A Double-Layer PBFT Consensus Algorithm Based on Node Evaluation and Schnorr Signature

This paper introduces an enhanced Practical Byzantine Fault Tolerance (PBFT) consensus algorithm, named Double-layer Behavior-scoring and Schnorr Signatures PBFT (DBS-PBFT), that addresses the inherent limitations of the traditional PBFT algorithm, such as high communication overhead, vulnerability to malicious nodes, and inefficient primary node selection. The algorithm integrates a node behavior-scoring evaluation model, a two-layer consensus process, and Schnorr aggregate signatures. Nodes are dynamically grouped based on performance scores derived from historical behavior, response speed, and communication success rate, with high-scoring nodes selected as group leaders. The two-layer consensus reduces message complexity by limiting broadcasts to intra-group and inter-leader communications, while Schnorr signatures streamline the verification process. Experiments were conducted in a simulated Python environment with 20 to 70 nodes to compare DBS-PBFT with the traditional PBFT algorithm. Results demonstrated a reduction in consensus latency by up to 40 % and a decrease in communication complexity from quadratic to linear scalability. The study demonstrates that DBS-PBFT significantly enhances efficiency and security for consortium blockchains, although scalability in ultra-large-scale networks warrants further investigation. The algorithm provides a practical solution for applications demanding low-latency, high-throughput consensus without compromising decentralization.