Mastering Cloud Scalability and System Designs Deep dive
summary
Shows concrete capacity math and architecture: estimate 46 million requests/day and size servers to handle 5x–10x spikes (example: ~90 servers), aim for 250 ms response and design for P99 tail latency, use queues as buffer and Redis timelines for hot reads, persist archives to SQL then shard to Cassandra + Kafka + Spark for heavy writes and recommendations, model Uber as event-driven with geo sharding and priority queues, and prove the stack meets Black Friday scale (millions of visitors, hundreds of terabytes) with these tradeoffs.
AI-written summary · about
chapters (14)
- 0:00System design fundamentals
- 3:36Requirement analysis
- 6:45Performance and caching
- 9:32Capacity estimation
- 13:07Database and storage choices
- 17:37Little's Law and queuing
- 21:05Selecting modern databases
- 24:28Case study: Uber
- 29:05Architecture and security
- 34:01Scalability techniques
- 37:03Optimization strategies
- 42:49E-commerce and marketplace
- 47:17Recommendation and state
- 52:58Conclusion and review
related talks
59mMastering the System Design Methodology
32mScaling Done Right
31mMastering the Balancing Act: Scaling Infrastructure for Growth and Stability
46mDigital scale-up architectural patterns
34mDistributed Systems - What Can Go Wrong and How to Get It Right
43mArchitecting for Speed & Scale
29mWhen Web Scale Doesn't Make Sense
41mFrom Zero to 10 million: Architecting for scale on AWS