A Practical Byzantine Fault Tolerance Algorithm Based on Reputation Scoring Mechanism and Dynamic Timeout Mechanism

Addressing the issues of random primary node selection, vulnerability to delay attacks, and sensitivity to network latency inherent in Practical Byzantine Fault Tolerance (PBFT), this paper proposes a Practical Byzantine Fault Tolerance Algorithm Based on Reputation scoring Mechanism and dynamic Timeout Mechanism algorithm (RT-PBFT). RT-PBFT, initiates by assigning reputation scores to nodes based on metrics including network stability, historical behavior, and contribution level, thereby quantifying node reliability. High-reputation nodes are then selected as primary nodes during view changes. Built upon the reputation scoring mechanism, RT-PBFT incorporates a dynamic timeout mechanism that allocates additional timeout periods to nodes dynamically based on their reputation scores. This design safeguards well-performing nodes from losing their primary status due to minor network fluctuations, reducing the consensus system's sensitivity to network delays. Simultaneously, nodes with lower reputation scores are allotted shorter timeouts, facilitating quicker view changes to replace nodes and evade delay attacks. The RT-PBFT algorithm, realized through these mechanisms, not only optimizes primary node selection but also mitigates delay attacks, tolerating delays for high-reputation nodes while promptly detecting faults in low-reputation ones. Experimental validation demonstrates that compared to PBFT, RT-PBFT significantly improves performance, marked by reduced consensus latency, communication overhead, and increased throughput under network volatility. These enhancements are pivotal for constructing large-scale, high-performance consortium blockchain systems, fulfilling the efficiency and security demands of scenarios such as Industrial IoT, finance, and supply chain management.