The Evaluations of Energy Management Methods in Supersaturated Cloud

Abstract

The cost of cloud is proportional to the degree of aggregation of servers. We propose supersaturated cloud which has higher degree of aggregation than conventional clouds and apply the supersaturated cloud to educational cloud. Here, we define super saturation of cloud as allocating much more logical resource than physical resource. Supersaturated cloud degrades the performance a little but improves the cost performance drastically, From the viewpoint of energy saving, the supersaturated cloud consumes less energy than conventional clouds because it contains a smaller number of running nodes. In this paper, we propose Cloud Energy Management System (CEMS) in order to decrease energy consumption of supersaturated cloud. In CEMS, we can select the policy of instance allocation methods such as performance first and energy first. We have evaluated them by comparing with each other via simulation. By predicting the life time of an instance, the best method will be chosen.