A high-capacity slicing PBFT protocol based on reputation evaluation model

Consortium blockchains, characterized by regulated blockchain technologies with limited authorization management, have gained popularity in various domains, including supply chain, Internet of Things (IoT), justice, and education, and have now become an integral part of production life. Ensuring information consistency in blockchains, the practical byzantine fault tolerance (PBFT) consensus algorithm is deemed more suitable for application in partially decentralized systems. However, existing consensus algorithms for consortium blockchains exhibit limited effectiveness. Particularly, with the number of nodes increasing, the performance of PBFT for consortium blockchains declines significantly, hindering the development of large-scale consortium blockchains. In this paper, we propose a novel consensus algorithm called reputation slice practical byzantine fault tolerance (RSPBFT) based on the reputation evaluation model. The RSPBFT adopts a slicing approach, achieving broad consensus through orchestrated micro-consensuses. Moreover, the reputation evaluation model assigns distinct reputation evaluation scores to nodes, with high-scoring nodes being preferred as leaders. A concept of "Gray-Area" is introduced to enhance the stability of the consensus model by distinguishing different node types. Simulation experiments demonstrate that in large-scale node deployment scenarios, RSPBFT outperforms other consensus algorithms. Furthermore, in consortium blockchain environments with an equal number of nodes as other consensus models, RSPBFT exhibits superior consensus stability.