不确定网络负载下虚拟机匹配调度仿真

Virtual Machine Matching Scheduling Simulation under Uncertain Network Load

针对传统的虚拟机匹配调度存在的内存占用率过高、调度过程所造成的系统负载均衡性较差等问题,提出一种不确定网络负载下的虚拟机匹配调度方法。首先构建虚拟机资源匹配调度模型。根据模型对目标网络不确定性负载属性进行分析,其次对于用户任务集合中的各个任务,计算其在虚拟机资源池中每个虚拟机资源上的信任效益函数值,选取效益函数值较小的任务映射到虚拟机上;最后采用遗传算法进行虚拟机匹配调度,依据虚拟机资源配置情况通过多次迭代过程确定最优染色体,得到虚拟机的最佳匹配调度策略。实验结果证明,所提方法可以有效降低内存占用率,提高负载均衡度。

This article puts forward a method to match and schedule the virtual machine under uncertain network load. Firstly, the resource matching scheduling model of virtual machine was constructed. According to the model, the uncertainty load attribute of objective network was analyzed. For each task in the user task set, the trust benefit function of each virtual machine resource in the virtual machine resource pool was calculated. Then, the task with the smaller benefit function value was selected to be mapped to the virtual machine. Finally, the genetic algorithm was used to perform the virtual machine matching scheduling. According to the virtual machine resource configuration, the optimal chromosome was determined through multiple iterations. Thus, the best matching and scheduling strategy of the virtual machine was obtained. Simulation results show that the proposed method can effectively reduce the memory utilization and improve the load balance.