云环境下副本优化放置策略研究  

Research on replica optimal placement strategy in cloud environment

作　　者：王鑫 孟雨 覃琴 蒋华 Wang Xin;Meng Yu;Qin Qin;Jiang Hua(College of Computer Science&Information Security,Guilin University of Electronic Technology,Guilin Guangxi 541004,China;School of Marine Information Engineering,Guilin University of Electronic Technology,Beihai Guangxi 536000,China)

机构地区：[1]桂林电子科技大学计算机与信息安全学院,广西桂林541004 [2]桂林电子科技大学海洋信息工程学院,广西北海536000

出　　处：《计算机应用研究》2020年第4期1111-1114,共4页Application Research of Computers

基　　金：广西高校中青年教师基础能力提升项目(ky2016YB150)。

摘　　要：为了提高云计算数据调度和副本访问的效率,对副本策略中的副本放置问题进行研究,提出一种基于蚁群算法的副本放置策略。根据自然界中蚁群觅食的原理,把蚁群算法应用于副本放置的整个过程;利用信息素的动态更新以及拉普拉斯概率分布改进的蚁群算法得出一组最优解进行副本放置。在CloudSim平台上进行了仿真模拟,实验结果表明,提出的方案在平均作业完成时间、网络利用率和负载均衡度上均优于原始蚁群算法,并在一定程度上降低了副本放置的时间消耗和网络负载。In order to improve the efficiency of data scheduling and replica access in cloud computing,this paper studied the replica placement problem in replica strategy and proposed a replica placement strategy based on ant colony algorithm.According to the principle of ant colony foraging in nature,the strategy applied ant colony algorithm to the entire process of replica placement.It used the pheromone dynamic updating and Laplace probability distribution to improve the ant colony algorithm,obtained a set of optimal solutions for replica placement.Simulation was on the CloudSim platform.The experimental results show that the proposed scheme is superior to the original ant colony algorithm in average job completion time,network utilization and load balance,and reduces the copy placement time consumption and network load to a certain extent.

关 键 词：云计算 蚁群算法 副本放置 负载均衡 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]