Design a Content Delivery for Twitter/X

Question

Design a fault-tolerant content delivery system that handles millions of requests. Discuss trade-offs in consistency, availability, and performance.

Codemia · Accepted Answer

### Functional Requirements:
1. **Content Retrieval**: Users should be able to retrieve tweets, retweets, and media content quickly.
2. **Rate Limiting**: Implement rate limits to prevent abuse and ensure fair usage across users.
3. **User Personalization**: Tailor content delivery based on user preferences and past interactions.
4. **Content Caching**: Frequent tweets should be cached to reduce load and serve requests faster.

### Non-Functional Requirements:
1. **Scalability**: The system must handle millions of requests per second as Twitter/X grows.
2. **Fault Tolerance**: The system must remain operational despite server or network failures.
3. **Low Latency**: Ensure that content is delivered with minimal delay to enhance user experience.
4. **Consistency**: While eventual consistency is acceptable, ensure critical data remains accurate immediately after updates.

Assuming Twitter/X has approximately 400 million daily active users, with an average of 5 requests per user per session:

- **Total daily requests**: 400M users * 5 requests = 2 billion requests per day.
- **Requests per second**: 2 billion / 86400 seconds = ~23,148 requests per second.
- **Peak load estimation**: Estimating a 10x peak during specific events or trending topics leads to around 230,000 requests per second.
- **Infrastructure needs**: If each server can handle 1,000 requests per second, we would need around 230 servers at peak load, considering redundancy and failover, leading to 300 servers in total.

### Component Diagram:
- **Load Balancer**: Distributes incoming requests to backend servers.
- **Content Delivery Network (CDN)**: Caches static content like images and videos closer to users.
- **Application Servers**: Handle dynamic requests and serve content from the database or cache.
- **Database**: A NoSQL database like **Cassandra** for storing tweets, user profiles, and interactions, ensuring horizontal scalability.
- **Cache Layer**: **Redis** for caching frequently accessed tweets and user timelines.
- **Queue System**: **Kafka** for managing asynchronous processing of content updates and notifications.

### Technology Choices:
- CDN: **Cloudflare** or **Akamai** for global content delivery.
- Load Balancer: **NGINX** for efficient request handling.

### Schema Design:
1. **Users Table**:
   - user_id (PK)
   - username
   - email
   - preferences (JSON)

2. **Tweets Table**:
   - tweet_id (PK)
   - user_id (FK)
   - content
   - created_at
   - engagement_metrics (JSON)

3. **Timeline Table** (for caching user timelines):
   - user_id (PK)
   - tweet_ids (List of tweet_id)

### Access Patterns:
- Frequently accessed tweets are cached in Redis to minimize DB hits.
- User timelines are generated and cached upon requests to optimize read times.

1. **Consistency vs. Availability**: Opting for eventual consistency allows for better availability during peak loads, but we must ensure that critical operations (like posting tweets) are consistent immediately.
2. **Read vs. Write Optimization**: The architecture favors read operations due to high traffic from users viewing tweets; however, we need to ensure write operations (tweet submissions) are efficiently handled without blocking reads.
3. **Caching Strategy**: Caching strategies must balance between cache invalidation complexity and read performance; we'll use TTL (Time-To-Live) for cached tweets to ensure freshness without overloading the database.
4. **Message Queuing**: Deciding to use Kafka introduces complexity but enables decoupled components for processing tweets and user notifications, enhancing system reliability.

Design a Content Delivery for Twitter/X

Twitter/X

What the Interviewer Expects

Key Topics to Cover

How to Approach This

Possible Follow-up Questions

Practice a Similar Problem on Codemia

Sample Answer

Requirements

Functional Requirements:

Capacity Estimation

Submit Your Answer

Twitter/X 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