Mapping out the next half a century of computing

Mapping out the next half a century of computing

Summary: It was 50 years ago the mainframe kickstarted the era of modern computing - so what should we expect of the next half century?


The IBM mainframe is 50 years old this week. The machine now celebrating its half century wasn't the first business computer of course — that was the LEO, which was built for the UK catering and cakes company Lyons way back in 1951. But unlike the LEOs, which were finally decommissioned in the early 1980s, the mainframe powers on. According to one recent survey, more than half of enterprise applications still call on a mainframe to complete their transactions, and CIOs say their mainframe workloads continue to increase.

There are plenty of reasons for that — the mainframe has turned out to be a solid and extremely reliable piece of enterprise computing infrastructure. Mainframes still sit at the back of most essential business systems, power ATM transactions, and as much as 80 percent of corporate data still sits on big iron. Few organisations wish to go through the pain of moving those essential systems and data away from mainframes unless they absolutely have to.

In the age of the cloud, some might consider the mainframe to be a dinosaur — but it's a dinosaur that you disturb at your peril.

Indeed, it's the humans, not the technology, that have proved to be the bigger problem for mainframe-using companies; the baby boomer engineers who programmed the mainframes are retiring and those skills are proving hard to replace.

It's fair to say that the arrival of the mainframe was also the birth of the enterprise tech industry as we know it. The widespread adoption of computing power was first used to automate accounting and then went on to create new ways of doing business.

A computer that took up most of a room was undoubtedly an impressive sight, but since then our computers have got far smaller while their impact on business and society has become far greater. It's unlikely that those early engineers ever imagined the vast computing power that we harness today as part of our everyday lives without even thinking about it.

Over the next 50 years, that processing power will be even less visible and even more powerful. The era of big iron is already giving way to the era of all-but invisible, pervasive, ambient computing, (the underlying processing power provided by datacentre that are out of sight and out of mind).

We're already seeing the start of that in the steps towards the internet of things and the gradual improvements in digital personal assistants — or IBM's own Watson cognitive computing project. All point to a future where computing is embedded in nearly everything — and probably even in ourselves.

None of this is certain of course, or without its own new problems, even if we're not overwhelmed by grey goo, exterminated by an unhinged AI or succumb to the singularity. Working out how to deal with the challenges created by technology, from automation-related unemployment to economic inequality, will be vitally important.

And so what will be the long-term impact of the mainframe era? It's tempting to echo the response of Chinese premier Zhou Enlai when asked about the impact of the French Revolution: it's way too early to tell.

ZDNet's Monday Morning Opener is our opening salvo for the week in tech. As a global site, this editorial publishes on Monday at 8am AEST in Sydney, Australia, which is 6pm Eastern Time on Sunday in the US. It is written by a member of ZDNet's global editorial board, which is comprised of our lead editors across Asia, Australia, Europe, and the US.

Previously on Monday Morning Opener

Topics: Servers, Big Data, Cloud, Enterprise Software

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
  • Yes we did!!!!!

    ". It's unlikely that those early engineers ever imagined the vast computing power that we harness today as part of our everyday lives without even thinking about it."

    The first "main frame" I programed was an IBM 650 in 1959. Yes, we did understand how much they could do in the future, but were laughed at by the ignorant managers.
    • Multix

      As an obscure contributor to Multix, I agree with you, gertruded. We saw the potential, managers didn't see the $$$ attached to it.
  • MainFrames are the 18 wheelers or railroads of computing.

    Yes, you can haul 10,000 tons of trash using 100,000 Prius, but is it effecient?
  • The mainframes have grown smaller and faster over the years, and we may end

    up having to alter them a little - as in Beowolf clusters, but I see them more as being clusters of small mainframe computers. We'll also see a lot more of fringe activities move out from the centralised computer centres as well. This is simply because there are many compelling business and management reasons to keep critical data in a centralised environment and processed in the one location.

    When you have multiple processing areas doing only part of the data there are opportunities for data to get lost, duplicated, or mismatched - think about a certain Mars probe where the data process of its design was done in multiple locations and ended up with mismatches causing issues.

    However, every business has a lot of computer work and data that can be done out in the organisation's fringe. A very good example of the trend for the future, I see, is that of the major retail stores where the minute to minute store sales and inventory is all processed in a centralised store computer (not at each point of sale) and the end of day summary is sent into head office to allow management to have a better grasp of the overall picture and to also create an off-site back-up of the store status at the end of the day. The mainframe is needed to handle all that data from the stores and work it over.
    Deadly Ernest
  • Predicting the future is a perilous business

    Humans are singularly bad at predicting the future accurately. We cannot even predict the weather in a reliable way. Who predicted 50 years ago that people would carry more computing power in their shirt pocket than was available to governments and big corporations? In 50 years from now, quantum computers may be as common as cell phones are today. That will engender a fundamental change in society, since it will no longer be possible to keep information secret if it can be intercepted by a third party.
  • Ethics Issues Coming!

    Major ethically issues are on the way along with new emerging technology. We already struggle with the issue of computers and automation replacing people in the work force. But what happens when we start adding technology internally to our bodies? What will that mean for those that are wealthy and are able to physically improve their capabilities to a point where the rest of us can't compete no matter how hard we try? It is already an issue, but it will be much worse. This new technology is exciting and frightening at the same time.
    • Been happening for decades - think pacemaker, bionic ear,

      bionic eye, prosthesis, peg leg etc.
      Deadly Ernest
    • Let us not be timid

      And let us not be timid about it...
  • Thank you Steve ....

    For the quiet acknowledgement of Leo's place in raising the flag for business computing applications. It's great to see an acknowledgement in a forum like this which is normally almost totally North American centric
  • Frankly, programming is becoming WORSE, so next 50 years are bad

    I just tried to do a simple thing in Windows 7; format a DVD. It wouldn't work right. I had to do it, because the new Win7 'system repair disc' cannot detect, reformat, and then install its files. That was no problem on XP, but then XP suffers from not being able to read/write DVD data disks at all, even and especially, those prepared in Windows 7. Which, doesn't work well. So now I have to shut down my Windows 7 computer, and boot in Linux, to get the DVD to properly format so I can make a Win7 System Repair Disc!

    That little anecdote is classic and symptomatic of the problem we all have today. TOO MUCH INCOMPETENT PROGRAMMING and TOO MUCH INCOMPATIBILITY. So, the trend now is to have our PCs become ever more like thin clients, and the mainframe function lives on, as if modern, 'in the cloud'. That's where it will have to do, since programmers today are so slipshod.

    Every piece of software I have is better in its OLDER version, than new. Acrobat 6 is far better than 9 or XI, for example. MS Office 2002 is far better than 2007 et seq. Windows 98 is better in some ways than XP, and XP is better than anything after it. So I have to keep whole machines running, since the later ones aren't compatible: so I have 10 DOS machines (for true DOS is still better in many ways), five Win98, five XP plus four unused licenses for future installations, six Win7, one Vista. And, some Linux distros (which brick after support dies for them, so you have to keep changing the distro), NONE of them on hard drives (lest you brick your machine, not merely the distro).

    This is silly. One person shouldn't have to maintain so many machines in order to protect and still run, old files.

    So the future, if like the present, is quite bad.
    • Update on the Windows 7 inability to create a repair disc

      The first time I tried to create it, I used a disc which had files on it. The Repair refused to function, because I had files on the DVD. So, I reformatted the DVD using Windows' own Quick Format, since that's what Format is supposed to do, erase files. Of course that is glitchy too, for when it tells you the format is complete, you still can't eject, for it says it must 'prepare' the disc for ejection. You wait and wait. But I did, it ejected, and I closed the drive.

      So back to System Repair, which kept on insisting that files were still there, even though the disc is blank, in Properties. So, I removed the disc, waited a few seconds, backed out of Repair and then inserted the disc, had to TYPE IN 'Create System Repair' to access the program in Win7 (since I can't create a shortcut), and guess what -- got the same refusal, and same error.

      BUT when I went to another machine, inserted another DVD that had files on it and then (right clicked) on ERASE THIS DISC, then and only then did System Repair work. Is this bad programming, or what?

      And it took all that time for Linux to separately blank a disc (in lieu of quick blank).

      Again, this is indicative of the state of computing, now. Incompetence everywhere.
      • Is this a rewriteable DVD?

        If not...
  • Re - the Chinese anecdote...

    For what it's worth, though it makes for an amusing quote, the Chinese premier was likely speaking in 1972 of the recent French unrest in 1968, not 1789.
    Too-Tired Techie
  • Let us not be timid

    I quote "...and probably even in ourselves.". Why not say hit the nail on its head and say "... and even in ourselves"? Why probably? Other than "people issues" again, the technology and will be there to turn us into cyborgs.

    If the old mainframe is here to stay, what name do you give to Watson? A server, a mainframe? Maybe we should rebase the mainframe such that we speak of it we mean a massively parallel computing platform.
  • 50 years? how about 33.

    Worked on mainframes and the chance of us guys foreseeing a computer in your pocket connected with everything else on the planet missed us by a mile.

    How about Arthur C. Clarke's Space Odyssey 2001 -- it was easy to draw a straight line from 1968 with the US roaring ahead with successes in getting to the moon to having Pan Am fly us there 33 years later and having an American in charge. In that case, noone would have predicted that our country would totally abandon the efforts and leave it to China(?) to pick up the ball 40 years later.

    One thing that is the same as in the first mainframes -- programming. We still enter a line of characters that is interpreted by a program into something that drives a bunch of electronics. Those lines of C# could have been entered on an 80 character Hollerith card and fed in via a card reader. Before we can get away from "mainframe head", we have to think up new ways to communicate with them.