Design And Application Of Anti-forecast Fault Tolerance In PBFT

With the development of hospital information system,electronic medical record(EMR)is gradually replacing the traditional paper-based medical record.As a distributed database,blockchain technology can effectively serve as the carrier of EMR management and sharing.At present,most of the systems using blockchain technology to manage EMR are improved on the basis of Fabric project,and in these improvements,there are problems such as the privacy of patients can not be protected,the.difficulty of medical record sharing between agencies and the failure to ensure the complete credibility of the data on the block chain.The practical Byzantine fault tolerant(PBFT)consensus algorithm used in Fabric project,because of the existence of the leader,makes the leader in the system very easy to be offline due to the attack of malicious nodes.When the leader is offline due to network reasons or attacked by malicious nodes,it cannot respond to the transaction on time.At this time,the system will initiate a view change to select a new leader to process the transaction.However,in the view change protocol of PBFT consensus algorithm,the leader in the new view can be predicted in advance,so it is very vulnerable to forecast attacks.At the same time,in the process of view change,it also affects the performance of the network.In this paper,firstly,aiming at the problem that the leader is vulnerable to predictive attack,an anti-forecast practical Byzantine fault-tolerant algorithm,afBFT(anti-forecast PBFT),is designed to resist the forecast attack.In this paper,in the laboratory environment,from the five aspects of normal traffic,average block time and traffic under attack,block time under attack and TPS under attack,we compare the afBFT consensus algorithm with PBFT consensus algorithm used in Fabric project and practical Byzantine fault-tolerant algorithm based on credit coefficient(CPBFT).Through the observation of the experimental results,we can see that the stability and network performance of the afBFT consensus algorithm is higher than that of PBFT consensus algorithm and CPBFT consensus algorithm due to the increase of the leader selection consensus,which can resist the leader from forecast attack to a certain extent and has strong security.Therefore,the afBFT consensus algorithm can solve the problem that the leader of Fabric project is vulnerable to forecast attack to a certain extent.On this basis,this paper presents the application of the afBFT consensus algorithm of EMR management program.In the scheme,two parts are designed:offline medical treatment layer and EMR upload layer.The offline medical treatment layer can protect the privacy of patients to a certain extent and ensure the accuracy of offline medical records.The EMR upload layer can solve the problem of sharing medical records among medical institutions to a certain extent.At the same time,afBFT consensus algorithm is used in the EMR upload layer,which ensures the authenticity and reliability of the data on the block chain and greatly increases the availability of the scheme.