RE: [dipal-discuss] dipal and WS-Agreement

[Jacques Durand <JDurand@us.fujitsu.com>]

Title: RE: [dipal-discuss] dipal and WS-Agreement

Anne:

This Agreement doc is an interesting material - and it appears to me that it is already using its own proprietary bits of "policy dialect" that clearly can be generalized, e.g. the Term Compositors (4.2.2), the notion of Variable (4.2.5.1). Possibly, the notions of restriction, group, "all", "choice" in the "Offer Item" (5.1.1) which could be abstracted in a more generic language, and certainly has been in other contexts.

SO I tend to see this as an indicator that a more general language is needed to cover the logic and operators behind expressing policy objects such as agreements.

Which leads me to my second comment: an "agreement" is clearly a notion that is related to "policy", but I have trouble to define how and how much. Is an agreement just an "agreed [set of] policy"? Is there more to it? Or a special case of policy? It is interesting that the WS Agreement doc only mentions twice the term "policy" and in a rather marginal way. It certainly would help to clarify a bit the modeling relationship, if possible.

I see two levels WS-Agreement could be "processed" by the policy-constraint gang (trying not to be  affecting its semantics):

- for abstracting the logic behind agreements (basic concepts, operators) and reusing a more general language for it.
- for aligning and clarifying the model of an agreement with the policy model, whatever it is.

Cheers,
Jacques

-----Original Message-----
From: Anne Anderson [mailto:Anne.Anderson@sun.com]
Sent: Friday, January 13, 2006 8:33 AM
To: dipal-discuss@lists.oasis-open.org
Subject: [dipal-discuss] dipal and WS-Agreement

Colleagues,

To restart the discussion for this new year, I would like to suggest
that DIPAL would be an ideal assertion/condition expression language for
use with the WS-Agreement specification being developed in the Global
Grid Forum Grid Resource Allocation and Agreement Protocol WG
(http://www.ggf.org/Public_Comment_Docs/Documents/Oct-2005/WS-AgreementSpecificationDraft050920.pdf).

The WS-Agreement specification specifically says that WS-Agreement will
use condition expression languages defined by other standards groups.
They mention XQuery as an example of such a language, but XQuery has the
disadvantage of not having a well-defined condition intersection
function: if two agreements contain conditions over the same
variable/schema element expressed using XQuery, it is not possible to
determine automatically whether the two conditions are compatible or
what their intersection is (i.e. what XQuery expression would satisfy
both agreement conditions).  The condition functions defined in
WS-PolicyConstraints, however, have completely defined, efficient
intersection algorithms.

Opinions?  Other ideas about how DIPAL might fit into other proposed
standards?

Regards,
Anne
--
Anne H. Anderson               Anne.Anderson@sun.com
Sun Microsystems Labs          1-781-442-0928
Burlington, MA USA

---------------------------------------------------------------------
To unsubscribe, e-mail: dipal-discuss-unsubscribe@lists.oasis-open.org
For additional commands, e-mail: dipal-discuss-help@lists.oasis-open.org