Content Authentication And Rights Protection Of Relational And XML Data Using Digital Watermarking

Use of digital data in relational and xml form, particularly for storage, exchange and distribution has increased significantly. Particularly over the internet, this exponential growth is unprecedented. Data in these two forms, of course, facilitates data owners to easily maintain and dispense their data besides providing users the ease of access and exchange. However, like other types of digital contents, relational and xml data also face security related threats particularly when the data resources are outsourced. Among other challenges in this context, two important concerns are about copyrights protection and content authentication. Digital watermarking has been used as a key technology to protect owners’ rights and to verify integrity of outsourced digital data.A watermark is a concealed message inside host digital data that can be used effectively for rights protection and integrity verification of digital contents. While there is a long history of using watermark for multimedia contents, it faces serious challenges when it comes to relational or xml data. These two formats of data have very limited room for hiding something inside as watermark. Furthermore, their usability requirements does not allow, most of the time, to embed even a single bit as watermark just by altering the original contents.In this thesis, we analyze these issues by conducting a thorough study that explores existing works and their shortcomings. As an outcome of this study, we propose new solutions to problems regarding authentication and rights protection of relational and xml data. We particularly target those issues that, previously, have not been addressed adequately. Our solution set employ digital watermarking technology to solve these problems. In this work we show that distortion free and query preserving watermarking method can be deployed for relational and xml data copyrights protection and integrity verification separately. We first present a robust query preserving watermarking scheme for relational data that does not depend on the presence of numeric attributes in the relation. The solution retains the semantic meaning of data after embedding, which makes this a distortion free and query preserving technique. Our proposed method provides better relational database watermarking solution for the database, which either has no numeric attribute, or has data with zero resilience to alterations caused by watermark embedding.We then explore a novel semi-fragile watermarking solution for relational data, which besides tamper detection, allows authorized users to modify their data that needs frequent updates, without requiring re-watermarking. Watermark embedding is distortion free, which helps to retains semantic meaning of data. Additionally, group based embedding ensures the localization of tampering.Finally, we present a combined rights protection and integrity verification solution for xml documents. Two watermarks are embedded in host xml document; one robust and the other fragile watermark, each for ownership verification and tamper detection. Watermark embedding is distortion free, imperceptible and secure with blind detection for both the watermarks. Experiments show that robust watermark can survive common attack and fragile watermark can detect even minor tampering in xml document.We also analyze our solution both theoretically and experimentally. The results of this analysis show that our solution set provides practically applicable mechanism for copyrights protection and integrity verification of relational data and xml documents.