Build a distributed Video Streaming Pipeline

Question

Design a distributed video streaming system that handles millions of requests. Discuss trade-offs in consistency, availability, and performance.

Codemia · Accepted Answer

### Functional Requirements
1. **Video Upload**: Users can upload videos seamlessly.
2. **Video Processing**: The system should transcode videos into multiple formats and resolutions for optimal streaming.
3. **Video Playback**: Users can stream videos with adaptive bitrate streaming to ensure playback quality.
4. **User Authentication**: Implement secure user authentication for access control.
5. **Analytics**: Track user interactions, video views, and streaming quality metrics.

### Non-Functional Requirements
1. **Scalability**: The system should handle millions of concurrent users with minimal latency.
2. **Availability**: Aim for 99.99% uptime globally.
3. **Latency**: Video start time should be under 2 seconds in most regions.
4. **Data Consistency**: Ensure eventual consistency for video metadata while maintaining high availability.
5. **Security**: Implement encryption for stored videos and secure streaming protocols (HTTPS, RTMPS).

Assuming Grubhub has about 10 million monthly active users and that each user streams an average of 10 videos per month:
- Total Video Streams per Month = 10M * 10 = 100M streams.
- Assuming peak traffic is 3 times the average, we can estimate around 10,000 streams per second at peak times.
- Each video stream requires an average bitrate of 5 Mbps, which translates to:
  - Total Bandwidth = 10,000 streams * 5 Mbps = 50,000 Mbps or ~6.25 TB/hour.
- For storage, assuming an average video length of 10 minutes and 100 MB per video:
  - Total Storage Requirement = 100M videos * 100 MB = 10 PB.

### Component Diagram
1. **Video Ingestion Service**: Handles video uploads using AWS S3 for storage.
2. **Transcoding Service**: Uses AWS Elemental MediaConvert for transcoding videos into various formats.
3. **Content Delivery Network (CDN)**: Leverage CloudFront for caching and distributing video content globally.
4. **Streaming Server**: Use NGINX with RTMP module for live streaming capabilities.
5. **Database**: Utilize Amazon DynamoDB for storing video metadata with high availability and low latency.
6. **Analytics Service**: Use AWS Kinesis for real-time analytics on video streams, tracking user engagement and performance metrics.

### Schema Design
1. **Videos Table**:
   - `video_id` (String, Primary Key)
   - `user_id` (String)
   - `title` (String)
   - `description` (String)
   - `upload_time` (Timestamp)
   - `duration` (Integer)
   - `thumbnail_url` (String)
   - `status` (String)
2. **User Table**:
   - `user_id` (String, Primary Key)
   - `email` (String, Unique)
   - `password_hash` (String)
   - `created_at` (Timestamp)
3. **Analytics Table**:
   - `video_id` (String)
   - `user_id` (String)
   - `view_time` (Timestamp)
   - `quality` (String)
   - `buffering_events` (Integer)
   - Composite Key: `video_id`, `view_time`.

### Architectural Decisions
1. **Consistency vs. Availability**: Opted for eventual consistency for video metadata to ensure high availability and low latency in video streaming.
2. **CDN Usage**: Using CloudFront minimizes latency for global users, but content freshness may lag due to caching. Implement cache invalidation strategies to balance.
3. **Transcoding Delay**: Real-time transcoding can introduce latency; opted for a pre-processing model where videos are transcoded asynchronously after upload.
4. **Fault Tolerance**: Implement retries and exponential backoff in the ingestion service to handle transient failures during video uploads.
5. **Monitoring**: Integrated AWS CloudWatch for monitoring system metrics and set up alarms for key performance indicators (KPIs) to ensure operational excellence.

Build a distributed Video Streaming Pipeline

Grubhub

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

Grubhub 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