Fuzzy Search Scheme In Cloud Encrypted Databases

Cloud computing on demand availability is extremely crucial for computer systems resources in terms of computing power and data storage.Since the separation between ownership and resources consumption,data protection and security have been confronted with serious challenges.Cloud data encryption is an effective method in open space to protect data security which is based on utilizing cryptographic techniques.However,despite of strengthening cloud databases security and privacy with cryptographic techniques and adding fuzzy search schemes to enhance them,it has been found that the existing fuzzy search schemes are unable to meet the increasing security demands of cloud computing services in terms of availability and security.In this paper,we focus on fuzzy search of cipher text in encrypted databases and design a fuzzy search scheme that can resist statistical attacks on the cloud.We also design an efficient cipher text fuzzy search scheme for Chinese characters,in order to improve the efficiency of the scheme as a whole due to the shortcomings in the existing schemes in searching for Chinese cipher texts.Our contributions are as follows.Existing fuzzy search schema,specifically Bloom Filters(BF)schema,based on vector index use bloom filters to build vector indexes for keywords.Interestingly,the deviation of the frequency distribution in vector index causes a statistical attack on the index and leads to the disclosure of sensitive information.This paper presents a fuzzy Secure-Bloom Filters(S-BF)scheme for encrypted databases that can resist statistical attacks.Noise vector is used to eliminate the deviation of frequency distribution of index value,and the statistical attack on vector index is solved.Compared with traditional schema,the experiments show that our proposed scheme can achieve privacy-preserving fuzzy queries over an encrypted database without losing its efficiency.When the fuzzy search scheme based on vector index is applied to Chinese characters,we found out that the scheme is highly inefficient,due to the high collision probability of Chinese character indexes.To further improve our proposed S-BF scheme,this paper proposes an efficient fuzzy query scheme for Chinese characters in a cloud encryption database(MS-BF scheme),which constructs Bloom filter column and minimum hashcolumn.The Bloom filter column is implemented based on the S-BF scheme,and the index conflict is reduced by adaptively selecting the Chinese character encoding method.In order to construct the minimum hash column,a fuzzy query scheme based on Mini Hash is proposed.In this scheme,multiple vector indexes are built for a single plaintext data to reduce index collision,and a minimal hash function is used to reduce the dimensions of vector indexes,which can save storage space.To resist statistical attacks against minimum hash columns,this scheme inserts a fake trap door in the index to hide the real index.Experiments show that comparing with the traditional vector index-based fuzzy search scheme,the average time for performing fuzzy search by MS-BF scheme is shortened by 31.4% while protecting cloud data privacy.Based on the above,a cloud encrypted database system that supports fuzzy search has been designed and implemented.The two fuzzy search schemes proposed in this paper are integrated into the system and tested.Moreover,we also applied verification and validation methods to ensure the proper operation and complete compatibility of our proposed system altogether.