针对目前常用的连续数据保护系统仅直接备份变化数据,从而需消耗大量存储空间的问题,设计并实现了一种基于TRAP-4的连续数据保护(CDP)系统。该系统通过卷过滤驱动捕获用户对数据卷的修改,并使用TRAP-4的方式对变化数据进行计算压缩,最终将压缩后的数据备份至备份中心;恢复时只需对压缩后的数据进行逆向解压重组便能将数据卷恢复至任意时间点。实验结果表明,在卷分块大小固定的情况下相对目前常用的CDP系统实现方法,该系统可有效的节省存储空间;并且,随着分块大小的增大以及修改文件的减小,该系统将进一步减少存储空间使用量。
Since the common continuous data protection system just backups the modified data directly and consumes a large amount of storage space, this paper presented a continuous data protection system based on Timely Recovery to Any Point-in-time 4 (TRAP-4). This system captured the modified data from the user by volume filter driver, and the data would be backed up to the backup center after calculation and compression. The recovery process can recover the data volume to any time-point through reverse decompressing and reorganizing the compressed data. Experiment shows this system can effectively save the storage space compared to common method. And, as the block size increases and decreases to modify the file, the system further reduces the storage space usage.