OGSA-WG teleconference minutes -- July 24, 2003
===============================================

* Participants
        Takuya Araki, Dave Berry, Bill Horn, Hiro Kishimoto, Fred
        Maciel, Takuya Mori, David Snelling, Andreas Savva(*), Jeffrin
        J.Von Reich

        (*)Note taker

* Last call's minutes approved without changes

* No AI update as Gridforge was down.

* Use case document 
  - Editor change approved:  Jeffrin to take over document from Hiro.
  - Jeffrin is drawing up a list of HP use cases and will post to the
    list. Will check that there is no overlap with existing ones. (No
    point having a large number of similar use cases).
  - New use cases
    - Some new use cases have been sent to the list. 
      - Hiro will draw up an updated list as part of handoff to
        Jeffrin
    - Not all of the new use cases are in our template format.
      - We (OGSA-WG) do an initial (simple) re-write and then ask the
        groups owning them to check and refine.
      
* Interim F2F arrengement
  - Hiro reminded people to book their room. If the reception does not
    recognize ogsa-wg interim as an event put Ian as contact.

  - Proposed F2F agenda
    - Looks ok.
    - Produce a package of documents to be used for the F2F and make
      them available before the meeting (gridforge seems to be back
      up). E.g.,
      - Platform
      - Use cases
      - CMM doc (Fred will prepare from the point of view of OGSA)
      - Jay's taxonomy
      - Factoring based on the taxonomy
      - ...

* Review Table of Contents straw for platform document
        > Subject: Table of Contents straw for platform document
        > Date: Thu, 17 Jul 2003 18:36:50 +0900

  - Sections 1-3 look ok. But it may be too soon to go into sections 4
    & 5 now. We should do some work on the content (service
    strawman, factoring) first.

* Platform document
  - Review of Jay's strawman

    - It is a taxonomy. It represents the OGSA user's point of view
      (service endpoints).
    
    - We need other points of view as well. For example, we should
      show what is common functionality across services, and how
      things are put together between OGSI and the upper level
      services (core, data, etc). This is the aim of the factoring
      email that Dave,Hiro,Andreas sent out.

    - We (OGSA-WG) need to offer more direction in the lower concept
      area to guide other groups doing OGSA work and avoid duplication
      of effort (each group re-discovering/re-defining things in
      different ways).

      - CMM and OGSI-Agreement are examples of work that has started
        to address the area between OGSI and upper level services.

      - OGSA-DAI is an example of a group trying to define both high
        level services (e.g., distributed query) and also the low
        level concepts. They are trying to decide what is fundamental
        and what is not.

      - GT3 implementation represents one factoring between OGSI and
        its higher level services.

  - Review factoring email
    - Used Jay's taxonomy as starting point. Divided it into two
      parts: infrastructure and services. It is not meant to be
      complete, but rather to show what kind of work needs to be done.

      - Infrastructure part contains things that are not full-blown
        services but are used as building blocks by higher level
        services.

        E.g., serviceDomain was moved here as it is a form of
        aggregation rather than a service on its own (its position
        under CMM may not be right.)

        Orchestration is here for similar reasons. There is a workflow
        service in Program Execution to represent the higher level
        service (accepting/managing workflow jobs perhaps). (Not sure
        whether dividing things this way fits the description of
        orchestration in the Platform document).

        Read OGSI-Info, OGSI-Storage as capabilities (names are really
        placeholders) --- we need something in that area but are not
        sure yet what it is. E.g., the motivation for OGSI-Info is the
        potentially common ground between notification and higher
        level information services like metering, logging, etc.
        
      - Services part keeps most of Jay's structure. It adds some
        extra categories such as Resource management and Autonomic. It
        also adds some services (fault handling, etc) that are present
        in the use cases/platform.
        
        "Extends" is used to show realations between the higher level
        services and the infrastructure bits. "Extends" in some cases
        may be better read as "Uses."  E.g., accounting "extends"
        ogsi-info but probably "uses" "ogsi-storage."  Other
        relationships may also exist.

        Maybe monitoring shouldn't be grouped together with logging.

        Note that "Job submission" "extends" CMM. This is to point out
        that jobs have a lifecycle; i.e., resources are not just
        hardware things.

        OGSA-Agreement as a high level service refers to services like
        resource reservation, etc.

        There is a temporary "Don't know" placeholder category. 
        - Policy is there because it seemed to look just like
          information at the time. This may be due to a
          misunderstading of the original text. What is the relation
          of policy {deployment, mgmt} to CMM? The policy services
          still don't look like high level services.
        - Some stuff from OGSA-SEC; not sure yet how they fit in.
   
* Any other business

- A service developer workshop was proposed for GGF9.
  - How to prepare for it? (OGSI/OGSA-WG are sponsors)