🎉 Celebrating 25 Years of GameDev.net! 🎉

Not many can claim 25 years on the Internet! Join us in celebrating this milestone. Learn more about our history, and thank you for being a part of our community!

of interest: faster browsers speeds

Started by
19 comments, last by frob 8 years, 4 months ago

of interest:

"scientists at MIT and Harvard have come up with a pair of technologies they claim can speed browser page loads on everyday web browsers by more than 30%"

the paper will be presented march 16th 2016 at the USENIX Symposium on Networked Systems Design and Implementation.

http://mashable.com/2016/03/09/mit-faster-web-pages/#_ni0QzccQSql

Norm Barrows

Rockland Software Productions

"Building PC games since 1989"

rocklandsoftware.net

PLAY CAVEMAN NOW!

http://rocklandsoftware.net/beta.php

Advertisement

Is it persuade web developers to make the pages 30% smaller?

If pages load 30% faster, ad providers will make their ads 43% more resource intensive.

In bandwidth, request-count, or local processing speed?

Well, disable Javascript... These MIT scientists complain that ESPN.com takes forever to load. Loads in under a second here. Besides, the "issue" and the solution that they've come up with (request dependencies) is just what Google designed spdy for, years ago. It's not for no reason that the server can push under spdy. Working implementation, not research paper.

Well, I read few pages then they lost me after mentioning async and defer tags along with "Javascript blocks rendering".

Apparently it is said to be faster in page load time so perceived speed imo therefore nothing under the hood like HTTP/2 . Main idea is apparently manifesting assets html parser can't be aware before parsing Javascript, therefore wasting time that could be used to download required assets.

I don't deny it may be helpful depending on your use case but in times people are aggresively looking for ATF (Above the Fold) to fit in single request using inline CSS and Javascript, it is at least not useful for landing pages.

mostates by moson?e | Embrace your burden

The graph (which is used in the link) reveals a large chunk of the problem. The real answer is direct enough:

1) Turn off ads.

2) Stop spreading render-required resources into hundreds of tiny requests.

The End.

Ads are well-established as slowing down the Internet. The real-time bidding of ads is slow. When loading the page it also requires transferring a huge ad infrastructure, communications to a bunch of unknown servers scattered across the globe, and that takes a lot of time and processing. That is especially true when the ads are mini-programs or javascript+canvas programs that themselves reach out and grab bunches of additional resources before they can display.

For the second, any time a page is designed to load 20, or 30, or 50+ requests before displaying then the page is going to be slow. It doesn't matter if those are a bunch of javascript files, or css files, or if the site is using one of the stupid "hypermedia client" systems where the page itself is blank and everything gets streamed in through AJAX and REST. A single large HTML file or a small number (e.g. under five) of source files can be loaded quickly, much quicker than hundreds of tiny little web requests.

The paper's description is a tool to help minimize and reorder dependencies. It is still more efficient and more memory friendly to remove the dependencies completely.

America is finally going to get Internet at speeds on par with the rest of the world?


L. Spiro

I restore Nintendo 64 video-game OST’s into HD! https://www.youtube.com/channel/UCCtX_wedtZ5BoyQBXEhnVZw/playlists?view=1&sort=lad&flow=grid

America is finally going to get Internet at speeds on par with the rest of the world?


*America may not be available in your area, please contact the world to see if you qualify for America.

America is finally going to get Internet at speeds on par with the rest of the world?

With a 34% increase, we'd almost be getting speeds at par with what chipmunks have in 3rd world countries.

This topic is closed to new replies.

Advertisement