Partitioning-Insensitive Watermarking Approach for Distributed Relational Databases

This paper introduces an efficient watermarking approach for distributed relational databases, which is generic enough to support database outsourcing and hybrid partitioning. Various challenges, like partitioning and distribution of data, existence of replication etc., are addressed effectively by watermarking different partitions using different sub keys and by maintaining a meta-data about the data distribution. Notably, the embedding and detection phases are designed with the aim of making embedded watermarks partitioning-insensitive. That means, database partitioning and its distribution do not disturb any embedded watermark at all. To the best of our knowledge, this is the first proposal on watermarking of distributed relational databases supporting database outsourcing, its partitioning and distribution in a distributed setting.