[GSA-RG] Proposal for a new structure of the GSA Requirements document

[Alexander Papaspyrou]

Philipp, all,

Joachim Lepping and I thought about a new structure for the GSA reqs
document, which has been around for a while now. We would suggest to put a
stronger focus on gap analysis regarding what's there in OGF and what's
needed for realizing certain (hopefully common and useful) Use Cases.

Our intention in doing so was to foster the process of proposing
recommendations (I use this OGF overloaded term loosely here) to other WGs
in the scheduling context in order to kick off discussion there. Hopefully,
this will then lead to an incorporation of our proposals into current
standards.

Since the interaction between us and the different WGs is probably going to
be an ongoing process, we would suggest to focus on simple use cases that
can be easily realized. This is also the reason why we left out certain
parts of Nicola's and Ramin's requirements stuff. As a first, very basic
example, we picked HPC Job Scheduling (this will be difficult enough to
fully cover). Additionally, we added what was formerly called "Complex
Workflow", assuming that this comprises jobs with (data) interdependencies.
Of course, these examples are just proposals. However, they could serve as
an example for how discussion on the recommendation part could be organized
in our group in Dortmund.

The general structure of the document is intended as follows: in the third
chapter (requirements), general requirements of Grid scheduling are
discussed on a pretty high level (list of desiderata). The use case part
should be very concrete on what is necessary to tackle for each of the
examples. This subchapter will then serve as a reference for the comparison
to what current efforts in OGF do or plan to do (chapter 5). We grouped the
WGs with respect to what is absolutely necessary for doing scheduling, what
would be nice to have and what is related in general, but pretty far away.
The final chapter (recommendations) should then again follow the structure
of requirements; there, general recommendations for other higher-level WGs
could be given in the first part. In the second part, very concrete
recommendations should be given to other WGs regarding the different Use
Cases, respectively.

Possibly, if we see that this is getting to much / too large, we could spawn
the Use Case-specific recommendations to separate documents; these, however,
should be created jointly with the affected WGs (think of the HPCP profile
approach).

We hope that our proposal is somewhat useful for GSA. Questions and rants
are, of course, very welcome.

Looking forward to a lively discussion in Boston and beyond, we remain with

Kind regards,
Joachim & Alexander

-- 
Dipl.-Inform. Alexander Papaspyrou
http://ds.e-technik.uni-dortmund.de/~alexp<http://ds.e-technik.uni-dortmund.de/%7Ealexp>

Robotics Research Institute                   phone : +49(231)755-5058
Information Technology Section                fax   : +49(231)755-3251
Dortmund University of Technology, Germany
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://www.ogf.org/pipermail/gsa-rg/attachments/20080222/36636621/attachment-0001.html 
-------------- next part --------------
A non-text attachment was scrubbed...
Name: Grid Scheduling Architecture - Requirements and	Recommendations.doc
Type: application/msword
Size: 130048 bytes
Desc: not available
Url : http://www.ogf.org/pipermail/gsa-rg/attachments/20080222/36636621/attachment-0001.doc