南京理工大学学报（自然科学版）Issue(5)：619-624,6.DOI:10.14177/j.cnki.32-1397n.2015.39.05.018
基于预拷贝的虚拟机内存状态动态迁移优化
Memory state transfer optimization for VM live migration based pre-copy approache
摘要
Abstract
As the pre-copy approaches of live VM migration have issues of low copy efficiency,long total migration time and large number of transferred pages,a method of VM memory state transfer op-timization based the pre-copy approaches is proposed here. The size of migration window is determined by the link bandwidth dynamically. The dirty probability of each dirty page is calculated in accordance with the temporal locality and spatial locality principles, and the pages to be transferred are determined by the dirty probabilities. The method can predict the dirty probability more fully and avoid unnecessary retransmission as much as possible. The experiment results show that the optimized approach is better than the existing pre-copy approaches in terms of the downtime, the total migration time and the number of transferred pages.关键词
虚拟机/预拷贝/动态迁移/脏页率Key words
virtual machines/pre-copy/live migration/dirty probabilities分类
天文与地球科学引用本文复制引用
费薇,李淑英,衷宜,徐建..基于预拷贝的虚拟机内存状态动态迁移优化[J].南京理工大学学报（自然科学版）,2015,(5):619-624,6.基金项目
国家自然科学基金(61300053) （61300053）
