Caffeine takes up nearly 100 million gigabytes of storage in one database
(100,000,000 * GB) / TB == 97,656.
So Caffeine’s current index could be stored non-redundantly on ~98 thousand 1TB drives. Newegg’s cheapest 1TB drive is currently $70, so that’s very-naively 6.8 million USD in HDD drive costs alone.
and adds new information at a rate of hundreds of thousands of gigabytes per day.
And it sounds like you’re on the hook for buying “hundreds” of 1TB drives each day, with a burn rate between $14,000/day (200 TB drives) to $63,000/day (900 TB drives).
My math is wildly naive as economies of scale push prices down and machine costs + redundancy + networking + labor + energy drive costs up, but it’s fun to think about.
The second ingredient that Matt discussed is that of site speed. From the sound of it, speed is going to be a huge factor in SEO moving in to 2010. He says that a number of people within Google consider speed to be very important to the web, and they are considering if that should play a role in the rankings of websites in search results.
According to Cutts, speed hasn’t played a role in rankings in the past, but that may very well change. Watch the video for more details about Caffeine and Google’s resources for helping webmasters improve site speed.
What a terrible idea — I hope it doesn’t go through.
This idea optimizes against great User Experience (sites with great graphics+ajax load slower than HTML 2.0-sites), technological progress (libraries+frameworks are getting bigger as web apps do more) and content-egalitarianism (often the best sites are some lone wolf’s hobby on a shared host while corporates sites can afford fat pipes and dedicated metal).