Large scale websites performance optimisation tricks
summary
Revamps a US airline website to survive Black Friday and Cyber Monday traffic by making PHP services return 204 No Content so heavy searches run asynchronously and results pre-cache into varnish + memcache, single-page apps poll memcache, and parallel product lookups use curl_multi. Deploys use siloed clusters behind an F5 with Ansible automation for zero-downtime upgrades, a one-year delivery target across teams in 10 time zones, built on a Symfony2 foundation.
AI-written summary · about
chapters (9)
- 0:00LARGE-SCALE WEBSITES PERFORMANCE OPTIMISATION TRICKS
- 10:12TECHNOLOGY STACK
- 21:50STATIC DATA
- 24:49ANATOMY OF A SILO
- 28:50component
- 31:06THE 204 TRICK
- 39:50ERROR HANDLING
- 42:09PARALLELISM
- 45:21LESSONS LEARNED
related talks
45mSurviving a Prime Time TV Commercial
47mFrom Legacy to High-Load: Evolution of a Web Application
29mModern Websites for E-commerce in the Real World
41mHigh Performance Kick-Ass Web Apps
37mScaling Rails for Black Friday and Cyber Monday
35mPHP: Performance and application development — what we diagnosed
45mSpeeding up Code, Webpages and Teams
50mBuilding and Operating Online Shops under Extreme Load
