A Simply Energy-efficient Migration Algorithm of Processes with Virtual Machines in Server Clusters

Virtual services on computation resources like CPUs and storages are supported to applications by using virtual machines in server clusters. Virtual machines with application processes can migrate from host servers to guest servers. In this paper, we discuss a process migration approach to reducing total electric energy consumption of servers by taking advantage of virtual machines in a cluster. We newly propose a Simple Globally-Energy-Aware Migration (SGEAM) algorithm. Here, a host server is first found to perform a process issued by an application in a cluster and then the process is performed on a virtual machine. Next, a virtual machine migrates from a host server to another guest server in order to reduce the electric energy consumption of servers. Here, the amount of computation to be performed by current processes on a virtual machine is simply estimated only by using the number of the current processes and then the total electric energy to be consumed by each server is estimated. Then, a guest server where a virtual machine migrates from a host server is selected so that the total electric energy to be consumed by not only the host server and guest server but also the other servers can be reduced. We show the total electric energy consumption of servers can be reduced in the SGEAM algorithm compared with other non-migration and migration algorithms in the evaluation.