Something that may be of interest...
The idea of this session is to particularly focus on issues of abstractions for workflows -- especially on techniques that could support better description of abstract workflows to aid their mapping and subsequent: (i) sharing; (ii) optimization across multiple workflow systems.
=================================================================
Minisymposium on Workflow Abstractions for Parallel and Distributed Computing in e-Science by José C. Cunha(*), Cecília Gomes(*), and Omer F. Rana(**)
(*) CITI/Dept. Informatics, Fac. Science and Technology, Univ. Nova de Lisboa, Portugal
(**) School of Computer Science, Cardiff University, UK
To be held at the PARA 2008 STATE-OF-THE-ART IN SCIENTIFIC AND PARALLEL COMPUTING Trondheim, Norway May 13-16, 2008 http://para08.idi.ntnu.no/
Abstract: Recent research in scientific applications has shown an increasing need to improve the software tools and environments that support complex computational processes for simulation, visualization, with high degrees of user interaction, and involving large data sets. In order to meet the new characteristics of such large-scale, dynamic and distributed computations, there have been efforts in developing more adequate parallel and distributed computing models and supporting environments, and in adapting them to increasingly powerful cluster and grid architectures.
Workflow-based environments provide a useful approach to provide adequate abstraction levels for scientists and engineers, and also allow flexibility in the mappings between the application abstractions and the lower layers of the underlying computing platforms. Such approaches can contribute to a new generation of more flexible and dynamic environments, that may ease the tasks of application development and the management of the life-cycle of scientific experiments.
This minisymposium aims to bring together experts working on the development of abstractions and models for the specification and coordination of workflows exploiting parallel and distributed computing, namely to handle workflow structure and behavior, manipulation and management of dynamic workflows, coordination and control of their parallel and distributed execution. The focus of this minisymposium is on how best to define abstract worflows before they are converted to concrete/executable versions. Case studies of scientific applications, with reports from experiments and lessons learned, illustrating the above and related dimensions, are also welcomed.
A limited number of contributed, 20 minute talks will be selected for this minisymposium. A tentative title and a short abstract should be submitted via the PARA'08 submission system before March 5, 2008 (max 150 words in plain text). An extended abstract (of 3-4 pages in Springer LNCS format) should be submitted before March 10.
Detailed instructions for authors can be found at:
http://para08.idi.ntnu.no/author-info.html
Make sure to submit your paper under "MS:Workflows for e-Science"
The abstract should be max 150 words + keywords typed into the submission system and the text of the extended abstract must be in .tex and PDF and follow Springer LNCS format
(http://www.springer.de/comp/lncs/authors.html).
After acceptance, full papers formatted according to the rules of the Springer LNCS and the PARA 2008 Workshop should be sent before May 2, 2008, in order to be reviewed for publishing in the Springer LNCS PARA 2008 volume.
Summary - Important Dates:
* March 5, 2008 - Title and short abstract * March 10, 2008 - Extended abstracts (3-4 pages in Springer LNCS format) * May 02, 2008 - Full paper deadline * May 13-16, 2008 - Tutorials and workshop * September 15, 2008 - Camera ready paper for publication in Springer LNCS
For specific information about this Minisymposium, please contact Jose C. Cunha (e-mail: jcc -at- di.fct.unl.pt)
School of Computer Science, Cardiff University, Queen's Buildings, Newport Road, Cardiff CF24 3AA, Wales, UK tel: 02920 879184 email: a.b.harrison@cs.cf.ac.uk