Show HN: An atlas of system designs with interactive architecture diagrams
2 days ago
- Precompute or compute on demand: trading instant reads for stale data versus fresh results at request cost
- System design trade-offs involve consistency, capacity, partition tolerance, and the skill of defending product requirements
- Key concepts include CAP theorem, PACELC, replication lag, sharding, caching patterns, and quorums
- Common patterns like queues, pub/sub, CQRS, push vs pull, and event sourcing are foundational
- Notable system designs include feed serving, chat, URL shortener, rate limiting, and video streaming
- Distributed systems problems often require balancing availability, consistency, and performance under constraints