There are few advantages to growing old, but if you were lucky enough to train, back in the hidden midst of times past, in programming languages COBOL or PL/I, you may have landed in later life into a well-paid role maintaining legacy mainframe systems.
A large part of the market for mainframes collapsed in the 1990s when the price of computing and mass storage fell dramatically. But mainframes still exist, and continue to run to this day, powering at least parts of essential systems in industries such as finance, insurance, government departments, and retail.
But as the last generation of COBOL programmers seek a well financially-cushioned retirement, companies are being left with the prospect of core systems failing, and no one qualified to keep them running. Now available in AWS Zones based in North Virginia, US and Frankfurt, Germany, and initially presented at re:Invent 2024, the AWS Transform for Mainframe can help companies transition from ‘big iron’ to cloud native platforms. The transition brings the advantages of scalability and extensibility, factors that limited the possibilities open to decision-makers in organisations whose core computing functions were based on mainframes like the IBM z series.
AWS Transform for Mainframe helps companies analyse COBOL and other code, extract business rules from often dense and opaque code, and split monolithic applications into functional domains. The agentic AI-powered transition promises to turn mainframe applications into software that can be extended, maintained, and further developed by today’s generation of programmers, without them having to learn the low-level languages of a previous era.
Teams in Europe and the USA can have the AWS platform’s AI agents create technical documentation based on the legacy code base, analyse code ready for semi-automated translation into Java, identify dependencies, and plan the complex task of migration.
Of course, IT teams are acquainted with many of the problems associated with replacing mainframes running languages like CICS, VSAM, and COBOL. Part of the issue is that mainframes are incredibly reliable for transaction processing, ensuring that (for instance) movements of finance take place in concord with other systems, with remote and local mainframe instances in strict lockstep. That underlying basis ensures that, effectively, monies leave one account to arrive in another, sub-processes (write to memory, write to disk, change database entries, etc.) surrounding which are dependent on the completion of each step by each party.
While many modern computers outperform mainframes in pure computing power, they can lack the surety that’s demanded by large industries such as aviation, financial market trading, and currency exchange. The AWS Transform for Mainframe platform help companies emulate and test their migrations, proceeding step-wise through processes written in languages that were optimised for speed and security, rather than readability and transparency.
Retiring technical debt is a constant for developers and IT professionals, given that technology evolves: What are today’s cutting-edge tools are tomorrow’s bottlenecks. In line with this ethos, companies wishing to move away from VMware workloads to container-based applications (or those saddled with sudden VMware price hikes) can also investigate AWS Transform for VMware, which uses the same agentic AI model to migrate from virtual machines.
It’s worth noting that IBM continues to manufacture and support mainframe hardware. In 2024, it drew $27 billion in revenue from software-related business, and the company has been fiercely protective of its intellectual property in the mainframe sector, successfully suing LzLabs in 2022 over the latter’s backward engineering of mainframe software. Sectors that rely on the absolute certainty of reliability (or as close as today’s computers can get to certainty) will take some persuading to migrate away from mainframes, and as past litigation suggests, simple software emulation is not an option.
But as maintenance and upkeep of legacy systems becomes increasing impossible, companies will be looking for alternatives to ensure their computing longevity. AWS Transform for Mainframe is an option that many will consider.
(Image source: Excerpt from “The Alchemist” by Celestin Francois Nanteuil, French, 1813 – 1873 and Eugene Isabey, French, 1803 – 1886. Public domain.)
See also: Taiwan builds AI supercomputer with NVIDIA and Foxconn
Want to learn more about cybersecurity and the cloud from industry leaders? Check out Cyber Security & Cloud Expo taking place in Amsterdam, California, and London.
Explore other upcoming enterprise technology events and webinars powered by TechForge here.