Deadline Based Performance Evaluation of Job Scheduling Algorithms

Abstract

Grid scheduling is one of the prime challenges in grid computing. Reliability, efficiency (with regard to time utilization), effectiveness in resource usage, as well as robustness tend to be the demanded features of Grid scheduling systems. A number of algorithms have been designed and developed in order to make effective Grid scheduling. Project management is the well known area of operation research. Here we are proposing a new prioritized deadline based scheduling algorithm (PDSA) using project management technique for efficient job execution with deadline constraints of users' jobs. An extensive performance comparison has been presented using synthetic workload traces to evaluate the efficiency and robustness of grid scheduling with respect to average turnaround times and maximum tardiness. Result has shown improved performance under dynamic scheduling environment. Based on the comparative performance analysis, proposed PDSA has shown the optimal performance as compared to EDF and RR scheduling algorithms under dynamic scheduling environment. In brief, PDSA is the true application of project management in grid computing.