Fawkes: Faster Mobile Page Loads via App-Inspired Static Templating

Authors: 

Shaghayegh Mardani, UCLA; Mayank Singh, IIT Delhi; Ravi Netravali, UCLA

Abstract: 

Despite the rapid increase in mobile web traffic, page loads still fall short of user performance expectations. State-of-the-art web accelerators optimize computation or network fetches that occur after a page’s HTML has been fetched. However, clients still suffer multiple round trips and server processing delays to fetch that HTML; during that time, a browser cannot display any visual content, frustrating users. This problem persists in warm cache settings since HTML is most often marked as uncacheable because it usually embeds a mixture of static and dynamic content.

Inspired by mobile apps, where static content (e.g., layout templates) is cached and immediately rendered while dynamic content (e.g., news headlines) is fetched, we built Fawkes. Fawkes leverages our measurement study finding that 75% of HTML content remains unchanged across page loads spread 1 week apart. With Fawkes, web servers extract static, cacheable HTML templates for their pages offline, and online they generate dynamic patches which express the up- dates required to transform those templates into the latest page versions. Fawkes works on unmodified browsers, using a JavaScript library inside each template to asynchronously apply updates while ensuring that JavaScript code only sees the state that it would have in a default page load despite downstream content having already been loaded. Across a wide range of pages, phones, and live wireless networks, Fawkes improves interactivity metrics such as Speed Index and Time-to-first-paint by 46% and 64% at the median in warm cache settings; results are 24% and 62% in cold cache settings. Further, Fawkes outperforms recent server push and proxy systems on these metrics by 10%-24% and 69%-73%.

NSDI '20 Open Access Sponsored by NetApp

Open Access Media

USENIX is committed to Open Access to the research presented at our events. Papers and proceedings are freely available to everyone once the event begins. Any video, audio, and/or slides that are posted after the event are also free and open to everyone. Support USENIX and our commitment to Open Access.

BibTeX
@inproceedings {246330,
author = {Shaghayegh Mardani and Mayank Singh and Ravi Netravali},
title = {Fawkes: Faster Mobile Page Loads via App-Inspired Static Templating },
booktitle = {17th {USENIX} Symposium on Networked Systems Design and Implementation ({NSDI} 20)},
year = {2020},
isbn = {978-1-939133-13-7},
address = {Santa Clara, CA},
pages = {879--894},
url = {https://www.usenix.org/conference/nsdi20/presentation/mardani},
publisher = {{USENIX} Association},
month = feb,
}

Presentation Video