesterday below and transfer. C) Create using soap message but with out of band transport for a (part of) the AA document and the uri is provided by the ACS (the ACS client acts as a file server). If we can drop the transfer from the interface and the above, we may ommit the asynchronous interface. On the other hand OASIS ASAP call for asynchronous interface to be provided for certain type of the interfaces, which may applies to ACS. Need more consideration. * 11:40 OGSA document status review by Pete We read through the OGSA Basic Profile. Agreed to keep eyes on OGSA Basic Profile, WS-I and other standards so that ACS won't violate them. * 11:45 Lunch * 13:00 What's AAF (continued) ** IUPF spec Tom explained briefly about IUPF specification which defines physical format of IU package. This spec was submitted to W3C along with IUDD spec. ACS may be able to refer it as physical format for AA. Whether SDD covers IUPF is not sure but Sun proposed SDD should. ** Relationship between AAD and SDD We recognized that there is some common information between AAD and SDD, for example, list of application contents. As for identification, Tom mentioned that the identification for AA and the IU identification are different. We need more discussion on what the AA identification is. We agreed that AAD should not define isolated spec but should extend SDD. We will put a requirement on SDD to make an extension point in their spec. Tom mentioned that the extension point for AAD will be made in IUPF. Keisuke pointed that the schedule since SDD spec will still be under discussion in October when we plan to finish our draft. ** What AAD contains. To realize ARI capability, AAD should contain such kind of information as identification, owner information, file list, and access control of each content. The detailed was not discussed due to time constraint. We will continue the discussion on the ML and telecons. * 14:50 break * 15:10 AA Replication The replication use case proposed by Pete and Mike was recognized as an effective usage scenario of ACS. To what extent ACS should define as normative spec concerning replication and synchronization was the argument point. - One time copy and no synchronization. - One way synchronization. - Two way synchronization. Keisuke said that the third party can take care of the synchronization, for example, management between geographically distributed application contents. There was a discussion the replication (including a one time copy, modification propagation and bi-directional synchronization) should ever to be a part of the part of ACS spec. Keisuke proposed simplest scope for version one for this to be out of scope. We will revisit this in the ML and telecons. There is an OREP-WG, to define the replication mechanism but they seems to be dormant. Sachiko will post some replication sequences to the ml so that we can clarify which properties and operations are needed (or not needed) to achieve the replication capability. Action: Sachiko, to send sequence diagrams as an input for further discussion by next call. * 16:25 Wrap up for day 2 The following topics were not discussed due to time restriction. They are to be discussed at the regular call. - AA Modification - AA Deletion - Version Management The detailed schedule for the spec writing will also be discussed after the all remained topics are cleared.