Load Balancing in Cloud Through Task Scheduling

Abstract

Cloud computing is defined as the combination of a parallel, distributed and multi-tenant computing model. Cloud computing provides the feature of delivering the resource-like computing and storage according to the demand of user by providing IAAS, PAAS, and SAAS services using pay- per use model. Due to increment in cloud users on industry the cloud usage increases. This drastic increase in usage of cloud highlights the load balancing need by utilizing some resources effectively according to the changing environment. Task scheduling is a fundamental issue in this environment. Task scheduling is a NP-hard optimization problem, and to resolve this problem many metaheuristic algorithms had been tried. Some good task scheduler changes their scheduling strategy according to the types of tasks and change in environment. In this paper, we improve the load balancing by minimizing the Makespan and Response Time QoS parameters by implementing the complete scheduling process i.e. Process scheduling and CPU scheduling. In our approach, we are using Metaheuristic algorithm ACO (Ant Colony Optimization) for process scheduling, i.e., allocating the processes according to the capacity of the processor and FCFS (First Come First Serve) or RR(Round Robin) for CPU scheduling and compare their performance with FCFS-FCFS and FCFS-RR Experimental results showed that our approach ACO-FCFS, ACO-RR outperformed FCFS-FCFS, FCFS-RR respectively by minimize the Makespan and Response Time and do better Load balancing.