Mainframes Without Experts?

The short blog on the long goodbye

Digitalization is roaring through the field of IT like a hurricane, upending everything in its wake. Business processes, business models, working relationships, even society at large – they are all succumbing to the upheaval. Only classical large computing systems – the insuperable mainframes – are weathering the storm. Their ranks may be thinning, but mainframes are still mainstays at most large companies, insurance firms, and banks. Mainframes are secure and reliable, making them popular systems for transaction-heavy applications, but the main reason they’ve stuck around so long is because of how expensive it would be to replace them.

Nonetheless, they are well past their expiration date. It’s now been over 15 years since the end of the mainframe was predicted. It was said to be unavoidable, unstoppable, and far preferable to deal with it today rather than wait until tomorrow. That was quite a few tomorrows ago. Technical wizardry made it possible to delay the end again and again – for example, by introducing object-oriented programming and Linux to the mainframe. Nonetheless, we always knew we would one day be forced to bid farewell to this once revolutionary technology. Digitalization’s prevailing technologies and processes make that abundantly clear. With a bit of acrobatic finesse, you might be able to work with AI, big data, or cloud computing, but once you hit apps and DevOps, the party is over. These modern technologies simply don’t have a place in the old mainframe landscape, and any additional tricks would just make systems more complex and more expensive. At some point, it is simply time to accept the fact that mainframe technology belongs to the (rather distant) past, and to finally bite the bullet and do what needed to be done 15 years ago – or ten at the very latest: develop strategies to replace the mainframes.

And this long goodbye from mainstream technology is not only causing technical problems. Since at least the turn of the millennium, we have been warned there would soon be fewer mainframe experts in the field. Slowly but surely, they are leaving the workforce and heading for retirement. There is no reinforcement in sight, because what young, ambitious IT fan is willing to devote a career to things like PL/1? They would much rather become blockchain experts or work in media. The increase in Germany’s retirement age from 65 to 67 has granted the mainframes a bit of a grace period – and I don’t doubt that desperate mainframe operators were one of the first things on politicians’ mind when that piece of legislation was passed.

Nevertheless, that relentlessly ticking clock will soon reach the day when there are simply not enough experts around to operate these systems. And because legislation that requires mainframe experts to work until the end of their lives and even beyond is bound to encounter both legal and biological hurdles, it is finally time to look squarely at the solution we’ve been avoiding for so long: It’s time to switch off the mainframe.