[Fundamental] System Design

Question 1

Main strategy

Answer 1

Requirement clarification

Main API

1. Single server
2. Add DB
3. Add multiple servers, add Load Balancer
4. Additional strategies: based on use case find bottleneck and appropriate solution

Question 2

Database Performance Improvement

Answer 2

• Replication
  
  • Writer instance
  • Reader instance
• Partiotioning: Divide columns
• Sharding: Divide Rows
• Indexing
• Materialized View: Pre-compute long-running query, ideally low writes
• Denormalization

Question 3

Caching

Answer 3

Redis (WIP)

Question 4

CDN

Question 5

Async Pattern

Answer 5

Message broker for long running tasks

Question 6

Real time use cases

Answer 6

• WebSockets: Ideal for low-latency, bidirectional communication but requires managing persistent connections.
• SSE (Server-Sent Events): Good for unidirectional, real-time updates but has scalability limits.
• HLS/DASH: Best for live video streaming with adaptive bitrate but has higher latency.
• MQTT: Efficient for lightweight messaging in IoT but needs a broker.
• Kafka Streams: Powerful for real-time data processing but more complex.
• UDP: Low latency for gaming but requires handling packet loss.