Energy-efficient Dynamic Consolidation Of Virtual Machines In Big Data Centers

Cloud-based data centers are under rapid development around the globe and have become a major consumer of energy.To improve the management process of virtual machines(VMs)for energy efficiency while providing required Quality of Service,we propose an optimization framework that seamlessly integrates a set of energy-aware resource management approaches including optimal VM allocation,energy-efficient overloaded physical machine(PM)consolidation,and energy-efficient underloaded PM consolidation.(1)We prove that the optimal VM allocation problem is NP-complete and design a heuristic algorithm,namely,Optimal VM Allocation(OVMA),which uses the upper threshold of the PM utilization ratio and the PM performance parameters to determine the target PM for VM allocation.(2)We design an algorithm for Overloaded PM Consolidation for Energy Efficiency(OPMCEE),which uses the quadratic exponential smoothing forecast function to predict the impact of a VM migration on the PM,and determines the VM that should be migrated from the overloaded PM to mitigate the overloading risk.(3)We design an algorithm for Underloaded PM Consolidation for Energy Efficiency(UPMCEE),which reallocates the VMs on the underloaded PM and shuts down the idle PMs to improve the resource utilization efficiency in cloud-based data centers.(4)We propose an optimized resource management framework that integrates the above three algorithms to perform energy-efficient consolidation of VMs.(5)We implement and evaluate the proposed resource management algorithms in CloudSim.The results show that the proposed VM consolidation solution yields up to 21.5%reduction in energy consumption,34.2%reduction in performance degradation due to migration,70.2%reduction in SLA violation time per active host,and 68%reduction in Energy and SLA Violations(ESV),respectively,in comparison with state-of-the-art solutions.