FinWhale DAG protocol achieves two-round consensus, beating prior limits
FinWhale slashes DAG consensus latency with a two-message fast path under partial synchrony.
FinWhale, introduced by Razya Ladelsky and Roy Friedman, is a new DAG-based Byzantine Fault Tolerant (BFT) protocol that achieves consensus in just two message delays during favorable conditions. Existing DAG protocols (like Narwhal, Mysticeti) require at least three rounds, even under optimal network conditions. FinWhale extends Mysticeti by adding a novel fast path commit mechanism that coexists safely with the original slow path rules. To maintain consistency across different local DAG views, the protocol introduces commit structures based on fast path evidence blocks, allowing validators to merge fast and slow path reasoning seamlessly.
FinWhale operates under the partially synchronous model and achieves optimal resilience with n = 3f + 2p - 1 validators, tolerating up to f Byzantine faults. The fast path succeeds whenever at most p validators are faulty (p between 1 and f). This matches known lower bounds for fast Byzantine consensus. The work demonstrates that optimal-latency fast paths can be integrated into uncertified DAG consensus protocols, potentially improving throughput and latency in permissioned blockchains, distributed ledgers, and other fault-tolerant systems.
- FinWhale is the first DAG-based BFT protocol to achieve consensus in two message delays via a fast path, compared to the usual three or more.
- It introduces fast path evidence blocks to safely combine fast and slow path reasoning across different local DAG views.
- Optimal resilience: tolerates f Byzantine faults with n = 3f + 2p - 1 validators, where p ≤ f is the number of faulty validators tolerated during fast path.
Why It Matters
FinWhale sets a new latency benchmark for DAG-based BFT consensus, enabling faster finality in blockchains and distributed systems.