An evolutionary approach to realizing the Grid computing vision

Over the last few weeks I worked together with Marvin Theimer, Marty Humphrey, Tony Hey, and Geoffrey Fox on a Grid related paper which has now been distributed to the community*.

The paper describes an evolutionary approach to designing and standardizing services specific to the Grid application domain. It promotes the use of stable and widely-accepted specifications while it stays clear of any arguments related to resource/state management or any other domain of functionality for which the key WS stakeholders have yet to reach broad agreement. The approach is illustrated through an example of how an HPC-specific set of Grid services could be designed and realized in terms of Web Services.

The positions expressed in this paper should not come as a surprise to those reading this blog or those having read papers I co-authored while in Newcastle. We have been advocating for the use of stable WS specifications since 2003. We now have more stable WS technologies (either standardized or very near completion) for which broad agreement has been reached between the major WS stakeholders but the design principles are still the same.

It was great working with my co-authors on this. Marvin and Marty have been in Athens to support the ideas and explain our motivation to the community. From the reports I have been receiving, they did a fantastic job at this.

 

* I think the attached PDF file must be renamed from “.bin” to “.pdf”.

3 responses to “An evolutionary approach to realizing the Grid computing vision”

  1. Steve Loughran

    I am sitting in Marty’s presentation as we speak. In the National Tech Uni of Athens. One of your old haunts?

  2. Savas,

    Agree entirely and I believe that this is also the position of the GGF. They are working with existing WS-* and providing domain specific extensions (i.e. OGSA-Naming, WSRF, WS-Notification etc) where required. I don’t currently have enough practical experience with the different specifications to comment on their value and applicability.

    cheers

  3. Buerta concosto per magnitum lambda.