ICA-MMT: A load balancing method in cloud computing environment

Energy consumption has become a major challenge in cloud computing infrastructures. Cloud computing data centers consume enormous amount of electrical power resulting in high amount of carbon dioxide that affects the green environment as well as high operational costs for cloud providers. On the other hand, reducing the energy consumption would negatively impact the SLA (Service Level Agreement) that is a crucial concern in any resource allocation policy. In this paper, we propose a novel power aware load balancing method, named ICA-MMT to manage power consumption in cloud computing data centers. We have exploited the Imperialism Competitive Algorithm (ICA) for detecting over utilized hosts and then we migrate one or several virtual machines of these hosts to the other hosts to decrease their utilization. Finally, we consider other hosts as underutilized host and if it is possible, we migrate all of their VMs to the other hosts and switch them to the sleep mode. The results indicate that our method as compared to the previously proposed resource allocation policies such as LR-MMT (local Regression-Minimum Migration Time), MAD-MMT (Median Absolute Deviation- Minimum Migration Time), Bee-MMT (Bee colony algorithm- Minimum Migration Time) and non-Power aware policy offers least power consumption and SLA violation.