Watermarking Relational Databases

Purdue University, West Lafayette IN

Investigators

Abstract

Watermarking is the technique of embedding un-detectable information into objects (e.g. images and video) mainly to protect the data from unauthorized duplication and distribution by enabling provable ownership over the content. Whereas considerable work has been invested in this topic, little has been done to enable the same concept in the area of relational databases. There is much to be gained from the ability to watermark relational databases, in particular considering current mainstream migration of business interactions towards distributed computing technologies. Effective watermarking can decrease the reluctance of vendors to provide public availability or sharing of valuable data. Compared to multimedia, watermarking in the area of relational data presents a whole new set of challenges. Relational data are characterized by value lying both in the actual values stored as well as the organization of these data. A major challenge for watermarking relational databases is simply the "lack of bandwidth", deriving from the inherent lack of a major noise component in that domain. To further add to the challenge, these constraints are data and application dependent. This project explores the issue of watermarking valuable outsourced relational data. In particular, it addresses: the design of a resilient watermarking method, a proof-of-concept implementation for numeric relational data, deployment on real outsourced commercial data, and an evaluation of the resilience of the algorithm.