An Adaptive and Hierarchical Approach to Configure Sharding Committees in Blockchains

The scalability challenge in consensus methods based on Proof of Work (POW) has led to the popularity of voting-based consensus methods. Practical Byzantine Fault Tolerance (PBFT) is one of the most desired approaches. However, the PBFT protocol faces communication complexity and possible faulty behaviors of nodes that can disrupt the consensus process. To reduce the complexity of PBFT, sharding is a useful technique. However, sharding may decrease the probability of successful consensus in PBFT, as the effects of faulty nodes become more destructive. In this paper, we introduce a shard-based Dynamic Balanced Consensus Tree (DBCT) that is well-suited for performing PBFT hierarchically. Additionally, we propose a distributed configuration method that gradually develops this structure while the PBFT consensus is in progress. Proper distribution of node weights in Sharding committees can provide a basis to propose appropriate ideas for aggregating votes in PBFT, control node clustering based on weight parameters, and provide a fair and proportional method to select the block-proposer node. Furthermore, the unpredictable dynamic configuration of this consensus structure could be considered as a defense strategy against some possible attacks. The simulation results confirm the quality of our proposed configuration method to make a dynamic and balanced consensus tree while the consensus is in progress.