Query Preserving Relational Database Watermarking

Abstract

In order to preserve the query results after watermarking relational data, it is necessary to keep the semantic value of data intact during watermark embedding process. A query preserving relational database-watermarking scheme is proposed in this paper. For watermark embedding, we use alphanumeric data as new embedding channel. The scheme retains the semantic meaning of data after embedding, which makes this a distortion free and query preserving technique. Watermark Embedding is done by adjusting the case of securely selected text data. 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 data alteration. To make this scheme more secure tuples and attributes selection is done using a secret key known only to the owner. Later the same key is used in detection process. Moreover, there is no need of original data for watermark detection so it is a fully blind scheme. The method proved (through experiments) to be robust against various kinds of attacks. An SQL Server database implementation has shown that our algorithms can be used successfully in real world applications.