Content characteristics based robust watermarking for relational database: a new approach to database security

Abstract

Digital data such as text, relational database, audio, video and software are intellectual property of creators/ writers/owners. The database services have become easily available and economical since the booming of internet. However, their outsourcing through the internet accompanies multiple threats like copying, modifying as well as unauthorized distribution. Relational Database has a wide-spread use in many real-life applications, hence, it is essential to maintain integrity and provide copyright protection. To counter the threats, watermarking techniques have been playing a vital role since the last decade. As a security measure, Relational Database Watermarking is becoming more popular and strengthened day-by-day. This is also one of the upcoming areas of interest among researchers of the Database Security. A technique earlier used for Image Watermarking is applied to watermark Relational Databases. In Image Watermarking technique, a pixel or a pair of pixels must satisfy certain characteristics. Usually, database watermarking techniques concentrate on hiding a watermark in database. Extraction and matching of hidden watermark with original watermark confirms ownership of database. This paper demonstrates the use of image watermarking technique for relational databases. Here we align some properties of attributes of database by changing some bit(s) in attribute value. Using secret key, we have ensured that values of two attributes of a tuples satisfy some bit-similarity property and to do so, we slightly alter values of attributes. Detection of such characteristic in a database using secret key can be done easily to verify the presence of a watermark.