Notes

Short reference notes — clarifications, deep dives, and questions worth keeping.

• System Design Capacity Cheat Sheet (Interview Ballparks)

  Practical ballpark throughput and latency numbers for Postgres, Cassandra, Redis, Kafka, app servers, CDN, S3, and more - plus how to use them in interview estimations.

  system-design interview scalability performance capacity-planning

• Payment Systems — 10k TPS, Database Crashes, and Duplicate Requests

  Three senior engineer interview questions in one: scaling to 10,000 transactions/second, handling database failures mid-transaction with Saga and WAL, and preventing double charges with idempotency keys.

  system-design payments saga redis interview durability

• Flash Sale — Solving the Inventory Race Condition

  50,000 requests, 1 item left. Why naive solutions fail and how Redis atomic ops, request queues, and optimistic locking solve correctness and scale separately.

  system-design redis kafka interview concurrency

• Protecting the Database from Traffic Spikes

  Why no database handles 10M writes/second directly — and the three-layer solution using Kafka, Redis, and capacity planning.

  system-design kafka redis cassandra scaling

• Why Kafka Sits Between Cassandra and the Fan-out Service

  What breaks in the feed system if you remove Kafka — four concrete failure modes, and the mental model for when async queues are necessary.

  system-design kafka feed-system message-queues

• Scaling the Feed System Beyond 100k Writes/Second

  What changes in the feed system architecture when write volume jumps 80x — stress-testing each component in the write path.

  system-design feed-system scaling redis

• Why Cassandra Over MongoDB, and How Databases Scale

  When to pick Cassandra vs MongoDB, plus sharding and replication explained — the concepts used implicitly in every system design.

  system-design databases cassandra mongodb