IMPROVING QoS IN ENERGY EFFICIENT CLOUD COMPUTING USING VM CONSOLIDATION, ADAPTIVE HEURISTICS AND PRIORITY BASED TASK SCHEDULING

Abstract

Cloud computing is a model where services (Infrastructure, Platform and Software) and a applications are provided as utility, that can be accessed using well known Internet protocols. Using this model the services like computing, applications, storage, networking etc. are given on rent and charged only for the amount of resources used. It has several benefits like elasticity, scalability and dynamic on-demand provision of services etc. but there is also the dark side of cloud computing. According to Gartner Report 2007, IT industry contributes about two percent of world's total CO2 production in the environment and according to U.S. EPA Report 2007; data centers uses consume about 1.5 percent of total U.S. electricity, which has been doubled since 2000 and costs $4.5 billion. According to McKinsey report on "Revolutionizing Data Center Energy Efficiency "; electricity equal to 25,000 households or a city is consumed by a typical data center. In a typical data center, the energy consumption is doubled every five years. So the objective of this work is to reduce power consumption in a data center while maintaining the quality of service (QoS). In this dissertation work, we are improving the algorithm 'Lago Allocator (LA)" [1] by removing the limitations of it, resulting in reduction of the completion time of workload and energy consumption and maintaining Service Level Agreement (SLA).