[Paper Review] Reaching Consensus in the Byzantine Empire: A Comprehensive Review of BFT Consensus Algorithms

Gengrui Zhang, Fei Pan+6 more|arXiv (Cornell University)|Apr 7, 2022

Distributed systems and fault tolerance4 citations

TL;DR

This paper presents a systematic taxonomy and comprehensive analysis of state-of-the-art Byzantine fault-tolerant (BFT) consensus algorithms, categorizing them into efficient, robust, and available families. It provides intuitive, step-by-step protocol descriptions with message-passing diagrams, compares their time and message complexities, and evaluates strengths and weaknesses to aid researchers and practitioners in understanding and selecting optimal BFT solutions for real-world distributed systems.

ABSTRACT

Byzantine fault-tolerant (BFT) consensus algorithms are at the core of providing safety and liveness guarantees for distributed systems that must operate in the presence of arbitrary failures. Recently, numerous new BFT algorithms have been proposed, not least due to the traction blockchain technologies have garnered in the search for consensus solutions that offer high throughput, low latency, and robust system designs. In this paper, we conduct a systematic survey of selected and distinguished BFT algorithms that have received extensive attention in academia and industry alike. We perform a qualitative comparison among all algorithms we review considering message and time complexities. Furthermore, we decompose each consensus algorithm into its constituent subprotocols for replication and view change backed by intuitive figures to illustrate the message-passing pattern. We also elaborate on the strengths and weaknesses of each algorithm as compared to the state-of-the-art approaches.

Motivation & Objective

• To address the growing complexity and inconsistency in understanding modern BFT consensus algorithms due to diverse assumptions and metrics.
• To provide a unified, intuitive, and comprehensive description of key BFT algorithms through structured decomposition into core subprotocols.
• To categorize prominent BFT algorithms into three main families: more efficient, more robust, and more available, based on design priorities.
• To compare the time and message complexities of surveyed algorithms systematically using a standardized framework.
• To support researchers and practitioners in evaluating and selecting BFT consensus protocols for real-world applications in blockchain, IoT, and machine learning.

Proposed method

• Propose a novel three-category taxonomy: more efficient, more robust, and more available BFT algorithms, based on design goals and system properties.
• Decompose each BFT algorithm into core subprotocols—normal operation, view changes, conflict resolution, and robustness mechanisms—using consistent structural descriptions.
• Use intuitive diagrams to illustrate message-passing patterns and workflows for each consensus instance across all surveyed algorithms.
• Conduct a qualitative comparison of all algorithms based on message complexity and time complexity, with tabulated results in Table 1.
• Highlight unique features such as pipelining (e.g., HotStuff), attack defense (e.g., Prime, Aardvark), fairness (e.g., Pompe, Fairy), and leaderless/asynchronous designs (e.g., Honeybadger, BEATS).
• Integrate discussions on emerging challenges such as post-quantum security, interoperability, and BFT in distributed machine learning to contextualize real-world applicability.

Experimental results

Research questions

• RQ1 How can state-of-the-art BFT consensus algorithms be systematically categorized based on their primary design objectives?
• RQ2 What are the key differences in message and time complexity across prominent BFT algorithms, and how do these affect performance in practice?
• RQ3 How do subprotocols such as view changes, conflict resolution, and leader election contribute to the correctness and liveness of BFT consensus?
• RQ4 What mechanisms do robust BFT algorithms employ to defend against timing, monopoly, and performance-based attacks?
• RQ5 How do emerging trends such as post-quantum security, interoperability, and integration with machine learning shape the future of BFT consensus?

Key findings

• The taxonomy successfully organizes BFT algorithms into three coherent categories—efficient, robust, and available—each addressing distinct system requirements and trade-offs.
• Algorithms like HotStuff and Zyzzyva achieve lower message complexity through pipelining and optimized quorum structures, reducing communication overhead.
• Robust BFT protocols such as Prime and Aardvark introduce effective defenses against timing and performance-based attacks, enhancing system security.
• Leaderless and asynchronous protocols like Honeybadger and BEATS improve system availability under network partitions and leader failures.
• The integration of reputation mechanisms (e.g., Prosecutor) and fairness mechanisms (e.g., Pompe, Fairy) enhances trust and transparency in transaction ordering.
• Emerging challenges such as post-quantum security and cross-chain interoperability are identified as critical frontiers requiring further BFT consensus innovation.