一种面向感知数据托管的虚拟机调度策略

A Strategy to Schedule Virtual Machines for Sensory-Data-Hosting

以虚拟化为主要特征的IaaS(基础设施即服务)云服务,可实现资源的池化管理和灵活配置.现有虚拟化方案下物联网的感知数据托管,由于在数据规模、更新速度以及应用需求的不同特征,往往会发生负载倾斜和负载不均的现象.针对这类问题,本文提出一种面向感知数据托管的虚拟机调度策略,基于贪心技术实现了2种模式下的调度算法.实验表明,本文方法可以动态调度虚拟机,达到物理集群负载均衡和优化节能的效果.

Cloud services on the basis of IaaS（Infrastructure as a Service）can flexibly manage and allocate IT resources in pools.The current virtualization solutions for the applications of sensory-data-hosting on the Internet of Things always bring about load unbalance due to the dynamic data scale,frequeut update and specific business requirements.In this paper,a strategy to schedule virtual machines is proposed for sensory-data-hosting,and the algorithms against two scheduling patterns are designed in greedy fashion.Our experiment shows that our method is effective in load-balancing and energy saving.