Transparent Data Deduplication in the Cloud

摘要: Cloud storage providers such as Dropbox and Google drive heavily rely on data deduplication to save costs by only storing one copy of each uploaded file. Although recent studies report that whole file can achieve up 50% reduction, users do not directly benefit from these savings-as there is no transparent relation between effective the prices offered users. In this paper, we propose a novel solution, ClearBox, which allows service provider transparently attest its customers patterns (encrypted) it storing. By doing so, ClearBox enables cloud verify space their occupying in cloud, consequently check whether they qualify for benefits price reductions, etc. secure against malicious rational provider, ensures files be accessed legitimate owners. We evaluate prototype implementation using both Amazon S3 back-end storage. Our findings show our solution works with APIs provided existing without any modifications achieves comparable performance solutions.