Using communicating Petri nets to implement parallel computation in single-tasking operating systems

摘要

An extended Petri net model, referred to as communicating Petri nets, is used for the operational specification of distributed systems, within a software engineering tool. This specification language naturally involves a parallel computational model, largely fitting the characteristics of parallel systems. Nevertheless, it can be efficiently matched even on single-tasking systems by means of appropriate light-weight scheduling policies. A number of such scheduling policies is devised and assessed both with respect to their impact on the semantics of the run-time code and with respect to the execution overhead.