基于差异数据的块级数据库备份系统

Design and implementation the Block-level backup for database based on differential data

针对传统数据库备份系统中备份方法存在的重复数据过多、磁盘存储空间利用率低等问题,本文提出了一种基于块差异数据的数据库备份系统.该系统通过实现基于rsync算法的块差异备份,将备份文件划分固定大小的块并利用索引表对块进行排序,Hash表对块进行标识,检测相邻时间点备份文件间及备份文件内部的重复数据,并利用块差异数据备份文件和块差异指示文件实现备份数据的差异传输.实验结果表明,与传统数据库备份系统的备份方法相比,该系统的块差异备份方法消除了各时间点备份数据中的重复数据,从而有效地提高了存储空间的利用率.

In order to improve the utilization of disk storage and reduce the redundancy of backups, a Block-based Differential Database Backup System is proposed. The block-based differential backup method in the system is based on rsync algorithm, which divides database backups into fixed size units, while takes advantage of hash tables and sort tables for quick information recording and locating, so as to make it able to detect duplicated data among blocks of the inner of backups and of backups of different time-points, and differential data transmission is implemented by using block differential data backup file （bddbf） and block differential indication file（bdif）. The experiment exposes that the block-based differ- ential backup method, compared with backup methods of traditional database backup system, eliminates data duplications among database backups in different time-points, therefore it utilizes the disk space more efficiently.