Research On Blockchain Consensus Mechanism Based On Byzantine Fault Tolerance

Blockchain is essentially a distributed ledger containing a variety of underlying technologies,in which there is no central node and all nodes are in an equal position.Nodes in distributed systems need to maintain the consistency of data,in blockchain technology,consensus mechanism determines how to reach consensus between nodes in the system,and it is also a very important core technology in blockchain,it enables all nodes in the system to jointly maintain the same data.Different from the traditional distributed system,blockchain system needs to consider the byzantine fault tolerance problem,so it mainly uses the byzantine fault tolerance consensus mechanism.Especially when the public chain system such as bitcoin appeared,the byzantine fault tolerant consensus mechanism was gradually widely used.In recent years,with the increasing research on blockchain technology,the technology has also been applied in various fields such as medical treatment,Internet of things and data privacy protection,different application areas have different performance requirements for the consensus mechanism.The existing consensus mechanism has many shortcomings in security and scalability,so it needs to be improved to meet the needs of different application scenarios.This paper analyzed several mainstream byzantine fault tolerant consensus mechanisms of blockchain system,focusing on the analysis of Delegated Proof of Stake(DPoS)mechanism in this consensus mechanism,points out the shortages existing in the consensus mechanism and improved.The specific work is as follows:A DPoS algorithm based on reward mechanism(RM-DPoS)is proposed to solve the problems of low voting enthusiasm of nodes and fail to identify malicious nodes in the DPoS mechanism.The algorithm by setting the new incentive scheme to inspire node vote enthusiasm,in the reward allocation scheme,by adding time factor into the income allocation algorithm based on shapley value and limiting the number of votes of nodes,not only can the benefits obtained by nodes be more reasonable,but also the condition of inequality between nodes in the system be improved.In order to make the voters vote for more reliable nodes,this paper establishes a trust evaluation model for the nodes in the blockchain system,which can effectively identify the malicious nodes in the system and reduce the probability of the malicious nodes becoming witness nodes.Experimental results show that the algorithm can identify the malicious nodes more quickly and improve the security of the system.A DPoS algorithm based on dynamic grouping(DG-DPoS)is proposed to solve the problems of the large number of tokens which affect the election result of the witness node and node failure in DPoS.The algorithm firstly dynamically groups the nodes according to their positions,and then carries out the election process of witness nodes in each group,which reduces the influence of the nodes with a large amount of coins on the election results to some extent.In view of the problems of node failure,the protocol of replacement of witness node is added in the consensus stage to improve the robustness of the system,and block synchronization process is completed by broadcasting block information in groups,which improves the performance of the system.At the same time,in the whole process of the algorithm,the reward allocation scheme and the node trust evaluation model proposed in RM-DPoS algorithm are used to ensure the reliability of the witness node.The experimental results show that the DG-DPoS mechanism has certain advantages over the original DPOS mechanism in security,delay performance and throughput performance.