An improved PBFT consensus algorithm based on grouping and credit grading

Area of Science:

• Blockchain Technology
• Distributed Systems
• Computer Science

Background:

• Practical Byzantine Fault Tolerance (PBFT) suffers from high communication overhead and limited network scalability.
• Existing consensus algorithms struggle to balance efficiency, security, and network size.

Purpose of the Study:

• To propose an enhanced Byzantine fault-tolerant consensus algorithm (CBFT) that addresses the limitations of PBFT.
• To improve communication efficiency, network size support, and security in blockchain consensus.

Main Methods:

• Developed a novel Credit-based Byzantine Fault-Tolerant (CBFT) consensus algorithm.
• Implemented a grouping model dividing nodes by response speed for separate intra- and inter-group consensus.
• Integrated a credit model to assign different responsibilities to node types, reducing malicious master node probability.

Main Results:

• CBFT demonstrated 3.1x higher throughput than PBFT and 1.5x higher than GPBFT with 52 nodes.
• Latency was reduced to 7.4% of PBFT and 38.8% of GPBFT.
• Communication overhead was reduced to 6.4% of PBFT and 87.3% of GPBFT.
• Byzantine fault tolerance improved by 59.3% with 300 nodes, with gains increasing with network size.

Conclusions:

• The proposed CBFT algorithm significantly enhances blockchain consensus efficiency and scalability.
• CBFT offers a more secure and performant alternative to traditional PBFT, especially in large-scale networks.
• The grouping and credit models effectively reduce communication overhead and mitigate risks associated with malicious nodes.