Efficient dynamic multi-keyword fuzzy search over encrypted cloud data

Abstract

Multi-keyword search of encrypted cloud data has attracted extensive attention worldwide in the recent years due to the increasing concern for data security and privacy in Cloud Computing. Fault-tolerance is important for multi-keyword fuzzy search which can provide accurate results even with the presence of minor spelling and typographical errors in the search keywords. But, existing fuzzy search schemes lack efficiency due to their high computational overhead and do not support file dynamic updates. This paper proposes an efficient dynamic multi-keyword fuzzy search scheme for encrypted cloud data to support dynamic file updates. Locality sensitive hashing (LSH) and Bloom filters are employed to generate index vectors and query vectors. Based on the generated vectors, a balanced binary tree is constructed as the index for the entire file set, and a Top-k search algorithm is developed to search k files that are most relevant to a given query with the help of the index tree. Extensive experiments conducted on real-world datasets demonstrate that our scheme is more efficient than existing similar schemes.