A Multi-Layer PBFT Consensus Algorithm with Inter-group Supervision

Yuting Zhao,Bing Guo,Chaoxia Qin,Mingjie Zhao
DOI: https://doi.org/10.1109/hpcc-dss-smartcity-dependsys57074.2022.00174
2022-01-01
Abstract:In order to solve the problem of poor scalability of the Practical Byzantine Fault Tolerance (PBFT) algorithm widely used in blockchain, many variant PBFT-based consensus algorithms reduce the communication complexity by running consensus protocol in a hierarchical grouping structure. However, it also brings greater challenges to ensuring consistency for the whole network. In this paper, we innovatively add the inter-group supervision mechanism to the multi-layer PBFT consensus process, the representative nodes of each consensus group check the consensus messages in other groups to prevent the malicious primary from breaking inter-group consistency. And a group number assignment algorithm based on VRF is designed to ensure the effective implementation of supervision. Next, to maximize the scalability of the system model, an objective function of optimizing the consensus structure is analyzed. This objective function can find a hierarchical grouping structure with minimum communication complexity while maintaining 1/3 Byzantine fault tolerance. At last, a hierarchical strategy based on node reputation is designed, which further strengthen the security of the multi-layer consensus model. The performance of the protocol is evaluated by simulations, and the results show that it can effectively reduce the communication complexity of the system while ensuring the Byzantine fault tolerance.
What problem does this paper attempt to address?