Improve the Performance of Data Grids by Cost-Based Job Scheduling Strategy

Grid environments have gain tremendous importance in recent years since application requirements increased drastically. The heterogeneity and geographic dispersion of grid resources and applications places some complex problems such as job scheduling. Most existing scheduling strategies in Grids only focus on one kind of Grid jobs which can be data-intensive or computation-intensive. However, only considering one kind of jobs in scheduling does not result in suitable scheduling in the viewpoint of all system, and sometimes causes wasting of resources on the other side. To address the challenge of simultaneously considering both kinds of jobs, a new Cost-Based Job Scheduling (CJS) strategy is proposed in this paper. At one hand, CJS algorithm considers both data and computational resource availability of the network, and on the other hand, considering the corresponding requirements of each job, it determines a value called W to the job. Using the W value, the importance of two aspects (being data or computation intensive) for each job is determined, and then the job is assigned to the available resources. The simulation results with OptorSim show that CJS outperforms comparing to the existing algorithms mentioned in literature as number of jobs increases