Eric Baldeschwieler interviewed by WSJ:
WSJ: What is the secret sauce of Hadoop? How does it achieve that level of performance?
EB: It doesn’t try to solve everything. [Traditional] SQL databases try to solve a set of problems that are very complicated. Hadoop only tries to solve a subset of them. By working on only a subset of the problems you can get a design that is inherently scalable, but can be broken up into chunks and distributed across a whole bunch of computers, and broken up in such way that you get recovery from failure.
Original title and link: Hadoop: The King of Big Data ( ©myNoSQL)