My third post on unconstrained analytics speaks to the point that analysts should be able to do whatever they want to do in the discovery process in order to achieve the most valuable results. The starting place for unconstrained analytics was the ability to run any analytics, any time. Beyond that, an analytic platform should be able to support any analyst, anywhere. Finally, analysts should be able to run any kind of analytics they want to run; and they should be able to run those analytics against any data they want to analyze.
The ParAccel Extensibility Framework extends the power of the fastest analytic database in the world in any direction, encompassing the full range of analytic capabilities. That means users can analyze any way, including all of the following kinds of analysis:
Traditional analytics: Start with basic SQL, doing analytics the old way.
Existing analytics: Work with any external tool running analytic models that already exist.
Extreme analytics: Test the limits of extreme SQL with 28K queries and up to 1000 joins, combining analytics from across the enterprise.
Statistical analytics: Run statistical, data mining, and simulation algorithms, for rapid results on normally long-running applications.
Advanced analytics: Bring in advanced functions to run within the database engine for maximum performance, taking the speed of analytics to a whole new level.
Now we have a picture of what we mean by unconstrained analytics: analyze any way; any analytics, anytime; and any analyst, anywhere.
