Concurrency exploitation in data-driven distributed systems

Studies the effect of the degree of task firing strictness (data availability) on the performance of a data-driven multiprocessing system. A prototype distributed system model, where inherent concurrency is exploited at the procedural level, is discussed. A set of behavioral programs has been simulated, and the speed-up versus the degree of task firing strictness (for different network delays) has been plotted. Simulation shows that, as the degree of strictness is relaxed, more concurrency is exploited in a computation, whence a better utilization of the distributed system resources is achieved