Distributed data parallelisms: specified in Descartes and programmed in SequenceL and Java

Abstract

This paper presents an overview of language approaches for the automation of distributed data parallel problem solutions. These data parallel problems are of interest to the oil industry and U.S. Federal Agencies involved in the analysis of various kinds of telemetry data. Special emphasis is placed upon major language issues facing the development of the information power grid. The paper presents an example of the type of parallelism desired in the grid. The problem is specified in Descartes. A concurrent Java solution is then presented. The Java solution contains artifacts of the design of parallel solutions. The same problem is then recast in the high level language SequenceL, in which parallelisms are implied. The SequenceL approach seems to be a good candidate for a grid oriented language, in that the abstraction relieves the problem solver of much of the burden normally required in development of parallel problem solutions.