Virtual Machines Placement Using Predicted Utilization of Physical Machine in Cloud Datacenter
Varun Barthwal, M. M. S. Rauthan, Rohan Verma · SSRN Electronic Journal · 2019
Cloud computing has enabled Information Technology to provide computing resources to the end user. Thousands of Physical Machines (PMs) or hosts are executing in datacenter for provisioning on demand delivery to the user. These datacenter requires power supply in large scale and produces the huge amount of heat to deliver the best quality of services. Virtualization plays the key role to obtain the objectives of quality of services, so management of Virtual Machines (VMs) is the major focused research area in this field. Management of VMs in datacenter is mainly done in three parts, VM selection from overloaded PM or host, migration of selected VMs and the placement of VMs into suitable PMs, while maintaining the consumption of energy and service quality. In the current article, a technique is used in which predicted value of CPU utilization is used for the selection of suitable PM for hosting the migrated VM from overloaded PM. The focus of our research is to minimize consumption of power and enhance the service level agreement.