High Performance in the Critical Rendering Path
Submitted by Nicolas Bevacqua (@bevacqua) on Thursday, 30 July 2015
Section: Full talk Technical level: Advanced
This talk begins by looking at the kind of performance optimizations we’re are already doing, such as bundling and minification, gzip compression, caching, and using progressive enhancement. Then I’ll cover things we should be doing such as measuring performance, setting up a performance budget, using nginx as a reverse proxy, inlining critical CSS while deferring the rest, deferring font loading, using a CDN, experimenting with WebP images, and maybe deferring image loading too.
Last, we’ll take a look at where HTTP is going with HTTP 2.0, and discuss the new multiplexing features that enable HTTP 2.0 to take advantage of a single TCP request, erasing the need for “hacks” such as bundling, spriting, CSS or image inlining, taking HTML back to the way it was originally designed to work, while actually improving performance.