Design a Chat System like Google Chat

Question

Design a real-time messaging system. Cover message delivery guarantees, presence indicators, group conversations, message search and history, and handling millions of concurrent connections via WebSockets.

Codemia · Accepted Answer

- **Functional Requirements:**
  - Real-time messaging with delivery guarantees (at least once).
  - Presence indicators to show if users are online, offline, or typing.
  - Support for group conversations with user roles (admin, member).
  - Ability to search through message history.
  - Support for attachments (images, files).
  - Notifications for new messages.

- **Non-Functional Requirements:**
  - Scalability to handle millions of concurrent users.
  - Low latency (<100ms for messages).
  - High availability (99.99% uptime).
  - Data consistency across multiple devices (e.g., mobile, desktop).
  - Security measures to ensure data privacy and integrity (encryption, authentication).

Assuming Google Chat targets 50 million active users:
- **Active Users:** 50 million.
- **Messages per User per Day:** Average of 20 messages.
- **Total Messages per Day:** 50M users * 20 messages = 1 billion messages/day.
- **Messages per Second:** 1 billion messages/day / 86400 seconds = ~11,574 messages/second.
- **Concurrent Connections:** Assuming each user is connected for 4 hours a day, the peak concurrency might be around 20% of active users, leading to:
  - 50M * 0.2 = 10 million concurrent connections.

This implies we need to design for handling at least 11,574 messages per second and 10 million concurrent WebSocket connections.

The architecture can be broken down into the following components:
- **Client Side:** Web and mobile applications using React and Flutter.
- **Load Balancer:** NGINX or Google Cloud Load Balancing to distribute WebSocket connections.
- **WebSocket Server:** Node.js with Socket.io for handling real-time messaging.
- **API Server:** Go or Java-based microservices for handling user management, message history, and search functionality.
- **Database:** Google Cloud Firestore for real-time data syncing and message storage.
- **Caching Layer:** Redis for presence indicators and frequently accessed messages.
- **Search Service:** Elasticsearch for fast message search capabilities.
- **Notification Service:** Google Pub/Sub for delivering message notifications.

The following schema is designed for efficiency and retrieval:
- **Users Table:**
  - `user_id` (String, Primary Key)
  - `username` (String)
  - `status` (Enum: online, offline, typing)

- **Messages Table:**
  - `message_id` (String, Primary Key)
  - `sender_id` (String, Foreign Key)
  - `group_id` (String, Nullable)
  - `content` (Text)
  - `timestamp` (Timestamp)
  - `attachment_url` (String, Nullable)

- **Groups Table:**
  - `group_id` (String, Primary Key)
  - `group_name` (String)
  - `members` (Array of user_ids)

This schema allows for quick lookups of messages by user, group, and efficient searching through timestamps.

- **WebSocket vs. HTTP Polling:**
  - Chose WebSockets for real-time messaging and low latency, but it increases complexity in connection management and state handling.

- **Eventual Consistency vs. Strong Consistency:**
  - Opted for eventual consistency for better performance and scalability, especially under high loads, but this may lead to temporary discrepancies in message delivery.

- **Database Choice (Firestore vs. SQL):**
  - Chose Firestore for real-time capabilities and horizontal scaling but at the cost of complex querying compared to a traditional SQL database.

- **Search with Elasticsearch:**
  - Integrated Elasticsearch for fast message search, which introduces additional operational overhead but significantly improves user experience in searching message history.

Design a Chat System like Google Chat

Google

How to Approach This

Sharpen Your Skills on Codemia

Sample Answer

Requirements

Capacity Estimation

Submit Your Answer

Google Software Engineer Interview Guide

Related Questions

Design a high-throughput Inventory Management System

Design a low-latency Rate Limiting System

Design a fault-tolerant Payment System

Design a fault-tolerant Messaging System

Design Walmart Product Search