面向虚拟机镜像的改进相似度分组去重优化方法  

An improved similarity grouping deduplication optimization method for virtual machine images

作　　者：梁小宇[1] 陈宁江[1] 闫承鑫 刘文斌 

机构地区：[1]广西大学计算机与电子信息学院,广西南宁530004

出　　处：《广西大学学报（自然科学版）》2017年第6期2154-2162,共9页Journal of Guangxi University（Natural Science Edition）

基　　金：国家自然科学基金资助项目(61363003;61063012);国家科技支撑计划项目课题(2015BAH55F02)

摘　　要：根据实验研究发现,云计算环境中虚拟机镜像备份之间存在大量的冗余数据。尽管传统的去重方法有较高的去重率,但需要花费大量时间,这对有时效性要求的海量镜像备份场景并不适用。考虑虚拟机镜像之间存在大量相同或相似的操作系统和应用程序等相似性特性,提出一种基于虚拟机镜像改进相似度分组去重优化方法。该方法利用镜像之间的相似性对镜像进行相似度分组,把相似度达到阈值的镜像归为一组,形成多个相似的镜像组。仿真实验验证了该方法减少去重过程中的索引空间范围,缩短去重的时间,提高了备份效率,特别适用于海量虚拟机镜像的快速备份场景中。According to the existed experimental results,there is a large amount of redundant data between the backups of the virtual machine image. Although the traditional deduplication method has a higher deduplication rate,it takes a lot of time. This does not apply to the massive virtual machine image backup scene with timeliness requirements. Considering the similarity feature with a large number of identical or similar operating systems and applications between virtual machine images,a deduplication method based on improved similarity grouping for virtual machine images is proposed. The method uses the similarity of images to segment images into similar groups,then the imageswhose similarity reach a certain threshold are classified into different sets,finally a number of similar image groups are formed. Simulation experiments show that the proposed method reduces the index space, shortens the deduplication time, and improves the backup efficiency during the deduplication process,which is especially suitable for fast backup of virtual machine images.

关 键 词：虚拟机镜像 重复数据删除 数据备份 相似性 

分 类 号：TP311[自动化与计算机技术—计算机软件与理论]