Efficient power management on a cloud system using two phases power convergence algorithm

Abstract

Infrastructure as a Service Cloud system is based on virtualization technology to enables the hosting of massive number of computing resources. The efficient power management can help lower the cost of operating a large cloud substantially. Power usage in a cloud system depends largely on the system load and the proper use of physical machines. This paper proposes a new algorithm, Two Phases Power Convergence (TPCC) Algorithm that can be used to decrease the system power consumption substantially. This algorithm is based on the converging of the power usage to system utilization using 2 phases strategy - scaling phase and balancing phase. The implementation on our cloud system, Maekin, shows the result of lower power consumption as expected. The contribution of this work is a better understanding of workload management in cloud computing environment to achieve a good power consumption. This is very useful for the improvement of a cloud system resource utilization.