[acs-wg] GGF13 minutes

[Sachiko Wada]

Folks,

Here are the minutes of the ACS sessions at GGF13.
Please review them and address your comments/observations, if any.
I apologize for the late post. 

They are also available at Gridforge:
https://forge.gridforum.org/projects/acs-wg/document/minutes_session-1/en/1
https://forge.gridforum.org/projects/acs-wg/document/minutes_session-2/en/1

Regards,

Sachiko Wada

-------------- next part --------------
GGF13 Application Contents Service (ACS) session #1

Date and Time: Mar. 14, 2005, 13:00 - 14:30
Place: Lotte Hotel Emerald Room 3
Lead:  Keisuke Fukui (Fujitsu)
Notes: Sachiko Wada (Ascade)

** Agenda

- Opening salutation
  Introduction & note takers 
  Agenda Bashing 
- "ACS introduction" by Keisuke Fukui of Fujitsu
- Use cases and visions
  "NAREGI PSE and ACS" by Prof. Kawata of Utsunomiya university in Japan
  "Enterprise Grid Appliance Operating Environment" by Peter Ziu of Northrop
  "Business Grid Project" by Keisuke Fukui of Fujitsu

** Summary

The session attracted about 30 people and succeeded to present the overall information to new comers to the session and share the current status of the WG, leaving the presentation about IUDD specification and the detailed planning to the next session.

** ACS overview 

Current status and history at GGF is presented by Keisuke. Below is questions and answers for the presentation.

 Q. What will and will not be defined in ACS? 
 A. We will define archive format and repository architecture.
    We are not going inside of the archive and repository. There are 
    some standards for each of the contents, and we are going to 
    include those standards as time matures. 

 Q. Has the milestone changed significantly since the last BOF session?
 A. No. The milestone presented is the one in the charter with updates to the 
    past milestones.

 Q. Do you have any reference implementation?
 A. We have two candidates now. The Business Grid Project and NAREGI PSE.

 Q. Since this activity will require changes for CDDLM, will ACS change its 
    spec as well?
 A. We have not yet decided anything. we are open to modification taking 
    into consideration of CDDLM.

** NAREGI PSE 
The outline and possible relationship to ACS are presented by Prof. Kawata.

 Q. How do you determine the granularity of archive? Are mathematical 
    libraries included in an archive?
 A. What should be included in archives depends on the situation or the Grid 
    systems. We expect ACS to have flexibility as to what the IT resources in 
    the Grid system should provide.

 Q. What extent is the definition of ACS(NAREGI PSE?) service? Do you have 
    schema?
 A. We already have some schema taken from the Business Grid Project. We 
    expect collaboration with SI, which has comprehensive schema already. So 
    we are planning to integrate these schemas. SI scope seems to be larger 
    than us. SI can handle Grid applications but not limited to the Grid. 
    They are trying to accommodate every software in a single standard. ACS 
    archive schema could be one part of that, specific to Grid applications.

** Grid appliance usecase (presented by Peter Ziu)

The future vision of the ACS, OGSA and beyond is explained. No questions.

** Business Grid Project (presented by Keisuke Fukui)

 Q. Do the three companies share development outcomes? Are you also going to 
    share IT resources in data centers?
 A. Business Grid Project is an experimental project to develop a new 
    framework for system management, and thus we are currently working very 
    closely to create a single Grid Middleware. As for a commercially-
    operating data center, we have not yet discussed about operating a single 
    or multiple data centers running that Grid Middleware.

 Q. Do people working in Business Grid Project agree that ACS will provide 
    this capability?
 A. The Business Grid middleware is now under development at their last 
    stage. We are talking closely so that two won't conflict with each other.
    We may refactor the ACS spec. or request modifications of the 
    implementation if it is appropriate.

 Q. Is the solution provided by the Business Grid Project independent of ACS 
    implementation?
 A. We are inspired from the Business Grid Project and I'd like to keep close 
    relationship with it. However, we are open to any discussion, since we 
    are the wg in GGF.

** wrap up.

We ran out of the time at the end and called for participation to the second
session to join the detailed discussion of the ACS-WG activities in future.
-------------- next part --------------
GGF13 Application Contents Service (ACS) session #2

Date and Time: Mar. 16, 2005, 11:00 - 12:30
Place: Lotte Hotel Onyx Room
Lead:  Keisuke Fukui (Fujitsu)
Notes: Sachiko Wada (Ascade)

** Agenda

- Opening salutation
  Introduction & note takers 
- Review of the fist session.
- Use cases and visions
  "Installable Unit Deployment Descriptor" by Thomas Studwell (IBM)

* "Installable Unit Deployment Descriptor" 
  The outline and current status were presented by Thomas Studwell. Below 
  questions and answered were exchanged during the presentation.

 Q. Do you have any plan to go inside the install artifact?
 A. No. The artifact is the one which is referenced from the IU and 
    SI(IUDD) activity do not go inside itself.

 Q. Do you think you can have platform-neutral install artifacts?
 A. Currently it is assumed that the artifact itself is platform specific. 

 A comment from Keisuke: 
   Requirements for IUDD are very similar to those for ACS, which is expected 
   to be a single standard for software installation. We hope Grid-specific 
   features will be considered in the future IUDD standardization. 

 Q. Are you going to plan for refactoring current IUDD spec?
 A. Yes. We are going to submit IUDD spec to a formal workgroup and 
    develop a totally new spec. The new workgroup will consider all of the 
    industry requirements. I expect that CDDLM, ACS, and any other interested 
    workgroups that need standard deployment descriptor to share the 
    specification. I expect contributions both from implementation 
    perspective from the participants.

* ACS specification
  The strawman specification is introduced to show what we have today. We are 
  going to have a single document that specify the ACS, as a WG draft after 
  the WG discussion by the GGF14 in June.
 - table of contents
 - goals and non-goals

* Wrap up

A tentative roadmap is shown and agreed to discuss about how we proceed our 
activity through either e-mails, tele-conferences and/or face-to-face 
meetings.