Performance Evaluation And Optimization For Blockchain Systems

Blockchain provides a secure and verifiable way to conduct transactions.In a blockchain system,parties share information and reach agreements with predefined rules.Blockchain has three main features: decentralized,traceable,hard to tamper with.The current blockchain technology is still at an early stage of development and is not quite mature in terms of performance,security,and privacy,as well as regulation,especially the performance issue,which is the main factor that restricts the wide application of blockchain.To address the performance issues,this thesis analyzes and tests the system architecture and system performance of blockchain to identify the bottlenecks of the blockchain system through analysis and testing,and then optimize them.Based on four mainstream blockchain systems(Ethereum,Hyperledger Fabric,Hyperledger Sawtooth,and Fisco-Bcos),this thesis compares the architectures,introduces the existing blockchain testing tools and common consensus algorithms,and points out the shortcomings of each of the four blockchains.The main work of this thesis includes:1.This thesis designs and implements a new blockchain performance testing tool-x BCBench Tool.x BCBench Tool can test different blockchain systems using predefined workloads,and by inputting different parameter configurations,the corresponding test results are obtained and visualized output.2.This thesis,four blockchain systems' throughput,and latency characteristics are extensively analyzed using x BCBench Tool to identify system bottlenecks and perform targeted optimization by adjusting parameter configurations and using different workloads.The experimental results show that the performance of the blockchain systems can be substantially improved by a reasonable system configuration.3.This thesis proposes a new consensus algorithm,COPBFT,which can reach consensus quickly,increase throughput,reduce latency,and improve scalability by introducing trust mechanisms and trust groups.The experimental results show that the COPBFT algorithm can substantially improve the performance of the blockchain system during the long-term operation of the system.Overall,the performance of a blockchain system is greatly affected by the system architecture,system configuration,and consensus algorithm.Adopting a suitable system architecture,configuring reasonable system parameters,and employing efficient consensus algorithms are important means to improve the performance of blockchain systems.