构建了在线迁移算法模型和虚拟机负载评估向量,并依据各物理节点的带宽占用率、内存占用率和CPU占用率,确定迁移时机和需要迁移的源物理节点,依据CPU在各节点的占有率、ROM的使用效率、带宽的利用率来确定目标物理节点。通过仿真模拟环境试验证明,文中算法在虚拟机在线迁移中能够大幅度地提高物理节点的负载均衡,同时SLA违反率最小和云计算数据中心能耗最小。
Firstly, build an online migration algorithm model and virtual machine load assessment vector, then, according to the bandwidth rate for each physical node, memory usage and CPU utilization to determine migration timing and source physical nodes need to migrate at the same time, according to the CPU in the share of each node, the efficient use of ROM, and bandwidth utilization to determine the target physical nodes. Through simulation environment test, showed that the proposed algorithm in a virtual machine live migration, and can greatly improve load balancing physical nodes, while the SLA violation rate minimum and cloud computing data center power consumption to a minimum.