Virtual Machine Placement using Analytic Approach in Data Center

Abstract

Cloud services are being used rapidly for various purposes and that leads to rapid growth of Data centers since Data Centers are serving as remote storage medium which is done by a group of servers. Virtual machines can also be considered as major reason for the success of cloud computing in such a large scale. Virtual machine requests are getting more day by day which directly implies that there is a need for allocation policy which can be efficient in dealing with the VM placement in data centers since placement of virtual machines plays a vital role in Data Centers. After a thorough consideration of the current scenario, it indirectly implies that a dynamic allocator which is aware of power needs to be studied. In this, every request for Virtual machine requests gets characterized based on four parameters. Those four parameters are disk, RAM, CPU and bandwidth. These allocators are designed in such a way that it accepts as many requests as possible. Here we tried various allocation strategies and these allocators are differing in terms of policies like different approaches. They are also differing in terms of policies like objective optimization or joint or disjoint selection of the resources. The behavior evaluation of this policy strategy is carried out in a way that varies the load in DC and also varies the quantity of virtual machines that are to be allocated. Furthermore, the results do clearly illustrate that disjoint approaches are outperformed by jointapproaches.