Wednesday, February 20, 2008

Google and the World Wide Web

I was fascinated to hear today how the web search engine Google actually works.

I'm told that if the hundreds of thousands of software "crawlers" used by Google to search web sites for key phrases were actually allowed to crawl over the real Internet, it would grind to a halt and crash in minutes.

Instead, in a secret location in America, over one hundred thousand computers sit in darkness in a collection of hangars. Each computer is tasked with downloading a small part of the world wide web, over which it runs the Google crawlers once downloaded.

A small team of maintenance staff patrol this facility, and when a computer malfunctions, it is immediately replaced with another.

What struck me was that the entire Internet could be downloaded onto just 100,000 computers. It seems amazing that - just fifty years ago - the processing power of the largest industrial mainframe computer could today be surpassed by the processing power of a single musical birthday card. Perhaps in fifty years the entire Internet will download onto a mobile phone?

No comments: