摘要
研究了云计算系统现有任务调度和运行管理机制,指出在云计算环境下,基础性支撑软件虚拟机和虚拟机管理器在长期运行过程中均会表现出老化现象,从而降低了其性能,进而影响了任务的调度和执行效果。在一定程度上影响了服务质量(QoS)。针对这种情况,提出了一种老化感知的任务调度和运行管理框架,该框架能在任务调度和运行时感知底层虚拟机和虚拟机管理器的老化状况,并通过及时再生手段恢复虚拟机性能,从而降低老化所带来的负面影响,提升系统可用性。基于CloudSim仿真工具,在该框架下实现了Max-Min调度算法,并对比说明了该框架的有效性。
The study analyzed the current task scheduling and management mechanism of cloud computing systems, and pointed that in cloud environments, aging of the basic software of virtual machines and virtual machine monitors oc- curs in their long time working, leading to their performance degradation, further, to the decline of the task sched- uling effect, and finally the quality of services （QoS）. In view of this problem, an aging-aware task scheduling and executing framework for cloud computing was proposed. The framework can sense the aging status of virtual ma- chines and virtual machine monitors during task scheduling and executing, and reduce the negative effect of soft- ware aging by rejuvenating the performance of virtual machines to improve the availability of cloud computing sys- tems. The task scheduling algorithm of Max-Min was realized under the framework based on the simulation tool of CloudSim to verify the effectiveness of the framework.
出处
《高技术通讯》
CAS
CSCD
北大核心
2015年第3期270-278,共9页
Chinese High Technology Letters
基金
973计划(2011CB302605)资助项目
关键词
云计算
虚拟机
软件老化
任务调度
资源管理
cloud computing, virtual machine, software aging, task scheduling, resource management
作者简介
男，1984年生，博士；研究方向：分布式计算，云计算；E-mini：liyan@ncic．ac．cn
通讯作者，E-mail：lxr@cert．org．cn
