Bullshark on Narwhal: Implementation-level Workflow Analysis of Round-based DAG Consensus in Theory and Practice

TL;DR

This paper presents Bullshark, a high-performance round-based DAG BFT protocol on Narwhal, analyzing its implementation workflow and demonstrating its ability to process 297,000 transactions per second with low latency.

Contribution

It provides an implementation-level workflow analysis of Bullshark on Narwhal, highlighting its performance and detailed protocol interactions, which were previously underexplored.

Findings

Achieves 297,000 transactions per second

Maintains 2-second latency

Provides detailed workflow analysis

Abstract

Round-based DAGs enable high-performance Byzantine fault-tolerant consensus, yet their technical advantages remain underutilized due to their short history. While research on consensus protocols is active in both academia and industry, many studies overlook implementation-level algorithms, leaving actual performance unclear - particularly for theoretical protocols whose practical performance cannot often be evaluated. Bullshark, a Round-based DAG BFT protocol on Narwhal mempool, achieves optimal performance: 297,000 transactions per second with 2-second latency. We analyze the algorithm's workflow, from transaction submission to blockchain commitment, breaking it down layer by layer at the functional level and delineating the key features and interactions of the Bullshark and Narwhal components. Future work aims to improve performance in Byzantine fault environments and optimize trade-offs in the CAP theorem.