[論文レビュー] It's not a lie if you don't get caught: simplifying reconfiguration in SMR through dirty logs

Allen Clement, Natacha Crooks+2 more|arXiv (Cornell University)|Feb 10, 2026

Distributed systems and fault tolerance被引用数 0

ひとこと要約

Gauss introduces a modular reconfiguration engine for state-machine replication that separates inner and outer logs, enabling arbitrary membership changes and protocol upgrades with minimal downtime.

ABSTRACT

Production state-machine replication (SMR) implementations are complex, multi-layered architectures comprising data dissemination, ordering, execution, and reconfiguration components. Existing research consensus protocols rarely discuss reconfiguration. Those that do tightly couple membership changes to a specific algorithm. This prevents the independent upgrade of individual building blocks and forces expensive downtime when transitioning to new protocol implementations. Instead, modularity is essential for maintainability and system evolution in production deployments. We present Gauss, a reconfiguration engine designed to treat consensus protocols as interchangeable modules. By introducing a distinction between a consensus protocol's inner log and a sanitized outer log exposed to the RSM node, Gauss allows engineers to upgrade membership, failure thresholds, and the consensus protocol itself independently and with minimal global downtime. Our initial evaluation on the Rialo blockchain shows that this separation of concerns enables a seamless evolution of the SMR stack across a sequence of diverse protocol implementations.

研究の動機と目的

• Motivate the need for modular, upgradable SMR components in production deployments.
• Propose a reconfiguration architecture that decouples reconfiguration from the underlying consensus protocol.
• Define formal safety and liveness properties for evolving SMR deployments across epochs.
• Demonstrate how inner/outer log separation enables seamless transitions between consensus implementations.

提案手法

• Propose Gauss, a three-stage reconfiguration protocol (prepare, handover, shutdown) that transitions between epochs with different memberships and consensus protocols.
• Introduce the log sanitizer that converts the inner log (consensus-specific) to an outer log (SMR-visible) to preserve SMR properties.
• Model epochs with evolving memberships and consensus protocols, with formal safety (assurance of consistent outer log) and liveness (progress) guarantees.
• Provide a worked example and sketches of proofs showing safety under arbitrary epoch transitions.
• Implement Gauss in Rialo and evaluate epoch-change latency across varying validator sets in a local testbed.

Figure 1 . RSM Node SubComponents

実験結果

リサーチクエスチョン

• RQ1 How can SMR reconfiguration be made modular so that membership changes and protocol upgrades can occur independently of the consensus engine?
• RQ2 Can a reconfiguration engine guarantee safety and liveness when transitioning between epochs with different memberships and consensus protocols?
• RQ3 What role does the inner/outer log separation play in enabling seamless reconfiguration across protocol evolutions?
• RQ4 What is the practical performance impact (latency) of epoch changes when validator counts grow?
• RQ5 How can reconfiguration downtime be minimized while preserving correctness across transitions?

主な発見

• Gauss enables arbitrary membership changes and protocol upgrades with minimal downtime by separating inner and outer logs.
• The log sanitizer translates the inner log of a chosen consensus protocol into a sanitized outer log used by other SMR components, enabling protocol-agnostic reconfiguration.
• Three-stage prepare–handover–shutdown protocol coordinates transition between epochs while preserving safety and liveness.
• In a Rialo implementation, Ready-to-Handover phase dominates reconfiguration latency, indicating the main bottleneck is committing Ready messages in the consensus layer.
• Worked example and proof sketches illustrate that outer-log safety is preserved even when different epochs use different consensus protocols and memberships.

Figure 2 . Epoch Transition Example