Velocity 2014
summary
Describes how a real production site scales its infrastructure to handle a billion users by redesigning request routing and load shedding, including the load balancing architecture that routes traffic across datacenters, the CDN and edge-cache tradeoffs, and a deploy pipeline that forces CSS delivery to the user in under 50 ms. Gives concrete numbers on latency and throughput, explains tools used (CDN, HAProxy, anycast, metrics and alerting stack), and what failed and what actually worked.
AI-written summary · about
related talks
27mVelocity 2010: Urs Holzle
18mScaling APIs to Billions: Extreme-Scale Reliability
06mVelocity 09: Alistair Croll
44mVelocity 09: Chris Bissell, Jeremy Custenborder, and Yadid
20mSoftware Performance Optimization and Scaling
25mSpeed at Scale: Optimizing The Largest CX Platform Out There
07mMike Loukides and Steve Souders discuss Velocity 2011
28mThe Path to Scaling a Digital Enterprise