Title: Don't Count on One to Carry the Ball: Scaling BFT without Sacrifing Efficiency

Abstract: With the emergence of large-scale decentralized applications, a scalable and efficient Byzantine Fault Tolerant protocol of hundreds of replicas is ideal. Although the throughput of existing leader-based BFT protocols has reached a high level of 10^5 operations per second for a small scale of replicas, it drops significantly when the number of replicas increases, which leads to a lack of practicality. This paper focuses on the scalability of BFT protocols and identifies a major bottleneck to leader-based BFT protocols due to the excessive workload of the leader at large scales. A new metric of scaling factor is defined to capture whether a BFT protocol will get stuck when it scales out, which can be used to measure the performance of efficiency and scalability of BFT protocols. We propose Leopard, the first leader-based BFT protocol that scales to multiple hundreds of replicas, and more importantly, preserves high efficiency. It is secure with the optimal resilience bound (i.e., 1/3) in the partial synchronous network model. We remove the bottleneck by introducing a technique of achieving constant scaling factor, which takes full advantage of the idle resource and adaptively balances the workload of the leader among all replicas. We implement Leopard and evaluate its performance compared to HotStuff, the state-of-the-art BFT protocol. We run extensive experiments on the two systems back-to-back in the same environments with up to 600 replicas. The results show that Leopard achieves significant performance improvements both on throughput and scalability. In particular, the throughput of Leopard remains at a high level of 10^5 when the system scales out to 600 replicas. It achieves a 5 times throughput over HotStuff when the scale is 300 (which is already the largest scale we can see the progress of the latter in our experiments), and the gap becomes wider as the scale further increases.