Attribute Level De-duplication Backup System for Relational Database

Abstract

Efficient storage utilization is very important for today’s business organizations to maintain business continuity and reduce the operational costs. Most of enterprises need large amount of data to operate their daily transactions. Data availability and integrity are becoming necessary for such systems for the extension of data retention periods in disk. So, we present the disk-based backup system in which relational database files are stored by using data deduplication technology. Our system applies attributelevel backup strategy in reducing the replicated data among records of a relation (intra-database) and also inter-database. It reduces more redundant values than record-level and file-level de-duplication and other traditional backup methods, especially in storage consumption. We only need lower space for a large amount of files including more redundant data because this system eliminates redundant data not only within a file but also among different files. Our storage mechanism also supports to the purpose of recovery from accidentally damages. We use MySQL data set of one food mart. The result depends mainly on how much data duplication in our data set used. Our result reduces at least 25% in a little amount of duplication and more for larger duplicated data volume.