QoS-aware Scheduling Algorithm Based on Complete Matching of User Jobs and Grid Services

Grid services scheduling is a challenging problem under open grid service architecture (OGSA). Grid services running on clusters must have mechanism and policies for establishing and supporting QoS. A graph theory formal description is introduced into the service grid model in this paper. The necessary and sufficient condition of complete matching of user job and service resources has been given and proved. Optimal solution to matchmaking of grid jobs and grid services is developed based on the running time weight matrix, and the arithmetic has been verified by simulation analysis which proved to be more efficient than the alike arithmetic. The arithmetic has been implemented and running well