基于概率预测的虚拟机动态迁移机制改进  

Improvement of Live Migration Mechanism for Virtual Machine Based on Pre-Copy

作　　者：方义秋[1] 陈玉洋 葛君伟[1] 

机构地区：[1]重庆邮电大学计算机科学与技术学院,重庆400065

出　　处：《微电子学与计算机》2016年第11期176-180,共5页Microelectronics & Computer

基　　金：重庆市教委科技项目(KJ090519)

摘　　要：提出在基于概率预测的pre-copy算法的基础上,引入带宽自适应算法,当脏页率突然增大时,带宽自适应算法能够自动提高迭代传输带宽,缩短传输时间.实验结果表明,在脏页率较高和脏页率变化较大时,该策略能够明显缩短迭代传送时间和停机时间,提高迁移性能和业务服务质量.An Adaptive Bandwidth Algorithm is introduced in the pre-copy algorithm based on probability prediction. When dirty pages rate suddenly increases to a larger value, that algorithm could improve the transmission bandwidth to shorten transmission time. Experimental results show that the rate of dirty pages increase suddenly or is great, that strategy can improve the transmission bandwidth, shorten the total time of iterative transfer and downtime, so as to improve migration performance and quality service.

关 键 词：云计算 虚拟机 动态迁移 脏页率 带宽自适应 

分 类 号：TP31[自动化与计算机技术—计算机软件与理论]