Research On Privacy-preserving Data Aggregation In Internet Of Things

With the widespread adoption of Internet of Things(IoT)technology,the number of IoT device connections is exploding,and the scale of data is expanding rapidly.Data is an important factor of production in the digital age,and their value needs to be released through publishing.However,in the publishing of IoT data,there are risks of exposure of privacy information contained in data.In recent years,IoT data privacy protection has become a research hotspot.Its purpose is to open and share data on the premise of guaranteeing data privacy.Currently,Privacy-preserving Data Aggregation(PDA)is a popular method for data privacy protection in IoTs.PDA methods not only preserve the statistical characteristics of raw data but also conceal the individual original data,which realizes the goal of providing the overall IoT data for the data users on the premise of ensuring the data privacy of terminal users.Therefore,PDA is considered as one of the best methods to strike a balance between data privacy and data availability.In order to improve the applicability of PDA technology in IoTs,researchers have conducted a lot of research on the security,privacy,functionality,robustness,efficiency,and other features of PDAs.However,there are still some problems in the existing PDA characteristic research,such as difficulties in compatibility between some of the features of the existing PDA solutions and shortages of features in some scenarios.In this dissertation,we analyze the compatibility of some features of PDAs and the new feature requirements of PDAs in some scenarios,and propose the coexistence problems of fault tolerance,accurate data aggregation and no dependence on online trusted authority in PDAs,as well as the problem of optional data sharing for terminal users in multidimensional data aggregation,and the privacy protection scope in PDAs based on attribute filtering.Finally,we put forward the corresponding solutions.The main contents and innovations of this dissertation are as follows:1)Privacy-preserving accurate data aggregation with fault tolerance.Traditional PDA protocols often require all terminals to be online,and the delay or failure of any terminal may cause the protocols to fail to continue.For this reason,fault-tolerant PDA protocols have been proposed,but the existing fault-tolerant PDAs cannot guarantee both accurate aggregated data and no reliance on online trusted entities.Aiming at the PDA with the characteristics of fault tolerance,accurate aggregation,and no reliance on online trusted entities,we take the smart grid as the background and propose a threshold fault tolerant PDA scheme named(k,n)-PDA.The scheme ensures terminals’ data privacy in the aggregation through the BGN homomorphism algorithm and occupies threshold fault tolerance and threshold collusion prevention with the threshold property of Shamir’s secret sharing algorithm.Therefore,in the(k,n)-PDA scheme,even if n-K terminals among n participating in the aggregation fail to send out information due to faults or offline,the protocol can still be executed correctly without any trusted entities to aggregate the data of online terminals and provide accurate aggregation values.2)Privacy-preserving multidimensional data aggregation supporting Optional data dimensions for sharing.Existing multidimensional PDA schemes do not support data owners to select the sharing data items.However,data owners have different privacy sensitivities to different types of data.Supporting end users to make decisions on sharing dimensions of data is of great significance for occasions where multiple types of data are generated by data terminals,such as smart health.Aiming at the open sharing and privacy protection of data in smart health,the OPERA(OPtional dimEnsional privacy-preserving data Aggregation)is proposed to support the demands of data owners on the autonomy of data sharing.The scheme adopts data vectors and selection vectors to express the multidimensional data and the corresponding selections,implements privacy protection in multidimensional data aggregation through the multiple secret sharing algorithm.In addition,benefiting from lightweight cryptographic primitives,such as the standard symmetric encryption algorithm and symmetric homomorphic encryption algorithm,OPERA is efficient in computation and communication.OPERA not only implements multi-dimensional data aggregation with privacy protection but also supports terminal users to make decisions on whether to share each item of data.Data users can also obtain aggregated data on various health attributes and the corresponding number of participating terminals.3)Privacy-preserving data aggregation that supports data source attribute filtering.In IoT data services that support interactive queries,when data users need to query aggregated data based on end-user attributes,they may have privacy protection requirements for query keywords.At the same time,data terminal users in the aggregation may also wish to keep their attributes private.However,existing PDA methods cannot achieve privacy protection for these two types of information.To this end,the Interactive Query Privacy-preserving Data Aggregation(IQPDA)is proposed for interactive query scenarios.The scheme uses the oblivious transmission feature of Garbled Bloom Filter to realize data filtering based on attributes and ensures the query keyword privacy of data users and the attribute privacy of terminals.The scheme also realizes data aggregation with data privacy through Paillier homomorphic encryption.Finally,the aggregation value of terminal data that meets the filtering conditions is provided to data users under the condition of satisfying all privacy.The above research work solves the coexistence problem of fault tolerance,precise aggregation,and no reliance on online trusted entities in PDAs,and broadens the scope of privacy protection in PDAs.These results will help PDA technology applicable in lots of IoT contexts,enrich the research content of PDA technology in IoT,and can be a positive reference for privacy protection research in IoT data publishing.