Dynamic Consolidation of Virtual Machines for Energy Efficient Management of Cloud Data Centers

Abstract

The thesis describes an effort to minimize energy consumption in Cloud Data Centers (CDCs). The general approach is to do so by smartly using user provided information. The research challenges herein were twofold. Firstly, finding the right benign information. Secondly, smart application of such information to significantly reduce the energy consumption. In this dissertation, we have addressed those research challenges by innovating novel algorithms. Our research exemplifies that if Virtual Machines (VM) are consolidated based on the time when a VM can be removed from CDC, then more physical machines can be turned into sleep state, yielding lower energy consumption.