Research On Similarity-based Secure Data Deduplication In Cloud Computing

With the cloud computing developed rapidly,cloud storage which is considered as an pattern of outsourced storage and centralized management has alleviated the burden of maintaining huge data in local.However,as an increasing number of individuals stored their data on the cloud server,how to maintain the ever-increasing data efficiently has become a bottleneck of cloud storage development.Moreover,the cloud user needs to encrypt their data before uploading to protect the data confidentiality,which makes it more difficult to maintain the cloud data efficiently.Therefore,secure data deduplication,a special type of cryptographic data compression technology has been widely studied by domestic and abroad scholars.However,there is an inevitable defect in the state-of-art secure deduplication scheme.That is,when people perform secure deduplication on the block-level,the server has to look for the duplicate block tag from all tags one by one.Certainly,it costs a huge computation overhead.What is worse,it causes a huge waste of computing resources to perform block-level deduplication in dissimilar files,because there is little duplicate block in these files.In order to solve these problems,we study existing secure deduplication schemes and propose a similarity-based secure duplication scheme by employing variable-sized chunking,counting bloom filter and a proof of ownership scheme based on the bloom filter,which has been accepted by an international journal,Pervasive and Mobile Computing(SCI Impact Factor: 1.719).Our contribution can be summarized in two folds:First,we propose a novel secure similarity-based data deduplication scheme in which blocklevel deduplication is only performed in similar files.Thus,a good balance between the computation overhead and the rate of deduplication can be reached.Otherwise,we also construct a novel content defined tag generation algorithm to compute the similarity of files effectively in secure deduplication scheme which can detect the similarity of files only with the file tag rather than the similarity coefficient based on the full-text.Second,we prove that the proposed scheme can ensure the data confidentiality,resist the duplicate faking attack and protect the Po W tag from being forged if the server is honest but curious,while it can significantly reduce the computation overhead of deduplication.Meanwhile,we perform an expriment based on our scheme on the server with Linux andMongo DB,which shows that our scheme is more time efficient than the state-of-art secure deduplication scheme.