Live Migration for Virtual Machine Based on Kalman Prediction of Dirty Pages

被引:1
作者
Sun, Zhihong [1 ]
Hu, Xianlang [1 ]
机构
[1] Jiangsu Automat Res Inst, Lianyungang, Peoples R China
来源
MECHANICAL COMPONENTS AND CONTROL ENGINEERING III | 2014年 / 668-669卷
关键词
virtual machine; live migration; dirty pages; Kalman prediction;
D O I
10.4028/www.scientific.net/AMM.668-669.1363
中图分类号
TH [机械、仪表工业];
学科分类号
0802 ;
摘要
The live migration of virtual machine (VM) is an important technology of cloud computing. Down-time, total migration time and network traffic data are the key measures of performance. Through the analysis of dynamic memory state of a virtual machine migration process, we propose a dirty pages algorithm prediction based on pre-copy to avoid dirty pages re transmission. Experimental results show that, compared with the Xen virtual machine live migration method adopted, our method can at least reduce 15.1% of the total amount of data and 12.2% of the total migration time.
引用
收藏
页码:1363 / 1367
页数:5
相关论文
共 6 条
  • [1] Chen Y., 2011, J COMPUTERS
  • [2] Clark N., 2005, P 2 S NETW SYST DES
  • [3] Jin H, 2009, P IEEE INT C CLUST C
  • [4] Liang H., 2013, APPL MATH INFORM SCI
  • [5] Michael R., 2009, ACM SIGOPS OPERATING
  • [6] Nathuji R., 2007, P ACM S OP SYST PRIN