基于生存期的云存储元数据缓存算法  

Metadata caching algorithm of cloud storage based on life cycle

作　　者：牛德姣[1] 蔡涛[1] 詹永照[1] 鞠时光[1] 

机构地区：[1]江苏大学计算机科学与通信工程学院,江苏镇江212013

出　　处：《江苏大学学报（自然科学版）》2012年第6期678-683,689,共7页Journal of Jiangsu University：Natural Science Edition

基　　金：广东省自然科学基金资助项目(S2011010006118);江苏省高校自然科学基金资助项目(09KJB520001);江苏大学高级人才启动基金资助项目(09JDG038);江苏省研究生创新基金资助项目(1221170029)

摘　　要：针对元数据管理子系统成为云存储中性能瓶颈的问题,研究了云存储元数据缓存算法.在分析元数据被访问特性的基础上,提出了元数据缓存生存期的概念;依据云存储的特性设计了元数据缓存生存期的计算规则,给出了基于生存期的元数据调出策略和元数据缓存写回策略,提高了云存储元数据管理的效率;分析了基于生存期元数据缓存算法适应用户访问特性的能力,讨论了使用基于生存期元数据缓存算法后如何保证元数据一致性的问题;使用基于生存期元数据缓存算法,实现了云存储元数据缓存原型系统,并使用通用数据集和测试工具进行了测试与分析.结果表明,该算法能提高云存储15%的I/O速度和16%的操作处理速度.To solve the bottle-neck caused by metadata management subsystem in cloud storage,a metadata ca-ching algorithm of cloud storage was proposed.Based on the characteristics analysis of accessed metadata,the concept of metadata caching life cycle was determined.The calculating rules of metadata caching life cycle was designed,and the metadata caching algorithms of metadata substitution and write-back were presented based on the life cycle definition.The metadata management efficiency of cloud storage was greatly improved by the caching algorithms.The metadata caching algorithms adaptability to user accessing characteristics was analyzed,and the metadata consistency guaranteed by the algorithms was discussed.A metadata caching prototype was realized based on the proposed algorithms.Analysis and experiments were completed on the benchmark dataset by general test tools.The results show that the proposed algorithms can effectively enhance I/O rate by 15% and operation processing rate by 16% in cloud storage,respectively.

关 键 词：元数据 云计算 分布式计算 文件组织 缓存 

分 类 号：TP334.5[自动化与计算机技术—计算机系统结构]