SLA Aware Approach for Virtual Machine Placement in Cloud Datacenter

Abstract

The massive adoption of cloud computing by IT industry has caused a dramatic increase in energy consumption and its impact on the environment in terms of carbon footprints. The high cost of energy and link between energy consumption and carbon emission have raised an issue of energy management. Recently, a number of approaches have been proposed to address this issue. Most of these approaches are based upon the proper utilization of the resources. On the other side, to avoid Service Level Agreement (SLA) violations, the servers deployed in data centres can be utilized up to a maximum threshold value. To avoid SLA violation, some virtual machines (VMs) will be migrated from over utilized hosts. Although there are many techniques to select such VMs that can be migrated, but no technique considers Service Level Agreement parameters while selecting such machines. In this paper, SLA based approach to select VMs to migrate from overloaded host has been purposed. The purposed approach will work on the basis of VM categories that are negotiated in service level agreement. The simulation results demonstrate that the proposed approach demands slightly more energy than the other efficient technique (Minimization of Migration). However, it reduces the cost paid in lieu of SLA violation as it reduces SLA violations belonging to VM categories for which SLA violation penalties are higher.