论文部分内容阅读
Cloudlet利用虚拟机来灵活处理来自不同移动设备的云应用服务访问请求,避免了移动设备直接访问云应用服务过程中的广域网延迟和带宽限制问题.但是基于虚拟机实现的Cloudlet,存在合成虚拟机耗时长的问题,对于用户体验影响很大.针对这个问题,本文提出一种基于统计预测的调度机制,该机制使用了两种预测模型,分别是基于Cloudlet切换统计模型和基于关键位置的统计预测模型,来预测移动中的用户将会使用的Cloudlet.然后,根据预测信息实现Cloudlet上应用服务虚拟机预先调度合成,以降低用户请求等待时间.实验表明,与其它方法相比,基于本文预测机制的方法可以有效降低用户的请求等待时间,提高用户体验.
Cloudlets utilize virtual machines to flexibly handle cloud application service access requests from different mobile devices, avoiding the WAN latency and bandwidth limitations of mobile devices accessing cloud application services directly. However, virtual machine-based Cloudlets have synthetic virtual machine consumption Which has a great impact on the user experience.To solve this problem, this paper proposes a scheduling mechanism based on statistical prediction, which uses two kinds of prediction models, which are respectively based on the Cloudlet switching statistical model and the key position based statistical prediction model To forecast the Cloudlet which the mobile users will use.Then, based on the forecast information, the application service virtual machine on the Cloudlet can be pre-scheduled to reduce the waiting time of the user request.Experiments show that compared with other methods, based on the prediction mechanism of this paper The method can effectively reduce the user’s request latency and improve the user experience.