International Journal of Computer Science Issues

More than a traditional journal...

In the Grid environment, the relationship between a customer and a service provider should be clearly defined. The responsibility of each partner can be stated in the so-called Service Level Agreement (SLA). A SLA is a formal contract between end-user and system to guarantee that customers\' service quality expectation can be achieved. In recent years, extensive research has been conducted in the area of SLA for utilizing computing systems and also, various SLA-based scheduling are proposed but the number of resources and tasks to be scheduled is usually variable and dynamic in nature. Most of proposed algorithms don\'t have flexibility in all situations, because every scheduling algorithm cannot improve all grid factors like resource utilization, load balancing, etc and cannot notice all parameters at the moment. In this paper, we propose SLA aware scheduling architecture which uses learning techniques for selecting best way to schedule resources in different situations. The proposed model causes increasing user satisfaction, number of completed tasks and system utilization and resource load balancing. At the end, we formulize relation between number of completed tasks and system utilization.