EFFICIENT VIRTUAL MACHINE SELECTION ALGORITHM FOR MINIMIZING UNNECESSARY MIGRATIONS IN CLOUD DATA CENTERS

Abstract

– Load balancing is an essential task in cloud data centers that is performed for proper utilization of resources. In addition to that, load balancing also helps in enhancing overall performance of the data centers in terms of availability and reliability. Load balancing is achieved through consolidation of virtual machines in data centers. The key task of load balancing is to continuously monitor the load of physical machines and to detect the overloaded machines and under loaded machines. In case of over loaded physical machine, few virtual machines are to be removed off to mitigate service level agreement (SLA) violation. In case of under loaded physical machine, all virtual machines from it are to be migrated to other suitable hosts. The proposed work devises an algorithm that selects appropriate virtual machines that are to be migrated from over loaded hosts. The virtual machine selection algorithm is found to reduce number of migrations that are unnecessary. The implementation of proposed algorithm has been done in cloudsim and results obtained are compared with those of existing benchmark algorithms of cloudsim. The results have shown that proposed algorithm has significantly reduced energy consumption.