Relational Databases Watermarking Technique Based on Randomized String Verification

Abstract

In the era of digital data transmission, relational databases also transmitted through internet for the web base environment and many other important areas, the unauthorized changes to database can lead to sarcastic losses for receiver of the Database. Owners of the database are need to protect and verify originality of their database. Digital watermarking is an effective technical solution to protecting the copyright of databases from wildcat copying by using the built-in properties of relational databases. In this paper, we propose a relational database watermarking scheme that partition the whole database in groups by using the tuple hash value of the database records. After the group formation, we randomize all the tuples of the group and fuse the use of hash byte function over the group tuples along with owner id to form a group string. Similarly, in detection process, we form a string to verify the suspicious watermarked database. The proposed scheme does not require any additional storage in the original database for copyright verification so the approach is robust against the any distortion in original database. Two experiments, Cast comparison and displacement attack conducted over watermark creation and detection algorithms with the group size variation and the results analyzed to show the running cost of the proposed scheme.