I just wrote this article for Info Week about Big Data, where outside of stealing the concepts from Julie Lockner, we redefine this absurdly ill defined term.
Simply put, big data applies to any data set that breaks the boundaries and conventional capabilities of IT. That's it. What kinds of things break the boundaries of IT? Big capacity does. Big transaction counts do. Big analytical requirements do. Big response time requirements do. Big bandwidth needs do.
Big Data is NOT just analytics. It's NOT just about storage. It's NOT just about anything - it's about everything. It's about tossing out the old way of doing things because those ways simply won't work in the world of BIG.
Think of it this way, take whatever it is that you are concerned with as a metric and ask yourself, "How will I be able to perform this function if I had to do it at a scale of 100X what it is today?" Whatever you come up with that breaks is Big Data, baby.
Why is Big Data way cooler than its buzz alter ego, Cloud? Because Cloud is a new marginally disruptive consumption model. Big Data is a massively disruptive force that will change every single thing in your current IT methods. Everything. It will literally disrupt the way you store, process, find, analyze, protect, secure,support, and deliver services and value to the organization you serve.
Big Data is not "transformational" (my 2011 IT word of the year) - it's CONFRONTATIONAL.
Other than that, it's no big deal.
You can read Steve's other blog entries at The Bigger Truth.