Take the 2-minute tour ×
Server Fault is a question and answer site for professional system and network administrators. It's 100% free, no registration required.

I've been thinking lately that hosting a site with server-side user input must be extremely expensive/difficult. Take something like facebook for instance. They probably have petabytes of information stored on their servers, generated by users. Isn't this expensive? Can someone explain how this is done?

share|improve this question

migrated from programmers.stackexchange.com Jan 4 '11 at 23:58

This question came from our site for professional programmers interested in conceptual questions about software development.

    
@DarkLightA : What do you mean by server-side user input? –  Abimaran Dec 29 '10 at 15:25
    
@Abimaran: I think he just means various data submitted by users (like status updates on Facebook). –  user43170 Dec 29 '10 at 15:26
    
That's not server side input, actually, the users don't have the access in the server side in a typical web site. –  Abimaran Dec 29 '10 at 15:30
3  
Try Google for size. They've got most of the Web on their servers, past and present, and oodles of indexes to everything. They've had massive research projects on how to run data centers as cheaply as possibly, including detailed studies on temperature vs. reliability and equipment lifetimes. –  David Thornley Dec 29 '10 at 15:32
    
So, it's an intranet site basically ? It's the same thing as a "web" site except users come from the local network, but from the server's point of view, it's the same thing. And as for the high storage and scalability, that's another thing, I'm sure someone will post a great answer on that matter. –  wildpeaks Dec 29 '10 at 16:00
show 2 more comments

2 Answers

Yes its expensive, there are many strategies (mostly around scaling OUT, as opposed to scaling UP)

If you want to understand the architecture behind such sites, you may be able to find about them on this blog: highscalability

I believe there was an article on Facebook

Edit:

There are numerous articles on Facebook on that site.

share|improve this answer
add comment

NoSQL databases may give you some insight on how they are doing this. David Thornley makes a great point about the Google hardware. Their biggest accomplishment may be their ability to manage tens of thousands of servers efficiently.

Some business models that offer free subscriptions have the most issues. If your next 1000 customers are paying a monthly fee, it's a lot easier to come up with the cash to buy more servers.

share|improve this answer
add comment

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.