IBM is never a stranger to innovation and being a leader in the market, especially when it concerns artificial intelligence and cognitive applications. The announcement that Python and the Anaconda stack is compatible with System z and the z/OS is just the beginning it seems, with a full-blown machine learning platform from IBM being released. And it’ll be available on System z before becoming generally available to x86 and of course their own POWER architecture.
IBM is making Machine Learning more friendly for System z and more
The great thing about this new platform is that it’s both framework and language agnostic. And of course, being released on System Z and z/OS first, it’s meant to help speed up and augment the kinds of insights that can be gleaned from the transactional data that the architecture is most suited towards. There’s tremendous value in data and being able to run true machine learning algorithms to extract that value where the data itself resides is tremendously useful. It’s also much more secure and much less complex than moving the data off of the mainframe, elsewhere.
One of the selling points of the new platform from IBM is in how it’ll help to find the best model to fit the data you’re giving it. That and it can be run rather easily within a Jupyter notebook to make creating the model more interactive and slightly more intuitive.
IBM Machine Learning will come to z/OS first, though will be available for other platforms in the future, including IBM’s own POWER-based systems. By deploying IBM Machine Learning on its POWER Systems, clients can leverage machine learning with greater efficiency, higher performance and cost effectiveness along with having full NVIDIA Pascal P100 support for faster training times.