Page 2 of 2

Posted: Thu Dec 01, 2005 12:21 am
by s.dot
OK so question here... :P

Obviously, I'm new to having my own server, how servers work, and all that good hoopla. Back in the beginning I thought learning PHP code was everything! Now, I feel so newb again. But that's OK, I'm learning. So onto my question.

How do the major sites IE: msn google yahoo handle their TREMENDOUS load of users? Obvsiously they have more than one server, but then that begs the question, if they have more than one server, is the actual site on all of those servers? Is it on just one server? If it is on just one server, how does the "internet" know where to find their site amongst their crapload of servers?

utterly confused,
-the scrot-

Posted: Thu Dec 01, 2005 8:51 am
by josh
google has thousands of medium size machines, probably around the power of your machine, they offload their load onto these machines, each one contains a different part of the google database, with the exceptions of duplicates for redundancy. This is called a server farm when implemented on a smaller scale, ever go to a website and get redirected to s1.example.com, refresh the page and reg redirected to s2.example.com?

Of course google also has their own team of geniuses [caged up in their headquarters] that work on algorithms 'round the clock for making what each machine has to do, take shorter. They even have their own filesystem I hear?

Try figuring out what is taking all the cpu cycles on your box, then we can point you in the new direction wether it means disabling un-needed services, or hunting down an infinite ninja loop in your code. And no worries I'm actually fairly new to this too, I've been hosting on linux for years but only got my first VPS a few months ago. Now I have a development box in my room, and the fact that I don't have to worry about taking it offline or killing my data allows me to experiment more, between Roja, and google you can find anything you need to know about linux is what I learned.