一种云计算环境下基于Xen的虚拟机调度机制

发布时间：2018-02-04 10:17

本文关键词： 调度虚拟机权值资源利用率　出处：《计算机测量与控制》2014年10期 　论文类型：期刊论文

【摘要】：目前Xen下的经典的虚拟机调度算法忽略了服务操作系统的I/O处理能力对系统整体性能的影响,同时未考虑增加延迟敏感型任务的响应速度,为此,设计了一种云计算环境下的改进的Credit调度算法;首先,设计了一种新的Credit值自适应计算方法,然后设计了一种根据服务操作系统的请求处理速度来确定各虚拟客户操作系统和服务操作系统CPU物理时间片的权值确定方法并描述了权值更新时机,最后,对改进的VCPU调度算法进行了定义和描述,在算法中通过优先考虑延迟敏感任务和缓存关联任务作为下一步调度的VCPU以进一步提高算法性能;实验结果表明:文中方法有效实现Xen框架下的虚拟CPU调度,较其它方法相比,将CPU利用率提高了8%以上,是一种有效的适用于云计算的虚拟机调度方法,具有很强的可行性。
[Abstract]:At present, the classical virtual machine scheduling algorithm based on Xen ignores the impact of the I / O processing capability of the service operating system on the overall performance of the system, and does not consider increasing the response speed of the delay sensitive task. An improved Credit scheduling algorithm in cloud computing environment is designed. Firstly, a new adaptive calculation method of Credit value is designed. Then we design a method to determine the weight of the physical time slice of the virtual client operating system and the service operating system CPU according to the request processing speed of the service operating system and describe the timing of the weight update. Finally. The improved VCPU scheduling algorithm is defined and described. In the algorithm, delay sensitive task and cache association task are considered as the next VCPU to further improve the performance of the algorithm. The experimental results show that the proposed method can effectively implement the virtual CPU scheduling under the framework of Xen. Compared with other methods, the CPU utilization rate is increased by more than 8%. It is an effective virtual machine scheduling method for cloud computing and has strong feasibility.
【作者单位】：呼伦贝尔学院计算机科学与技术学院;
【基金】：校级一般科研项目(YJYB201208)
【分类号】：TP302;TP316
【正文快照】： 0引言据IDC公司的市场调查显示:服务器的平均利用率只能达到自身能力的10%~15%[1-2],因此,尽管目前计算机硬件的计算能力非常强大,但其绝大部分资源仍处于闲置状态[3-4],为了进一步提高硬件资源的利用率,虚拟化技术近来已经获得越来越多的重视[5]。虚拟计算的概念由IBM在20世

【参考文献】