[Paper Review] Cogsworth: Byzantine View Synchronization

Oded Naor, Mathieu Baudet+2 more|arXiv (Cornell University)|Sep 11, 2019

Distributed systems and fault tolerance7 citations

TL;DR

Cogsworth presents a Byzantine view synchronization algorithm that achieves linear communication complexity and constant latency in the partial synchrony model, significantly improving upon quadratic-cost existing approaches. It enables efficient, scalable consensus in BFT systems by decoupling view synchronization from consensus, ensuring all honest nodes converge on the same view with minimal overhead under normal conditions.

ABSTRACT

Most methods for Byzantine fault tolerance (BFT) in the partial synchrony setting divide the local state of the nodes into views, and the transition from one view to the next dictates a leader change. In order to provide liveness, all honest nodes need to stay in the same view for a sufficiently long time. This requires \emph{view synchronization}, a requisite of BFT that we extract and formally define here. Existing approaches for Byzantine view synchronization incur quadratic communication (in $n$, the number of parties). A cascade of $O(n)$ view changes may thus result in $O(n^3)$ communication complexity. This paper presents a new Byzantine view synchronization algorithm named Cogsworth, that has optimistically linear communication complexity and constant latency. Faced with benign failures, Cogsworth has expected linear communication and constant latency. The result here serves as an important step towards reaching solutions that have overall quadratic communication, the known lower bound on Byzantine fault tolerant consensus. Cogsworth is particularly useful for a family of BFT protocols that already exhibit linear communication under various circumstances, but suffer quadratic overhead due to view synchronization.

Motivation & Objective

• To formally define the Byzantine view synchronization problem as a prerequisite for liveness in state machine replication (SMR) under partial synchrony.
• To address the quadratic communication overhead of existing view synchronization mechanisms in Byzantine fault tolerance (BFT) protocols.
• To design a solution that achieves linear communication complexity and constant latency in expectation and under broad conditions, improving scalability.
• To decouple view synchronization from consensus cores, enabling modular optimization of BFT protocols.

Proposed method

• Cogsworth introduces a view synchronization abstraction that operates independently of the consensus protocol, enabling modular design.
• It uses a reliable broadcast primitive to ensure all honest nodes agree on the next view, forming a consensus-like synchronization layer.
• The algorithm employs a threshold-based view advancement mechanism, where nodes advance to the next view only after receiving a sufficient number of view-change messages.
• It leverages a PaceMaker-style mechanism inspired by HotStuff, but optimized for Byzantine fault tolerance with bounded latency and communication.
• The protocol ensures all honest nodes remain in the same view for a minimum duration, enabling progress in the outer consensus protocol.
• It achieves optimistically linear communication by minimizing redundant message exchanges and using efficient message aggregation and validation.

Experimental results

Research questions

• RQ1 Can view synchronization in Byzantine fault-tolerant systems be achieved with linear communication complexity and constant latency, independent of the consensus core?
• RQ2 What is the minimal communication cost for achieving view synchronization in the presence of Byzantine faults under partial synchrony?
• RQ3 How can view synchronization be decoupled from consensus to allow for modular, efficient BFT protocol design?
• RQ4 What conditions enable linear communication complexity in view synchronization under benign and Byzantine failure models?
• RQ5 Can a view synchronization mechanism be designed that maintains constant latency while ensuring liveness and safety in the face of Byzantine faults?

Key findings

• Cogsworth achieves linear communication complexity in expectation and constant latency under a broad set of conditions, including benign failures.
• The protocol ensures all honest nodes remain in the same view for a bounded duration, enabling progress in the outer consensus protocol.
• Cogsworth reduces the overall communication complexity of BFT systems from O(n³) to O(n²) in the worst case, matching the known lower bound for Byzantine consensus.
• The algorithm maintains constant latency during synchrony, with message delays bounded by δ, and performs optimally under partial synchrony.
• It outperforms existing view synchronization mechanisms that incur quadratic communication costs, especially under cascading view changes.
• Cogsworth is particularly effective for BFT protocols that already have linear communication in normal operation but suffer from quadratic overhead due to view synchronization.