Bycon: a Byzantine-Fault-Tolerant Consensus Algorithm

In a leader-based Byzantine-fault-tolerant consensus algorithm, the view change protocol replaces a leader when the cluster detects the faulty behavior of the leader. Practical Byzantine fault tolerance (PBFT) algorithm generates a new leader in a round-robin pattern and inefficiently deals with the log inconsistency in the view change protocol. In this paper, we present Bycon, a Byzantine-fault-tolerant algorithm with an efficient view change protocol. To ensure efficiency of the view change protocol, the Bycon cluster elects a leader with the complete log list in the view change protocol and deals with the log inconsistency in the normal log replication protocol. The experimental results show that the view change protocol of Bycon has a constant cost among time, CPU and network flow. But the cost of PBFT has a positive correlation with the number of prepared logs since the last checkpoint.