基于萤火虫群优化的虚拟机放置方法

Approach of glowworm swarm optimization based virtual machine placement

利用虚拟机放置策略对云数据中心的物理资源利用效率进行优化十分必要。提出了基于萤火虫群优化的虚拟机放置(glowworm swarm optimization based VM placement,Gso-wmp)方法。GSO-VMP方法将物理主机的处理器使用效率表示为荧光素值,当一个虚拟机被放置到物理主机上时,该物理主机的荧光素值都要进行更新;能够在局部径向范围内搜索到更多的可用物理主机,完成虚拟机放置,减少了虚拟机的迁移次数,从而间接地节省了物理主机的能量消耗。使用CloudSim作为GSO-VMP的仿真环境进行仿真,实验结果表明,GSO-VMP方法使得云数据中心的能耗降低、多维物理资源利用率提高。

In a cloud data center,one of the most important problems is using novel virtual machine placement strat-egy to promote the physical resource utilization.An approach of glowworm swarm optimization based virtual ma-chine placement for cloud data centers called GSO-VMP was proposed.In the virtual placement,GSO algorithm was used to find a near-optimal solution.Each physical host had a luciferin value which represented the available CPU utilization.Whenever a VM was placed to a physical host,luciferin value of this physical host was updated.GSO-VMP algorithm could search the more available physical host within local range and thus the virtual migration numbers had been decreased and low energy consumption had been obtained.GSO-VMP had been evaluated using CloudSim with real-world workload data.The experimental results show that GSO-VMP has good performance in resource wastage and energy consumption.