Keyword Semantic Extended Top-k Ciphertext Retrieval Scheme Over Hybrid Government Cloud Environment

Li Yong,Liu Hefei,Yuan Bin,Wang Kun,Shen Xiujuan

doi:10.1109/access.2021.3128933

Li Yong, Liu Hefei + Show 3 more

Open Access

https://doi.org/10.1109/access.2021.3128933

Copy DOI

Abstract

Privacy protection is a key problem that must be solved when building government cloud system to realize data storage and sharing. In order to meet the privacy protection requirements of data security storage and document security sharing between data publishers and authorized data access users over the government cloud platform, this paper proposes a keyword semantic extended Top-k ciphertext retrieval scheme over hybrid government cloud environment. Firstly, the scheme uses the hybrid cloud mode to build the government cloud platform to realize the storage and sharing of government documents. Then, the key technologies such as mechanical segmentation method, term frequency-inverse document frequency statistics method, keyword semantic expansion method, homomorphic matrix encryption method and vector space model are used to retrieve the ciphertext documents in a completely confidential state. According to the correlation score calculated by ciphertext retrieval, the Top-k ciphertext target documents with the highest correlation score are returned. And then the scheme decrypts and restores to obtain the Top-k plaintext target documents. The analysis of security and experimental test results shows that this scheme can not only meet the data storage and sharing requirements over government cloud environment, but also prevent the privacy leakage risk of data over government cloud. It is an effective solution to promote the construction and development of government cloud.

Highlights

In recent years, with the rapid development of emerging IT technologies such as big data, cloud computing and 4G or 5G Internet high-speed communication, governments at all levels have successively built a number of government cloud information sharing platforms based on cloud computing technology in order to achieve efficient storage and sharing of data, improve the quality of government services and the collaborative efficiency of business handling among departments [1]–[3]
3) RETRIEVAL TIME OVERHEAD According to the above description of the implementation process and core algorithm of the system scheme and the implementation principle of MRSE scheme, the analysis shows that the time overhead of ciphertext retrieval in this scheme and MRSE scheme mainly depends on the number of document indexes, the length of security index vector and the number of query returned results, that is, it depends on the size of document set m Keyword dictionary size n and return target document k value
The existing research results of searchable encryption technology do not have a mature scheme with strong applicability over the government cloud environment

Summary

INTRODUCTION

With the rapid development of emerging IT technologies such as big data, cloud computing and 4G or 5G Internet high-speed communication, governments at all levels have successively built a number of government cloud information sharing platforms based on cloud computing technology in order to achieve efficient storage and sharing of data, improve the quality of government services and the collaborative efficiency of business handling among departments [1]–[3]. It is mainly composed of Cloud Server Provider (CSP), Private Cloud Server (PCS), Data Owner (DO) and Data Accesser (DA). THREAT MODEL According to the hybrid government cloud system architecture shown, its DO, data accesser and PCS are trusted, but the CSP is generally considered to be an ‘‘honest and curious’’ semi trusted entity It will work honestly in accordance with user data hosting and communication protocols and will not deliberately disclose user privacy information; On the other hand, it will analyze and mine users’ retrieval requests out of ‘‘curiosity,’’ which will inadvertently cause privacy disclosure in users’ data in the process of analysis and mining. In the subsequent segmentation work, this paper intends to use the Jieba mechanical word segmentation tool for document segmentation

TERM FREQUENCY INVERSE DOCUMENT FREQUENCY STATISTICAL METHOD

CONCLUSION