[ogsa-wg] Teleconference minutes - 26 July 2007 - Workflow
Hiro Kishimoto
hiro.kishimoto at jp.fujitsu.com
Mon Aug 6 23:06:04 CDT 2007
Hi all,
Draft workflow minutes from 26 July 2007 attached.
It is also uploaded onto the GridForge.
- https://forge.gridforum.org/sf/go/doc14702
Thanks,
--
Hiro Kishimoto
---------------------------------------------
OGSA Teleconference - 26 July 2007 - Workflow
=============================================
* Participants
Michel Drescher (Fujitsu)
Chris Kantarjiev (Oracle)
Fabio Benedetti (IBM)
Bahare Rahmanzadeh
Donal Fellows (UMAN)
Hiro Kishimoto (Fujitsu)
Andreas Savva (Fujitsu)
ZackK
Minutes: Hiro Kishimoto
* Agenda bashing
Michel proposes the following as today's session;
- Open discussion
- Donal report out about Taverna
- Discuss a schedule
* June 14 minutes accepted
* New action item
AI-0726a: Show typical/existing use cases for cyclic
dependency declarations
(spawned from AI-0508a, owner unspecified)
* Action items review
AI-0508a: Donal to get in contact with Taverna Drill
use cases for cyclic dependency declarations
Donal has sent out overview of Taverna to the list and
also will explain at today's session. AI closed.
Since "cyclic dependency" is not covered, new AI specific
to "cyclic dependency" is created but no one takes charge
of it.
AI-0726a (new): Show typical/existing use cases for cyclic
dependency declarations
(spawned from AI-0508a, owner unspecified)
AI-0511b: Michel to pull workflow strawman proposal together
No progress, keep it open.
AI-0614a: Steven Newhouse will check whether Windows Workflow can
also be used with CCS / BES style services
No progress, keep it open.
AI-0614b: Andreas will put the initial presentation points on the
wiki
Done.
AI-0614c: Andreas will put the initial questionnaire points on the
wiki
Done.
AI-0614d: Steve McGough will compile the list and put it on the wiki
This is a list for presentations.
Keep it open. We will clarify at the August F2F.
AI-0614e: Steve McGough will look for the workflow material from
the Berlin GGF
Keep it open.
* Taverna workflows by Donal Fellows
Taverna is not service oriented workflow but data oriented workflow.
It is actually data & control oriented workflow.
Taverna is scientific oriented, underneath medical or pharmaceuticals
resource managements are expected to be free to use.
There is no run time authorization mechanism.
However a workflow should go through ethics committee and it requires
human intervention.
How much work will be necessary to Tavarna people adopt standards?
It should not be BPEL, Taverna is very very different from BPEL.
BPEL workflow document is easy to becomes very large but
Taverna is much sophisticate than BPEL.
Taverna has very unique view for how to model. They don't depend
specific execution environment but use specific meta date.
* Next step
Donal thinks it is remote chance to succeed workflow standard.
No one wants to achieve workflow interoperability.
Chris Kantarjiev says interoperability is out of our scope.
Donal agrees.
Control job dependency is in-scope.
Donal thinks it is rather components dependency.
Exapmle is just run posix jobs.
Donal thinks Steve wants to stay simple cases.
Hiro says we work on community practice document as fast track.
Consensus is "focus community practice document" (fast track).
Job dependency (2nd track) should be postponed and focus community
practice document.
* OGF21 session plan
OGSA-WG will request 5 sessions including workflow.
If it is not necessary, we can cancel it later.
Chris K, our AD, is ok to submit 5 session request. Hiro will do so
shortly.
More information about the ogsa-wg
mailing list