Clownfish DAG protocol slashes communication costs with sparse edges
Quadratic complexity and multiple leaders make DAG-based BFT finally scalable.
Traditional DAG-based BFT protocols achieve high throughput but suffer from poor scalability because each vertex includes a linear number of edges (references) to previous rounds. That O(n²) communication complexity per round limits the number of validators. Clownfish tackles this bottleneck head-on: it selectively reduces the number of edges in DAG vertices, slashing the total communication complexity per round to quadratic (O(n²)) when using a communication-optimal consistent broadcast.
Beyond the core sparse-edges idea, Clownfish introduces two practical improvements. It allows multiple leaders per round, which lowers average latency while preserving the reduced communication cost. It also optimizes the round advancement rule to cut additional latency in failure cases (e.g., when leaders are Byzantine). Experimental results confirm that Clownfish scales significantly better than existing DAG-based BFT protocols, making it a strong candidate for large-scale decentralized systems.
- Achieves quadratic total communication complexity per round using communication-optimal consistent broadcast.
- Supports multiple leaders per round to reduce average latency without increasing communication overhead.
- Optimized round advancement rule reduces additional latency in failure cases compared to prior DAG-based protocols.
Why It Matters
Enables DAG-based BFT consensus to scale to hundreds of validators, critical for next-gen blockchains and distributed ledgers.