Jan 4, 2003 | Q
Over the past few years, I’ve heard rumors that the reason Internet Explorer loads web pages so much faster than its competitors was that it takes liberties with the way that it requests the pages. Finally, someone put the effort into analyzing the conversation that the web browser has with servers, proving that the rumors are true, and that as a result, there’s a built-in advantage to using Internet Explorer with Microsoft’s own web server. And as much as I love IE (and generally defend the actions of Microsoft in the software market), I don’t like that the company is playing loose with the fundamental specs that govern how machines talk to each other on the Internet.
IE 5 is much faster than its competitors (and IE 6) when loading pages from Apache running on Linux. I know this from having measured browser load performance at a previous job. There’s something more to the IE 5 performance advantage than the cheat described in the article.
It’s very possible that writer was observing HTTP keep alive at work. The packet sequences described in the article are exactly what you would expect with keep alive.
• Posted by: Bravada Zadada on Jan 5, 2003, 1:12 AMInteresting, that it could be HTTP Keep-Alive. I hadn’t thought of that. It shouldn’t be hard to test…
• Posted by: Jason on Jan 5, 2003, 1:22 AMSome posters at Slashdot agree with the keep-alive analysis.
But hey, it’s more fun to think that Microsoft is cheating in some way!
• Posted by: Bravada Zadada on Jan 5, 2003, 6:13 PM