IBM Plans AI-Native 'Z Atlas' Tool to Tame Mainframe Complexity

Your daily brief on leadership and business

IBM, which presents its Z-series mainframes as the most reliable and cost-effective environment for certain workloads, acknowledged on Monday that the platform can slide into complexity that is hard to manage and wears down resilience. Its answer is a forthcoming product called IBM Z Atlas.

The company laid out the plan in a Statement of Direction. It calls Z Atlas an AI-native solution that uses system intelligence to ground AI skills and agents in the context of key workflows, such as planning and validating changes or modernizing applications and infrastructure.

According to the document, Z Atlas intends to map a client's entire z/OS stack across infrastructure, middleware and applications. IBM would then use that map to cut manual effort in tasks like patch management, regulatory response and sharing knowledge among teams, from systems programmers to application developers.

The statement also targets what it names the leading cause of unplanned z/OS outages, which is changes that have not been proven safe. Z Atlas is meant to show how a change affects the stack, test it and validate it. IBM also hopes the tool helps mainframe technicians and operations teams understand how components connect, interact and depend on one another.

The same document says this design speeds up clients' modernization work by keeping experts in control. It also gives every practitioner, in operations or development, the same shared view of the Z system, which IBM says reduces risk.

What IBM expects it to do

IBM says Z Atlas will help mainframe shops with four things:

• understand systems and dependencies before making decisions or starting changes;
• assess potential impacts, risks and operational considerations;
• speed up planning for modernization, deployment and provisioning;
• provide trusted context for AI-native experiences and automation.

AIOps is an emerging practice in which agents either keep infrastructure running or find problems and draft fixes without human help. IBM is not going all the way there: the Statement of Direction says AI-assisted insights and recommendations are intended to preserve appropriate human review and decision authority.

Modernization, the IBM way

Rivals often argue that the best way to modernize a mainframe is to drop it and move its applications to another platform. Vendors have also proposed that generative AI rewrite mainframe applications in other languages. Analyst firm Gartner considers that idea highly overrated.

IBM's revenue spikes each time a new generation of Z-series machines arrives. Last week, CEO Arvind Krishna said the current Z17 generation has outsold all others by $1 billion.

Source: The Register