Silly Google, do you realise what kind of gold mine you are sitting on for the hard drive industry? It's a pretty good bet that at least, oh 80 per cent of PC owners of laptops and desktops don't back up stuff on a regular basis and an unhealthy chunk of them don't back up things at all. Regrettably, this writer must confess he's in the former category, conducting sporadic backups when big money projects with many hours are on the line.
With all sorts of empty disk space sitting around and growing, hard disk manufacturers are going to need a better excuse to sell more drives. If you're sticking a 400GB drive into a desktop box, it's going to make a tidy profit, but being able to bootstrap the sale of three more 400GB drives because a desktop-tailored GFS will transparently salt copies away on other drives is just, well, silly.
Given their druthers and natural habits, people don't want to worry about backing up their home machines. They just want it to naturally happen. And despite all the hype about backing up your storage remotely through a broadband connection, most people would prefer to keep the family financial information and photos on a hard drive at home, not at a data centre that can be snooped at by Big Brother or the local dime store hacker.
Google has some other little code treasures in its closet that could be useful for PC applications. MapReduce simplifies code for parallel processing, just the sort of kit you want for dual-core and coming quad-core processors Of course, Microsoft says it has something called Driad that's better. Google Work Queue allows a big pool of services to be assigned to projects as needed, not exactly a tool for the home, but can it compete in the enterprise arena?
If you believe all of the reporting on the Times, Google is also on the road to designing its own CPUs, hiring many of the old DEC Alpha crew so they can get faster processors that suck less energy. I'm not sure if this is a great idea on Google's part, since it's going to be difficult to leapfrog Intel's pipeline of new designs and manufacturing processes for the next three years. On the other hand, Google is likely to buy processors in such vast quantities that it might actually make sense for them to commission its own optimised silicon. µ