Reseaches On Privacy-Preserving Data Aggregation And Secure Two-Party Computation

Data aggregation is common in the Internet of Things(IoT)scenarios,which is a technique that combines different fine-grained data from different source,and is a hot research issue for strict requirements of data legal compliance.The thesis focuses on data aggregation in IoT scenario of smart grid.There exist many problems in existing data aggregation schemes in smart grid,like high overhead in multidimensional data aggregation,the low-efficiency error detection after the failure of batch verification,and the lackness of verification of user report shares in aggregation framework supporting complex computation based on secret sharing multi-party computation(MPC).To deal with these problems,the thesis focuses on the research of efficient,lightweight and privacy-preserving multi-functional data aggregation.Data aggregation that enables complex calculations is a research trend and its efficiency is dependent on the efficiency of the underlying technology,like MPC.As a special example of MPC,secure two-party computation(2PC)has a wide range of application scenarios,and can be extended to become a general MPC for data aggregation supporting complex computation.Cloud servers are introduced to share the participants' burden and improve the efficiency of 2PC.The higher security requirements for complex calculations have been put forword with the development of the increasingly complicated Internet of Things.Therefore,to deal with the problems of high participant computation overhead,destruction of the fariness of participants in classical 2PC by server,high protocol overhead unbeneficial to complex computation applications in the existing schemes,this thesis also focus on the research of server-assisted 2PC under malicious model in order to implement lightweight data aggregation supporting complex computation.The main work of this dissertation is as follows:1.An efficient and privacy-preserving multidimensional data aggregation scheme with fast error detection in smart grid.We propose a novel,efficient and privacypreserving multidimensional data aggregation scheme supporting fast error detection,named EPMDA-FED.A novel multidimensional data packing method is proposed via Chinese Remainder Theorem and then encrypted without a trusted authority through an efficient,indistinguishable under chosen plaintext attacks and collusion-resistant encryption scheme.Besides,users'consumption reports can be verified in a batch way to guarantee the consistency of negotiated key,authenticity and integrity.In addition,we propose a novel fast general error detection algorithm combine?divide with sublinear complexity based on uniform(k,n)set.The algorithm is a general group testing method and applicable to any aggregate signature scheme in any pragmatical scenarios.This provides a new idea for error detection in other application scenarios.Compared with the corresponding multi-dimensional data aggregation scheme,EPMDA-FED has less computational overhead,and the communication overhead is only 30%of the counterpart schemes.2.A verifiable privacy-preserving data aggregation scheme supporting flexible computation in fog-based smart grid.A verifiable privacy-preserving smart grid data aggregation scheme VPDC-MPC supporting multi-party computation is first proposed in fog-based smart grid.The scheme can support data aggregation and function query from customers and the control center simultaneously through MPC of secure cloud servers.VPDC-MPC realizes batch verification of users'data secret shares.VPDC-MPC can ensure the consistency and correctness of secret shares during distribution and transmission and efficient detection of error reports with invalid secret shares via tree data structure.Besides,even if some malicious internal enemies collude with each other,honest individual privacy cannot be leaked.Performance assessment shows that even though the comparative tradeoff are made even in computation and communication overhead,our scheme is more practical than counterpart schemes.3.Secure two-party computing assisted by cloud computing server under malicious model.The secure two-party computation protocol under the malicious model,named as MSa-2PC,is creatively proposed with the optimization of the authenticated garbled circuits based on the half-gate Garbled Circuit and informationtheoretic message authentication code.In the protocol,the part of evaluation work is securely outsourced to cloud server to reduce the operational overhead of evaluator party.Through security analysis,we prove that our scheme resists malicious adversaries,and meets proposed privacy,correctness,and fair goals.The performance analysis shows that MSa-2PC generates less communication overhead for AND gate and preprocessing phase can reduce the number of interactions in online phase,making the protocol more efficient.