Shortage of mainframe skills looms but companies remain in denial

Shortage of mainframe skills looms but companies remain in denial

Summary: Like the rabbit caught the a car's headlights the IT industry does not know which way to jump to sort skills shortage.


Despite a looming mainframe skills shortage few companies have worked out how they are going to deal with the issue.

According to the survey by Compuware, "the impending retirement of the mainframe workforce continues to be a major cause of concern for CIOs", but few companies have created a formal plan to address these risks. When compared to Compuware’s 2011 survey, the last time it was done, the situation has hardly changed.

The issue is a simple one: many of the people with core mainframe skills are now approaching retirement age and it seems that most companies are waiting for somebody — they do not know who — to step in and solve the issue for them. It is one they seem incapable of solving themselves.

Some 66 percent of CIOs fear that the impending retirement of the mainframe workforce will hurt their business by reducing their ability to support legacy applications. Many core applications in areas like banking, finance and even telecommunications still run on mainframe systems.

The loss of people with mainframe development skills is seen as a primary concern by 61 percent of the CIOs surveyed. On a slightly better note, while some 40 percent of CIOs admitted that they had no formal plans in place to address the mainframe developer/operator shortage, this was slightly better than the 46 percent who back in 2011 said they had no formal plan.

Compuware's global survey of IT skills in the mainframe workforce covers 350 companies around the world.

Further Reading:

Topics: Big Data, Banking, Data Management


Colin Barker is based in London and is Senior Reporter for ZDNet. He has been writing about the IT business for some 30-plus years. He still enjoys it.

Kick off your day with ZDNet's daily email newsletter. It's the freshest tech news and opinion, served hot. Get it.


Log in or register to join the discussion
  • All mainframe stuff should be replaced with modern systems.

    Really there's not anything that couldn't be replaced not that technical challenges don't exist. Let it die. The rigor needed to get the job done is needed in the modern systems and we don't need to perpetuate the legacy forever nor should we. Should we keep the steam engine just because its there?

    Rather than trying to replace mainframes in the past modern systems took some different approaches to gain separation. They should now add any necessary gap features while also maintaining the new approach and go directly after mainframes to replace them.

    Mainframe free in 2020-2025.
    • A mainfram still has better uptime than Windows.

      And a smaller footprint for the work performed.

      A single mainframe can fill in for hundreds of the "replacements".
    •'re an idiot

      Had it ever entered you little pea-brain that maybe some mainframe operators simply do not have the funds to replace their mainframes? Obviously not.

      One can not simply discard millions of dollars of equipment. Especially if it is still productive.

      Back under your bridge.
      • already done it so take you offensive diatribe and shove it.

        You should be fired.
      • Rude post

        IT_Fella, your post is rude and classless. Furthermore, while a lack of funds may be an issue for some, lack of expertise and the lack of what is necessarily an intimate knowledge of the legacy application code base may be the real reason some organizations can't easily move off of their big iron.
      • money?

        Do you know how expensive mainframes are to own, license and operate? Yes, there would be conversion costs, but it's not like the mainframe is free just because you already own it.
      • Can I be childish for a moment?

        Perhaps it takes one to know one? It isn't the equipment. It's the custom software that only runs on the mainframe.
  • No easy fix, but...

    If an organization opts to replace mainframes with server farms it should be approached incrementally, identifying points where existing software systems can be severed into pieces that can then be relinked with loose coupling to permit ongoing evolution down the road.

    This would have been the ideal strategy beginning back in the 1990s, but much time and money has been squandered by persuing body-shop vendor promises of "big bang" conversions to another mono-cultural technology. So many years and millions of dollars later the recent past is littered with failures that left things back at Square One. At best a function or two may have been moved to public-facing Web applications, at worst the vendor merely reimplemented existing parts of these systems that had already been moved to the Web a decade earlier.

    The vendors bear a large share of the responsibility, but the majority belongs on middle management hired post-1980s with no roots in technology who insisted on trying to drive the train. This was even worse in organizations that "consolidated" operations and piled everything high and then added layers of parasitic bureaucracy. There the technology folks were rendered completely voiceless in the decision making process and hung out to dry when ill-conceived projects went off the rails.
    • MIS-management in most companies has never thought ahead

      or, perhaps, been allowed to fund their thinking ahead by higher-up executives.

      The time to think about replacing mainframe applications was when server farms first became capable of doing the same jobs. The time to retire mainframe programmers would have been AFTER the mainframe itself was obsolete, not forcing them to leave in their 50s to "save money" on younger, cheaper people. And nobody ever thought that investing in cross-training mainframe people, rather than letting them go, would give them several years to a decade of service by people who knew BOTH systems, rather than trying to teach the young folks enough to be competent in systems they would only be using for a few years.

      One example: in the early 1970s, two firms on opposite coasts in the same business were bought by the same holding company. For TEN YEARS they continued to operate separately (except for mutually advantageous coordination of services), with incompatible proprietary software running on incompatible operating systems (one was DOS-360, the other was OS-360 MFT). They wasted those ten years NOT EVEN STUDYING how to combine their systems, much less making a plan to do so, even though it was nearly inevitable that they would. Then when the holding company decided to merge their operations, the ENTIRE JOB had to be done in a year. Actually it took almost two years, but the second year was "tested in production" at the expense of the customers and the company bank accounts, with the potential for many transaction crashes and maybe a system crash every night.

      Management only thinks of upfront cost (not as an investment but as a burden) and short term return, not creating stable organizations and systems that will be efficient over many years.
    • Mainframe screen scraping using Attachmate

      A lot of companies are not doing any new programming on the Mainframes. Many of them are tying them into desktops using Attachmate or similar screen scraping software. Not ideally the best solution, but it at least keeps it going. I know of one particular project where they used a desktop client server application to fill data and functionality gaps of the mainframe interacting with it through Attachmate. Over the years the desktop CS application ended up having more functionality and data over the mainframe that they finally moved all the data and functionality over to the desktop CS application.
  • Oh, for crying out loud.

    Here's what's going to happen: manager to mainframe dude: we're doing a reorg and adding Bill to your team. He's a sharp kid. Get him up to speed and teach him everything you know.
    • That's what should have happened 15 years ago

      But apparently training people when there's a shortage of qualified personnel is called "stealing from the stockholders". Much better to get Big Bad Government to do the training instead.
      John L. Ries
      • It did happen fifteen years ago

        It's happening right now. And training existing personnel is called smart business and will raise your profitability which is what most stock holders like. That's a second post from you showing zero critical thinking skills.
        • You're sarcasm detector isn't working

          I actually agree with you on this point. What this really is is an appeal for corporate welfare, either in the form of higher H1B quotas, or subsidized training by universities. Neither is necessary or appropriate.
          John L. Ries
    • Re: Oh for crying out loud

      I was Bill at someplace I used to work. Learned our mainframe fairly well within a few years. But there's stuff that the old guy new from doing it for 30 years that honestly takes a long time to pick up.
      • That's true with everything,

        Including any system you may replace your mainframe with.
      • Solved by IBM

        IBM is a huge services company. Once Bill knows 80%, rely on a good services contract with IBM for the other 20%. Eventually Bill will know 90% and you'll need IBM less. Of course I recommend that Bill is really Bill, Bob, Butch, Barney, Barry and Brad. If Bill meets with an unfortunate demise or leaves for a competitor, you have a pool of talent on which to rely.
        • You forgot Barbara and Betty!

          And in a profession practically started by Grace!
  • Just another 'the sky is falling' article

    I've been hearing about the 'shortage' of mainframe (or COBOL) skills for years now. Nothing has changed. If there really was a problem, companies would be doing something about it such as hiring and training new people. The other thing that would be happening is that rates and salaries for mainframe skills would be increasing. I was contacted a couple years ago about a mainframe programming job. They would go 'as high as $35/hour'. That is what I was making 30 years ago.
  • Mainframe shortage

    I suspect many companies will TRY to outsource mainframe programming and, out of sight, out of mind. The people assigned to actually work on these dinosaurs will be hitting the online forums and Googling their little asses off trying to learn JCL or File-Aid.

    Sad, really...