Highly Scalable Website Architectures

I’ve always wondered why my SQL Server 2000 database with only 15,000 records would be so slow to display search results compared to Google search.  While it’s practically impossible to recreate the custom hardware and web servers Google uses to power Google.com, there are many alternatives that can still get you high performance at relatively low cost. Following are some links to how Instagram, which was recently bought by Facebook for $1B, scaled their service on a budget.

http://instagram-engineering.tumblr.com/post/13649370142/what-powers-instagram-hundreds-of-instances-dozens-of

http://techcrunch.com/2012/04/12/how-to-scale-a-1-billion-startup-a-guide-from-instagram-co-founder-mike-krieger/

http://highscalability.com/blog/2012/4/9/the-instagram-architecture-facebook-bought-for-a-cool-billio.html

To learn more about how other companies scale their websites, visit http://highscalability.com.

Cross-Browser CSS Gradients

CSS gradients are great and much better and easier to use than gradient images. However, the CSS code to created them can get a bit complicated. Here’s a CSS Gradient Generator that makes this super easy and supports color stops and many preset gradients.

http://www.colorzilla.com/gradient-editor/

If you already have some CSS for your gradients, you can copy and paste it (import it) to generator cross-browser CSS for your gradients.

Also, make sure to enable IE 9 support if you still need it.

Also, remove “filter: progid:DXImageTransform.Microsoft.gradient( startColorstr=’#a3d2f6′, endColorstr=’#ffffff’,GradientType=0 ); /* IE6-8 */” since that messes up IE 9.

Free, Open Source Web Fonts

There’s no doubt that the type of fonts you choose for your website can make a big impact on its look.  The problem is many nice fonts are expensive and embedding them in a website requires converting them to the right formats and including them in your CSS correctly. One free alternative is to use Google Web Fonts which, at the time of this posting, has 501 font families. It’s super easy to add to a website by just copying and pasting.

To learn more, visit http://www.google.com/webfonts

Title Case

Title Case

This happens so often that I’m actually tired of seeing it, especially coming from Americans and Brits. Apparently, many people don’t know how to title case words that are used in a title. While there are specific rules like you’re not supposed to title prepositions, articles, and conjunctions, instead remembering those are, you can just use a title case converter.

http://titlecase.com

There, now you have no excuse for misspelling titles.

Unicode / Foreign URLs

Recently, I came across a need to create a redirect for a URL that contained a unicode character. The url was something like

www.somedomain.com/réseau (notice the accented character)

While this would be easy to type for French people, it isn’t so for others, like Americans. To accomodate different possible spellings, a good practice is to create a redirect for 2 alternative spellings so that any one of the following URLs resolve:

www.somedomain.com/réseau
www.somedomain.com/r%C3%A9seau
www.somedomain.com/reseau

This way, percent-encoded URLs will work and people who don’t feel like finding out how to type “é” can just type “e” and that would work as well 🙂

Javascript Screen Scraper / Website Downloader

Recently I needed to download the HTML source code from 9640 pages behind a login. At first I used PHP and cURL which normally works just find. With cURL, I first cURLed to the login page passing my login info. Then, in the same script, I cURLed to the pages I wanted to download and save to my local machine.  That didn’t work so I inspected the headers and set cURL’s header options to be exactly the same. That didn’t work either and I just got a 301 Moved Permanently error. It seemed like I needed a way for the remote server to think I was no different than a regular web user browsing in a web browser. So, I tried Javascript. I logged into the website, fired up Firebug, and pasted the following code into the console tab’s command line:

[cc lang=”js”]
function wait(msecs) {
var start = new Date().getTime();
var cur = start
while(cur – start < msecs) {
cur = new Date().getTime();
}
}

Continue reading Javascript Screen Scraper / Website Downloader