摘要
尽管Linux的ext2文件系统在处理较小文件时有很高的效率 ,但对大文件的读写却并不理想 .在分析了ext2内部实现的基础之上 ,指出了ext2采用的inode结构及块分配策略在处理大文件上的不足 .为了提高其处理大文件的效率 ,通过以下两种方法来加以改进 :a .利用B+ 树来组织文件数据块和空闲数据块 ;b .利用连续块分配策略来进行文件的扩展 .最后 ,从理论上论述了改进后的ext2在处理大文件上比原有的文件系统有着较好的读写效率 ,并给出实验数据加以佐证 .
After analyzing the implementation of ext2, this paper points out that the inode structure and block allocation scheme used by ext2 have shown some deficiency when dealing with the large files. In order to improve the performance on large files, two ways are adopted. They are: a. using the B + tree to organize the file data blocks and free blocks and b. using the contiguous block allocation scheme to scale the files. The paper addresses itself theoretically that the improved ext2 has a better I/O efficiency than the former one, proved by some experimental data.
出处
《华中科技大学学报(自然科学版)》
EI
CAS
CSCD
北大核心
2002年第5期1-3,共3页
Journal of Huazhong University of Science and Technology(Natural Science Edition)
基金
国防科技预研基金资助项目 .