摘要
针对虚拟机(VM)迁移过程中,迁移时间和停机时间对用户使用感受的影响,从提高网络服务质量降低网络运行成本角度出发,提出一种基于联邦云马尔可夫链多虚拟机并行迁移策略.首先,对传统的虚拟机顺序迁移过程进行研究,并研究构建出多虚拟机情况下的并行迁移过程,给出其迁移时间和停机时间计算模型.然后,将多虚拟并行迁移构思为一个泊松分布过程,并基于马尔可夫链构建多虚拟并行迁移的数学分析模型.最后,通过实验,给出模型参数对多虚拟并行迁移过程的影响,并且对比了顺序迁移过程的迁移时间、停机时间及网络成本,验证了并行迁移过程在处理大数据迁移中的优势.
According to the influence of users using feel with migration time and downtime in the virtual machine （VM） migration process ,in order to improve the service quality of the network and reduce the cost of running the network, here proposed the parallel migration analysis of multi virtual machine based on Federated cloud Markov chain. Firstly, here studied and constructed the sequential migration model of virtual machine, and presented the migration time and the down time of the model. Then, the multi virtual parallel migration is conceived as a Poisson distribution process, and a mathematical analysis model of multi virtual parallel migration is constructed based on Markov chain. Finally,the effect of the model parameters on the multi virtual parallel migration is presented,and the network cost,migration time and downtime of the sequential migration is compared , and the advantage of the parallel migration is verified.
出处
《小型微型计算机系统》
CSCD
北大核心
2016年第8期1878-1882,共5页
Journal of Chinese Computer Systems
基金
国家自然科学基金项目(61462069)资助
关键词
云计算
马尔可夫链
泊松分布
并行计算
多虚拟机
cloud computing
Markov chain
Poisson distribution
parallel computing
multi virtual machine
