IBM Destination Z - Group home

What does the future of the mainframe look like?

By Trevor Eddolls posted Mon March 16, 2020 05:25 AM

mainframe future

What does the future of the mainframe really look like? This is a question that came up in a discussion with colleagues the other day. It followed a semi-humorous discussion where we joked/hoped that mainframes would last until we both retired! But thinking seriously about it, what is the mainframe’s future?


It occurred to use that, in reality, the mainframe definitely has a future. The worst-case scenario is that the number of mainframe sites decreases, and the majority of those sites continue to run the software that they’ve always run. It’s the situation where there aren’t enough staff or enough time for anything new to be trialled and adopted. There’s no budget for anything other than what was in last year’s budget, and there’s no appetite to risk trying anything that might be better. And, quite likely, the people who really understood what was going on under the bonnet have all retired and the new people can do little more than keep things going.


The alternative scenario is where the mainframe plays a major part in the success of business. Now the big issue with this second scenario is not so much that the opportunities aren’t there – because they definitely are – no, the problem is that the mainframe experts aren’t sitting round the table with the other IT guys (those Linux, Windows, and cloud people) and selling the benefits of the mainframe to them. And they are not then, as a group, sitting down with management and showing how the mainframe can be part of the solution to the question: “how can we ensure we are not only still in business in five or 10 years, but also making lots of money?” For any company, if you want to grow, you need infrastructure that will scale up successfully and won’t cost a fortune in order to do so. The mainframe can be a substantial part of that infrastructure solution.


So, what will that future mainframe look like? It will be amazingly secure. Every document will be encrypted, not only while it’s stored on the mainframe, but also when it’s sent from the mainframe to a recipient. There will be no way that anyone other than the intended recipient will be able to read what the document says. In that respect, it will be hacker proof. I imagine that there will come a time in the near future when insurance companies will demand that level of security in order to keep the premium at an acceptable level.


Mainframes are currently very good at keeping out hackers provided the security software, like RACF is properly installed. But what is there to stop a trusted employee from wreaking havoc on the mainframe either intentionally or accidentally. At the moment, most sites rely on SMF reports run every night to suggest a breach. This leads to organizations taking 206 days to discover a breach and a further 73 days to respond and recover. That’s according the IBM/Ponemon report in 2019. They also suggested that the average breach costs $4.3 Million. And, they informed us, that nearly 50 percent of all breaches were caused by human error or system glitches. Clearly, this can’t go on, and our mainframe of the future will be running some kind of File Integrity Monitoring (FIM) software that will be able to identify changes and alert the appropriate staff – and do all that very very quickly. And, at the same time, be able to identify and potential false positives and prevent alerts for them being sent.


I think the word hybrid will be commonplace. People will be driving hybrid cars, fashionable pet dogs will be hybrids (Labradoodles, Peekapoos, etc), and so will mainframe computing. It makes sense to move some workloads off the mainframe. For example, you could move from tape backups, restores, archives, etc to cloud storage, which would be much cheaper. Big Data analytics works better in the cloud – apparently. Users may consider moving other workloads to distributed platforms, eg security information and event management (SIEM), again saving on workload on the mainframe. And integration with non-mainframe staff would help with this.


And, much like now, mainframes would be integrated into apps running on phones and tablets. Information about mainframe applications would be available on those phones and tablets. And reports from FIM and SIEM, and other monitoring software would be available in encrypted format to authorized staff. The mainframe would be treated as just another server used by a company, but one that comes with a number of advantages over many of the others.


The mainframe of the future would look like any other server. We have rack-mounted mainframes available now. I’m sure this is a trend that will continue. And the screens will look much like the screens associated with any other operating system. There will be a layer of software between the operating system and the main subsystems (CICS, IMS, Db2) and the end user’s screen that will make it possible to control what is happening on the mainframe without using arcane commands or any of the things that current systems programmers can use so quickly and efficiently at the moment. It may take slightly longer, but it will be intuitive and usable by IT specialists no matter what their background.


And with this renaissance in the fortune of the mainframe, we will see a growth in development on it. New applications will make use of what the mainframe does well. And that’s where all the current work on DevOps or DevSecOps comes in. No more waterfall methodologies. New software will be developed that exactly matches the needs of its users at that time. And updates will be delivered at regular and short intervals. Care will be taken that anything that doesn’t work will be backed out without the end users noticing the change. And the DevOps methodologies in use will work on Z, Windows, Linux, and Android in exactly the same way. A developer trained on one platform will be able to work on every other.


Open source software will rule the land – although proprietary versions will be available so that users will be confident that a vendor has used their expertise to check it out in a mainframe environment. In fact, mainframe software vendors will be releasing versions of their software into the open source community. Things like Github and Jenkins (or whatever comes next) will be as familiar to mainframers as anyone else.


Management will understand words like secure and hybrid, and will use them as frequently as they do the phrase, value for money. This will be the future of mainframes. And my friends and I need never have to retire!