Kognitio Enables Computationally Intensive Processes via Massively Parallel ‘R’

Bookmark and Share

Big data, in-memory analytics and cloud computing vendor Kognitio has announced that the Kognitio Analytical Platform enables new fully parallel not-only-SQL (NoSQL) capabilities, including the R language for statistical computing and graphics.

Kognitio’s parallelization of R is part of the greater external scripting functionality in the latest release of the Kognitio Analytical Platform. “External scripting allows you to write a program or script, or take an existing binary such as R, and use Kognitio as a parallelization engine. We allow you to use R against every single core that you can bring to bear against the data, and we handle the parallelization,” Roger Gaskell, CTO of Kognitio, tells 5 Minute Briefing. This is cited by analysts as a principal aspect of the “logical data warehouse,” which provides a data science lab functionality that enables distributed processing, an emerging best practice for analytical data management.

Processes such as product forecasting, Monte Carlo scoring and robust clustering are computationally intense and difficult to execute over rapidly growing data volumes. Kognitio Analytical Platform can deliver results in near real-time as opposed to the hours or days required by competing solutions. “We’ve put in a framework that allows you to execute anything as a script or a binary and run it in parallel; R is just one example. That is an important differentiator and provides a huge degree of flexibility,” Gaskell explains. Data scientists can design R jobs as they normally would, using the Kognitio Analytical Platform to parallelize and run thousands of jobs from a single query. The external tables functionality allows data to be brought directly into RAM from a range of sources including Hadoop.

The new capabilities will be fully available as part of the Kognitio version 8 release in June. For more information, visit