Architecture & Teams at Less-than-Google Scale
summary
Compares concrete tradeoffs at non-Google scale: Uber builds an append-only schemaless key-value store that gains linear scaling but loses joins, read-after-write guarantees, and hireable familiarity; Amazon enforces service interfaces over the network and discovers pager cascades, universal rate‑limiting and a new service for service discovery; Twitter’s hash-bang single-page apps delay render and strand low-end devices; recommends humane ops like ~once‑a‑month on‑call, and cites 10,000 searches per second as a growth milestone.
AI-written summary · about
chapters (11)
- 0:00Start
- 2:39CASE STUDIES
- 4:49AMAZON AND SERVICE ARCHITECTURE
- 16:05PROJECT ECOSYSTEM
- 16:55DOCUMENTATION
- 17:27LICENSING LANDMINES
- 21:06APPROPRIATE AVAILABILITY AND SCALABILITY
- 23:16PSYCHOLOGICAL SAFETY
- 23:45SET GROUND RULES
- 28:26MEASURE SUPPORT PERFORMANCE
- 28:44DISASTER RECOVERY
related talks
53mWhat I Learned About Going Fast at eBay & Google
55mFollowing Google or Don't Follow the Followers, Follow the Leader
51mHow Your Organization is Killing your Software
49mSoftware Architecture, Team Topologies & Complexity Science
48mMoving Fast at Scale
52mThe Future of Software Engineering
58mScaling like a Silicon Valley startup
53mKeynote - Software Architecture, Team Topologies and Complexity Science