Research On Privacy-preserving Identity Authentication Protocol For Big Data

With the rapid development applications of big data,many sensitive information can be found using big data mining tools and analysis tools,such as health data and health care information etc.of user privacy information.Therefore,the security and privacy of big data have attracted much attention of the people,and authentication protocol is an important part of it.Currently,many authentication protocols focus on authentication and secrecy,but pay little attention to privacy protection.Therefore,this thesis studies the authentication protocol which preserves user privacy and provides user authentication for big data.The mainly works are as follows:1)Present a User Privacy-Preserving Identity authentication protocol for Big Data,UPPIABD(User Privacy-Preserving Identity Authentication for Big Data,UPPIABD).UPPIABD protocol takes the user's keyboard and mouse usage habits,system processes,network behavior and other characteristics as authentication factors.According to the characteristics of these authentication factors,homomorphic encryption and fuzzy hash functions are carried out respectively.User privacy is not leaked and identity authentication is realized.At the same time,referring to HTTP protocol standard message structure,the UPPIABD protocol message structure is presented.2)Analyze the secrecy,authentication and privacy of UPPIABD protocol.Firstly,UPPIABD is formalized with Applied PI calculus.And then the secrecy and the authentication are modeled by non-injective correspondence and Query respectively.Next the Applied PI calculus formal model of the UPPIABD protocol is transformed into the input of the security protocol analysis tool ProVerif,which is formally analyzed and proved through ProVerif.The results indicate that UPPIABD protocol has the secrecy and authentication.Secondly,Secrecy of User Privacy Information by UPPIABD protocol is analyzed.The results show that UPPIABD protocol has confidentiality to user's privacy data and can protect user's privacy.3)Design,implement and test UPPIABD authentication system.UPPIABD authentication system mainly includes configuration file generation module,authentication module and database module.The configuration file generation module runs on the user side and need not any interaction with the user,it can automatically complete the generation and encryption of the user profile;the authentication module is responsible for storing and verifying the user password.The verification results are returned to the user,while passing the user profile to the database module;based on the ciphertext of user profile,database module performs user authentication and user profile storage,and the results of authentication are also returned to the user.UPPIABD authentication system is tested through recall rate and the FPR rate in Hadoop.