Now, facing the existential crisis of the cloud, it's trying to do the same thing with Watson, its supercomputer. Watson, IBM hopes, will generate the software and services revenue it needs to remain relevant in the cloud era.
The company plans to invest $1 billion in a new Watson unit. There is reason for optimism. That's because Watson isn't just another way to play Jeopardy. It's also another way to do cloud analytics.
Cloud analytics is the big software challenge of our time. Cloud analytics takes the huge fire hose of data created by the Internet and turns it into intelligence, into answers to questions that may not have been asked yet.
Cloud analytics is the difference between the National Security Agency collecting all that data on what everyone is doing online and actually doing something about terrorism. It's the difference between reacting to what customers may be doing and being pro-active, getting things to them before they ask.
The primary method for analyzing big data sets has been Hadoop, which was developed at Yahoo (YHOO), refined by Google (GOOG) and built through an open-source process. It's okay. If you ask Hadoop to analyze data along certain parameters, it will do that. But if you ask Hadoop a question in English that relates to a big data set, you won't get an answer back in English, or even in real-time.