摘要
针对资源虚拟化环境中的混合型负载调度问题,提出一种基于能耗比例模型的虚拟机调度算法.该算法利用处理器的"性能计数器"机制来评估各个虚拟机的近期能耗状态,并采用"最近最小能耗比例优先"的策略进行调度.理论分析给出了该算法的有效性证明和相关特性.实验结果显示,当系统面对混合型负载时,基于能耗比例模型的调度算法在"调度偏差"和"相对能效"两方面明显优于现有的虚拟机调度算法.
To address the issue of scheduling mixed-workloads in resource virtualization environments,a novel virtual machine scheduling algorithm based on energy-consumption ratio model is proposed. In the proposed algorithm,the recent energy consumption status of individual virtual machines are evaluated by using the processor ＇s ‘performance monitor counters＇ mechanism,and the scheduling policy is‘most recent minimal energy-consumption ratio first＇. Theoretical analysis presents the validation and characteristics of the proposed algorithm. Extensive experiments showthat when the system is in presence of intensive mixed-workloads,the proposed algorithm significantly outperforms existing scheduling approaches in terms of scheduling deviation and normalized energy-efficiency.
出处
《电子学报》
EI
CAS
CSCD
北大核心
2015年第2期305-311,共7页
Acta Electronica Sinica
基金
国家自然科学基金(No.61402163)
湖南省教育厅科学研究项目(No.13B015)
湖南省自然科学基金(No.13JJ9022
No.14JJ7071)
湖南省科技计划项目(No.2012GK3075)
关键词
云计算
虚拟机
能耗
调度算法
cloud computing
virtual machine
energy consumption
scheduling algorithm
