Attached. https://forge.gridforum.org/sf/go/doc13903?nav=1 -- Andreas Savva Fujitsu Laboratories Ltd OGSA Teleconference - 25 September 2006 ======================================= * Participants Mike Behrens (R2AD, LLC) Mukund Buddhikot (SUN) Chris Jordan (SDSC) Hiro Kishimoto (Fujitsu) Tom Maguire (EMC) Mark Morgan (UVa) Andreas Savva (Fujitsu) Alan Sill (TTU) Paul Strong (eBay) Jem Treadwell (HP) Minutes: Andreas Savva * Summary of new Actions AI-0925a: Hiro to setup wiki page with names of people who have expressed interest in the security team AI-092625b: Jem to pull definitions together and produce a new merged first draft. - Send a mail to the ogsa list about this work - Set Oct. 9 as the target for the first Glossary draft AI-0925c: Hiro to ask Alan Clark to ask EGA TC member who want to contribute to join the OGSA list and this convergence work. - Members may have changed but organizations have not AI-0925d: Paul Strong to send EGA Glossary MS-Word file to Jem Treadwell AI-0925e: Hiro to talk with Andrew Grimshaw and Mark Morgan and bring back a concrete proposal for the HPC Profile use case as a convergence candidate. AI-0925f: Hiro to talk with Ellen Stokes and add more items in DMTF register in the context of this convergence work AI-0925g: Tom will send a note to the list (to: Ellen S and Fred M) on why the current separation of service/info model may not be quite right. * Minutes approval ** September 21 minutes approved with no changes ** F2F minutes for September 15 (two parts) approved with minor changes: - Fix Michael Behrens affiliation - Correct AI number overlap * AI review [Only items with updates/comments are listed below] - AI-0921b: Pending. (Really a sub-item of AI-0921c) - AI-0921c: Pending. New Security ADs not selected yet. - AI-0921e: Closed. Alan has sent information to the list - AI-0921f: Closed. Already discussed on the Interop call. - AI-0921g: Closed. - AI-0921i: Close once the agenda is fixed - AI-0914b: Closed. There was a call last week. - AI-0925a: Hiro to setup wiki page with names of people who have expressed interest in the security team - AI-0912a: Closed. Today's discussion. * EGA Reference Model Discussion ** Glossary - Jem gave an overview of his glossary presentation from GGF18; number of terms and overlaps (or non-overlap) - Explanatory (EGA) vs more formal/committee style (OGSA) definitions - Should also look for less obvious overlaps, beyond just same/different terms - Should probably define some new ones (e.g., Enterprise) - Possible editors for this work: Jem Treadwell & Paul Strong - Glossary convergence should be done in the context of other convergence work (i.e., Reference Model and OGSA) - The EGA reference model structure and verbosity of the Glossary is driven by the expected consumers of the documents. One reason that more text was pushed in the Glossary was to avoid excess overlap with definitions. - Might need to have taxonomony/ontology discussion to sort out things - Is a single Glossary possible? - Probably no problem doing it. - There is no way but to have a single Glossary - The OGSA Glossary is specific to the OGSA Architecture document---for scoping reasons. (In other words, it is not a general Grid or GGF Glossary.) - Occasionally there are 'battles' over common definitions of terms. Having an OGSA definition should not stop others from having their own definitions of terms. - Definitions of terms perceived to fall in this category are often prefixed with "In OGSA, " in the Glossary to make the context clear. - How to proceed? Wait for broader work to start? - No need to wait. Could start small, using email and then use the occasional call/F2F meeting to resolve remaining problems - In terms of process the OGSA list should be used AI-092625b: Jem to pull definitions together and produce a new merged first draft. - Send a mail to the ogsa list about this work - Set Oct. 9 as the target for the first Glossary draft - Mike Behrens also volunteered to help out on the Glossary - Is everyone from EGA on the OGSA list? Maybe. - Are the EGA lists still active? The EGA alias does work and people have been invited to join AI-0925c: Hiro to ask Alan Clark to ask EGA TC member who want to contribute to join the OGSA list and this convergence work. - Members may have changed but organizations have not - Is the EGA Glossary available as an MS-Word file? It would simplify the work for the first draft. (The pdf version is otherwise sufficient.) AI-0925d: Paul Strong to send EGA Glossary MS-Word file to Jem Treadwell - From a more general perspective, it would be nice to have the EGA documents available on the OGF site. - This was assigned by OGF to the OGF Editor (Greg Newby). He is now reading the EGA documents. ** Reference Model and OGSA convergence - At GGF18 the proposal was to consider OGSA as a rendering of the EGA reference model. Agreed that this seems a good starting point. It should be tested by picking a candidate topic and see if it is really the case. - Hiro proposed the OGSA HPC Profile as a test case (a concrete use case) - The HPC Profile is more detailed when compared to the style of the use cases in the EGA Reference Model. Therefore it would help flesh out what is needed in the context of OGSA. Doing UML would be good. - It was noted that EGA graphs (DAGS) are specific to instantiation - Tentative agreement on the HPC Profile proposal but need more information before making a decision. AI-0925e: Hiro to talk with Andrew Grimshaw and Mark Morgan and bring back a concrete proposal for the HPC Profile use case as a convergence candidate. - What are the deliverables? - If the use case approach works then the refinement could be put in the next version of Reference Model. - Should this work be in the same WG? Should a new OGF Reference Model WG be formed? - Forming a new WG is a possibility, especially since the reference model work is at a different level from other OGSA work - Initially this work can start in an OGSA design team. It could be spun out later when the scope is clearer. A decision could be made by the end of the year. - Does this mean that the result is a converged model? - Yes, the result would be a reference model that also covers the HPC space, in addition to the enterprise space. - Can we expect more use cases from EGR-RG? - Maybe but their use cases are more abstract - Paul Strong may bring more use cases - Part 2 of the Reference Model document includes use cases. Should these be in a separate document? Or should the work be done in a different WG? - Doing the use case work in the same WG is ok - Typically the use case document would be a separate document - Another open issue is how does the Information Model work (in context of DMTF) comes together with the EGA reference model (dags). - A possible viewpoint is that the EGA Reference Model is the overarching model: - With specific/concrete stuff mapping to CIM - But some more abstract stuff do not map to CIM easily - There is some more abstract work in the SMIS space and this will end up at some point in DMTF schema - So the DMTF might be an eventual consumer for some of this work AI-0925f: Hiro to talk with Ellen Stokes and add more items in DMTF register in the context of this convergence work AI-0925g: Tom will send a note to the list (to: Ellen S and Fred M) on why the current separation of service/info model may not be quite right.