Exchange · CEX

Southeast Asia spot + contract trading system

Full-stack delivery from zero to 1 month

PeriodAbout 1 month

Business LineCEX spot / perpetual

Core StackGo· Rust / PostgreSQL / Redis / React

RegionSoutheast Asia

Project Background and Challenges

The client is a Southeast Asian overseas team planning to launch a trading platform that supports both spot and perpetual contracts. The core requirements are that matching performance does not degrade under high concurrency, and that the fund account system is secure and auditable.

The client had previously evaluated the white-label solution, but since it could not meet its customized risk control and settlement needs, they turned to custom development.

The scope we undertake

Matching engine, account and clearing system, contract risk control module, market push feed, and management backend. The front-end trading interface collaborates with the client team to complete the process.

Scope 1

Matchmaking engine

Scope 2

Account and clearing and settlement systems

Scope 3

Contract risk control module

Scope 4

Market updates

Scope 5

Manage the backend

Technology stack

Matchmaking

Go/Rust memory matching + persistence separation

Data

PostgreSQL (ledger) + Redis (market caching / matching snapshots)

Real-time

WebSocket market and order status pushes

Front-end

React + TypeScript

Key technical decisions

Decision 01

Matching engine: Self-developedvsOff-the-shelf engine

Choosing self-developed memory matching. Customers need to embed custom risk control rules at the matching layer, making the expansion cost of off-the-shelf engines higher than self-developed ones, and long-term maintenance is more dependent on others.

Decision 02

Ledger consistency

Uses dual bookkeeping + asynchronous reconciliation to ensure strong consistency in fund accounts without blocking the main matching chain.

Decision 03

Market push structure

Matching and push are decoupled, allowing market data to operate on an independent channel to avoid high-frequency market trends crowding matching resources and ensuring stable matchmaking delays.

Delivery Success (Desensitization)

<1ms

Single matching delay (0.1 concurrent sending)

3000+

Peak throughput (pens/second)

<99.99%

System availability after launch

Teams and cycles

The total cycle is about 1 month, involving 10 people, and progressing in four stages:

Requirements and architecture

Core matchmaking

Sub-control and settlement

Joint debugging and going online

Want to see the complete Figma file?

Schedule a 30-minute technical consultation where we walk you through the most relevant Figma / Lanhu files for your project via remote demo, including all pages, component libraries, and interactive prototypes.

Book Technical Consultation Telegram