An empirical evaluation of energy-aware load balancing technique for cloud data center

Load balancing is one of the main challenges in cloud computing, to dynamically distribute the workload across multiple nodes to ensure that no node is either overloaded or underloaded. This paper presents a novel energy-aware load balancing technique that uses an amalgamation of the Artificial Bee Colony and the Firefly algorithms. This technique aspires to balance the load of the cloud infrastructure while trying to maximize the energy efficiency through the efficient usage of the cloud resources. The proposed load balancing technique has been executed in the actual data center of BSNL, Chandigarh. The competence of the proposed technique is exhibited by comparing it with the three standard techniques namely RR, FFD and ACO. The experimentation results show that the proposed algorithm outperformed the existing approach, followed in the data center and the other two approaches. It saved 40.47% of the average energy consumption, which is accomplished by improving CPU utilization level by 49.68%, memory utilization level by 24.41%, reducing VM migrations by 63.10% and saving 53.21% of nodes. The improved results illustrate that the proposed technique effectively balances the load, thereby curtailing the energy consumption and enhancing the performance levels of the cloud data center.