More Google Guessing

I wrote about Google's purchase of "dark fiber", the unused super high speed connections that form the guts of the Internet. At the time nobody knew Google's motivation, the heart of the Internet is fast enough now. The last mile connections to the home are the limitation. During my last post I had an idea that might explain what Google is up to.

Distributed applications work on a common problem but live on separate computers. Two of the more famous examples are SETI@Home and Folding@Home. They can use your computer to tackle a small piece of a really hard problem like scanning radio signals for signs of life or simulating the movements of human protein. Millions of Internet connected computers when combined act as one big supercomputer. So what does that have to do with Google? Well for starters they embedded Folding@home in some copies of the Google Toolbar which has been installed on millions of computers to block popups and provide easy access to search features. Why? My theory is that they were trying to learn about distributed computing over the Internet and not being friendly to Folding@Home out of the goodness of their heart.

Google has a huge threat to its revenue because it is just software and not a lot of software at that. It happens to have a huge server farm which runs that software but here is the problem: Hardware gets cheaper and software gets Open Source. Google is rich because people trust them and their product is simple to use. Users are fickle. Users will use whatever their trusted tech friend suggests and those tech friends are increasingly fans of Open Source. Google's search algorithm is not open source.

Competition is a big problem for Google. Flickr is better than the Google image search and it may only be a matter of time before something beats their web search. A9 seems to have some ideas about how to do it. So like any rational investor Google is trying to diversify. Why dark fiber?

If you sell software and someone creates a free, open source version of it you will go out of business. If you sell hardware people can copy the design and manufacture it in a country with liberal intellectual property laws and undercut your prices because they didn't need and R&D budget. Google's purchase of Internet infrastructure might prevent the threat of open source software and the threat of cheap knock-off hardware. Basically I'm suggesting they're going to get into the router business. No government wants the infrastructure of the global economy(E-commerce, VOIP, email) running on cheap hardware and they probably also don't want that same infrastructure running on software whos potential flaws are available to "many eyeballs". Obscurity isn't a sound basis for security but Google could argue that diversity is one form of security. The SQL slammer worm flooded the Internet with traffic and shut down all sorts of services including ATMs. It had no effect on Linux or Mac machines, only Windows machines running certain database software.

Google could write an operating system dedicated to routing which wouldn't be vulnerable to flaws found in Linux or BSD based routers. They could install it on proprietary Google Routers which they'll use to turn their dark fiber into lit fiber. They might sell it to governments that wish to prevent terrorist attacks on the infrasturcture which would cause major economic damage. Maybe that's not so far fetched considering Bin Laden wants to attack ports to disturb commerce. If their next generation distributed search algorithm can take advantage of some proprietary technology that's built into their routers then they will have no competition.