Pretty interesting graph extracted from Presentation: Scale at Facebook

Relative Performance of Language Runtime

I knew PHP has pretty huge performance overhead but boy, just look at Ruby and even Python!

But I am not so sure that Python’s performance cost is as equal to PHP because Python has the ability to generate compiled files also which certainly boost performance.

Excerpt from The Art of Scalability Chapter 27,  Too Much Data

1. Storage costs to store data
2. People and software to manage data
3. Power and space to make storage work
4. Capital to ensure the proper power infrastructure
5. Processing power to traverse the data
6. Backup time and costs

Some quite interesting stats of the cost of monetization with the ratio of scaling in different platforms.

But I’m not so sure, why it includes web servers (Nginx, Apache), languages (PHP) and web frameworks (Rails, Django, Mochiweb, Jetty) as a single group, doesn’t make sense.

http://journal.dedasys.com/2010/01/12/rough-estimates-of-the-dollar-cost-of-scaling-web-platforms-part-i