【摘要】 云平台中传统虚拟机部署策略通常是依据单一性能指标来选择目标宿主机,而忽略了承担不同业务的虚拟机对不同资源依赖性的差异,从而导致因资源得不到合理分配而引起的过度迁移现象。针对以上问题,提出了云平台下一种虚拟机的个性化部署机制,该机制通过时间序列预测机制对宿主机的负载进行预测,并结合虚拟机的个性化负载需求,选择并部署在适当的宿主机上。实验结果显示,利用该机制部署虚拟机,在一定程度上能够有效降低虚拟机迁移频率,延长发生首次迁移的时间。更多还原

【Abstract】 The traditional method of deploying virtual machines depends on the single performance indicators to select the target host and ignores the difference of business features revealed by virtual machines.Owing to lack of rational allocation of resources,the phenomenon of excessive migration of virtual machines is often brought.This paper presents a model that automatically deploys virtual machine with personalized indicators in cloud.Personalized load features and load forecasting are used in the mechanism.Experimental results show the effectiveness of the mechanism.更多还原