Modeling and Performance Analysis of Grid Task Scheduling Based on Composition and Reduction of Petri Nets

Abstract

Task scheduling is one of the most important and difficult issues in grid computing environment. Although some algorithms for grid task scheduling were proposed, formal modeling and analysis for task scheduling in grid computing have not given more attention. In our previous work, a high-level timed Petri net to model the workflow of grid tasks was proposed. But it only focused on independent tasks. In this paper, we tackle the scheduling problem of task that consists of a set of communicating subtasks. First, an extended timed Petri net model (ETPN) for a subtask is constructed. In order to reduce the complexity of model and the state explosion problem in reachability analysis of Petri nets, a Petri net model for a task is gotten using composition and reduction of Petri nets. A construction algorithm for reachable task graph (RTG) of ETPN is given in this paper. Finally, we give the methods to verify the correctness of the grid task scheduling and analyze time performance of system by analyzing RTG of ETPN