We had many discussions about Big Data and some circled arround a definition of the Big Data problem.
In the end we concluded that a Big Data problem is described as a data collection, processing and / or analytics problem you seriously have to think about the solution before building it.
Although this sounds vage, I think it is a good conclusing because a Big Data problem does not only exist if you have a lot of data, the type of data (many rows and/or many columns), the query type (how frequently, how complex) and the real-time requirements (latency in import and querying, query response times) are at least some factor which influcene the solution design.
One thing should be clear to all who have to deal with a Big Data problem for the first time. Oracle and Hadoop are not a solution by definition - just because they have a big name and are well known, the don't fit every problem and there are much more alternatives (and many are much better) who have to be considered.
Best
Mike (parstream.com)
The best of ZDNet, delivered
ZDNet Newsletters
Get the best of ZDNet delivered straight to your inbox
