In "Bending Browser Connection Limits", we showed how some people will move various objects like images, scripts, and CSS files to other related URLs like images.theirdomain.com to improve the number of connections the browser establishes. Well, one of the paranoid amongst our staff demonstrates that, like everything else in Computer Science, this technique comes with a trade-off.
For those who are interested in privacy, image URLs sourced to another site could be a telltale sign of a tracking GIF or 'Web bug'. Please note that some browsers, particularly Firefox, provide an interesting feature to squash such third party images. So, if you turn that setting on. you will find that you will need to make quite a number of exceptions to have a reasonable experience. See this screen grab for the Firefox setting:
As an example, you might go to a site like Wells Fargo for your online banking. Normally, it looks like this:
If you attempt to improve your privacy by blocking non-originating server images, you find that the page becomes pretty much unusable. Not only do you lose most of the branding indicators, but you lose the all-important 'sign on' button!:
Now, a closer look at this example reveals the big trade-off. You'll notice that in the case of Wells Fargo, they use a CDN (Content Distribution Network) called Akamai which uses modified URLs. The idea with this is that the heavy payload is cached towards the edge of the network to improve delivery. Good stuff -- if you can afford it. Unfortunately, the images get blocked because of your browser settings. Now, you can go add the exception as suggested earlier, but getting sites to work could be a pretty involved process.
Well, there you have it, folks. Try to speed your site, and you might accidentally lock out some privacy-minded users. Like a host of other things in technology and life, each kiss may have its related curse.
Cheers,
Port80