To HTTP/2, or Not To HTTP/2, That Is The Question
Abstract
As of February, 2015, HTTP/2, the update to the 16-year-old HTTP 1.1, is officially complete. HTTP/2 aims to improve the Web experience by solving well-known problems (e.g., head of line blocking and redundant headers), while introducing new features (e.g., server push and content priority). On paper HTTP/2 represents the future of the Web. Yet, it is unclear whether the Web itself will, and should, hop on board. To shed some light on these questions, we built a measurement platform that monitors HTTP/2 adoption and performance across the Alexa top 1 million websites on a daily basis. Our system is live and up-to-date results can be viewed at http://isthewebhttp2yet.com/. In this paper, we report our initial findings from a 6 month measurement campaign (November 2014 - May 2015). We find 13,000 websites reporting HTTP/2 support, but only 600, mostly hosted by Google and Twitter, actually serving content. In terms of speed, we find no significant benefits from HTTP/2 under stable network conditions. More benefits appear in a 3G network where current Web development practices make HTTP/2 more resilient to losses and delay variation than previously believed.
Keywords
Cite
@article{arxiv.1507.06562,
title = {To HTTP/2, or Not To HTTP/2, That Is The Question},
author = {Matteo Varvello and Kyle Schomp and David Naylor and Jeremy Blackburn and Alessandro Finamore and Kostantina Papagiannaki},
journal= {arXiv preprint arXiv:1507.06562},
year = {2015}
}