Research Of Dynamic Consensus Mechanism Based On PBFT Algorithm

As the underlying technology of bitcoin and other virtual currencies,blockchain has the characteristics of decentralization,tamper resistance and traceability.It is the current research hotspot to solve the real social credit problem by using blockchain technology.Based on the existing Practical Byzantine Fault Tolerance(PBFT),this thesis proposes a Dynamic Network Byzantine Fault Tolerance(DNBFT),which improves the shortcomings of the PBFT consensus algorithm that cannot dynamically change the number of nodes.At the same time,the operation of clearing out the network is performed on the malicious node.The main contents of this thesis include:Firstly,the problem of low throughput,high latency and the inability to dynamically change the numbers of nodes in the consensus algorithm affecting the blockchain bottleneck is studied.It is found that the PBFT algorithm is compared with the Proof of Work(PoW)algorithm and Proof of Stake(PoS)algorithm has the advantage of high system throughput,but cannot realize the dynamic change of the network node and does not punish the malicious nodes in the network.This thesis proposes a dynamic network consensus mechanism based on PBFT,which provides the same security and activity as the PBFT consensus protocol,allowing network replica nodes to join or exit the consensus network without stopping the network operation;The mechanism provides a mechanism to remove malicious and long-term downtime nodes to enhance the robustness of the system;introduce the concept of participation to record whether the network node is sufficiently active,by increasing the cost of refusing to participate in the consensus,the enthusiasm of nodes to participate in the consensus is ensured,and the system security is effectively improved;malicious nodes are pulled into the blacklist of the system network to ensure the security of the blockchain network to a greater extent.The dynamic network consensus algorithm DNBFT is tested in terms of throughput and delay.The experimental results show that the throughput can reach 10,000 or more,and the delay can reach milliseconds.By comparing the transaction delay and throughput in different block intervals.The interval time is given,and the optimal block interval for dynamic network consensus is 20 s.The delay is between[200ms-250ms],which can meet the requirements of most blockchain applications.Then,for the mutual credit recognition mechanism,there is insufficient credibility,a lot of auditing resources and computational cost,and various thesis certificate frauds can not guarantee high reliability and authenticity.The dynamic network consensus protocol DNBFT and index network topology are adopted,by IBM's Hyperledger Fabric project framework quickly builds a blockchain network.A performance test model is proposed for the blockchain based on dynamic consensus protocol.The most time-consuming three parts of the blockchain network consensus process are selected to model the average consensus time.It is verified that the blockchains based on DNBFT can support effective operation under the network scale of 50 system nodes.The realization of this model can improve the credibility of the mutual recognition mechanism of credits,and provide a reference for mutual recognition of Mooc credits.