Mainframes Aren't Dead, They're Just Learning AI Tricks

Even mainframes can't escape the AI bug, with a report finding that big iron is becoming a prime candidate to host and run AI workloads, while enterprises are increasingly integrating their mainframes with modern infrastructure rather than seeking an off-ramp.

IBM spin-off Kyndryl says in its second State of Mainframe Modernization survey that big iron has become the cornerstone of a hybrid IT strategy supporting the most mission-critical workloads, at least for those companies still operating them, although skills remain a problem.

From a survey of 500 senior IT decision makers, the infrastructure services firm says it found many organizations are integrating their mainframe with public and private cloud platforms and rationalizing their modernization programs – moving some workloads off the mainframe while updating others in place to continue to benefit from the security and reliability of the platform.

Given that Kyndryl is all about IT infrastructure and integrating systems, color us shocked that it put out this conclusion.

Naming 2024 as "the year of AI adoption on the mainframe," Kyndryl says 86 percent of those surveyed are deploying or planning to deploy generative AI tools and applications in their mainframe environment.

IBM itself said that GenAI was a factor in its mainframe biz bringing in more revenue in its most recent quarterly results.

The majority (80 percent) of respondents are still at the stage of exploring the possibilities, but the survey found that 41 percent of respondents hope to use GenAI to speed operational actions and make them less prone to human error. Meanwhile, 33 percent are targeting improvements to customer experience through increased personalization, for example.

Other goals for AI are said to be unlocking mission-critical data and transforming unstructured data into actionable information, while a third of respondents expect to use GenAI to uncover business insights from their mainframe-managed data to aid the development of new products or services.

According to Kyndryl, GenAI also has the potential to aid modernization efforts by "illuminating the inner workings of monolithic applications," which it claims could offset the shortage of mainframe skills in the current workforce.

However, a minority of big iron users are not planning to use GenAI in any shape or form, citing security and regulatory concerns for their reluctance, or simply that other priorities are taking precedence.

The survey also finds that organizations still operating mainframes are increasingly seeking to use them to best effect while also enjoying the flexibility offered by cloud services. Thus, while 96 percent of respondents said they are moving some workloads off their big iron (typically about a third), 89 percent agreed that these systems are still extremely or very important in their business operations.

In the report, Kyndryl says it distinguishes three basic patterns to mainframe modernization, with most organizations following a combination of them. One is to migrate away, transferring some or all applications and data to the cloud or to an on-premises alternative.

The second is to integrate mainframe data and applications with other platforms, allowing them to be accessed by new cloud-based applications, for example. This pattern is forecast to increase with the continuing evolution of AI and GenAI, Kyndryl says.

Number three is to modernize workloads on the mainframe, defining which applications to keep, replace, or retire. This involves modernizing application source code or using more modern languages, as well as exploiting new technologies such as AI and containerization.

According to the report, more respondents this year said they are focused on modernizing in situ or integrating with cloud, while fewer are choosing to move workloads away from the mainframe as their primary strategy.

• IBM reveals upcoming chips to power large-scale AI on next-gen big iron
• You're not hallucinating: Generative AI is helping IBM's mainframes grow
• AI query optimization in IBM's Db2 shows you can teach a tech dinosaur new tricks
• BMC's $1.6B victory over IBM is TKO on appeal

Some 53 percent of respondents said their use of the mainframe increased this year, with 49 percent saying they expect it to increase again over the next 12 months – words that will no doubt be music to IBM's ears.

The return on investment for these modernization projects is impressive, according to Kyndryl, yielding one-year returns of between 114 and 225 percent for organizations proceeding with them. The company has a vested interest in this, of course, as it styles itself as the world's largest IT infrastructure services provider and, as an IBM spin-off, has much big iron knowledge and experience.

Perhaps the growing prominence of cyber incidents is having an effect, but 66 percent of respondents in the survey said that the level of security offered by mainframes is their most important feature, and almost half cited security as the number-one driver of their investment in modernization.

One head of IT at a wholesale enterprise in the US quoted in the report said they adopted a hybrid cloud approach because it allows them to maintain sensitive data within a secure mainframe environment while leveraging the cloud for less critical workloads.

But longstanding issues with a shortage of staff with relevant experience remain, and this is being compounded by rapidly evolving areas such as AI and cybersecurity.

The report finds that 18 percent of business leaders integrating their big iron with other platforms said that insufficient expertise has been the main hurdle to project success, while more than a quarter of respondents are concerned their organization lacks the right level of skills to effectively modernize their mainframe.

Enterprises still face the twin-pronged difficulty of new hires entering the workplace with a lack of mainframe skills while experienced staff are retiring and taking their skills with them.

These are difficulties that IBM has been seeking to address, unveiling a pair of initiatives earlier this year to tackle the shortage of engineers with mainframe know-how.

Kyndryl says that generative AI can help by assisting developers in modernizing and converting complex legacy code, identifying dependencies between applications, and even generating technical documentation.

But wouldn't you know it, 43 percent of respondents indicated they lack the skills to use AI and GenAI capabilities, leading more organizations to turn to external providers – like Kyndryl – to deliver their modernization projects.

Overall, the report concludes that the mainframe remains a critical pillar of the IT environment for many enterprises, and is even becoming more relevant again thanks to its security, reliability, and performance.

"Mainframes continue to occupy a central role in the hybrid world and evolve to serve new use cases, with technologies such as AI and security increasingly influencing modernization plans," said Kyndryl's Global Practice Leader for Core Enterprise & zCloud, Petra Goude.

But it is vital the skills shortage does not become a roadblock to progress, she added. ®