摘要
研究了云计算环境下的分布式文件系统KFS的系统架构,对于海量数据存储的云存储系统来说元数据管理效率是关键,通过分析KFS文件系统的元数据模型,提出了基于KFS分布式文件系统元数据的改进模型,即利用内存缓冲策略,对待插入的元数据进行预处理并批量插入,减少查找和分裂次数,大大提高了KFS文件系统的数据访问效率.最后通过算法复杂度的分析,证明该改进算法能有效提高分布式文件系统KFS的元数据服务器的效率.同时该改进模型对于采用B+树索引机制来集中管理元数据的类似系统同样适用.
The architecture of distributed file system in large distributed storage systems is studied.For efficient metadata management,a new improved model of metadata management of KFS is presented that caches a array of items in memory which are preprocessed to bulk insert into index tree.It provides efficient access to data due to that the new model can reduce search times and split times of B^ ＋ tree node.Complexity analysis of the algorithm shows that the improved model can provides very high performance of KFS metadata server,and it can be applied to similar file system of B^ ＋ tree index of simple metadata server.
出处
《大连交通大学学报》
CAS
2014年第4期97-101,共5页
Journal of Dalian Jiaotong University
基金
辽宁省教育厅科学研究计划资助项目(L2011077)
