I asked AI to write tech commics.
| DDIA, as a live comic | complete, growing | Replication, partitioning, consensus — twelve short comics, each built around one misconception it exists to kill. Then met again inside Kafka, Postgres, Redis, RabbitMQ, the web tier and S3, in nine-chapter deep-dives with animated traces and a hardware envelope you can overload. Plus two whole applications you push until they break. |
| The Papers That Broke the Database | being written | The history of distributed systems through the papers that made it. One chapter live of eighteen; the season's map is on the page. |
| The calculator | shared instrument | Put in a workload, get machine counts back — and check every number against the division that produced it. |
npm install
npm run dev # http://localhost:5173
npm run build # typecheck + build + per-route HTML
npm test # the suite
npm run check:diagrams # geometry lint (needs the dev server + Chrome)
npm run og # re-render social cards after editing scripts/routes.mjsBook one is an unofficial companion to Designing Data-Intensive Applications, 1st edition, by Martin Kleppmann — chapter numbers follow that edition, since the 2nd renumbers. Unofficial and not affiliated with the author or O'Reilly. Read the book; this is a lens on it, not a replacement. Primary sources are linked at the point of use throughout.
Every limit here is an order-of-magnitude rule of thumb, useful for intuition. Verify against your own load tests before committing production capacity.
Code is MIT (see LICENSE).
The written and illustrated content — comics, prose, diagrams and runbooks — is © Samuel Xing, all rights reserved. Quote it with attribution; please ask before republishing wholesale.
If this was useful, a ⭐ is the whole compensation plan.