Analysis of Two Watermarking Protocols Based on Commutative Encryption

We analyse two watermarking protocols proposed by Wang, Leung, Cheung and Wang, and by Cheung, Leung and Wang. These protocols claimed to provide fair content tracing using watermarking schemes and commutative encryption, whereby a content provider can trace an illegal copy to a customer that distributed this copy. Simultaneously, an honest customer is assured that he or she will not be falsely accused of illegal content distribution. We show that the two protocols contain flaws by examining how a dishonest content provider can frame an honest customer by falsely accusing the customer of illegally distributing content. Similarly, we show how a dishonest customer may possibly redistribute the digital content without being traced by the content provider. Therefore, the protocol does not fulfill the aims as claimed. We further put forward our thoughts on how best to avoid these security weaknesses.