Evaluation And Analysis Of The Utility Of Asynchronous Consensus Protocols

Blockchain technology has been widely applied due to its advantages of decentralization.high security,reliability,immutability,and traceability.The consensus protocol is a critical component of blockchain technology as it ensures the security and reliability of the blockchain network,directly affecting its performance.Asynchronous consensus protocols have proven to achieve high performance,making it possible for their application in blockchains.JD Chain and Ant Chain in China,for example,have adopted asynchronous consensus protocols,However,research on practical asynchronous consensus protocols and their applications in blockchains is still in the early stages in China and around the world.Based on the current research on asynchronous consensus protocols,this article proposes AsyncTool,a performance-testing tool for asynchronous consensus protocols that can be applied in different modes,stages,and scenarios.Unlike traditional and monotonous testing methods,this tool has both theoretical and practical loads,with the comparison of the results highlighting the difference in performance between actual and theoretical scenarios.The testing results are divided into broadcast and consensus phases to identify performance bottlenecks in asynchronous consensus protocols.The tool also enables testing under various scenarios,such as transaction duplication and node failure,to simulate real-world testing conditions.Using AsyncTool,this article conducts performance tests on high-efficiency BEAT.PACE.and FIN protocols on Amazon Web Services.The results show a significant difference between the actual and theoretical throughput,indicating that the scalability of asynchronous consensus protocols in real-world scenarios is not ideal.Moreover,the article demonstrates that the cost of achieving high throughput in asynchronous consensus protocols is high latency,which is unacceptable in practical applications.When considering transaction duplication,the performance of asynchronous consensus protocols is significantly reduced,with a maximum reduction of up p to l/n,resulting in a considerable waste of bandwidth.In response to the problems identified through the testing and analysis,this article proposes solutions based on existing technologies.Parallelizing the broadcast and consensus phases can fully utilize bandwidth and address the practical scalability and high-latency issues in asynchronous consensus protocols.Additionally,the article proposes a super-proxy node to filter out duplicate transactions and send batch transactions to increase the number of transactions per consensus round.This method can also effectively alleviate resource competition among consensus nodes.