Improvement And Research Of Blockchain Consensus Algorithm Based On Byzantine Fault Tolerant

The gradual rise of blockchain has aroused great attention and widespread concern from all walks of life.As the core component of blockchain technology,consensus algorithm has a direct impact on the efficiency of blockchain,so it has gradually become the research focus in the field of blockchain.The throughput of blockchain is mainly determined by consensus algorithm,so improving the throughput has become the main research purpose of consensus algorithm.If the throughput of blockchain can be improved,it can make the blockchain execute more transactions or requests,which is of great significance to further promote the development and application of blockchain.This thesis summarizes the common blockchain consensus algorithms,and makes some improvements based on DBFT algorithm and Fast-Hotstuff.The main work of this thesis is summarized as follows:(1)Delegating Byzantine Fault Tolerant(DBFT)algorithm is a consensus algorithm in the public blockchain.DBFT algorithm selects a part of consensus nodes from normal nodes to perform the consensus process,which can adapt to the network environment of large-scale nodes in the public blockchain.The consensus process of DBFT algorithm is that nodes broadcast messages to each other and vote for a block twice.Therefore,there is the problem of too many communication steps between nodes.This thesis proposes Improved-DBFT algorithm.In view of the high tolerance of temporary forks in public blockchain,Improved-DBFT adopts a hybrid consensus process of one round of voting and two rounds of voting,the DBFT algorithm is changed to eventual consistency algorithm.In which one round of voting is used to reduce the communication steps between nodes,and two rounds of voting is used to uniquely determine the next consensus group.Secondly,the forks solution method is improved to deal with the problem of blockchain forks in the case of one round voting.If two conflicting blocks are added to the blockchain at the same height,the branch of the block with the largest view number is taken as the effective branch,and then a new block is generated according to the branch to extend the blockchain.Then,the only effective chain is determined according to the longest chain principle.Algorithm analysis shows that the Improved-DBFT algorithm can meet the persistence and liveness requirements of the public blockchain,and effectively reduce the communication steps between nodes.Simulation results show that when the number of nodes is 7,the throughput of DBFT is about 6000 TPS,and that of Improved-DBFT is about 8000 TPS.The results show that Improved-DBFT can effectively improve the throughput.(2)Fast-HotStuff algorithm is a common consensus algorithm in consortium blockchain.Because the consortium blockchain does not allow forks,the algorithm adopts the consensus process of two rounds of voting to ensure the strong consistency of the blockchain.In the FastHotstuff algorithm,if the consensus can be executed normally,or if the leader node fails after the second round of voting,it can provide high throughput.However,if the leader node fails after the first round of voting,the throughput will be greatly reduced.To solve this problem,an Improved-Fast-HotStuff algorithm is proposed.A new block generation strategy is added to the proposed algorithm.In the consensus process of a block,if the leader node fails after the first round of voting and causes the view change,the replica node delivers its voting message to the new view,and if the leader node in the new view receives enough voting messages,then it expands the block of previous view to generate a new block and initiate consensus process.Therefore,Improved-Fast-HotStuff algorithm can make more blocks added to the blockchain,so as to improve the throughput.Algorithm analysis shows that Improved-Fast-HotStuff has security,liveness,optimistic responsiveness and optimistic extensibility.Simulation results show that when the number of nodes is 7,if the leader node fails after the first round of voting,the throughput of HotStuff and Fast-HotStuff is less than 4500 tps,while the Improved-FastHotStuff is more than 8000 tps.The results show that the Improved-Fast-HotStuff algorithm can still provide high throughput even when the leader node fails.