IBM z/OS 3.1 Sets the Stage for an AI-Infused Smart Operating System
Move over ChatGPT. IBM is creating its own AI buzz with IBM z/OS 3.1. In March this year, IBM previewed this momentous release, which introduces AI to the operating system (OS). The company confidently calls it the “AI-infused operating system for the next generation of computing.”
IBM z/OS Product Manager Mike Young says, “Including AI in this release signals the shift in IBM’s mindset and strategy for the Z platform now and into the future. It’s setting the stage for a fully AI-infused smart operating system.”
Young recently gave us an insider’s look into how AI is preparing z/OS 3.1 for the future and what it means for IBM and IBM clients.
AI’s Role in the z/OS and zSystems Strategy
IBM has long had an AI vision for its z/OS and zSystems. As the company prepares for AI’s big debut in 3.1, Young explains it is part of the natural progression.
“We realized the next logical step was to bring AI into our own products and the operating system itself. Our clients are already using AI to solve business problems. By having it in their OS, they can use it to help solve core underlying IT concerns or issues that crop up,” Young says. “Our mantra was to enable AI seamlessly at the system core to make it proactively resolve problems before they occur.”
With that, IBM introduced AI on IBM z16 in April 2022 with the addition of the Telum processor industry’s first on-chip AI accelerator. This release included support for several frameworks with AI library capabilities so clients can deploy AI to z/OS where their data resides. By having this function, clients can avoid the complexity, security concerns and expense associated with moving data to another platform or location, which are both costly and risky.
The advantages of the acceleration chip are core to the AI capabilities of z/OS that carry over to the z16 platform. Young says, “The big impact will be in the synergy between the operating system and platform in taking advantage of the hardware function provided by the on-chip accelerator. By integrating it throughout the whole operating system, we’ll help improve AI performance on z/OS so our clients can take advantage of AI and the operating system benefits too.”
AI Foundation Augmented With Sponsor User Data
The AI foundation in z/OS 3.1 relies on the same technology that underpins Watson Machine Learning. It’s further enhanced with development data from members of the IBM zSystems Sponsor User program.
“We’ve had this program for a few releases now. It’s grown and continues to be crucial to our development. We invite clients across the install base to apply to join us,” Young says. “These clients come with broad coverage across geographies, organization size and industries, including financial services, government agencies, retail and insurance. The diversity of our clients enables us to understand various use cases and factor in our clients’ insights.”
As part of the Sponsor User program, many clients were invited to help provide some of their operational data to train the models to validate that they work logically. This information will help the z/OS team ensure the models operate as expected on different client environments by first proving they work in the Sponsor User environment or with their data.
Also critical to Version 3.1 is the use of IBM Design Thinking. “This process has been amazing in helping simplify our efforts to ensure seamless integration with external features or functions, such as from the cloud space,” Young added. “Our clients have been working with our design and development teams every step of the way, from ideation to features, functions, usability, interfaces and even validation. So far, they’re pleased with the ability to apply AI to their data, which was the real driver for incorporating AI right into the OS platform.”
Confidence in a Secure and Reliable AI Framework
The biggest addition in z/OS 3.1 is the secure framework to support the system-level services with all the necessary AI capabilities. The team plans to build on that framework with further AI support in future V3 releases to improve individual functions.
“This framework will enable the operating system components and IT management tools to take advantage of AI together,” Young says. “It enables us to instill trust in the OS by delivering secure and non-biased AI results without the risk of changes by outside influencers. The centralized location is also helpful when teams must justify and explain results to upper management and auditors.”
In fact, the first component to take advantage of this framework is Workload Manager (WLM). z/OS will be able to predict the upcoming workload on the system and prepare its resources in advance. This way, when the workload runs, z/OS will be ready to process it, without delays, latency or the need to bring in additional resources.
From there, IBM plans to keep adding new functions as part of its continuous delivery strategy for z/OS.
Key Features Beyond AI
As ground-breaking as AI is to z/OS 3.1, the preview teases of quantum-safe encryption, resiliency, OS simplification and client application modernization.
Quantum-Safe Encryption and Resiliency
In addition to AI in z/OS 3.1, Young highlighted the release’s quantum-safe security and resiliency features. Building on discussions based on z16, IBM is making strides to improve resiliency and security—the fundamental tenets of IBM zSystems—to stay ahead of potential adversaries.
“To ensure our clients remain protected, we’re taking advantage of the security hardware acceleration in z16 and adding quantum-safe signature algorithms,” Young says. “They’ll be available in applications, such as Integrated Cryptographic Services Facility (ICSF), and in the resiliency space.”
Another new optional, priced feature is the z/OS Authorized Code Scanner (z/OS ACS). With this tool, users can scan their system and monitor it for potential privilege escalations in their applications. It looks for programming paradigms that are potentially unsafe or otherwise problematic. When it detects one, it alerts the system, user and administrator so they can take any necessary action.
“Our clients have expressed a lot of interest in this capability lately, considering security concerns in the market,” Young says. “We hope it will go a long way in helping them validate everything runs on their system as expected and gives them confidence in knowing potential privilege concerns aren’t present.”
OS Simplification and Client Application Modernization
In this space, IBM is further enhancing z/OS Container Extensions (zCX), which runs Linux applications inside z/OS. In 3.1, zCX will support OpenShift and include several performance integration points that take more advantage of the hardware.
“These updates enable application developers to run the Linux frameworks they’re familiar with alongside the core heritage z/OS applications that run their business. They‘ll save significant time over setting up an entirely different environment for that one application service they want to use,” Young explains.
Also in the simplification category for 3.1 is the ability to access and write to data sets on z/OS using not only SQL (NoSQL) APIs. This feature will enable clients to access their critical core data directly through a NoSQL API instead of using extract, transform and load (ETL). As Young mentioned, moving data somewhere else comes with cost and risk, which this feature avoids.
More to Come
IBM is leading the way in providing a secure, trusted AI-augmented z/OS platform. By leveraging AI technologies in Watson and customer development data, 3.1 is elevating z/OS to a higher standard of operating systems. “This version will be foundational for the future of IBM Z Systems and z/OS,” says Young.
While the release of z/OS 3.1 is still in preview, IBM aims to make it available during third quarter this year. Stay tuned to TechChannel to keep up on news and further developments about this release.