A Java-Based Environment for Adaptive, Fault-Tolerant Parallel and Distributed Computing--Scientific Computing Associates, Inc., One Century Tower, 265 Church Street, New Haven, CT 06510-7013; (203) 777-7442
Dr. Robert Bjornson, Principal Investigator
Dr. Andrew H. Sherman, Business Official
DOE Grant No. DE-FG02-98ER82663
Amount: $74,972

Meeting DOE's advanced technology goals depends increasingly on the effective use of widely-distributed high performance computing resources. The ability to access computational power easily and effectively will give researchers an invaluable resource. To support this, it is essential to have easy-to-use middleware tools for specification and management of distributed computations that provide features such as portability, flexibility, and dynamic adaptivity. While Java addresses some low-level execution issues, there is a clear need for higher-level integration tools and environments. This project will investigate a new paradigm for wide-area distributed computing based on Java and the Piranha model for adaptive execution of parallel programs. The result will be an environment offering not only high performance, portability, and conformance with modern object-orient standards (due to Java), but also dynamic adaptivity and fault tolerance (due to Piranha). In Phase I, designs will be developed for a new distributed computing environment and its main components (the Applet Piranha Loader, the Applet Java Piranha Class Library, the Java Piranha Application Scheduler, and the Java Feeder Class Library).

Commercial Applications and Other Benefits as described by the awardee: Java-Piranha would enable users to leverage the tremendous computing power latent in the internet, allowing governmental and commercial organizations to increase productivity and efficiency without acquiring or supporting large numbers of new computers. It would also facilitate a new "electric-utility-like" marketplace for computing power by making it possible to aggregate distributed resources dynamically on a metered or "leased" basis.