一种基于蚁群算法的云存储副本动态选择机制研究

发布时间：2018-01-15 10:04

本文关键词：一种基于蚁群算法的云存储副本动态选择机制研究　出处：《计算机应用研究》2015年11期 　论文类型：期刊论文

【摘要】：针对云存储技术中副本选择优化问题,提出一种基于蚁群原理的云存储副本动态选择算法。构建基于蚁群的副本动态选择模型,建立副本选择度量标准(如带宽占用、网络路径时延和平均访问时间等)与蚁群信息素的映射,并对虚拟机实例负载状况和虚拟机集群资源利用状况进行量化评估,感知所监控的云节点的资源度量情况;最后利用副本信息素概率计算式得到一组选择副本资源的最优解,最终达到优化负载均衡的目的。经Open Stack模式的云平台对新算法仿真实现,实验结果表明新算法成功实现了副本的有效分发和虚拟机集群的负载均衡,与Round Robin和Server Load算法相比,新算法有更好的负载均衡效果。
[Abstract]:Aiming at the replica selection optimization problem in cloud storage technology, a dynamic selection algorithm of cloud storage replica based on ant colony principle is proposed, and a replica dynamic selection model based on ant colony is constructed. Build replica selection metrics (such as bandwidth occupancy, network path delay, average access time, etc.) and map them to ant colony pheromones. The load status of virtual machine instance and the resource utilization of virtual machine cluster are quantitatively evaluated to perceive the resource measurement of the monitored cloud node. Finally, a set of optimal solutions for selecting replica resources are obtained by using the probability formula of replica pheromones. Finally, the purpose of optimizing load balancing is achieved. The new algorithm is simulated on the cloud platform of Open Stack mode. Experimental results show that the new algorithm has successfully implemented the effective distribution of replica and load balancing of virtual machine cluster, compared with Round Robin and Server Load algorithm. The new algorithm has better load balancing effect.
【作者单位】：河南大学软件学院;
【基金】：国家自然科学基金资助项目(U1304615) 河南省科技厅基础前沿项目(132300410149,122300410170) 河南省教育厅科学技术研究重点项目(14B520003,12A520009) 河南省博士后资助项目(2013-8)
【分类号】：TP333
【正文快照】： 0引言随着云计算的普及,云存储逐渐成为一种无缝、集成的资源共享应用技术。在云存储中,数据中心需要将需求量大的资源复制到多个云存储物理设备上,再以虚拟主机为基本单位向外提供服务。由于这些资源极其庞大,且虚拟主机动态变化,加之网络负载的不确定性和物理服务器处理能力

【参考文献】