All three were right
- Netflix, Instagram and Twitter made different database choices and all were correct (
10:03) - They were not picking the best database — they were picking the best database for their specific access pattern (
10:03)
The real system design skill
- It is not memorizing which database is fastest or most scalable (
10:03) - It is understanding your data, your queries, your trade-offs, then matching them to the right tool (
10:19,10:24)
Summary table
| Company | Database | Driving access pattern | Accepted trade-off |
|---|---|---|---|
| Netflix | Cassandra | 3M+ writes/sec, key-based reads | No joins, aggressive denormalization |
| PostgreSQL | Read-heavy relational queries | Requires heavy scaling engineering (pooling, replicas, partitioning, indexing) | |
| Redis (cache) + durable store | Instant pre-computed timelines | Redis is not durable; hybrid fan-out needed for celebrities |
Outro
- The author points to a system design master course covering these architectural patterns in more depth with hands-on examples (
10:24)