Delta - An Option Trading Platform

Delta is a high-performance, microservices-based option trading platform designed for speed, scalability, and reliability. It features a real-time matching engine, live market data streaming, and a responsive React frontend.

System Architecture

Delta uses an event-driven microservices architecture to handle high-frequency trading operations efficiently.

The Flow

1. Client Request

   • The user places an order via the FrontendClient.
   • The request hits the HTTP Server (behind Nginx/Docker Proxy).

2. Order Queuing

   • The HTTP Server validates the request and pushes it to a Redis Stream (trade).
   • Instead of blocking, the HTTP server subscribes to a specific response channel and waits.

3. Order Processing (The Engine)

   • The Engine (in-memory) polls the trade stream.
   • It matches orders, updates balances, and manages positions in microseconds.
   • Market Data: The Poller service continuously feeds live prices from Backpack Exchange into the Engine via Redis.
   • Liquidation: If prices move against a user, the Engine auto-liquidates positions within 3-5 seconds.

4. Response & Feedback

   • Once processed, the Engine publishes the result back to Redis Pub/Sub.
   • The HTTP Server picks up the result and sends the final response to the user.
   • Simultaneously, the Engine broadcasts updates to the Frontend via WebSockets (through the Poller/HTTP layer).

5. Persistence (DB Worker)

   • To keep the Engine fast, it never writes to the database directly.
   • It pushes completed trades to the engine-response stream.
   • The DB Worker asynchronously reads this stream and saves data to PostgreSQL.

Why Microservices?

• Independent Scaling: We can scale the HTTP Server to handle millions of connections without touching the Engine.
• Performance: The Engine remains single-threaded and lock-free (in-memory), unburdened by slow database writes.
• Reliability: If the DB Worker crashes, the Engine keeps running. Data piles up in Redis and is processed once the worker restarts.

---

Services Overview

Service	Technology	Role	Documentation
Frontend	React, Vite, Zustand	User Interface, Charts & Trading Panel	Read More
HTTP API	Express, Bun	REST API, Auth, Request Gateway	Read More
Engine	Node/Bun	Order Matching, Risk Management	Read More
Poller	Node/Bun, WS	Market Data Aggregator	Read More
DB Worker	Prisma, Bun	Async Database Writer	Read More

---

Project Setup

Prerequisites

• Docker & Docker Compose
• Bun (for local development)

Quick Start (Local)

1. Clone the repository:

   git clone https://github.com/SamadeshPoudel/Option-Trading-Platform.git
   cd Option-Trading-Platform

2. Setup Environment Variables: Copy .env.example to .env in the root and each service directory.

   cp .env.example .env
   # Repeat for frontend/, http/, engine/, etc.

3. Start Infrastructure (Redis & Postgres):

   docker compose up -d redis db

4. Run Services: You can run services individually using bun run index.ts in their respective directories, or use Docker Compose to spin up the whole stack:

   docker compose up --build

5. Access the Platform:

   • Frontend: http://localhost:5173
   • Backend API: http://localhost:5000

---

Tech Stack

• Frontend: React 19, TypeScript, TailwindCSS v4, Lightweight Charts
• Backend API: Node.js/Bun, Express, Better-Auth
• Core Engine: In-Memory Typescript Engine
• Database: PostgreSQL (Prisma ORM)
• Message Broker: Redis Streams & Pub/Sub
• DevOps: Docker, Docker Compose, Nginx

---

Live link: delta.samadesh.com