GARWM: Towards a generalized and adaptive watermark scheme for relational data

Abstract

Watermarking relational data is to protect the intellectual property of sensitive and valuable relational data that is outsourced for sale. Based on analysis of the characteristics of relational data and the corresponding watermarking techniques, the relational data watermarking problem is formalized, and a generalized and adaptive relational data watermarking framework (GARWM) is proposed. The GARWM framework exploits the properties of relational data in the semantics of watermarking, such as preservation of logical relationship in usability preserving attack, discrimination in significance of attributes, and local constraints/global metrics, to strengthen existing methods. The insertion/detection algorithms are presented to insert/detect watermarks into/from non-numeric attributes besides numeric attributes. We identify and classify potential threats to the watermark of relational data including simple attacks, detection-disabling attacks, and ambiguity attacks. We show that GARWM is resilient to these threats, and experiment results quantitatively demonstrate the robustness of our techniques. © Springer-Verlag Berlin Heidelberg 2005.