From PBFT to the present: a thorough overview of blockchain consensus protocols

The emergence of Bitcoin has introduced blockchain technology, which has been widely applied in various scenarios, prompting the continuous pursuit of more efficient, energy-saving, and secure blockchain consensus protocols. Byzantine fault tolerance (BFT) consensus serves as an effective solution in specific contexts. From the Byzantine generals problem in 1982 to the first practical BFT protocol, PBFT in 1999, an increasing number of BFT protocols have been proposed and applied to blockchain systems. The surge in protocols has made it difficult to navigate the BFT landscape, let alone discern which specific protocols optimally align with diverse application requirements. To facilitate a better understanding, design, and application of blockchain consensus protocols, this paper adopts PBFT as a baseline to critically assess the design of BFT variants in this paper. We categorize the design philosophies of existing blockchain BFT consensus into two primary types: the redesign of fundamental consensus processes (RFCP) and the addition of S optimization mechanisms (ASOM). For RFCP, taking the consensus processes of PBFT as a reference, we analyze the design considerations such as the number of commitment phases, the number of replicas, and view-change strategies. These structural elements interact with each other, affecting the protocol's fundamentals and correctness. For ASOM, we outline several optimizations, such as committee-based consensus, optimistic mechanisms, and pipelining, which can be layered without affecting the fundamental consensus processes. Finally, according to the proposed design philosophies, we present a detailed framework to rapidly analyze or design a specified blockchain BFT consensus protocol.