Research And Application Of Practical Byzantine Consensus Algorithm Based On Directed Acyclic Graph Structure

With the rise of digital currency,many countries and enterprises have turned their attention to the underlying blockchain technology.Due to the technical advantages of nontampering,non-forgery,traceability and decentralization,blockchain technology has developed rapidly and has been gradually applied to military,financial,medical,entertainment and other fields.The popularity of blockchain not only makes a large number of users feel the advantages of blockchain,but also exposes some disadvantages.The most important problem is that the throughput capacity of blockchain system can no longer meet the needs of current user transaction volume.At present,there are two mainstream ways to improve the throughput of blockchain system: one is to improve and optimize the traditional blockchain consensus mechanism,and the other is to change the underlying data structure of blockchain.In terms of consensus algorithm,Practical Byzantine fault tolerance(PBFT),as a widely used consensus mechanism under the alliance chain,has its unique advantages,but there are also some defects,such as low throughput,excessive communication overhead,consensus needs to select the master node,and can not be applied to the fully asynchronous network model.In terms of modifying or replacing the block organization,a consensus mechanism of directed acyclic graph(DAG)structure based on workload proof or weight has been proposed.This structure can effectively improve the throughput capacity of the system,but there are still some defects,such as waste of computing power,uncertain transaction confirmation time,and possible cancellation of confirmed transactions.Based on the above problems,this thesis proposes a new consensus algorithm-PBFT consensus algorithm based on DAG structure(DAGs-PBFT).The research work of this thesis is as follows:(1)This thesis applies the DAG structure to the traditional PBFT consensus algorithm and proposes the DAGs-PBFT consensus algorithm.Each node in the consensus network has its own chain,the consensus node records the proposed new block on its own chain,and the new block refers to its latest block and the block synchronized with it by other nodes as a pair.The voting of the referenced block finally constitutes the DAG structure,and the local DAG copy is agreed through the Gossip protocol.In the DAGs-PBFT consensus algorithm,each node reaches an agreement on the consensus order through the local DAG copy and virtual voting,which greatly reduces the communication complexity of the traditional PBFT algorithm and improves the throughput of the blockchain system.(2)Combined with the actual demand in the field of supply chain and logistics,this thesis applies the proposed DAGs-PBFT blockchain scheme to the actual environment to solve the problem,and adopts the principle of transaction segmentation to optimize the DAGs-PBFT algorithm according to the characteristics of low transaction relevance in the supply chain industry.Under the condition of ensuring the security and stability of the blockchain system,the characteristics of the optimized DAGs-PBFT algorithm are analyzed through experimental simulation,so as to further improve the performance of the blockchain and reduce the computational complexity.Further,by means of security proof and experimental simulation analysis,the DAGs-PBFT algorithm and its optimization algorithm are compared with the traditional PBFT consensus algorithm and the DAG consensus algorithm based on proof of work or weight based,and are tested the performance of the two algorithms.