Research On Anonymous Method For Private Blockchain Transaction Data Publication

In recent years,blockchain technology has been developed rapidly,involving cryptocurrency,finance,Internet of things,medical and other fields.A large amount of data has been accumulated in these applications,which provides an unprecedented opportunity for researchers to analyze blockchain data,understand and solve related problems.However,blockchain technology not only brings convenience to us,but also causes some privacy problems.Therefore,how to make researchers conduct data analysis under the condition of protecting personal privacy has become an urgent problem to be solved.In the public blockchain,the current methods to protect the privacy of transaction data are mostly achieved by modifying the underlying mechanism of the blockchain,and the typical methods are mixed currency mechanism and encryption-based methods.These methods bring additional overhead and inefficiency to the blockchain,which is not allowed in many application scenarios.In addition,the processed transaction data is difficult to use for data analysis,which loses the availability of data;In private blockchains and consortium blockchains,the ownership and management of data are owned by the blockchain administrator.In this case,the anonymous method based on data publishing can be used to process the data without modifying the blockchain mechanism.In this way,the data can be used reasonably while ensuring that the performance of the blockchain is not affected.However,due to some characteristics of blockchain transaction data,the traditional anonymous method used in relational data publishing cannot be directly applied to blockchain transaction data.This paper conducts research on the unique data format of the blockchain combined with the traditional data anonymization method,and proposes an anonymous method of transaction data applied on the private blockchains and consortium blockchains.The main work of this paper is summarizedas follows:(1)First of all,this paper introduced the privacy problems existing in the blockchain in detail,then we divide the privacy of the blockchain into identity privacy and transaction privacy,based on the previous l-diversity model,we propose an extended model—l-address diversity and l-transaction diversity model to protect the identity privacy and transaction privacy of blockchain users,so that the published data not only meets the requirements of anonymity,but also meets the needs of data analysis.(2)Next,we represent the transaction data of the blockchain in the form of table data.In the processing of identifier attributes,this paper uses the hash algorithm to anonymize the identifier attributes.In this way,both the privacy information of the individual and the structural information of the blockchain is protected.When anonymizing quasi-identifier attributes,it is divided into numerical attributes and classification attributes to be discussed separately,and processed in different ways.(3)Then,this paper proposes an anonymous method CP-BK(clustering partition based on Bisecting K-medoids).We use the degree of difference in quasi-identifier attributes to measure the distance between tuples,and cluster the tuples according to the distance formula.When clustering,the improved K-medoids algorithm is used to continuously divide the data set,so that the number of tuples in each group is as small as possible,thereby reducing information loss.We also analyze the privacy and time complexity of the algorithm.(4)Finally,this paper implements the proposed anonymous method and gives a detailed system analysis and the implementation process of each module.We conducted simulation experiments using real transaction data sets in the blockchain,and evaluated the algorithm from three aspects: information loss,algorithm execution time,and clustering effect.We compared our method with the greedy algorithm which is widely used in the previous anonymous method based on clustering.By analyzing the experimental results,we verified the feasibility of the method proposed in this paper.The published data has high utility under the condition of satisfying identity anonymity and transaction anonymity,the clustering-based anonymous method proposed in this paper is superior to the previous clustering anonymous methods based on greedy strategy in terms of clustering performance and information loss.