Joint Reputation Based Grouping and Hierarchical Byzantine Fault Tolerance Consensus Protocol

Abstract

Consensus protocol has challenges in terms of low consensus efficiency and centralization, as well as poor fault tolerance. A joint reputation model based grouping and hierarchical byzantine fault tolerance consensus protocol has been proposed. It is composed of both grouping and hierarchical models. The grouping model uses joint reputation values to balance node grouping, minimize overall differences in joint reputation values among groups. Some nodes with a joint reputation value ranking in the top 50% of the group are randomly selected as function ones, which improves the degree of decentralization compared to some election strategies of leader nodes. In the hierarchical model, nodes in each group are layered and the supervision nodes are mapped to the upper layer, which improves consensus network security and efficiency. Besides a new communication structure has been designed to improve fault tolerance and reduce communication complexity by improving the three phases of Practical Byzantine Fault Tolerant (PBFT). Comparative experiments have shown the superiority of the developed protocol over other existing protocols.