| As an extension of service-oriented government,smart government needs data to help it make decisions more accurately and objectively.In order to make decisions more rationally,government departments usually need to share certain types of data,which can provide an auxiliary reference for the management or decision-making analysis of the department.However,government data sharing is still faced with challenging issues of varying degrees of difficulty in trust transmission,difficulty in attribution of responsibilities,and high risk of privacy leakage.Blockchain provides a new direction for the safe and efficient sharing of data among government departments.In view of the privacy protection issues in the context of government data sharing,the research work of this paper can be summaried as:(1)This paper firstly analyzes the needs of government data sharing and the existing sharing mode,and further discusses the privacy issues in government data sharing and the current research on privacy protection methods for privacy issues.From a service-oriented perspective,a blockchain-based privacy protection government data sharing framework is constructed based on the idea of "services on the chain and data off the chain",and the off-chain data privacy protection process and on-chain access shared data are designed separately.Furthermore,we analyze proposed framework function and advantage in promoting data sharing among government departments and improving the management or service level of government departments.(2)The development history and classification of smart contract and the operation process of smart contract on the blockchain platform are reviewed.In order to better describe the business logic of government data sharing,drawing on the idea of FSM(Finite State Machine),a formal definition scheme of smart contracts for government data sharing is proposed.Through a case of government data sharing,the working mechanism of this scheme is explained in detail,and the realization idea is given on Hyperledger Fabric,the currently widely used consortium blockchain platform.It overcomes the shortcomings of Fabric not recording the Chaincode state and realizing state transition and enriches the application of blockchain in government data sharing.(3)In order to solve the problem of large statistical errors in the current privacy protection algorithms with large data domain size and small amounts of data,a BCS(Binning Count Sketch)approach for sharing government statistics based on LDP(Local Differential Privacy)is designed.This approach uses the idea of equi-width binning to divide sensitive attribute values into smaller data domains,and the data provider shares the relevant parameters obtained after the perturbation to the data requester.The data requester then performs statistical corrections through the count sketch matrix to obtain the required statistical data.The proposed BCS is compared with the classic CMS(Count Mean Sketch)and HCMS(Hadamard Count Mean Sketch)algorithms in terms of frequency estimation,data size,privacy budget,data domain size,time complexity,and transmission cost.Experimental results show that the algorithm effectively reduces statistical errors and enhances its utility of data after privacy protection with both various distributions and data domain sizes.Furthermore,extending the idea of data binning to two typical LDP algorithms,GRR(Generalized Randomized Response)and RAPPOR(Randomized Aggregatable Privacy-preserving Ordinal Response),also achieved good results. |
