To push, or not to push? The future of HTTP/2 server push
HTTP/2 server push gives us the ability to proactively send assets to a browser without waiting for them to be requested. Sounds great, but is this new mechanism really a silver bullet? Using new research and real-world examples, Patrick Hamann leads a deep dive into server push and attempts to answer the question we're all asking: Is it ready for production?
Talk Title | To push, or not to push? The future of HTTP/2 server push |
Speakers | Patrick Hamann (Fastly) |
Conference | O’Reilly Fluent Conference |
Conf Tag | The Web Platform in Practice |
Location | San Jose, California |
Date | June 12-14, 2018 |
URL | Talk Page |
Slides | Talk Slides |
Video | |
The emergence and embrace of HTTP/2 in recent years has given us a new set of tools to optimize the delivery of our web pages. One notable addition was server push—the ability to proactively send assets to a browser without waiting for them to be requested. Server push promises to put an end to the performance optimization hacks we’ve developed to overcome HTTP/1’s inefficiencies, such as inlining and resource bundling. However, is this new mechanism really the silver bullet we all thought it was? Is it time to abandon our build systems and stop bundling our assets entirely? Or are lack of server support and browser inconsistencies holding us back? Lastly, what are new specifications such as cache digests, origin frame, and the 103 status code doing to improve the situation? Using new research and real-world examples, Patrick Hamann leads a deep dive into server push, exploring current and future best practices for loading assets in the browser. This will ultimately allow us to make better decisions when loading our web pages and lead to faster, more resilient experiences for our users. Topics include: