Research And Implementation Of Blockchain Asynchronous Consensus Algorithm

The consensus mechanism is a vital component of blockchain technology,ensuring the integrity and reliability of the entire system.From the perspective of network communication model,Byzantine Fault Tolerance(BFT)consensus algorithm can be divided into synchronous BFT consensus algorithm and asynchronous BFT consensus algorithm.Since the synchronous BFT consensus algorithm relies on the global clock,selectively delaying the communication between the node and the leader can make the algorithm inactive,thus restricting the application scenarios of the synchronous consensus algorithm.The asynchronous Byzantine consensus algorithm can be applied to real network scenarios such as network communication failures and unpredictable network environments,and the leaderless model can fully release the system’s concurrency capabilities,so it has broader application value.However,the existing asynchronous BFT consensus algorithm still has the following problems when applying the blockchain system:(1)The communication(time)complexity is relatively large.There is no timeout mechanism in the asynchronous Byzantine consensus algorithm,and it usually needs to run more rounds of communication to reach a consensus,so it has a greater communication(time)complexity.(2)The throughput is relatively low.The asynchronous Byzantine consensus algorithm needs to select block producers multiple times before it can be terminated.The efficiency of block generation is low,resulting in low overall system throughput.This paper is oriented towards two types of blockchain systems applicable to public chains and alliance chains,and studies low-communication overhead and highthroughput asynchronous consensus algorithms.The main research contents and innovations are as follows:(1)To solve the problem of high communication(time)complexity of asynchronous consensus algorithms,an efficient three-phase commit asynchronous consensus algorithm LCOHBBFT is proposed.LCOHBBFT reduces communication overhead by first determining the order of transactions and then requesting missing transactions.When broadcasting a transaction,the original text of the transaction is broadcast directly to avoid the transmission of redundant data.After the order of transactions is determined,the use of a two-stage transaction recovery mechanism enables nodes to seek help from 21)+ 1 nodes to recover missing transactions,which reduces the communication overhead of transaction recovery while satisfying Byzantine fault tolerance.Finally,experiments show that the communication overhead of the LCOHBBFT algorithm is 60% of the first practical asynchronous consensus algorithm Honey Badger BFT.(2)To solve the problem of low throughput of asynchronous consensus algorithm,an asynchronous consensus algorithm Tortoise BFT based on the reputation model is proposed.Tortoise BFT introduces a reputation model to select block producers with relatively stable network conditions to reduce the number of block producer extractions.Tortoise BFT adopts a heterogeneous transaction execution mode so that each node can concurrently submit a block request to fully release the concurrent execution capability of the system.Tortoise BFT reduces the communication overhead of the algorithm and improves the throughput of the algorithm while satisfying the security constraints.Finally,experiments show that when the number of nodes in the network reaches 80,Tortoise BFT has more than 40% improvement in communication overhead,throughput and transaction confirmation delay compared with Dumbo BFT.