Research On Virtual Resource Evaluation And Scheduling Technology For Load Balancing In Cloud Computing

With the rapid development of cloud computing technology, there have been moreand more large-scale data centers providing cloud computing services. Cloud computingservice providers can create multiple virtual hosts on a single node throughvirtualization technology. Each virtual host has its own operating system andapplication environment, so that users can get their application environmentconveniently. The number of virtual host on physical host increases as demandincreases. The load of physical host was diversity changes because of every virtual hostrunning different applications, so that there will inevitably be a single physical machinewhose load is too high to meet the demand of users. Scheduling virtual host effectively,which is important to achieve dynamic load balancing of the entire data center andimprove resource utilization to meet the needs of users, becomes a research focus.On this situation, we propose a new solution for virtual resource load evaluationand load balancing in the cloud computing environment. Experiments show that themethod can evaluate the load of virtual resource accurately and schedule virtualmachine efficiently. This paper’s work and innovations are summarized as follows:(1) Propose a method of load evaluation based on fractal theory. The strategy fullyconsiders the characteristics of the physical node load in cloud computingenvironment; predicts load changes according to the node load historical. Theexperiments show that the strategy can forecast node load, determine the timeof virtual machine migration and improve the efficiency of the system byreducing the number of unnecessary virtual machine migration.(2) Propose a dynamic scheduling strategy based on load evaluation of virtualresources. The strategy summarizes the inadequacies of traditional loadbalancing algorithm in cloud computing environment and judges the time ofvirtual machine migration through the load evaluation strategy based on fractaltheory. It can schedule high load virtual machines to a lightly loaded physicalnode. The experiments show that the stratege can be effective in migrating ofvirtual machines, make a package of different load types of virtual machine toachieve load balancing of the system.(3) Deployment of cloud computing platforms; Program load monitoring module,load evaluation module, virtual machine scheduling module and test thecorrectness and effectiveness of the strategy through simulation experiments.