# Distributed High-Concurrency Ticketing Engine

An asynchronous, queue-backed distributed system designed for high-throughput flash sales and event ticketing. The engine guarantees zero race conditions, zero overselling, and sub-2ms inventory gatekeeping under heavy multi-threaded traffic.

---

## The Core Problem Solved

During high-demand flash sales (e.g., concert ticketing), thousands of concurrent requests hit the database simultaneously. Standard database transactions often fail to handle race conditions, leading to overselling, heavy database locks, and system crashes.

This project solves that by shifting state validation completely to in-memory atomic operations and offloading heavy database writes to asynchronous background queues.

---

## Architecture & Key Features

* Sub-2ms Gatekeeping (Redis Atomic Operations):

Utilizes Redis atomic DECR commands for instant stock validation, bypassing relational/document DB read bottlenecks entirely.

* Dynamic Multi-Event Idempotency:

Employs Redis Sets SADD / SISMEMBER) to enforce strict per-user, per-event booking idempotency, eliminating duplicate bookings.

* Asynchronous Queue Persistence (BullMQ):

Decouples the HTTP response cycle from database persistence. Accepted bookings are queued instantly into BullMQ background workers for asynchronous MongoDB Atlas document creation.

* Zero-Dependency Execution (Docker):

Containerized multi-service ecosystem (Node.js API, Redis instance, BullMQ worker process) managed cleanly via docker-compose.

---

## Benchmarks & Load Testing Proof

System reliability was tested using a multi-threaded asynchronous load generator firing 500 simultaneous requests against a limited stock of 100 tickets:

* Total Simultaneous Requests: 500

* Successful Bookings (202 Accepted): 100

* Rejected Requests (400 Sold Out): 400

* **Double Bookings / Race Conditions:** 0

* **Average Redis Validation Latency:** < 2ms

---

## Tech Stack

* Runtime & Backend: Node.js, Express.js

* Caching & In-Memory Lock: Redis

* Message Queue & Worker: BullMQ

* Database: MongoDB Atlas, Mongoose

* DevOps & Infrastructure: Docker, Docker Compose

* Testing: Custom Asynchronous Multi-Threaded Load Scripts