
Briefing
This paper presents a formal model and machine-checked proof for a DAG-based Byzantine Fault Tolerant (BFT) consensus protocol that supports dynamic validator sets and stake-weighted decisions. It rigorously demonstrates that blockchains built on this model maintain their nonforking property, even with participants continuously joining and leaving. This foundational work establishes a new standard of provable security for adaptive, long-lived decentralized systems, critical for future blockchain scalability and resilience.

Context
Prior to this research, many DAG-based BFT consensus protocols often relied on correctness proofs that assumed a fixed set of network participants. This theoretical limitation presented a significant challenge for long-lived blockchain systems where validator sets dynamically evolve. The absence of comprehensive machine-checked proofs also meant subtle flaws could persist, as evidenced by errors discovered in some published proofs.

Analysis
The core mechanism involves modeling the protocol as a labeled state transition system within the ACL2 theorem prover. The authors define and prove a set of interdependent state invariants, including certificate nonequivocation and anchor nonforking, culminating in the proof of blockchain nonforking. This formal approach extends traditional quorum intersection arguments to accommodate dynamic stake, where decisions are weighted by validator contributions. A “lookback” mechanism is integrated into the protocol’s design, delaying committee changes to ensure consistent decision-making across rounds.

Parameters
- Core Concept ∞ Blockchain Nonforking Proof
- Formalization Tool ∞ ACL2 Theorem Prover
- Consensus Protocol Basis ∞ DAG-based BFT (Narwhal, Bullshark, AleoBFT)
- Key Feature ∞ Dynamic Stake-Weighted Validator Sets
- Proof Type ∞ Inductive Invariant Proofs
- Authors ∞ Alessandro Coglio, Eric McCarthy
- Publication Date ∞ April 23, 2025
- Primary Property Verified ∞ Consistency (Nonforking)

Outlook
This work provides a robust theoretical framework for designing and analyzing highly dynamic blockchain consensus protocols. Future research will build upon this foundation to explore additional properties such as liveness and more intricate syncing mechanisms. The practical application of this verifiable design extends to next-generation decentralized networks, ensuring their foundational integrity in environments characterized by continuous participant evolution and high-stakes operations.

Verdict
This research decisively advances the foundational principles of blockchain security by providing the first machine-checked proof of nonforking for DAG-based BFT consensus with dynamic, stake-weighted validator sets.
Signal Acquired from ∞ alessandrocoglio.info