Task Consolidation Algorithm for Heterogeneous Cloud Computing

Abstract

As the recent advancements are going in the field of computer technologies like network devices, hardware capacities and software applications, cloud computing has emerged as an important paradigm that provides scalable and dynamic virtual resources to the users through the internet. Energy consumed by modern computer systems, particularly by servers in a cloud has almost reached at an unacceptable level. Also the energy consumed due to under utilization of resources accounts almost 60% of the energy consumed at peak load. It has resulted into reduced system reliability, extremely large electricity bills and environmental concerns because of resulting carbon emission. Therefore there is a great need to optimize energy consumption. Methods like memory compression, request discrimination, task consolidation among virtual machines are developed to enhance resource utilization. Task consolidation problem has been addressed as an optimization problem in heterogeneous cloud computing environment. Task consolidation maps user service requests to appropriate resources in cloud computing environment. The resource allocation problem in cloud computing environment is NP- complete. This thesis presents resource allocation problem as LPP to optimize energy consumed by the computing resources in cloud computing environment. We have used greedy algorithms to obtain sub-optimal solution for task consolidation problem. The performance of the task consolidation algorithm has been simulated with in-house simulator developed by us using Matlab. The simulation is carried out with three different arrival patterns and the result shows in favor of our proposed EATC(Energy Aware Task Consolidation) algorithm.