Dynamic load balancing in cloud based multimedia system with genetic algorithm

Nowadays cloud computing is the most advanced paradigm promising to show the vision of computing applicability into reality. It provides a versatile and straightforward way to store and retrieve immense information left-out concern the hardware required. Cloud based Storage (CS) incorporates a resource manager, cluster head and server clusters. Inside the resource manager the client's request for information service tasks to server clusters in line with the assignment features. And each cluster head distributes the assigned task to the servers inside its server cluster. In Cloud based Storage every server clusters completely handles a particular kind of information service and receives the client's request dynamically in different time steps. It is a research challenge to design an efficient load balancing algorithm which can assign the multimedia message jobs with minimum cost between server clusters and clients while not overloading the server clusters. Differing from previous works, this paper takes into account a more practical dynamic services scenario in which each server clusters handles only a particular type of multimedia task and each client request different type of multimedia services at different time. Such scenario can be handled as an integer linear programming problem which is computationally in feasible in general. In this paper an attempt to solve the problem by an efficient GA in immigrant scheme Simulation results exhibit that the proposed genetic algorithm can efficiently cope with dynamic multi-service load balancing in CMS.