基于负载趋势预测的虚拟机动态迁移策略研究

Research on Dynamic Migration Strategy of Virtual Machine Based on Load Trend Prediction

虚拟机动态迁移技术是云计算环境下实现负载均衡和绿色节能的有效方法。迁移时机的确定、迁移对象和目标主机的选择是实现虚拟机动态迁移的关键所在。采用二次指数平滑法结合加权概率算法设计出基于负载趋势预测的虚拟机动态迁移策略。该策略能够有效避免因短期负载上升造成的无谓迁移,减少迁移次数,同时也避免了迁移时群聚效应的出现,能有效实现云数据中心的负载均衡。仿真实验表明,该策略在负载均衡、减少迁移次数方面有一定的改善。

Dynamic migration technology for virtual machine is an effective way to achieve load balance and energy saving in cloud computing. The key to realize the dynamic migration of virtual machine is the determination of the time of migrating, the choice of migrating object and the target host. By adopting the second exponential smoothing method and weighted probability algorithm, we design the dynamic migration strategy of virtual machine based on load trend prediction. The strategy can effectively avoid the unnecessary migration caused by short-term load rise, reduce the frequency of migrating, but also avoid the emergence of a critical mass of migration, which can effectively achieve the load balance of cloud data center. The simulation test shows that the strategy can improve the load balance and reduce the frequency of migrating.