Research And Implementation Of Data Centers Resource Dynamic Scheduling Method

With the rapid development of cloud computing,more and more public clouds and private clouds are built up,which has caused many cloud service providers to focus on building data centers.The number and scale of data centers are showing a trend of high speed growing,which brings the most direct impact that is the energy consumption of data centers are becoming serious.The energy consumption problem is mainly caused by the resource utilization of servers,which is caused by the low load or idle state.In view of the above problems,this paper analyzes and studies the problem of resource scheduling in data centers.Firstly,the object of data center resource scheduling is analyzed,and the distribution and the on-line migration of virtual machine are analyzed.Secondly,the performance of the resource scheduling is analyzed comprehensively,and the resource scheduling is considered from the perspective of the tradeoff between energy consumption and performance.Thirdly,through the research on the technical features of the existing cloud platform and the process of resource scheduling,the design of dynamic resource scheduling framework is proposed.Finally,this paper presents a data center resources dynamic scheduling method and gives a detailed description of its three steps.In this paper,through the online virtual machine migration by using virtualization technology,the data center resources dynamic scheduling method integrates resources between the different physical servers and then shut down idle servers to achieve the goal of scheduling.By means of the method of energy saving and resource load balancing and minimizing migration targets,it focuses on a unified consideration,mainly solves when is triggered to migrate,target virtual machines and servers selection problems.In the process of solution,it uses multi resources factors comprehensive determination and continuous variation of the threshold with load forecasting methods to complete delaying trigger of migration,and combined with the saturation factor of the data center,the number of physical servers in the running state is considered from the overall point of view,and the unified scheduling is realized.