排序可验证的语义模糊可搜索加密方案

Semantically Searchable Encryption Scheme Supporting Ranking Verification

为解决现有语义模糊可搜索加密方案无法取得排序可验证的问题,提出一种支持语义模糊搜索并能对搜索结果的排序进行验证的方案。首先引入相关度分数和域加权得分,构造精确度更高的倒排索引结构。通过非线性保序加密,云服务器可以对搜索结果进行高效排序,减少了用户的计算开销和网络资源浪费。接着基于WordNet词典集对查询关键词进行语义拓展。在构造出语义拓展树后,只选取最相关的语义拓展词进行查询。通过引入语义相似度,设计双因子排序算法对搜索结果进行精确排序。然后为创建验证信息,将关键词集合插入布隆过滤器中,并计算布隆过滤器、安全索引、密文文档的消息认证码。在搜索阶段,通过布隆过滤器可以验证索引中是否存在查询关键词,并使用消息认证码对索引、文档的完整性和正确性进行验证。通过引入索引树结构进一步提高方案的搜索效率。最后,将本方案与相关方案在功能、存储开销、搜索开销等方面进行比较,并对创建索引、创建验证信息、语义拓展、加密查询、查询、验证等方面的开销进行仿真实验。方案对比分析显示本方案在语义搜索、排序、验证等功能上具有优势。实验结果表明,本方案不仅实现了语义模糊搜索与结果排序,而且实现了排序结果可验证功能。

In order to solve the problem that the existing semantically searchable encryption schemes could not realize ranking verification,a novel method was proposed in this paper which could not only support the semantic search,but also verify the ranking of the search results.Firstly,the relevant score and the weighted zone score were introduced to construct a more accurate inverted index structure.Using the non-linear order preserving encryption mechanism,the search results were sorted efficiently by the cloud server,which reduced the computation and communication overheads.Then,the query keywords were semantically extended based on the Word Net.Based on the semantically extended tree,the most relevant semantically extended words were selected.Combining the semantic similarity and relevance score,a two-factor ranking algorithm was designed to accurately sort the search results.Moreover,a set of keywords were inserted to the bloom filter.The message authentication codes of the bloom filter,security index and encrypted files were calculated to build the verifiable information.In the search phase,the bloom filter was used to verify the existence of the query keywords in the index.And the integrity and correctness of the index and files could be verified by the message authentication codes.The index tree was constructed to improve the search efficiency.Finally,the proposed scheme was compared with the related schemes in terms of functions,storage overheads and search overheads.And simulation experiments were carried out on the cost of building index and verifiable information,semantic extensions,encrypting queries,querying and verifying.The comparative analysis showed that the advantage of the proposed scheme in terms of semantic search,ranking and verifying.The experimental results demonstrated that the proposed scheme not only realizes the semantic search and the search results ranking,but also supports the ranking result verification.