If I asked you for the defining characteristic of a big data customer, you'd probably say they're sitting on large amounts of data. If I asked for the defining characteristic of a NoSQL customer, you might answer they require high levels of concurrency.
Well, if that's the total market for NoSQL and big data, then both MongoDB, Inc., as well as the various companies supporting Hadoop should probably shut their doors and call it a day.
[ 18 essential Hadoop tools for crunching big data. | Work smarter, not harder -- download the Developers' Survival Guide from InfoWorld for all the tips and trends programmers need to know. | Keep up with the latest developer news with InfoWorld's Developer World newsletter. ]
In truth, opting for Hadoop is in many ways an economic decision. If a company has deep pockets and daunting amounts of data, then it can throw money at a high-end MPP solution from IBM, SAP, or Teradata -- in fact, most large companies have already made that sort of investment. But not all of us hang out with the 1 percent and light our cigars with $100 bills. Even those that do then have to make business decisions "up front" on whether the exorbitant costs of keeping data and deciding what to do later.
For the rest of us, Hadoop provides analytics capabilities we couldn't access before. Even the cost of commercially supported "enterprise" distributions of Hadoop amounts to nickels on the dollar compared to, say, IBM Netezza.
NoSQL technologies like MongoDB or Neo4j are also, in effect, economic decisions. If you buy a fat enough server and pay for enough developer time, you can indeed run nearly any document or graph database job in your favorite RDBMS. But developer time is not cheap and server licenses get expensive -- plus, the infrastructure to scale up an RDBMS so that it supports high availability and disaster recovery costs a bundle. No wonder the brighter operations folks like the sound of the NoSQL alternatives: Save money by using commodity hardware, and snap on more servers as needed.