Building Scalable Real-Time Chat Systems: WebSockets, Redis Pub/Sub, and Message Queue Infrastructure

Real-time communication platforms require ultra-low latency, bi-directional socket state management, and reliable message persistence. Scaling chat infrastructure from a single node to millions of concurrent active connections across React.js, Next.js, and React Native mobile apps demands a decoupled event-driven backend architecture.

1. The Core Real-Time Stack: WebSockets + Redis Pub/Sub

Because stateful WebSocket connections reside on specific backend server instances, horizontal scaling requires a distributed message bus. Redis Pub/Sub acts as the inter-node broker, allowing server instance A to broadcast a message to server instance B where the target recipient is connected.

// Node.js WebSocket Gateway with Redis Adapter
import { WebSocketServer } from "ws";
import { createClient } from "redis";

const pubClient = createClient({ url: process.env.REDIS_URL });
const subClient = pubClient.duplicate();
await Promise.all([pubClient.connect(), subClient.connect()]);

const wss = new WebSocketServer({ port: 8080 });

subClient.subscribe("chat:room:general", (message) => {
  wss.clients.forEach((client) => {
    if (client.readyState === 1) {
      client.send(message);
    }
  });
});

2. Message Schema & Dual-Database Storage Strategy

Chat systems use a dual-storage model: ScyllaDB/Cassandra or MongoDB for ultra-fast time-series append-only chat history, and PostgreSQL for relational user metadata and room permissions. Redis Hashes maintain real-time user presence ('online', 'idle', 'offline') with auto-expiring TTL heartbeats.