I saw on nytimes.com today “I.B.M. to Push ‘Cloud Computing, ’ Using Data From Afar“. The idea is to make it easier to process large volumes of data using centralized computing capacity in the “cloud”. Despite having several hundred phds working on this, they’re offering the exact same software (Hadoop) and vaunted “google programming model” as you can already use, (at 10 cents per server-hour) on Amazon’s Elastic Compute Cloud.
When you’re out innovated and out maneuvered in bringing a (open source) technology to market by a book company, its no wonder your stock price is basically stagnant.
I recently had the opportunity to use Amazon EC2 to process ~35 gigs of log files for someone and it was quite easy, costing me about $15 in data transfer and CPU time fees, and I bet it would have been cheaper/faster if I’d taken the time to use something like hadoop.
It’s clear from the article that IBM doesn’t yet have a firm grasp on cloud computing. They still see it as number crunching.
Meanwhile most users of EC2 or 3tera’s own AppLogic aren’t running Hadoop. They use these utility computing services as a more flexible alternative to colocation and managed services.