Fuzzy Keyword Search Of Encrypted Data Across Multiple Servers Over Cloud Computing

To enhance the efficiency of data searching,most data owners store their data files in different cloud servers in the form of cipher-text.Thus,efficient search using fuzzy keywords becomes a critical issue in such a cloud computing environment.This paper proposes a method that aims at improving the efficiency of cipher-text retrieval and lowering storage overhead for fuzzy keyword search.In contrast to traditional approaches,the proposed method can reduce the complexity of MinHash based fuzzy keyword search by using MinHash fingerprints to avoid the need to construct the fuzzy keyword set.The method will utilize Jaccard similarity to rank the results of retrieval,thus reducing the amount of calculation for similarity and saving a lot of time and space overhead.The method will also take consideration of multiple user queries through reencryption technology and update user permissions dynamically.Security analysis demonstrates that the method can provide better privacy preservation and experimental results show that efficiency of cipher-text using the proposed method can improve the retrieval time and lower storage overhead as well.As for the above problems,this paper proposes a fuzzy keyword search scheme for multi-server multi-user,which preserve the security and privacy.Contribution of the proposed scheme can be summarized as follows:(1)This paper presents a multiple server searchable encryption scheme,which realizes the task of executing cipher-text retrieval in parallel with multiple servers.Compared with single-server model,our method improves the processing ability of big data and the security of data in the cloud.(2)Fingerprint index storage.This program uses the MinHash function to generate the fingerprint index for the keywords without having to set the index storage space and the predefined dictionary library in advance,thus greatly reducing the complexity of the search and saving a lot of storage overhead.(3)By the Jaccard distance to sort the retrieval files reduces the amount of similarity calculation,therefore saving a lot of time and space overhead and improving the retrival efficiency.(4)Efficiency of the proposed scheme.Theoretical analysis of the specific functions and principles implemented in the scheme was performed on the Erron real data set.The experimental results prove that the proposed scheme not only guarantees the security of the retrieved data,but also improves the retrieval efficiency and saves a lot of storage overhead.