基于层次分析法权重和灰色服务器负载预测的云计算on-line迁移策略

On-line Migration Strategy in Cloud Computing Based on AHP Weight and Gray Server Load Predicting

为了克服云计算环境下由于实时用户任务的不确定性到来和服务器性能差异而导致的云计算环境的负载不均衡问题,提出了一种层次分析法权重获取和灰度算法预测服务器负载的云计算on-line虚拟机迁移策略;首先,设计了基于层次分析法(AHP)和灰色服务器预测的虚拟机on-line迁移模型,提出了采用层次分析法获取虚拟机各资源需求权重,然后,采用灰色模型预测下一时刻的服务器负载,采用此权值向量与各无需迁移的服务器的空闲资源向量进行加权得到加权和,将具有最小加权和的物理服务器作为迁移的目标宿主机;最后,定义了基于层次分析法权重和灰色服务器负载预测的云计算on-line迁移算法;在CloudSim环境下进行实验,结果表明文中的迁移策略使得云计算在响应用户任务时,具有任务失败次数少、SLA违约率低和迁移成功率高的优点,同时与其它方法相比,具有负载均衡程度高的优点,具有较强的可行性。

In order to solve the load unbalance problems in Cloud environment such as uncertain arrival of in time user task and the difference of server predicting, an on--line virtual migration strategy based on AHP weight obtaining and gray algorithm to predict the server load at the next time was designed. Firstly, the on--line migration model based on AHP and gray server predicting was proposed. Then the gray model is used to predict the server load at the next time, the weight vector is weight with the server free resource vector to get the sum, and the server with the smallest weight sum was regarded as the goal server. Finaliy, the virtual machine on--iine algorithm based on AHP weight and gray server predicting was defined. The experiment is operated in the CloudSim environment, the result shows the method in this paper has the advantage such as relative litter failure times, low SLA violating rate and high migrate success rate, and compared with the other methods, it has priority of more even load balance, therefore, it is a feasible migration method in cloud environment.