"Innovation at Scale"
summary
One platform handles hundreds of millions of users, hundreds of thousands of hosts and petabytes of data so a generic toolset fails. Spam on groups sends 20 million emails a day to Taiwan so a custom rule framework is built then parallelized with Hadoop + Pig, cutting full scans from days to hours and letting rule iterations outpace spammers. For global performance a network of edge Y! CPI proxy pods halves coast‑to‑coast latency (about 300ms to 150ms) and trims page downloads accordingly.
AI-written summary · about
chapters (3)
related talks
01h 00mBuilding Safe Systems at Scale - Lessons from Six Months at Yahoo
27mLessons Learned From Architecting the Highest-scale Operational Systems in the World
28mMicroliths: how to avoid traps in your stack and your culture
51mScaling Sustainable Innovation through Team Autonomy
30mProduction › Prototype: How a Small Team Replaced a Core Platform in 90 Days
21mRun Less Software
50mScaling for Global Growth with Modern Cloud
22mScaling Culture of Resiliency in the Enterprise at Charter Communications