Application Research Of Data Privacy Protection Based On Homomorphic Encryption

Massive data and powerful computing resources have become the main driving force for the development of intelligent technology today.However,in recent years,an increasing number of data abuse and privacy breaches have reminded people that improper use of big data technology can also bring catastrophic consequences.The data sharing is the only way to make use of data,so protecting data privacy and security in specific computing processes has become a practical need.Homomorphic encryption is an encryption scheme that allows to compute with ciphertexts.Users can encrypt private data and then send the program ciphertexts to perform computation in ciphertext field,and in the view of attackers,the ciphertexts does not contain any private information.Above all,homomorphic encryption can perform computation without revealing the data.In this dissertation,we apply homomorphic encryption to protect data privacy,provide secure computation for data,and design security solutions in the several scenarios of data privacy protection.The main work is as follows:(1)Based on the El Gamal and SM4,we design a ciphertext retrieval mode.Besides,we proposed a cloud ciphertext retrieval scheme by combining the model with SM3 algorithm and applied it to an image retrieval system,which can retrieval data securely.The cloud cannot obtain any information during the retrieval process.Analysis of efficiency and simulation show that the system based on this scheme has the advantages of lightweight,high efficiency,and strong practicality.(2)Aiming at enhancing the security and performance of multi-party Privacy Set Intersection protocol,we designed a multi-party privacy set intersection protocol based on full homomorphic encryption.Firstly,the design framework of the two-party privacy set intersection protocol based on partial homomorphic encryption is given,and the application of the privacy set intersection protocol based on threshold homomorphic encryption in the multi-party and multi-party threshold scenarios are described respectively.Then,based on the unbalanced APSI protocol,we design a multi-party privacy set intersection protocol based on full homomorphic encryption.Analysis shows that the protocol is correct,secure,and efficient.Finally,we summarize the problems of homomorphic encryption-based privacy set intersection protocols and look into the future development direction.(3)We proposed a multi-party privacy set intersection protocol based on the Lifted El Gamal threshold homomorphic algorithm and Bloom filter.In additions,we applied it to the multi rider carpooling scenario,protecting the privacy data security of drivers and riders.Firstly,the user inserts the data into the Bloom filter and encrypts the elements of the Bloom filter using a threshold homomorphic algorithm.Then,the riders send the ciphertext to the driver.Finally,the driver completes the ciphertext matching calculation,obtains the intersection size together with the rider decryption calculation result,and compares it with the threshold to determine whether the carpool is successful.Theoretical analysis and simulation show that this scheme can resist collusion attacks in the semi honest security model,and has high efficiency and availability.