IBM set to deliver mainframe AI services, support

As it previewed in March, IBM is set to deliver an AI-infused, hybrid-cloud oriented version of its z/OS mainframe operating system.

Set for delivery on Sept. 29, z/OS 3.1, the operating system grows IBM’s AI portfolio to let customers securely deploy AI applications co-located with z/OS applications and data, as well as a variety of new features such as container extensions for Red Hat and Linux applications that better support hybrid cloud applications on the Big Iron.

In this release of the mainframe’s OS, AI support is implemented in a feature package called AI System Services for IBM z/OS version 1.1. that lets customers build an AI Framework that IBM says is designed to support initial and future intelligent z/OS management capabilities.

It includes support for key AI lifecycle phases including data ingestion, AI model training, inferencing, AI model quality monitoring, and retraining services, IBM says.

“AI System Services is intended to offer a seamless and simplified installation, setup, and management experience of the AI-infused capabilities without requiring additional data science or AI skills,” IBM wrote in the Announcement Letter detailing the new release.   “It is designed to pave the way for AI use case providers that can harness the foundational AI capabilities to address AI model operationalization requirements, simplify the process to put future AI use cases to work, and accelerate time to market.”

The AI System Services package includes:

• A Data Collection engine that delivers a common way to ingest, filter, and collect IT data intended to be used for AI model training and to perform inference. It parses and collects the raw IT data, and streams it to the data store, an AI Framework for IBM z/OS component that holds data to train the AI models and data to be used for inference purposes, IBM wrote.
• An AI model server that provides a set of capabilities used to operationalize AI models by use case providers. It focuses on training the prebuilt models or AI model training applications with your data to take your data patterns into consideration, deploy the trained model, and apply inference and AI model management, including model quality monitoring, retraining, and versioning capabilities. The AI model server uses IBM Watson Machine Learning for z/OS Core edition, IBM wrote.
• An AI-powered Workload Manager (WLM), designed to predict upcoming batch workload and react accordingly to proactively optimize system resources.

While supporting AI applications is a core component of the z/OS strategy (which runs on older /14 and z/15 models as well as the new z/16) the operating system also continues the cloud transformation support of the mainframe, IBM stated.

Utilizing the mainframe to support cloud environments is a growing trend according to industry watchers.

A recent mainframe customer survey by Aracti asked mainframe sites whether they operate in a hybrid cloud environment that found 42% of respondents currently used their mainframe in a hybrid cloud environment (up from 21% last year).

A further 16% think that they will run a hybrid cloud environment at some time in the future (down from 21% last year).  Still not everyone wants to support clouds on the Big Iron.  Some 42% of the sites surveyed (down from slightly from 47% last year) don’t use hybrid cloud and don’t have any plans to do so.

“z/OS 3.1 is designed for clients to leverage industry standard technology to increase development velocity and modernize their applications with new environments and APIs to consistently build, deploy, and manage workloads, both Linux and z/OS, across a hybrid cloud environment,” IBM wrote.

Along those lines z/OS 3.1 includes Container Extensions (zCX) and support for zCX Foundation for Red Hat OpenShift enhancements designed for improved performance and security, with new features and capabilities to support NFS, HTTPS, IBM WebSphere Hybrid Edition, and IBM Storage Fusion OpenShift Shared Persistent Storage, IBM wrote.

It also includes a new physical file system designed for z/OS UNIX utilities to provide transparent access to data in a secure and consistent manner, IBM stated.

IBM has added other cloud-related services.  For example, last year IBM added IBM Z and Cloud Modernization Stack 2022, and Wazi as-a-Service that lets customers more easily develop and test mainframe applications as-a-service in a public cloud environment.

IBM Z and Cloud Modernization Stack 2022.1.1 offers industry-standard tools to modernize z/OS applications on a pay-per-use basis. For example, the service includes support for features including:

z/OS Connect, which utilizes a JavaScript Object Notation (JSON) interface to tie into and link with existing applications to make Z applications and data part of a hybrid cloud strategy.

z/OS Cloud Broker, which integrates z/OS-based services and resources into Red Hat OpenShift, which can let organizations create, modernize, deploy, and manage applications, data, and infrastructure.

Developer support for IBM Wazi, the cloud-native development environment. A Wazi Sandbox lets developers build and test z/OS application components in a virtualized, containerized, self-service, personal sandbox running on Red Hat OpenShift that runs on x86 hardware.