A Study on Energy Consumption of DVFS and Simple VM Consolidation Policies in Cloud Computing Data Centers Using CloudSim Toolkit

The exponential growth in cloud services has led to an all-time high use and expansion of cloud computing frameworks. The frameworks operating in data centers, reportedly account for about two-hundredths of total energy consumed around the globe. Numerous methods have been proposed to curb or reduce the consumption. One such method is Virtual Machine consolidation. In this paper, we simulate the operation of a data center with varying number of hosts across different operating hours with support of CloudSim and estimate the energy consumption of non-power aware hosts, dynamic voltage and frequency scaling-enabled hosts, and two popular VM consolidation policies, namely-local regression minimum utilization and static threshold random selection. We then compare the above techniques to show the various levels of energy consumption. We also have a brief look at the SLA violation rates of the two consolidation policies.