Research On Privacy-Preserving Query Algorithm Over Encrypted Data

With the development of cloud computing,cloud-based data storage,computing and query effectively meet the massive data storage and query services in different fields in the Internet environment,reducing the burden of local data storage and maintenance.While cloud computing provides efficient services,its user privacy,data leakage and other security issues also need to be solved,and data encryption and uploading data to the cloud is a commonly used solution,and data encryption affects user query,data analysis efficiency and accuracy.Range query and reverse k-nearest neighbor query are two relatively efficient,accurate and widely used query methods in encrypted data privacy protection query services.Two issues worth paying attention to are how to effectively merge similar range queries and improve the efficiency of range queries in a given time.On the other hand,on the encrypted dataset,how to improve the dataset privacy,query result accuracy and support personalized reverse neighbor query of the reverse neighbor query service.In response to the two questions above,this paper has carried out the following two tasks:First,in order to reduce the total execution time of multiple range queries,this paper proposes a privacy-preserving group range query algorithm.The algorithm performs Hilbert ordering on all range queries,proposes an effective range merging determination protocol,and merges similar range queries based on Hilbert curve to reduce the total number of retrieval nodes and improve query efficiency.This paper defines all datasets as R-trees,and designs the maximum calculation protocol,minimum value calculation protocol,range intersection judgment protocol and range determination protocol based on symmetric homomorphic encryption technology,and realizes the privacy protection group range query based on R-tree.In this paper,the algorithm is experimented on the US Census Data dataset from three aspects: data outsourcing,trapdoor generation and computational overhead of group-range query processing,and the experimental results show that the algorithm can effectively reduce the total execution time of multiple range queries.Secondly,based on the number of neighbors independently set by the user,this paper proposes a personalized inverse k neighbor query algorithm for privacy protection.By improving the tree structure to represent the cube,the algorithm proposes an encryption distance comparison protocol and an equality test protocol for privacy protection,and realizes a personalized reverse neighbor query algorithm for privacy protection.This paper proves that the privacy-preserving personalized inverse neighbor query algorithm is safe in terms of dataset and query results,and verifies experiments from three aspects: data outsourcing,trapdoor generation and query processing on the US Census Data dataset.