Hi team, Here are the minutes of the GGF15 sessions. Please review it and point out where the record is not correct, if any. I will upload the minutes to the gridforge later. Sachiko ====================================================================== GGF15 ACS Session #1 ----------------------------------------------------------------------- Date and Time: Oct 5 14:00-15:30 EDT 2005 Lead: Keisuke Fukui (Fujitsu) Notes: Sachiko Wada (Ascade) ------------------------------------------------------------------------ We reviewed the ACS specification version 0.51. https://forge.gridforum.org/projects/acs-wg/document/draft-ggf-acs-spec-0.51... * TOC no comments (Chapter 1 and 2 was pushed after the rest of the draft.) * Chapter 3 We reviewed chapter 3 in the draft and OGSA EMS section posted by Mike. The comments from Pete and Sachiko and TBD items are covered and addressed in the draft. ACTION: - Mike, to reconfirm about comment D10 to Pete. (section 3.1.4) - Mike, to draw diagram for section 3.1.4 which describes multiple job creation from a single AA. Issues: - Consider if the digest values can be used to verify cached reference. ====================================================================== GGF15 ACS Session #2 ----------------------------------------------------------------------- Date and Time: Oct 5 18:00-19:30 EDT 2005 Lead: Keisuke Fukui (Fujitsu) Notes: Sachiko Wada (Ascade) ------------------------------------------------------------------------ We continued reviewing the ACS specification. * Chapter 3 (continued) We continued reviewing chapter 3. Hiro sent comments on section 3.3.4 and we adopted the comments. ACTION: - Mike, to change the diagram simpler since we need not describe whole EMS architecture. (section 3.3.4) Issues: - Revisit comment S18 and S20. (section 3.2.2) - Decide how we can address to the naming issues in the draft, i.e. which of the WS-Name or EPR to be used. OGSA seems to be moving to use WS-Name instead of EPR but have not reach consensus yet. * Chapter 4 ACTION: - Mike, to confirm to Pete weather the description of section 4.5 replication is sufficient. * Chapter 5 Steve suggested specifying an ACS Fault type as base of other faults. Stuart suggested that whole tree structure of AAs should be retrieved from Application Repository. Currently users of ACS can retrieve the immediate parent and children of an AA. ACTION: - Mike, to rephrase the Steve's suggestion as an remained issue in ACS-WG. - Mike, to rephrase the Stuart's suggestion as an remained issue in ACS-WG. Issues: - Detailed discussion about resource properties. Should we add manageability properties ? This is related Manageability section. - TBD remained that whether we should use getAAD instead of GetResourceProperty in order to retrieve AAD. - Make clear that who (ApplicationRepository or ApplicationArchive) sends the Notification when AA is created and updated. The list of the changes to the draft will be posted later. Also, the updated version of the draft will be uploaded to the document manager. The review of the remained chapters will be continued on the tele-conferences aiming at the submission by the end of October.
Folks, Attached is a resulted change list from GGF15 review sessions. The draft is also uploaded as 0.52 in below: https://forge.gridforum.org/projects/acs-wg/document/draft-ggf-acs-spec-0.52... Sachiko, could you please procced to register the meeting minutes in the doc manager? I created the below folder for the GGF15 materials. https://forge.gridforum.org/docman2/ViewCategory.php?group_id=147&category_id=1156 Root Category Open Folder Meeting Materials and Minutes Open Folder GGF15 materials and minutes As you see, I have uploaded the change list there, so you don't need to include it. Thanks in advance! -Keisuke Sachiko Wada wrote:
The list of the changes to the draft will be posted later. Also, the updated version of the draft will be uploaded to the document manager. The review of the remained chapters will be continued on the tele-conferences aiming at the submission by the end of October.
Page 13: Deleted An Application Content may be Page 13: Inserted In order to ensure its integrity, AA may include Page 13: Deleted real entity in the AA or a reference to an external storage. An external reference would be useful if the content file is huge in size or owned by other entity than creator. Note that ACS is not responsible for resolving the external reference or ensuring the existence of referenced entity through the AA lifespan, if it is described in AAD as a reference. If it is provided in a real entity with the Create request but attached as a reference, ACS repository will verify and retrieve (resolve) the reference at the creation of the AA instance. If that fails, it will return a fault for the Create request. In order to ensure its integrity, AA may include Page 13: Comment [6] 2005/10/11 17:37:00 I changed the figure and removed external reference. Page 13: Deleted jar, Page 13: Comment [7] 2005/10/11 17:37:00 should be removed since jar is a special type of zip including META-INF directory Page 13: Deleted Zip will do both while others do the former only. Page 13: Comment [8] 2005/10/11 17:37:00 I think we need not this sentence. There are many formats other than zip which do both bundling and compression. Page 13: Comment [9] Perhaps we could say a unit of gconfiguration managementh Page 14: Deleted Different styles of the DD may be needed for deployment in other system that accepts AA and ARI Page 14: Inserted ACS doesnft parse the DD, therefore the ACS is neutral to any language describing DD Page 14: Comment [10] 2005/10/11 17:37:00 Page : 14 This is being discussed at OGSA-EMS and has not reach consensus yet. I suggest to drop off this diagram because the whole EMS architecture is beyond our scope. I believe the relationship to EMS is explained later section 3.3.x.. Page 14: Comment [11] An element? Page 15: Comment [12] Agreed to be ok in the ACS-WG#1 Page 15: Deleted Page 15: Inserted Page 15: Comment [13] Of the same ID? (one modified from the other?) Page 15: Inserted . Both versions are still available Page 15: Comment [14] This is confusing. I canft tell if you incrementing a version of a particular AAID, or making a copy with a new AAID and creating a new version number. ( I am pretty sure it is the former, but maybe we can clarify, since Andrew Grimshaw (OGSA) requested the capability to create new AAfs from existing ones. Page 15: Comment [15] We should allow deletion. The result can be vague on the specification. At least, the repository must keep the consistency of the information. Page 15: Deleted Page 15: Inserted Page 16: Comment [16] 2005/10/11 17:37:00 Page : 16 I think it helps to understand this section if "data transport" is explained prior to this section. Page 16: Deleted (URLs) Page 16: Comment [17] 2005/10/11 17:37:00 Page : 16 I supporse the refernces are not limited to URLs. It may be EPR or any domain specific reference. Page 16: Comment [18] 2005/10/11 17:37:00 Page : 16 I supporse this means to use extended transport method rather than SwA and Embedded that are defined in ARI spec. Page 17: Deleted In either case, the Page 17: Inserted The Page 17: Inserted For the case 1), Page 17: Inserted For the case 2), it is out of scope of the ACS specification and is left to the implementations how those are presented inside the Application Contents. Page 17: Deleted with Page 17: Inserted With Page 17: Inserted , for example, embedded or Page 17: Comment [19] 2005/10/11 17:37:00 Attaching to SOAP message is not mandatory. SwA is one of transport methods as described in later paragraph. The discrete files may be "embedded" into the Create message body or transported by GridFTP or anything. Page 17: Inserted , for example, embedded or Page 17: Comment [20] 2005/10/11 17:37:00 Same as comment [S17] Page 17: Deleted a Page 17: Deleted jar, war Page 17: Comment [21] 2005/10/11 17:37:00 Should be removed since jar and war are special types of zip. Page 18: Deleted Page 18: Inserted Page 20: Deleted OGSA Page 32: Inserted s Page 34: Inserted s Page 37: Inserted s Page 39: Inserted s Page 41: Inserted s
Folks, I have added items in issues in spec tracker from the GGF15 review sessions. (This time I added "GGF15" as a group. It doesn't show up in the column, but the below link lead you to the filtered with the value.) https://forge.gridforum.org/tracker/index.php?group_id=147&atid=805&filter=basic I also added remaining TBD items in the spec to the tracker as you can see following the below link (leads you to the "TeleCons" group): https://forge.gridforum.org/tracker/index.php?group_id=147&atid=805&filter=basic If you are to see the entire tracker items, select "Any" for Gruop. Please help us to close those in time by Oct. 27. Thanks! Mike, In the GGF15 sessions, we got two comments form CDDLM forlks. Could you please create the tracker items for those listed below?
* Chapter 5 Steve suggested specifying an ACS Fault type as base of other faults. Stuart suggested that whole tree structure of AAs should be retrieved from Application Repository. Currently users of ACS can retrieve the immediate parent and children of an AA. ACTION: - Mike, to rephrase the Steve's suggestion as an remained issue in ACS-WG. - Mike, to rephrase the Stuart's suggestion as an remained issue in ACS-WG.
I appreciate your efforts on this very much!! -Keisuke Sachiko Wada wrote:
Hi team,
Here are the minutes of the GGF15 sessions. Please review it and point out where the record is not correct, if any.
I will upload the minutes to the gridforge later.
Sachiko
====================================================================== GGF15 ACS Session #1 ----------------------------------------------------------------------- Date and Time: Oct 5 14:00-15:30 EDT 2005 Lead: Keisuke Fukui (Fujitsu) Notes: Sachiko Wada (Ascade) ------------------------------------------------------------------------ We reviewed the ACS specification version 0.51. https://forge.gridforum.org/projects/acs-wg/document/draft-ggf-acs-spec-0.51...
* TOC no comments
(Chapter 1 and 2 was pushed after the rest of the draft.)
* Chapter 3 We reviewed chapter 3 in the draft and OGSA EMS section posted by Mike. The comments from Pete and Sachiko and TBD items are covered and addressed in the draft. ACTION: - Mike, to reconfirm about comment D10 to Pete. (section 3.1.4) - Mike, to draw diagram for section 3.1.4 which describes multiple job creation from a single AA. Issues: - Consider if the digest values can be used to verify cached reference.
====================================================================== GGF15 ACS Session #2 ----------------------------------------------------------------------- Date and Time: Oct 5 18:00-19:30 EDT 2005 Lead: Keisuke Fukui (Fujitsu) Notes: Sachiko Wada (Ascade) ------------------------------------------------------------------------ We continued reviewing the ACS specification.
* Chapter 3 (continued) We continued reviewing chapter 3. Hiro sent comments on section 3.3.4 and we adopted the comments. ACTION: - Mike, to change the diagram simpler since we need not describe whole EMS architecture. (section 3.3.4) Issues: - Revisit comment S18 and S20. (section 3.2.2) - Decide how we can address to the naming issues in the draft, i.e. which of the WS-Name or EPR to be used. OGSA seems to be moving to use WS-Name instead of EPR but have not reach consensus yet.
* Chapter 4 ACTION: - Mike, to confirm to Pete weather the description of section 4.5 replication is sufficient.
* Chapter 5 Steve suggested specifying an ACS Fault type as base of other faults. Stuart suggested that whole tree structure of AAs should be retrieved from Application Repository. Currently users of ACS can retrieve the immediate parent and children of an AA. ACTION: - Mike, to rephrase the Steve's suggestion as an remained issue in ACS-WG. - Mike, to rephrase the Stuart's suggestion as an remained issue in ACS-WG. Issues: - Detailed discussion about resource properties. Should we add manageability properties ? This is related Manageability section. - TBD remained that whether we should use getAAD instead of GetResourceProperty in order to retrieve AAD. - Make clear that who (ApplicationRepository or ApplicationArchive) sends the Notification when AA is created and updated.
The list of the changes to the draft will be posted later. Also, the updated version of the draft will be uploaded to the document manager. The review of the remained chapters will be continued on the tele-conferences aiming at the submission by the end of October.
participants (2)
-
Keisuke Fukui -
Sachiko Wada