GGF8 GRAAP Meeting Minutes

GGF IP Statement and sign-in
----------------------------

Agenda
------

Charter
-------
- WG Context, milestones, scope clarification (all on web site)


Current State of the WG and Activities
--------------------------------------

- Requirements document updated after GGF7, submitted to mailing list,
  submitted to GGF editor for review as informational document.

- State of the art document submitted to mailing list, *consensus that
  it should be submitted to GGF editor for review as informational document.*

- WG web site moving entirely to GridForge
  (http://forge.gridforum.org/projects/graap-wg/) 

Liasons
-------

Missing Groups?  
CMM WG (Jim Pruyne on interim basis, would like to have a new volunteer)
DAIS Newly added (Karl Czajkowski (KarlC) volunteers)
Grid High Performance networking group (Volker Volunteers)
GSI, KarlC suggests dropping that relationship
ACE-RG (Stephen Pickles in place of Michael Daw for right now)
WRT to ACE-RG: There was discussion on the mailing list about ACE
looking at GRAAP as an approach to meeting room reservation.  Decision
that this should be captured in another use case document since the
current use case document is already submitted.  *Need author(s) for
this document.*

Storage Resource Management Group might be a good link, but not a full
GGF WG/RG at this time.  So no liason assigned.


Operations Document (available on GridForge web site)
-------------------

Discussion of where the SLA content/language is coming from.  Policy
BoF/RG is possible, must see how it evolves.  JSDL BoF/WG is likely
for the job/task sla type language.

In discussion of the state diagram, Jon MacLaren (JonM) suggests further
scenarios or a more complete (set of) state diagrams because one
diagram does not seem to fully capture the full set of possibilities.
He volunteers to give this clarified diagram or description a shot.

Discussion of need for polymorphism and distinction between SLA types.
General questions: Will all cases fit our current view of the types of
SLAs?  Can we model composition and dependency cleanly on its own?

Clarification that port types (i.e. WSDL) are simply out of scope of
this document by design.  Prefer higher level first view before diving
to that level of detail.

Question: without port types, what's the future of this document if later
lower level description becomes inconsistent.  Does it serve as
primer?

Volker Answers: Ultimately, the idea was to set milestones that are
doable between GGF dates.

Security concerns discussed.  Consensus that we should push out a
document describing security concerns for consumption by OGSA Security
related WGs.  *Volunteers for this document are needed.*

Open Discussion
---------------

Discussion of OGSI-Agreement (slides available on web site)


Clarification of agreement components (T., S., R.A. in the diagram in
the slides)

Closenss and relevancy also to the work in the GESA-WG was noted.
Using GESA needs as a test of OGSI-Agreements power/generality was suggested.

Discussion to clarify relation of lifetime of agreements to the
lifetime of any agreed upon underlying service.  These are actually
orthogonal and need not be specifically related.

Request for use cases from OGSI-Agreement to be pushed into GRAAP use
cases.

Monitoring portion of agreement applies largely to the management
system and perhaps a third-party auditor.

No one aware of anything that can be pulled from web services work that's
related.  This seems to be new ground in general.

Large consensus (no dissention) that we bring the OGSI-Agreement document
within scope of GRAAP.  *To be discussed explicitly on the mailing
list.*

Expectation that a security model around the terms of the agreement
are needed beyond the OGSI-Security. model.  *Kate Keahey and Sam Meder
volunteer.* Push request on to the list.  *Update our milestone to
include this.*