2 replies to this topic

[Shanpav]

Hello,
I am little bit confused about the Starting of Google Search engine, In wikipedia they say that Larry and Sergey rammed their dormitory room with inexpensive computers and applied Brin’s data mining system to build a superior search engine, but Search engine uses a database (If I am correct) and to process a lot of data they need powerful PC's and how then lowcost PCswill handle the data, Can anybody explain how they did it, As today microsoft Uses so much Powerful Computers but still Strive to beat google,but cant till date beated it.
Is Search engine not a Question of a good computing Powerhouse also inside a big University like Stanford.
Please help to guide me.

[Orjan]

I have no knowledge in this particular case, but you can make many computers work as one, as a cluster, where the work load is divided by the computers the cluster is containing.

What about great power, they has obviously changed computers since their development stage at university, and they use a special database, they store vertically instead of horizontal, has I got it explained to me. I'd google a bit about it, maybe read wikipedia.

[WingedPanther]

A Beowulf cluster is an example of a super-computer built out of cheap computers. Processing power is NOT measured in the speed of the individual components. If a single processor can perform 1million operations per second (ops), then 100 of these computers can perform 100million ops. If you need more processing power, you can gain it by either upgrading your computers, or by getting more computers. Since a 1 GHz processor is much cheaper than a 4 GHz processor, for example, you can get ten 1GHz processors and actually have more processing power for the price of one 4GHz processor. The question ends up being: how fast is the system?