Ontology-based Job Scheduling Using Mobile Agent Technology in Grid Computing

This paper proposes an ontology-based job scheduling system for grid computing. The system infers candidates, which are grid resources suitable for user requirements, from resource information with semantic meanings. The system selects the most reliable candidate among candidates and allocates a job to the candidate. In this system, job scheduling and allocation are controlled by three agents, i.e., the job scheduling agent, the job mobile agent, and the information mobile agent to decentralize the workload of grid middleware and overcome job failures. This paper also proposes a grid ontology, which defines base grid concepts and their relationships so as to give semantic meanings to resource intimation. For performance evaluation, we compared the proposed scheduling algorithm with existing scheduling algorithms. Experimental results demonstrate that the proposed scheduling algorithm outperforms other scheduling algorithms.