Skip to main content

The legacy mainframe crisis: Why COBOL is not the issue

The legacy mainframe crisis
(Image credit: Pixabay)

The challenges organizations face with legacy systems are not in fact a result of COBOL, or any other programming language; the language is just a syntax for expressing business rules. A programming language like any other which anyone can learn. 

So why are so many organizations – including the state governments of New Jersey, Kansas, Connecticut and Colorado – experiencing such issues with legacy applications written in COBOL? I argue that the problem is not with COBOL itself, but with the arcane mainframe environment the programmers must inhabit to maintain the programs.

Sweeping issues under the rug

It is common for an emergency to shed light on an otherwise darkened corner of our technical infrastructure, where many suddenly realize that sweeping issues under the rug was bound to catch up to them at some point. This has certainly been the case for multiple US state government departments who found themselves in the news at the beginning of April as their mainframe systems failed to handle a surge in unemployment claims processing – a result of the COVID-19 pandemic. 

New Jersey Governor Phil Murphy even put out a public plea for volunteers who know how to code the decades-old programming language COBOL in which many legacy mainframe applications are written. This was the first time that much of the media became fully aware of the legacy technology issue, but to those of us in the business of IT modernization, the mention of a COBOL skills crisis came as much less of a surprise. 

The pandemic was merely a tipping point for this eventuality which we have seen on the horizon for some time, and the truth is that this challenge poses a threat to more of society than most commentators know.

The legacy mainframe crisis

The legacy mainframe crisis extends far beyond state governments or even just the public sector. As many reading this will acknowledge, it is endemic within corporate America and in established industries globally. For example, financial services providers, insurers, investment and retail banks are some of the most embedded and shackled organizations to this platform. 

An estimated 70% of the world’s commercial transactions are processed by a mainframe application at some point in their cycle, which means US state governments are merely the canaries in the coal mine. These industries in addition to telecoms providers, airlines, manufacturers and global retailers to name but a few, are all suffering from the same acute skills crisis as state governments. 

If this issue is not addressed at its core, we will see far reaching consequences for businesses on a global scale, compounding the economic battering they have taken already from the pandemic.

The solution?

But so far, the one solution purported to tackle this issue has been to attempt to coax ageing COBOL programmers out of retirement to prop up these systems. These are the programmers who built and maintained crucial banking applications in their heyday, and their knowledge of the complex web of inter-dependencies and programming idiosyncrasies left the industry with them. 

Whilst this had helped to tide the US state governments over in the initial crisis, unfortunately this will do nothing to solve the deeper issue. The next time we face a similar emergency, this solution may no longer be an option.

The focus on COBOL

There is a reason why young programmers are not flocking to learn COBOL and mainframe skills in their droves. Navigating the reality of maintaining old programs, with none of the creature-features they are used to in a conventional environment, is at the heart of this crisis. 

The core IT of large corporations has been built up over years like geological layers. The deeper one digs into an organization’s core processes; the more ancient fossils will be found. These are usually the parts integral to supporting the layers above and, over decades, the oldest have become buried beneath layer upon layer of inter-dependency and algorithm.

These challenges are not unique to the mainframe environment, but many more facilities exist to solve them in more modern and familiar computing environments. Facilities that the current generation of programmers see as table stakes in any programming exercise. To put it simply, there are more fundamental issues than the language, and hiring or training more COBOL developers simply papers over the cracks and does nothing to solve the deeper issue at scale.

Mainframe platform

CIOs and other IT leaders have no problem with COBOL itself, but ask them about the mainframe platform and you find that almost three quarters believe the inflexibility of their mainframe limits innovation.

Rather than recruit as many septuagenarians as the HR department can find, perhaps a more strategic answer is to put the programs into an environment where 30-year-olds can work on them as easily as any other application they are used to supporting. Whilst this may have been considered a pipe dream 20 years ago, innovations from the world of cloud services have advanced to the point where truly viable open-source and cloud-based alternatives to the mainframe platform are feasible options for even the most conservative industries. 

The mainframe modernization conundrum has been a long running issue for decades, but with cloud and open source technologies advancing and gaining traction in the enterprise, as well as having the tools now available to move critical applications out of the mainframe environment, freedom is now a reality.

Delaying the inevitable

What I hope the US state government’s experience has shown is that all businesses hold on to legacy mainframe technology are simply delaying a similar and inevitable fate. This can has been kicked down the road for far too long; and the impact of the current crisis has revealed just how fragile this critical part of our infrastructure has become as a result. 

Mainframe modernization will not be an overnight job for any company, but there is a fork in the road, and we have now seen what the path of doing nothing leads to during times of crisis. Cloud providers such as AWS, Google Cloud and Microsoft Azure have all recognized the potential of the public cloud to host mainframe workloads and drive benefits for customers reliant on legacy mainframes, with the support and collaboration of the vast global Open Source world. 

Those organizations reliant on legacy mainframes should see the US state governments’ experience as the warning on the horizon, understand they’re not alone in their journey to modern and sustainable IT, and take action now.

  • Mark Cresswell is the Chief Executive Officer of LzLabs.