Research On Multi-keyword Retrieval Over Encrypted Data

With the development of the times,the data to be processed is growing day after day.In order to save local storage overhead and increase economic benefits,more and more data users choose to outsource their data to a third party,such as cloud servers.Data users usually encrypt the data before outsourced to protect the security of data.However,this approach is not convenient for data retrieval,it also hinder the development of cloud storage technology to some extent.How to retrieve data efficiently while ensuring data security is urgent need for data users.The appearance of ciphertext retrieval technology can effectively meet the need and become the research focus in recent years at home and abroad.This thesis research on multi-keyword retrieval over encrypted data based on index and searchable encryption algorithms.The main work and achievements are as follows.(1)Firstly,we research on the index technology of ciphertext retrieval.Based on the traditional word weighting framework TF-IDF,and considering multi-keyword application scenarios,we propose a new weight scheme named P~2KW.Then,we build an inverted index with word weighting.In order to improve the safety and retrieval efficiency of the index,we encrypt the keywords and pointers,use the skipping pointer and grading index technology.We also give the index building and updating operations on the Hadoop platform.(2)Then,with the aid of mature encryption technology,we design the encrypt scheme and key management strategy of the ciphertext index and the original document.Then,we comparative analyze symmetric and asymmetric searchable encryption algorithms.Combined with lagrange interpolation polynomial,we improve the asymmetric encryption algorithms and propose a scheme which supporting for multi-keyword retrieval called PKEMS.Through comparative analysis,this scheme can reduce the bilinear pairings computation and improve the efficiency of the algorithm.(3)Finally,we study the system on core functions and encryption module.We also design and implement a ciphertext retrieval system for multi-keyword retrieval.Then,we test the system with the collected experimental data.The security analysis and test results show that the improved ciphertext index increased the retrieval efficiency,our scheme achieves a good balance between security and retrieval performance.