Efficient fuzzy keyword search scheme over encrypted data in cloud computing based on Bed-tree index structure

Abstract

In cloud computing, enabling search directly over encrypted data is an important technique to effectively utilize encrypted data. Most of the existing techniques are focusing on fuzzy keyword search as it helps achieve more robust search performance by tolerating misspelling or typos of data users. Existing works always build index without classifying keywords in advance. They suffer from efficiency issue. Furthermore, Euclidean distance or Hamming distance is often chosen to evaluate strings’ similarity, ignoring prefixes matching and the influence of strings’ length on the accuracy. We propose an efficient fuzzy keyword search scheme with lower computation cost and higher accuracy to address the aforementioned problems. We employ the sub-dictionaries technique and the Bed-tree structure to build an index with three layers for achieving better search efficiency. With this index structure, the server could locate the keyword and could narrow the search scope quickly. The Jaro-Winkler distance is introduced to qualify the strings’ similarity by considering the prefixes matching and string length. The secure privacy mechanism is incorporated into the design of our work. Security analysis and performance evaluation demonstrate our scheme is more efficient compared to the existing one while guaranteeing security.