基于二元组索引的低内存开销快速文件管理方法

Low Overhead Memory and Fast File Management Method Based on Binary Group Index

在航天器上采用FLASH设备作为存储介质,需要采用文件系统来管理存储设备.在航天嵌入式系统中,数据存储的规模越来越大,进行采集任务产生的数据量将达到TB级别.出于安全性的考虑,航天器嵌入式系统内存容量受限,并且实时性要求极高.普通嵌入式系统在内存容量和实时性的要求同航天嵌入式系统相差较多,因此不适用于航天嵌入式系统.本文针对航天器系统资源受限的特殊需求,对嵌入式系统中常用的YAFFS2文件系统中存在的内存中文件索引结构占用空间随文件数据量线性增长的问题,提出基于二元组索引的低内存开销快速文件管理(Low Overhead and Fast File Management based on Binary group,BLOF)方法,通过基于B+树引入二元组链表结构对连续FLASH页进行统一索引,并依据索引结构设计文件读、文件写操作算法,具有文件系统内存占用小,读写文件速度快的特点.实验结果表明,BLOF相比于YAFFS2索引方法,对任意数据量的文件:1)当文件数据所在闪存页达到平均6个页以上连续时,BLOF方法内存占用低于YAFFS2方法,最高节省55.55%;2)在文件规模大于512KB时,BLOF方法的文件读写时间始终小于YAFFS2的文件读写时间,并且YAFFS2文件写时间和BLOF写时间之差随文件大小呈线性增长趋势.BLOF方法更适合于文件数据量大且存储在大量连续页的情境下.

Flash equipment is used as the storage medium on spacecraft,and file system is needed to manage the storage equipment.In the space embedded system,the scale of data storage is becoming larger and larger,and the amount of data generated by the acquisition task will reach the level of TB.For the sake of security,the memory capacity of spacecraft embedded system is limited,and the real time requirement is very high.Common embedded system is different from aerospace embedded system in memory capacity and real time requirements,so it is not suitable for aerospace embedded system.In order to meet special requirements of limited spacecraft system resources,and to resolve the problem that the space occupied by file index structure in memory increases linearly with the amount of file data in YAFFS2 which is a commonly used embedded file system,a new index method called BLOF(low overhead and fast file management based on binary group)is proposed in which continuous FLASH pages are indexed uniformly by introducing binary group structure based on B+tree,and the file read and file write algorithms according to the index structure are designed.So that the file system can implements fast file reading and writing,and low memory footprint.The experimental results show that compared with the YAFFS2 index method,BLOF is more effective in the following cases:1)when the number of FLASH pages for storing file data reaches an average of six or more consecutive pages,BLOF takes less memory footprint than YAFFS2,and maximum saving is 55.55%;2)when the file size is greater than 512KB,the file reading and writing time of the BLOF method is always less than that of YAFFS2,and the value of YAFFS2 file writing time minus BLOF file writing time increases linearly with the file size.BLOF method is more suitable for the situation where the file data is large and stored in a large number of continuous pages.