云环境下基于迁移的虚拟机集群优化算法

针对云环境中虚拟机集群负载不均衡问题，提出一种基于虚拟机迁移的集群优化算法。通过对节点负载的实时监测，动态调整各种资源的权重，根据资源权重选择可最大程度降低主机负载的虚拟机进行迁移。该算法利用预测机制，消除主机资源利用率的临时越界引起的不必要的虚拟机迁移。在选择目标节点时，采用多目标决策法，兼顾多资源匹配率，服务级目标违背率(SLA)等多种管理目标。实验结果表明，与同类型的负载均衡算法相比，该算法能减少迁移次数，降低SLA违背率。

Virtual Machine Cluster Optimization Based on Migration in Cloud Environment

An optimization algorithm based on virtual machine migration is proposed for the load imbalance in virtual machine clusters. The algorithm uses the prediction mechanism to eliminate unnecessary virtual machine migration caused by temporary cross border of host resource utilization. It dynamically adjusts the weights of various resources through real time monitoring of the load of the nodes. It can select the virtual machine which will reduce the maximum degree of the host load to migrate according to the weight of the resources. In selecting the target node, it combines the multi objective decision method with the multi resource matching rate, SLA (service level agreement) violation rate and other management objectives. Experimental results shows that the algorithm proposed enjoys smaller number of migration and lower rate of SLA violation in the premise of load balancing than other load balancing algorithms of the same type.