Energy efficient virtual machine placement algorithm with balanced and improved resource utilization in a data center

Abstract

Powerful data centers are the essential supporting infrastructure for mobile, ubiquitous, and cognitive computing, which are the most popular computing paradigms to utilize all kinds of physical resources and provide various services. To ensure the high quality of services, the performance and cost of a data center is a critical factor. In this paper, we investigate the issue of increasing the resource utilization of data centers to improve their performance and lower the cost. It is an efficient way to increase resource utilization via resource sharing. Technically, server virtualization provides the opportunity to share resources in data centers. However, it also introduces other problems, the primary problem being virtual machine placement (VMP), which is to choose a proper physical machine (PM) to deploy virtual machines (VMs) in runtime. We study the virtual machine placement problem with the target of minimizing the total energy consumption by the running of PMs, which is also an indication of resource utilization and the cost of a data center. Due to the multiple dimensionality of physical resources, there always exists a waste of resources, which results from the imbalanced use of multi-dimensional resources. To characterize the multi-dimensional resource usage states of PMs, we present a multi-dimensional space partition model. Based on this model, we then propose a virtual machine placement algorithm EAGLE, which can balance the utilization of multi-dimensional resources, reduce the number of running PMs, and thus lower the energy consumption. We also evaluate our proposed balanced algorithm EAGLE via extensive simulations and experiments on real traces. Experimental results show, over the long run, that EAGLE can save as much as 15% more energy than the first fit algorithm.