📝 Problem Description
Design a music streaming service like Spotify. Users can search and play songs, create playlists, follow artists, and get personalized recommendations. Handle audio streaming, content delivery, and offline mode.
👤 Use Cases
1.
User wants to searches for a song so that sees search results
2.
User wants to plays a song so that audio streams with low latency
3.
User wants to creates playlist so that playlist saved and shareable
4.
User wants to downloads for offline so that can play without internet
✅ Functional Requirements
- •Search songs, albums, artists, playlists
- •Stream audio with quality selection
- •Create and share playlists
- •Follow artists and users
- •Personalized recommendations
- •Offline playback
- •Cross-device sync
⚡ Non-Functional Requirements
- •Playback start < 200ms
- •No buffering during playback
- •Support 100M concurrent users
- •99.99% availability
⚠️ Constraints & Assumptions
- •Audio licensing requirements
- •Multiple audio qualities (128, 256, 320 kbps)
- •Limited offline storage on devices
📊 Capacity Estimation
👥 Users
500M users, 100M concurrent streams
💾 Storage
100PB (50M songs × 3 qualities)
⚡ QPS
Play: 100K/sec, Search: 50K/sec
🌐 Bandwidth
10Pbps peak
📐 Assumptions
- • 50M songs
- • 3 quality versions per song
- • Average song: 5MB (320kbps)
- • 100M concurrent streams
💡 Key Concepts
CRITICAL
Adaptive Bitrate Streaming
Adjust quality based on network conditions.
CRITICAL
CDN for Audio
Edge caching reduces latency and origin load.
HIGH
Pre-buffering
Buffer next song in playlist for gapless playback.
HIGH
Collaborative Filtering
Recommendations based on similar users.
💡 Interview Tips
- 💡Start with the audio streaming architecture
- 💡Discuss the recommendation system in detail
- 💡Emphasize CDN usage for audio delivery
- 💡Be prepared to discuss royalty tracking and licensing
- 💡Know the difference between streaming and downloading
- 💡Understand the catalog management and metadata