Enhancing Relational Database Security by Metadata Segregation

Although many prominent Relational Database Management Systems provides inbuilt security controls and mechanisms, the information resided in the data-store are at great risk. This research aims to reduce the risk of unauthorized data access by providing an extra layer of security. This research proposes a novel method for incorporating information security while designing the relational database by segregating the information on the basis of its sensitivity level and creating referential integrity constraints dynamically at run time. Different techniques to identify and quantify sensitive attributes and restructuring database architecture have been discussed for the proposed approach. The primary keys of the restructured tables and most critical information attributes were secured using Transparent Data Encryption utility provided by Oracle 11 g to prohibit illegitimate use of information. The performance of the proposed architecture was evaluated with 1,000,000 records which shows that by increasing the number of records, the response time of Select statement increased dramatically whereas it increased gradually for Insert, Update and Delete operations.