In an effort to expand beyond the increasingly challenging legacy data warehousing market, Teradata Corp. is rolling out a package of technologies that it claims remove the scalability barriers of the R graphical and statistical computing platform while significantly expanding its range of possible uses.

R is the second most widely-used tool in the analytics community according to the latest KDnuggets Software Poll, but it suffers from severe limitations in the amount and size of objects it can handle.

Originally introduced in 1993, the open-source statistical modelling language is designed to analyze only information that’s kept in the memory of the machine on which it runs, whether it be a desktop or a server. That not only restricts how much data the user can work with but also limits the available processing power, a shortcoming that renders the technology practically unusable beyond a certain threshold.

R adoption has slowed as the amount of unstructured information users need to analyze has  it exploded, according to KDNuggets research. It dropped from the number one position in 2012 to second place in 2013 and slipped off the top ten fastest growing technologies this year, although it managed to cling to the runner-up spot thanks to a comfortable from earlier years. Their trend is not encouraging, though.

Teradata Aster R lifts the processing and memory limitations that users have historically struggled with by leveraging the firm’s homegrown parallel in-database execution capability, thereby offloading analysis from applications running on top. That functionality complements what partner Revolution Analytics offers in its ScaleR package, which pulls objects into memory from a server-deployed R implementation or a Teradata backend and distributes the load among the available cores.

Teradata has seen its fortunes erode as the market has shifted away from  database machines and toward distributed Big Data analytics. Its stock price has fallen by nearly half since its September, 2012 high.

In addition to supporting larger scale,  Teradata Aster R comes with more than 100 pre-built R functions for data management, access, exploration and machine learning that run in parallel out of the box, a feature the company says can save analysts days of tweaking. Also included in the package is its patented nPath engine, which adds another 70 ready-to-use queries that allow developers to work with information stored in Hadoop using familiar SQL syntax.

On top of that, the platform packs a Parallel Connector that makes it possible for users to build their own algorithms using the more than 5,500 R available packages as well as any new ones that emerge from the open-source community. Perhaps most significantly, it provides integration with the Aster Snap Framework, which enables multiple analytic engines and databases to be “snapped” together and orchestrated centrally. That functionality can be useful  for advanced applications such as fraud detection, which draw on multiple sources of information.

photo credit: krazydad / jbum via photopin cc