Resource Scheduling Algorithms for Grid Computing and Its Modeling and Analysis Using Petri Net

Yaojun Han, Changjun Jiang, You Fu, Xuemei Luo

Abstract: A resource scheduling algorithm called XMin-min is proposed in this paper. In the XMin-min algorithm, we consider not only the expected execution time of tasks, but also expected communication time when calculating expected completion time. In the paper, the execution cost of tasks and budget of application are selected as QoS and an algorithm XMin-min with QoS is also proposed. An extended high-level timed Petri net (EHLTPN) model suiting scheduling of resource in grid computing is presented in the paper. In the EHLTPN, the firing times assigned to transitions are functions of the tokens of input places. We construct a simple model for the resource scheduling in grid computing using EHLTPN. A definition of Reachable Scheduling Graph (RSG) of EHLTPN to analyze the timing property of the resource scheduling is given in this paper. Two algorithms can be use to settle the “state explosion” problem while constructing RSG of EHLTPN.