Wednesday 6 August 2008

Distributed Programming Abstractions @ e-Science 2008


The 4th IEEE International Conference on e-Science takes place in December 2008 at University Place Conference Center (University of Indiana) in the US. The conference has a special theme on "Cloud Computing" this year, and is a unique event that brings together computational scientists with computational infrastructure developers. The conference features an excellent list of workshops and tutorials -- covering areas such as workflow, Problem Solving Environments and a focus on "community engagement" to get better uptake of computational infrastructure within the day to day practices of researchers. Along with colleagues, I am also involved in organizing a workshop at this event on Abstractions for Distributed Applications and Systems. The overarching aim of this workshop is to highlight and understand the gap between: (i) the increasing complexity and availability of distributed computational infrastructure; (ii) the lack of takeup and effective use of such infrastructure within computational science applications. It may be argued that applications have not been able to make effective use of distributed infrastructure and tools because of the gap of suitable abstractions/patterns to utilize infrastructure efficiently -- the effort required has been too high for many scientists. Interestingly, the design and implementation of the most ambitious distributed infrastructure are not consistent with a clear theoretical understanding of distributed applications -- the different classes of application, specific application level requirements and common usage modes. Additionally, this workshop aims to determine where programming abstractions are important and where non-programmatic abstractions are likely to make greater impact in enabling applications to effectively utilize distributed infrastructure. In addition to covering traditional applications and systems, this workshop will also address emerging infrastructure and application areas, such as programming and system abstractions for Clouds and data-intensive applications. Paper submission deadline is September 1, 2008.