An energy efficient virtual machine placement algorithm with balanced resource utilization

Virtualization is a key technology for resource sharing in IaaS/PaaS cloud infrastructures. One primary issue in virtualization is the virtual machine placement (VMP) problem, which is to choose proper physical machine (PM) to deploy virtual machines (VMs) in runtime. In this paper, we study the VMP problem with the goal of minimizing the total energy consumption. We first present a multi-dimensional space partition model to characterize the resource usage states of PMs. Based on this model, we then propose a virtual machine placement algorithm, which can balance the utilization of multi-dimensional resources, reduce the number of running PMs and thus lower down the energy consumption. We also evaluate our proposed balanced algorithm via extensive simulations. Simulation results show that our approach can save as much as 15% energy compared to the first fit algorithm over a long run.