4 additional system design katas

1. Social Media Feed (e.g., Twitter)

Problem Statement:
Design a system to display a real-time feed of posts from users a person follows. Prioritize freshness, scalability, and low latency.

Key Requirements:

• Functional: Post creation, follow/unfollow, feed generation, trending topics.
• Non-Functional: Handle 10M+ users, <1s latency for feed refresh, 99.99% availability.

What’s Being Practiced?

• Fan-out Strategies: Push vs. pull models (e.g., write-time fan-out to followers’ caches).
• Caching: Redis for hot feeds, CDNs for static content.
• Sharding: Distribute users by user_id to balance load.
• Real-Time Analytics: Track trending hashtags with Apache Kafka + Flink.

Solution Sketch:

• Hybrid Fan-out: Push posts to followers’ caches for active users, pull for inactive.
• Storage: Cassandra for posts (time-series data), Redis for feed caching.
• APIs:
  • POST /tweet → Kafka → Feed Service → Cache.
  • GET /feed → Fetch from Redis.

---

2. Ride-Sharing App (e.g., Uber)

Problem Statement:
Design a system to match riders with nearby drivers in real time. Include surge pricing and trip tracking.

Key Requirements:

• Functional: Driver/rider location updates, ride matching, fare calculation, payment.
• Non-Functional: <500ms match latency, handle 100K+ concurrent rides.

What’s Being Practiced?

• Geospatial Data: Use Redis GEO or QuadTrees to find nearby drivers.
• Real-Time Messaging: WebSocket/GRPC for location updates.
• Dynamic Pricing: Surge pricing algorithm based on demand/supply.
• Idempotency: Ensure payment transactions aren’t duplicated.

Solution Sketch:

• Location Service: Store driver locations in Redis GEO.
• Matching Engine: Use grid-based spatial partitioning.
• Payment Service: Idempotent API with Stripe/PayPal integration.

---

3. Distributed File Storage (e.g., Dropbox)

Problem Statement:
Design a cloud file storage system with real-time sync across devices, versioning, and conflict resolution.

Key Requirements:

• Functional: File upload/download, sync, version history, sharing.
• Non-Functional: Handle 1PB+ storage, <200ms sync latency.

What’s Being Practiced?

• Delta Sync: Only transfer file chunks that changed (like rsync).
• Conflict Resolution: Use vector clocks to order edits.
• Distributed Storage: Shard files across S3/GlusterFS.
• Metadata Management: Store file metadata in PostgreSQL.

Solution Sketch:

• Chunking: Split files into blocks, hash for deduplication.
• Versioning: S3 for file history, DynamoDB for metadata.
• APIs: WebSocket for real-time sync notifications.

---

4. E-commerce Inventory System (e.g., Amazon)

Problem Statement:
Design an inventory management system that prevents overselling during flash sales.

Key Requirements:

• Functional: Product catalog, inventory tracking, order processing.
• Non-Functional: Atomic inventory decrements, 10K+ orders/sec.

What’s Being Practiced?

• Distributed Locks: Use Redis or ZooKeeper to reserve stock.
• Idempotent APIs: Prevent duplicate orders.
• Event Sourcing: Kafka to log inventory changes.
• Caching: Cache product details with write-through to DB.

Solution Sketch:

• Inventory Service: Use Redis for atomic DECR operations.
• Order Pipeline: Kafka → Order Service → Payment Service → Inventory.
• DB Sharding: Split products by product_id.

---

5. Video Streaming Platform (e.g., Netflix)

Problem Statement:
Design a video streaming service supporting adaptive bitrate streaming for 50M+ users.

Key Requirements:

• Functional: Video upload, transcoding, CDN delivery.
• Non-Functional: <2s startup latency, 99.9% uptime.

What’s Being Practiced?

• Transcoding Pipeline: FFmpeg + Kubernetes for parallel processing.
• CDN Optimization: Edge caching with Cloudflare/Akamai.
• Adaptive Streaming: HLS/DASH to switch bitrates dynamically.
• Load Balancing: Round-robin vs. least connections for traffic.

Solution Sketch:

• Upload Service: Store raw videos in S3.
• Transcoding Service: Convert to multiple resolutions.
• Delivery: CDN for cached videos, API Gateway for metadata.

---

Summary: Core Skills Practiced

Problem	Key Skills
Social Media Feed	Fan-out strategies, caching, real-time analytics.
Ride-Sharing	Geospatial indexing, surge pricing, idempotency.
File Storage	Delta sync, conflict resolution, chunking.
E-commerce Inventory	Distributed locks, event sourcing, atomic ops.
Video Streaming	Transcoding, CDNs, adaptive bitrate streaming.

---

Why Practice These Katas?

1. Broaden Pattern Recognition: Each problem teaches unique strategies (e.g., fan-out for feeds, geospatial indexing for ride-sharing).
2. Interview Readiness: These mirror real FAANG system design questions.
3. Tradeoff Analysis: Learn when to prioritize CAP theorem aspects (e.g., availability for feeds, consistency for inventory).

Next Step: Pick one kata, sketch a solution, and compare it with a reference architecture! 🚀