Google

Posts you want to find years later go here.
Post Reply
quantus
Tenth Dan Procrastinator
Posts: 4891
Joined: Fri Jul 18, 2003 3:09 am
Location: San Jose, CA

Google

Post by quantus »

I'm just opening up this thread to track the evolution of Google... We've sorta done it up 'til now, but it's been dispersed through many threads and it'd be nice to bring it together some. Feel free to post links to old threads as they become relevant again.

http://www.iht.com/articles/2006/06/13/ ... search.php

So, Google has even more computing power than I thought @ 450,000 servers up from 100,000 in 2003 and 8,000 in 2001. By comparison, it will take MS 'til 2011 to reach 800,000 servers. Even if Google gets another 450,000 computers in 4 years, that's gonna be 100,000 more than MS and a year sooner. To boot, I'm sure Google's utilizing those computers more efficiently than MS. As this world become more and more based on information processing, who's really gonna be able to compete with them?

(yes yes, I know that not all computers are created equal so sheer number shouldn't matter, but I think it will in this case since it's throughput that will win this war, not really computational speed.)
Have you clicked today? Check status, then: People, Jobs or Roads

Martin
Chump
Posts: 124
Joined: Fri Jul 18, 2003 12:30 am
Location: Los Gatos, CA
Contact:

Post by Martin »

Hey dude, recall Moore's law. It applies to basically all hardware. Anyone can catch up.

Jonathan
Grand Pooh-Bah
Posts: 6722
Joined: Tue Sep 19, 2006 8:45 pm
Location: Portland, OR
Contact:

Post by Jonathan »

I think the hidden assumption is that scaling to another 10x machines requires a great deal of software complexity. The cost and expertise involved in the software might make it possible to stay ahead of the curve for quite a while.

quantus
Tenth Dan Procrastinator
Posts: 4891
Joined: Fri Jul 18, 2003 3:09 am
Location: San Jose, CA

Post by quantus »

Exactly. Node management and fault tollerance become huge factors as you continue to grow the system, both of which require a significant software investment. Then throw in the heterogeneous nature of this huge system and you've just multiplied your problems some more. You need to manage all that data as well by keeping it distributed, redundant and close to the areas more likely to use it.
Have you clicked today? Check status, then: People, Jobs or Roads

Dave
Tenth Dan Procrastinator
Posts: 3483
Joined: Fri Jul 18, 2003 3:40 pm

Post by Dave »

I wish i bought google stock for $90.

and now i own all 5 latest posts!

I deserve a new title I believe.
It takes 43 muscles to frown and 17 to smile, but it doesn't take any to just sit there with a dumb look on your face.

Post Reply