High Scalability

I ran across a link to this site today at work:
http://highscalability.com/. A site dedicated to tracking architectures of very high traffic sites on the Net. Lots of good information if you are curious about how these sites work. The only thing that seems to be missing is pictures.

Some key points I have noticed.

• Lots of horizontal scaling. Eventually, having once uber database server always breaks. So lots of little cheap machines instead of 1 big machine.


• Lots of caching. You will see lots of references to memcached and similar software.


• Language is not your limiting factor. You see sites that run on Perl, PHP, Java, Ruby, C++, C, etc... If done right, any of these languages should work.

LAMP Scaling by the Big Boys

I ran across this at work and thought I should share. There are a lot of LAMP (Linux, Apache, MySQL, Perl/PHP/Python) sites out there on the Internet. One of the bigger ones is Wikipedia. Alka Gupta has posted this blog entry on how that site is scaled.

A very nice high level explanation. Good information to have in case you ever have a site that gets this big. Seven billion page views a month is a lot!