ENERGY EFFICIENT DYNAMIC VIRTUAL MACHINES ALLOCATION WITH PREDICTING HYBRID FACTORS(CPU,MEMORY,AND BANDWIDTH) USAGE FOR CLOUD INFRASTRUCTURE

Abstract

Cloud computing is emerging advance technology used to put our resources both equipment and programming applications on the web. It utilizes web, remote central servers to keep up information and applications. With incredible increase in cloud computing technology and services, the demand for cloud computing has also grown extremely. This extreme demand for cloud based data storage, services and processing forces cloud providers to optimize their platforms and facilities. Reducing energy consumption is one of the most important issues in this optimization effort. Dynamic virtual machine allocation and migration is one of the techniques to achieve this goal. This technique requires constant measurement and prediction of usage of machine resources (centeral processing unit(CPU),memory,and bandwidth) to generate migrations at right times. In this thesis, we present a dynamic virtual machine allocation and migration method to improve energy efficiency while maintaining agreed quality of service levels in cloud datacenters. Our proposed method, called hybrid local regression dynamic selection method (HLRDSM), tries to estimate short-term CPU utilization,Memory utilization and Bandwidth utilization of hosts based on their utilization history by sum of them as one dimension matrix by using absolute summation formula. This estimation is then used to detect overloaded and under loaded hosts as part of live migration process. If a host is overloaded, some of the virtual machines (VMs) running on that host are migrated to other hosts to avoid service level agreement (SLA) violations; if a host is under loaded, all of the VMs in that host are tried to be migrated to other machines so that the host can be powered off. Our simulation experiments show that our method HLRDSM with smoothing parameter(span parameter) method generalized cross validation (GCV) is achievable to apply and can significantly reduced 17.73% and 18.87% for random and planetlab workload dataset respectively compare to existing algorithm of overall power consumption for cloud infrastructure.We did extensive simulation experiments using CloudSim3.0.3 by extending some of core classes like VMs allocation policy to evaluate the efficiency and effectiveness of our proposed method.