Lemonshark: Asynchronous DAG-BFT With Early Finality
summary
Lemonshark is an asynchronous DAG‑BFT that reinterprets the DAG at the transactional level to deliver early finality for blocks before leader commitment by combining a reversed breadth‑first causal sort and key‑space partitioning into N shards with a single-writer-per-shard schedule. It formalizes a persistence condition and recursive reference chains to guarantee safety, is implemented by forking Bullshark, and shows up to 65% latency reduction (25%+ under heavy cross‑shard) across three-run averages while tolerating crash and Byzantine behavior.
AI-written summary · about
chapters (10)
- 0:00Lemon Shark protocol intro
- 1:01DAG-based BFT protocols
- 2:46Leader-based finality delays
- 4:54Key insight for early finality
- 7:00Transaction types and sharding
- 8:24Early finality in single-shard
- 11:32Extending early finality
- 12:37Evaluation and performance
- 13:42Resilience under crash faults
- 14:22Conclusion and future
related talks
16mShoal++: High Throughput DAG BFT Can Be Fast and Robust!
01h 12mScaling distributed ledgers: Achieving arbitrary throughput and sub-second latency
15mAlea-BFT: Practical Asynchronous Byzantine Fault Tolerance
29mBuilding Probabilistic BFT Protocols Through Subsampled Voting
18mChitu: Avoiding Unnecessary Fallback in Byzantine Consensus
22mOptimizing network topologies...
16mDispersedLedger: High-Throughput Byzantine Consensus on Variable Bandwidth Networks
33mLibraBFTv2: Optimistically-linear BFT Consensus with Concrete Latency Bounds