Deduplication-friendly watermarking for multimedia data in public clouds

Abstract

To store large volumes of cloud data, cloud storage providers (CSPs) use deduplication, by which if data from multiple owners are identical, only one unique copy will be stored. Deduplication can achieve significant storage saving, benefiting both CSPs and data owners. However, for ownership protection, data owners may choose to transform their outsourced multimedia data to “protected formats” (e.g., by watermarking) which disturbs deduplication since identical data may be transformed differently by different data owners. In this work, we initiate research of resolving the fundamental conflict between deduplication and watermarking. We propose DEW, the first secure Deduplication-friEndly Watermarking scheme which neither requires any interaction among data owners beforehand nor requires any trusted third party. Our key idea is to introduce novel protocols which can ensure that identical data possessed by different data owners are watermarked to the same “protected format”. Security analysis and experimental evaluation justify security and practicality of DEW.