Automatic distribution of reactive systems for asynchronousnetworks of processors

摘要

The paper addresses the problem of automatically distributingnreactive systems. We first show that the use of synchronous languagesnallows a natural parallel description of such systems, regardless of anyndistribution problems. Then, a desired distribution can be easilynspecified, and achieved with the algorithm presented here. Thisndistribution technique provides distributed programs with the samensafety, test, and debug facilities as ordinary sequential programs.nFinally, the implementation of such distributed programs only requires anvery simple communication protocol (“first in first out”nqueues), thereby reducing the need for large distributed real timenexecutives