IBM Rolls Out Machine Learning Product on Mainframe


IBM sees benefits in using machine learning where most enterprise data resides.

IBM has extracted the core machine learning technology from IBM Watson and will place it on the z System mainframe, the company said in a Feb. 15 announcement.

The new machine learning product, which allows data scientists to create analytic models on the private cloud, includes support for: 

  • Any language (Scala, Java, Python).
  • Any popular machine learning framework (e.g. Apache SparkML, TensorFlow, H2O)
  • Any transactional data type. 

IBM said the z System mainframe is the “operational core of global organizations where billions of daily transactions are processed by banks, retailers, insurers, transportation firms and governments.”

The mainframe is capable of processing up to 2.5 billion transactions — the equivalent of roughly 100 Cyber Mondays — in a single day. By moving machine learning to the mainframe, IBM said it could extract greater value from z Systems data by minimizing latency, reducing costly processing, and avoiding security risks associated with traditional ETL processes

IBM Machine Learning also offers “Cognitive Automation for Data Scientists” from IBM Research to assist data scientists in choosing the right algorithm for the data.  The cognitive automation feature will score data against the best available algorithms and provide the best match for their needs.

“The service also considers various circumstances – such as what the algorithm is needed to do and how fast it needs to produce results,” IBM stated in a press release.

The company cited several use cases for the product:

  • Retail: A sales forecasting system that takes into current market trends and offers real-time personalization.
  • Financial services: Product recommendations based on current interests, trends, and market movements.
  • Healthcare: Personalized healthcare offerings must be tailored to an individual and their unique circumstance. Healthcare and personal fitness devices connected via the Internet of Things can be used to collect data on human and machine behavior and interaction.

Argus Health hopes to leverage advanced analytics from IBM Machine Learning with members across various scenarios, including in the doctor’s office and the pharmacy.

“We are excited about the possibilities and the potential we have seen from IBM Machine Learning working in concert with our RxNova claims processing platform, clinical solutions, and applied analytics in creating models that are constantly improving by using new data and enabling real-time results to the benefit of members, their caregivers and physicians,” said Marc Palmer, president of Argus Health, in a press release.


With Apache Spark, old mainframes learn new tricks

Cognitive computing

Machine learning

Chris Raphael

About Chris Raphael

Chris Raphael (full bio) covers fast data technologies and business use cases for real-time analytics. Follow him on Twitter at raphaelc44.

Leave a Reply

Your email address will not be published. Required fields are marked *