Efficient,Verifiable And Trustworthy Privacy-preserving Aggregation For Mobile Traffic Data

Mobile phone has become one of the most important parts in human life today,since the services it provides to mankind are countless.Statistical analysis of mobile phone traffic data can not only provides a better user experience for mobile phone users,but also can help Operators make better marketing plans.However,the traffic data of mobile phone users carrying users’ privacy information,the premise of sharing traffic data is users’ privacy getting guaranteed.In this case,we propose an efficient,verifiable and trustworthy privacy-preserving aggregation scheme.The main challenge is to find a solution that enables the data Aggregator to efficiently compute the set of individual inputs that are processing safely under some confidentiality protocols.In this paper,we will present an efficient,verifiable and trustworthy privacy-preserving data aggregation scheme.Through the aggregation scheme,Aggregator can compute statistical functions of users’ mobile phone traffic data in a privacy-preserving way to offer various services,so as to provide users and Operators more user-friendly services.In particular,this scheme can manage dynamic users more conveniently and efficiently by introducing a semi-trusted collector for collecting obscure key information.We first propose a basic scheme,called EVTA,which can support the functions of average.Then we present an advanced scheme,called A-EVTA,which improves the calculation of discrete logarithm,and can also support for data aggregation.Besides,we consider a strong adversary,who is aiming at knowing user’s privacy information.We prove the scheme is security enough to resist all kinds of attacks initiated by adversary using the attack game.In this paper,adversary can learn nothing even though he has already compromised a set of users or colluded several malicious users,as our scheme uses blind signature and data value vector before each task sensing to make sure that the uploading data is verifiable and trustworthy.Finally,the performance evaluation illustrates that our scheme promotes efficiency comparing to those previously reported competitive approaches,especially,the applications require large plaintext space and highly accurate data.The efficiency of the advanced scheme is significantly improved by solving the discrete logarithm problem.In addition,the scheme of this paper can finely support fault tolerance,resistance to brute attack,and is applicable to dynamic user management.