区块链以其去中心化、分布式共识、信息不可篡改等特点被广泛应用，Hyperledger Fabric是一个开源的企业级许可区块链平台，系统性能对于Hyperledger Fabric区块链有效性评估显得至关重要。以Hyperledger Fabric为框架基于广 义随机Petri网(Generalized Stochastic Petri Nets，GSPN)进行建模对系统进行评估，给出了以Hyperledger Fabric v1。2为框架的系统结构图，并对系统交易流程的5个阶段进行了详细描述。针对系统结构建立GSPN模型，对每个交易流程进行了状态和数值分析，给出了参数配置的方法，通过虚拟机搭建实验平台和Matlab仿真。结果表明，区块交易量N值对系统性能有很大影响，N值大于5时系统性能瓶颈从提交阶段变为背书阶段，对复杂的区块链所提性能分析方法具有可扩展性。
Modeling and Analysis of Blockchain System Performance Based on GSPN
Blockchain is gradually and widely used due to its decentralization,distributed consensus,and tamper-proof feature.Hyperledger Fabric is an open-source enterprise-level permissioned blockchain platform,system performance is critical to the evaluation of the effectiveness of the Hyperledger Fabric blockchain.Therefore,the article uses Hyperledger Fabric as the framework to model and evaluate the system based on Generalized Stochastic Petri Nets(GSPN).First,a system structure diagram based on Hyperledger Fabric v1.2 is given and five phases of the system transaction flow are described in detail.Secondly,a GSPN model is established for the system structure.In the model,the status and numerical analysis of each transaction flow are carried out,and the method of parameter configuration is given.Finally,virtual machine building platform experiment and Matlab simulation results show that the value of the block transaction volume N has a great impact on system performance.When N is greater than 5,the system performance bottleneck changes from the submission stage to the endorsement phase.The performance analysis method proposed by the block chain is scalable.
国家科技期刊平台
NSTL
万方数据
