Load balancing with Job Migration Algorithm for improving performance on grid computing: Experimental Results

• Published in: Advances in distributed computing and artificial intelligence journal 2019-01, Vol.8 (4), p.5-18
• Main Authors: Wided, Ali, Okba, Kazar, Fatima, Bouakkaz
• Format: Article
• Language: English
• Subjects: Algorithms; Computational grids; Computer networks; Computer simulation; cpu utilization; Distributed processing; Dynamic loads; Geographical distribution; grid computing; job migration; Load balancing; memory utilization; Overloading; Parameters; queue length; Queues; Resource management; Resource utilization; Response time; threshold
• ISSN: 2255-2863; 2255-2863
• DOI: 10.14201/ADCAIJ201984518

Grid is the collection of geographically distributed computing resources. For efficient management of these resources, the manager must maximize its utilization, which can be achieved by efficient load balancing with Job Migration techniques.Job Migration from overloaded resources to underloaded is an attempt to load balancing across all processors, thus reduce average response time. The decision of migration is based on the information exchange between resources.In this paper, the authors propose a novel Job Migration Algorithm for Dynamic Load Balancing (JMADLB), in which parameters such as CPU load and queue length have been considered and have been used for the selection of overloaded resources (or underloaded ones) in Grid. Here, the overloaded resources do not accept any new job; but, the new jobs are migrated to underloaded resources, even though this mechanism migrate extra jobs to obtain load balancing. The performances of the proposed algorithms were tested in Alea 2 simulator by using different parameters like response time, resources utilization and waiting time in the global queue. In addition, they were compared with other scheduling algorithms such as First Come First Served (FCFS) and Earliest Deadline First (EDF).