L1 Overview

Introduction

The Hyperliquid L1 is custom built around a performant derivatives exchange as the flagship native component. A perpetuals order book exchange was chosen for the following key reasons:

1. It is a realistic real-world application with more infrastructure demands than any existing L1 can handle.

2. It is the most valuable vertical in defi upon from which most user built applications would benefit.

3. It drives real users to interact with the underlying L1 infrastructure.

The Hyperliquid L1 is continually pushed by the flagship native perps DEX, leading to crucial optimizations that general purpose chains miss in their design.

Note that the L1 state includes all margin and matching engine states. Importantly, Hyperliquid does not rely on the crutch of off-chain order books. A core design principle is full decentralization with one consistent order of transactions achieved through BFT consensus.

Latency

Consensus currently uses an optimized consensus algorithm called HyperBFT, which is optimized for end-to-end latency. End-to-end latency is measured as duration between sending request to receiving committed response.

For an order placed from a geographically co-located client, end-to-end latency has a median 0.2 seconds and 99th percentile 0.9 seconds. This performance allows users to port over automated strategies from other crypto venues with minimal changes and gives retail users instant feedback through the UI.

Throughput

Mainnet currently supports approximately 100k orders/sec. The current bottleneck is execution. The consensus algorithm and networking stack can scale to millions of orders per second once the execution can keep up. There are plans to further optimize the execution logic once the need arises.