Study On Privacy-preserving Data Aggregation And Its Applications

With the development of wireless sensor networks and big data technology,the amount of data has increased dramatically.The development speed of transmission capacity in network,computing power and storage resources held by devices are slower than the growth of data.Data aggregation technology can aggregate a large amount of data within the same type into one data,which greatly reduces the cost of data transmission,storage and processing.Therefore,data aggregation technology is a powerful data processing tool for the massive data.In addition,the massive growth of data promotes the birth and rapid development of new technologies and services,and brings convenience to people's lives,however,it also inevitably threatens people's privacy.Therefore,how to realize the application demand of processing a large amount of data without disclosing user privacy is particularly urgent.Data aggregation technology with privacy protection is a subject that is worthy of research.Aiming at the problem of privacy leakage in traffic analysis applications,a business address recommendation scheme based on multi-dimensional aggregated traffic analysis for user demand is proposed.This scheme realizes a privacy preserving data aggregation method by homomorphic encryption technology.Based on this method,the scheme realizes the security statistics of users' mobile traffic,analyzes the users' needs,and recommends the appropriate location area for businesses.In addition,in the study of the existing data aggregation schemes for smart grid privacy issues,we find that these schemes fail to fully consider the risk of privacy leakage caused by the time sequence of data in the process of data aggregation.The smart meters collect data on the power consumption at regular interval and send it to the aggregator after encryption.The aggregator aggregates all the received ciphertexts and send them to the control center that is responsible for obtaining the total power of the user within the time interval after decryption.If the adversary with the ability to invade the database of the Control Center,it can launch regular blocking attacks on the target users so that it will obtain the total power of the time-series users including the target users or not.By analyzing a series of total power of the time series users,the adversary can obtain the privacy information of the target users.We define the above attack process as malicious data mining attack,and propose a privacy-preserving data aggregation scheme against this malicious data mining attack for smart grids.