Bayesian workload scheduling in multimedia cloud networks

In this paper, the resource optimization problem in multimedia cloud networks is considered. Firstly, we discuss the general three tier architecture of cloud data centre where the resource optimization is the critical task for the multimedia service provider (MSP). Then, we present a general overview of objective and quality of service (QoS) parameters which are essential for resource optimization in multimedia cloud networks. A comparative analysis of resource optimization problems in terms of nature of the problem, constraints, solution approaches, allocation procedure are discussed. Furthermore, we formulate a new optimization problem which incorporates a weight update factor into task based scheduling problem. Finally, we apply Bayesian theory to identify the path and update the weight of each path, and evaluate the scheme with simulation. The response time performance of Bayesian workload scheduling scheme is same as heuristic one. Moreover, the scheduling weight of Bayesain scheme is more robust and universal because it depends on the relationship with tasks.