Technical Report Number
Our ultimate goal is to develop the software support needed for the design, analysis, understanding, and testing of programs involving many thousands of concurrent processes running on a highly parallel multiprocessor. We are currently evaluating the use of a shared dataspace paradigm as the basis for a new programming language supporting large-scale concurrency. The language is called SDL (Shared Dataspace Language). In SDL, a content-addressable dataspace is examined and altered by concurrent processes using atomic transactions much like those in a traditional database. Associated with each process is a programmer-defined view. The view is a mechanism which allows processes to interrogate the dataspace at a level of abstraction convenient for the task they are pursuing. This paper provides an overview of the key SDL features. Small examples are used to illustrate the power and flexibility of the language. They also serve as a backdrop against which we discuss programming style implications of the shared dataspace paradigm.
Roman, Gruia-Catalin; Cunningham, H. Conrad; and Ehlers, Michael E., "A Shared Dataspace Language Supporting Larger-Scale Concurrency" Report Number: WUCS-88-09 (1988). All Computer Science and Engineering Research.