Clustering PBFT Improvement Based on Trusted Lists

Abstract

Blockchain has great application value in the field of finance, Internet of Things, digital copyright, decentralization is the most fundamental feature of blockchain. Decentralization increases the fault tolerance and security of the whole blockchain network, ensuring that the data cannot be tampered with, but decentralization greatly increases the amount of message communication in the whole system, so the whole blockchain network is very inefficient in the case of consensus of multiple nodes. The consensus algorithm is the core technology for implementing decentralization.In this paper, we propose a CG-PBFT consensus algorithm based on credit list grouping hierarchy, which aims to solve the problems of wasteful communication resources and inefficiency caused by too many broadcast messages in PBFT (Practical Byzantine Fault Tolerance algorithm). The implementation of the CG-PBFT algorithm is divided into three main steps: First, by selecting nodes with higher credit value as group leader nodes, the harm of malicious nodes to the whole consensus system can be significantly reduced. Second, the optimal grouping according to the communication delay between nodes can improve the efficiency of the whole consensus system. Finally, the intra-group consensus is performed first in the consensus phase, followed by the global consensus, which can significantly reduce the network overhead. Experiments demonstrate that the CG-PBFT consensus algorithm has lower network overhead, lower network latency, and higher Byzantine node fault tolerance.