@JMiles... Re: Big Data is not like a Big Mac
I'd use a flame icon but because I have to post this anonymously, I can't. :-(
You clearly don't know anything about the 'big data' market space, nor anything about big data in general.
Seems you have a chip on your shoulder....
Many early adopters moved to 'Big Data' because it was a cheaper way to handle large amounts of data than alternatives. Keep in mind that traditional data warehouses top out around 100TB. Note that this number changes based on improvements in disk density, speed and networking capabilities. The issue though it that for the cost of the Oracle licenses alone, you could stand up a hadoop cluster.
Then there's the issue of 'democratizing' the data. Which means you can find more value from your data by joining it with other data sets which reside in a different warehouse.
Then there's the ability to run a map/reduce that would allow processing of data in hours that used to take weeks if it would complete at all on traditional data warehouses. The whole basis for Big Data is that it can solve problems over massive data sets as long as the process can be acted upon atomic structures within the data. 'Big Data' isn't HPC (albeit there's now a project that lets you solve HPC problems on your hadoop cluster.)
As to success stories, there are many success stories out there.
However there are also a lot of failures. Not because of the technology but because those tasked with implementing the solution don't know what they are doing.
To be clear, there is a clear correlation between having senior experienced staff and the success rate of Hadoop projects.
The issues with Hortonworks is their business model. Of course there's more, but some of what I would like to say in public, I can't.
And yes, I've been called an expert, though I loathe that term.