Research On Energy Consumption-Performance Balance Scheme Of Cloud Data Center

With the proliferation of cloud computing,cloud data centers need to be equipped with thousands of compute nodes to provide good quality of service.The operation of large-scale cloud data centers is bound to consume a lot of power.Therefore,under the premise of ensuring the Qo S of cloud computing services,it is crucial for cloud computing service providers to control the energy consumption of data centers within a reasonable range.Server consolidation is an effective means to reduce energy consumption and increase physical host utilization.However,in the case of sparse computing tasks,the consolidation ratio will be too high,and in the case of a large number of computing tasks,the Qo S of the cloud computing service will not be effectively guaranteed.Therefore,this paper achieves the relative balance between cloud service quality and energy consumption through a combination of virtualization technology and server consolidation technology.The main work of this paper is as follows:1)Building a dual-manager consolidation model for heterogeneous cloud environments for server consolidation issues.The model combines a centralized master manager with a self-heuristic slave manager.The master manager controls the scheduling of global operations,and the slave manager assists in executing the consolidation instructions.2)Based on the dual-manager consolidation model,two host detection algorithms and three virtual machine selection algorithms are proposed.In the host detection phase of the consolidation process,an Empirical Forecast Algorithm(EFA)and a Newton Interpolation Algorithm(NID)are proposed for the host load state and the CPU utilization threshold respectively.In the virtual machine selection phase,the Maximum CPU Usage Selection(MCU),Minimum Memory Usage Selection(MMU)and Maximum Weight Priority Selection(MWP)algorithms are proposed by considering the CPU and memory resource usage.3)Design a virtual machine placement strategy(VMP)that can optimize the performance of the data center system twice.The placement scheme adopts an iterative migration and a package migration placement strategy for the overloaded and underloaded two abnormal physical machines respectively.4)The Ef-algorithm group and the Ni-algorithm group are obtained by implementing the proposed algorithms for each sub-phase of the consolidation problem.In order to test the effectiveness of these consolidation algorithm,this paper carried out simulation experiments on the Cloud Sim platform,and selected the Planet Lab dataset to simulate the use of virtual machines.Simulation experiments show that the proposed method has certain advantages over traditional methods.