Balancing Performances in Online VM Placement

Abstract

Optimal usage of data center resources has become one of the most important topics in the cloud computing research community. Increased efficiency alongside decreased power consumption becomes a desired goal. Crucial point in achieving this goal is the process of virtual machine placement. In this paper, we analyze and compare several heuristics aiming to evaluate their capabilities with special attention to balanced resource usage versus total number of used physical machines. The presented results identify the preferred placement heuristic that achieve maximum balancing performances based on the data center characteristics, size of the cloud services and their diversity.