Design a real-time Content Delivery System

Question

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

Codemia · Accepted Answer

### Functional Requirements
1. **Real-Time Content Delivery**: The system must deliver content (images, videos, documents) in real-time to users across the globe.
2. **User Authentication**: Users must authenticate to access personalized content.
3. **Content Management**: Admins must be able to upload, update, and delete content seamlessly.
4. **Analytics Tracking**: The system should track user interactions with content for analytics purposes.

### Non-Functional Requirements
1. **Scalability**: The system must handle millions of concurrent requests with low latency.
2. **High Availability**: The system should be operational 99.99% of the time with minimal downtime.
3. **Performance**: Average response time for content delivery should be less than 200ms.
4. **Data Consistency**: Content updates should be consistent across all nodes with eventual consistency being acceptable for non-critical data.

Assuming Rippling has a user base of 10 million users with an average of 5 content requests per user per day:

- **Total Daily Requests**: 10 million users * 5 requests = 50 million requests/day.
- **Peak QPS (Queries Per Second)**: Assuming peak usage is 10% of daily requests occurring in a 2-hour window:
  - 50 million requests/day / (24 hours * 3600 seconds) = ~578 requests/second (average).
  - During peak, this could spike to ~5,000 requests/second.

### Storage Estimation
- Assuming average content size is 1MB, we would require:
  - 50 million MB = ~47.7 TB of storage for the content. Considering a 30% growth over the next year, we estimate needing around 62 TB of storage.

### Component Diagram
- **Load Balancer**: Distributes incoming requests across multiple content delivery servers.
- **Content Delivery Network (CDN)**: Caches and serves static content globally.
- **Application Servers**: Handles user authentication, content management, and analytics processing.
- **Database**: 
  - **Primary Database**: PostgreSQL for relational data (user profiles, content metadata).
  - **NoSQL Database**: MongoDB for user-generated content and analytics data.
- **Message Queue**: Apache Kafka for processing content uploads and analytics events asynchronously.
- **Cache**: Redis for caching frequently accessed content to reduce load on databases.

### Schema Design
- **User Table**: `User(id, username, password_hash, email, created_at)`
- **Content Table**: `Content(id, user_id, title, description, file_url, created_at, updated_at)`
- **Analytics Table**: `Analytics(id, user_id, content_id, action_type, timestamp)`

### Access Patterns
- Frequent access to user profiles and content details will require efficient indexing. We will index `user_id` and `content_id` in the `Analytics` table to optimize query performance.

1. **Consistency vs. Availability**: By using eventual consistency for analytics data, we ensure high availability but may introduce slight delays in reporting.
2. **Database Choice**: Choosing PostgreSQL for relational data offers strong consistency but might not scale horizontally as easily as NoSQL options. The trade-off here is between ACID compliance and scalability.
3. **Caching Strategy**: Using Redis for caching improves performance but adds complexity in ensuring cache invalidation is correctly handled during content updates.
4. **Monitoring and Logging**: Implementing detailed logging and monitoring using Prometheus and Grafana introduces overhead but is essential for maintaining high availability and performance.

Design a real-time Content Delivery System

Rippling

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

Rippling 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