Death of the Mainframe

Much ado has been made of the talent gap in the mainframe space. Enterprise mainframe clients have been metaphorically wringing their hands as the aging-out of their mainframe-savvy workforce metaphorically wrings their necks.

A friend of mine, now an influential figure at a major mainframe software vendor, used to be my teammate for many years at IBM. During our long mutual tenure there, we saw the mainframe “die” at least three separate times, only to be resurrected once the world realized they couldn’t really replace it with the flavor of the month. Almost ten years ago, I said to him, “Wouldn’t it be funny if what actually killed the mainframe was that there just wasn’t anyone left to work on it?”

Now, the lack of a pipeline of qualified mainframe workers will soon put great pressure on mainframe clients and software vendors as current employees trickle, then stream out of the workforce forever.

As this crisis insidiously infects the medium- and long-term prospects of the mainframe platform in a way that viruses never do, much has also been suggested and written about what should be done to stem the tide before it becomes the tsunami that wipes out the platform. For all that worry, and all that talk, there has been shockingly little action.

Even worse, the suggestions made and the approaches used reflect, in my view, a catastrophic misunderstanding about what motivates the young people so desperately needed to fill this mainframe talent gap.

The tactics taken have been ineffective, and some more novel approaches I suggested myself have been ignored.

Conclusion: It is now too late. The mainframe as we know it is dead. Period.

Well, it’s not dead if dead means that mainframes will stop being used and disappear from the planet. So let me explain exactly what dead means to me.

The mainframe is still a viable and important platform, and for many of the world’s most important business and government entities, it is more than the preferred option for their most critical workloads. It’s the only one. It will continue to be needed, and will certainly still be around for years to come.

So how can mainframes be both dead and very much alive at the same time?

Before I answer that, let’s review what’s currently being done to address the situation, just in case some increased focus and resources are magically allocated as some form of Hail Mary game-saver.

Generally, the problem can be simplified into three main components, and all three deal solely with people—the next generation of mainframe workers that will be crucially needed tomorrow, if not today.

  1. The platform must be attractive and “fun” to work on—it must have at least a minimal “cool factor.”
  2. It must have a clear and foreseeable longevity sufficient to sustain an entire career.
  3. To those who might choose it and are capable, it must be an option that is better than, or as good as, other available career options.

Ostensibly, every young person entering the tech workforce in any capacity makes their career decision based on these three things.

So what kind of efforts are being made to attract new employees to the platform? Do they showcase the mainframe career path as a cool, sustainable, and superior choice?

One increasingly common strategy is the characterization of the mainframe as being important. As the mainframe community belatedly comes to grips with the fact that our beloved platform is not as attractive to young people as it is to us, the sales pitch has shifted from how awesome the technology is to how important it is.

This is a small step in the right direction, but unfortunately, most young people are not particularly moved by this prospect, and for the ones who are, this message is completely undercut by the woefully below-market compensation structure in this space, as I pointed out earlier this year

There are still far too many misguided mainframers out there who think their unbridled enthusiasm for the platform will somehow transfer by osmosis into the psyches of young graduates. But take a look at the kinds of discussions involving the mainframe that take place on any public forum and it’s almost comical to see how oblivious the posters are to the fact that the very comments they feel are so compelling are sending the next generation running and screaming in the opposite direction.

Compounding the problem are those who mildly chastise anyone who doesn’t see things their way. As I suggested in this 2022 article,mainframe experts tend to view themselves as part of an elite and exclusive club while failing to see that no one is interested in seeking membership.

An effort gaining some traction are the grassroots movements aimed at universities and businesses to form agreements and partnerships to get mainframe topics reintroduced into college curricula. These efforts are being met with resistance, and whatever successes they achieve (and there have been some) are far too few and far too small to make a dent in the personnel shortfall.

Notably, a sizable portion of these initiatives are taking place at small midwestern schools and southern HBCUs. This is likely by design. To those students, a mainframe career might be seen as a big opportunity. In Silicon Valley, it is not.

Some last-ditch efforts are underway to create technology to mask the perceived unfriendliness of the mainframe. Under this view, it seems mainframe companies hope to be able to run their mainframes without any employees who understand mainframe technology. These technologies are still a long way from completely abstracting the underlying mainframe infrastructures, and even if they did, people who truly understand how mainframes work would still be needed, even if the number was fewer than today. These “solutions” also circumvent the primary need, which is, and will continue to be, people.

There is yet another problem. Taken together, all these efforts, and the problems they are meant to solve, suggest the mainframe is not long for this world. Imagine a young, bright person with plenty of good options available. Even if they buy into the “importance” argument and don’t care whether their job is cool or not, and even if they are willing to accept less money for doing the same kind of work that they could earn on other platforms, is their choice to go all in on mainframes going to continue to be a viable one at the tail end of their careers? From the way these issues are framed, and even based on the solutions being attempted, the conclusion would almost have to be no, and this perception will likely become a self-fulfilling prophecy. The belief that there is no future in mainframe careers will cause there to be no future in mainframe careers.

Eventually, for all these reasons, and the unsuccessful attempts to change things, the mainframe career path will die. We will soon fall well below critical mass on the mainframe talent needed to sustain operations at existing mainframe installations throughout the world.

So what then? What does the post-mainframe world look like? Here is my prediction.

At least in the United States, the mainframe career path will all but cease to exist, even as mainframe usage continues and, in some sense, thrives.

Eventually, with no available workers, nearly every mainframe shop will outsource their operations. There will be a dozen or so major players, mostly in India where mainframe education is still widespread, who will remotely serve the needs of mainframe clients. First some, and then all of most enterprise entities’ operations will be outsourced and handled completely by one of these companies, each in a separate “partition” of the outsourcer’s business. Staffing at these businesses will be consolidated with only limited dedicated staff devoted to each of their clients. Any remaining staff at the mainframe client companies will work in partnership with the outsourcing company they hire, until they are no longer needed.

Laws will need to be changed. For example, regulatory requirements for the physical location of data and/or people will have to be modified to match (future) current practice, and that will happen because there will be no other choice.

There will be breaches and scandals involving security lapses by the outsourcing companies. They will be forgiven, because there will be no other choice.

And there will be privacy concerns and political fallout. That will all get resolved too, because there will be no other choice.

Some very large companies have already made this move, and more will follow.

In the end, mainframes will survive. But if your career is in front of you and you are evaluating your options, your view should be this:

The mainframe is dead.

David Herlich is a Senior Mainframe Software Developer who most recently worked at DataKinetics—the global leader in Data Performance and Optimization. He started his career at IBM, where he spent over 25 years in the Db2 for z/OS development organization as a developer of the product and as an evangelist for the platform. He earned his degree in Computer Science and Mathematics from the University of California, Davis, along with a double-minor in Communications and Psychology.