Customization of virtual machine allocation policy using k-means clustering algorithm to minimize power consumption in data centers

Abstract

Cloud Computing provides rapid provision of computing resources like processing power, memory, network resources, storage, etc. Running computing resources for longer time, leads energy consumption, increase the emission of Carbon Dioxide (CO2) and increase the expenditure cost for the resources usage. Hence there is a necessity to minimize the execution time to reduce energy consumption in the cloud environment. One of the existing approaches to reducing energy consumption is based on Migration and Placement Policy for Virtual Machine, but still improving placement technique we can further minimize power consumption. In our proposed architecture for cloud resource allocation based on Clustering method, we do map a group of tasks to virtual machines. For clustering, we work on task usage of CPU, memory, and bandwidth. This proposed clustering technique further decreases energy consumption by efficient resource allocation.