Research On The Strategy Of Load Balancing About Cloud Computing Resources Based On Virtual Machines Scheduling

Cloud computing, which is the integration of parallel computing, grid computing and other technologies, can provide scalable, cheap and distributed computing for users. Since it comes into being, it has attracted many people and has developed rapidly. And the virtualization technology is as an important driving force for the development of cloud computing. The same physical machine is divided into a number of virtual machines in resources to offer the users. Virtualization technology can improve the utilization of resources, enhance the scalability of the system and make the data center dynamic and heterogeneous. However, the problems are also emerging with the rapid development of cloud computing. The problem of load balancing is one of them. The heavy load will have a serious effect on the performance of the application, and even produce a mistake to lead to poor experience for users, while the light load will result in a waste of the system resources and increasing the additional cost in the data center. Good load balancing can play a significant role in the operation efficiency of the node, improving the robustness of the system and increasing users’ satisfaction and so on.Researches of load Balance mostly have done during the initial stage of tasks allocation, while the load balance of resources is few studied during the stage of tasks running. The load balance of resources in cloud computing should be paid attention to always. Only in this way can sustained and stable environment be ensured to bring good experience to users. Researches about scheduling of virtual machines resource are mostly concentrated in the dynamic migration and researches are fewer that achieve the goal of load balance by the way of migration. Dynamic migration of virtual machines will consume a large of system resources. Unnecessary migration not only does not balance the system load, but also will aggravate the load imbalance of resources. So the decision about the migration of virtual machines is particularly important. In addition, it will directly affect the load of the system that which virtual machine is chosen to move out and that which physical node to chosen to move into. And the current researches are relatively simple and the accuracy is not high.In this paper, a load balance strategy about cloud computing resources is proposed based on the scheduling of the virtual machines, which makes use of the migration of virtual machines to make the load of resources balanced during the process of the application in cloud computing, so as to improve the utilization of the system resources and maintain the stability of the system. Three algorithms are achieved in this strategy. First, the algorithm of scheduling decision with multiple thresholds and multiple targets is achieved based on prediction, in which thresholds about physical machines and virtual machines are considered at the same time, and the prediction with historical data is used to filter out the situation of instantaneous shock and finally make the correct decision of virtual machines scheduling:dynamic migration or resource reconfiguration. Second, the algorithm of choosing virtual machines to migrate is achieved, in which the function of the migration benefit is the standard and then heap sort is used to sort the virtual machines and finally the cost-effective virtual machines are chosen to migrate. Third, the algorithm of choosing target node is achieved, in which the improved firefly algorithm is used to achieve the choice of multiple virtual moving into nodes that avoids the local optimal solution and improves the accuracy of the result.