A model of the solution of the task distribution and load balance among parallel multiprocessors based on the class 2 of Petri Net

If a computing service is composed by several runtime tasks among which dependence relationships exist, the structure of the class 2 of Petri Net can be used to describe the dependence relationships and how much data is depended among runtime tasks; it can be regarded as a model for discovery of parallelism. By determining computing service, constructing Petri Net, calculating the waiting time priority (WTP) for each task, the distribution mapping can be resolved to improve the parallelism, balance the load, and reduce data communication. In this paper, we will investigate the features of the solution and under which circumstance it is appropriate.