A Practical Byzantine Fault Tolerance Algorithms Based on Randomized Mean Clustering, Trust and Credibility
摘要
The Practical Byzantine Fault Tolerance (PBFT) consensus protocol, while robust, faces efficiency challenges with the escalating number of network nodes. Various enhancement algorithms employ grouping strategies to bolster performance in large-scale networks. However, as nodes consolidate into smaller clusters, their capacity to accommodate Byzantine nodes diminishes markedly, prompting the need for a systematic approach to distributing Byzantine nodes evenly across consensus regions. In order to address the above issues, this paper proposes a Practical Byzantine fault tolerant algorithm based on Randomized Mean Clustering, Trust, and Credibility (CTPBFT). Nodes possess two key attributes: trust and credibility. Trust determines the priority of node cluster allocation, while credibility evaluations ensure cluster uniformity. A customized random equalization algorithm assigns nodes within the federated chain to diverse clusters. Post-partitioning, the consensus process is streamlined, with each cluster autonomously conducting consensus activities. A well-designed reward function incentivizes nodes to execute the consensus protocol, aiding the system’s smooth and efficient operation.