Skip to Main Content
Watermarking databases is an emerging technological interest, especially within the frame of a vastly growing number of publicly available web-databases. In order to ensure integrity, detect malicious modification and protect ownership rights, we propose a watermarking algorithm based on parameterized tuple partitioning and whitespaces, using a public watermark. The watermarking scheme is non-intrusive, resilient, blind, reversible and suitable for databases of any size with reasonable performance on embedding and extraction. Moreover, we emphasize locatability of malicious modifications within the scope of predefined tuple sets, and support incremental watermarking to cope with the dynamic nature database systems are subject to.
Date of Conference: 2-5 Oct. 2012