IBM Destination Z - Group home

Needed Innovations for the IBM Z Ecosystem

By Destination Z posted Mon December 23, 2019 03:30 PM


“The better you get, the more room you have for improvement.”
If you Google that phrase, you’ll get this article. Why? Well, first, because I’m the one who said it. But also, because we are too used to being told that, as we get better, we have less room for improvement. In other words, the worse your score, the more room you have to improve—as if there’s some theoretical 100% out there that you’re aiming at in life. Of course, you can achieve 100% on tests and assignments in school—but that’s theory, not practice, and, as I’m fond of quoting, “In theory, practice and theory are the same thing; in practice, they’re not.” It’s time to start recognizing that, in the real world, we want more room for improvement, not less.
That’s important, because one of the things that often restrain excellent inventions such as the the IBM Z platform is the idea of achieving perfection, as if there were some sort of ceiling and limit to further progress, rather than a foundation for a next wave of advancement.
The prospect of unresolvable imperfection has often been a bit of a bugbear for us mainframers, as we have sought to create, install, configure and maintain systems with at least five nines of perfection according to unidimensional measurements such as uptime. And, yes, these are important things to achieve and maintain. 
But if our insistence on fixed degrees of perfection has limited progress in some ways, other factors have always found their way into our ecosystem and pushed us forward, even when the visionaries didn’t see it coming. To quote Isaac Asimov, “The most exciting phrase to hear in science, the one that heralds new discoveries, is not 'Eureka!' but 'That's funny...' ”
The Measure of IBM Z
One of the challenges with innovation is that it’s non-linear—just the opposite of established, measurable functionality. So much of our journey on the IBM Z platform has been about achieving numbers, from memory to MIPS to uptime—all commodity, unidimensional measures. And yet, when distributed computers trot out similar numbers, we’re forced to look deeper and compare qualities that resist simple measurement such as architecture, culture, integration and legacy (which has never been a bad thing, despite the tone of voice in which distributed computing sales people may pronounce it).

That solid foundation of innovation, which has been measured every which way, continues to enable newer emerging features and applications, surpassing the distributed computing world in every way but one: humanity.
Oh, the humanity! And me, writing a thesis about the humanity of the mainframe! 
Yet, I would assert, everything that eventually gets recognized as genuine progress has in some way helped advance humanity, not merely obsolete it.
The best of the IBM Z platform has indeed done that—from a business perspective. Yet we still need workstations to front end the IBM Z platform if we want genuinely friendly workspaces.
That reminds me of a funny vignette from the early days of personal computing. I think it may have been on a sitcom. A salesperson was expounding on the virtues of a new PC, and the prospective customer asked, “What colors does it come in?” That was funny because, just like Ford’s Model T, which was available in any color as long as it was black, likewise, early PCs took their cue from business computers in terms of monotonous aesthetic design—at least, until Steve Jobs started gushing fonts and colors and such.
Today, you still don’t get to choose how your mainframe looks, except for the fact that each successive model looks cooler, especially since IBM introduced that “stealth” design. If you want a different-looking mainframe, get a different generation (or manufacturer, if you go far enough back).
High-End Business Computing: Pocket Protectors
Of course, the form of distributed computers has substantially remained function oriented as well, and case color is generally only modifiable by adding stickers or paint. On the other hand, these days, everything is in some way a computer case, with embedded computing in everything from cars to appliances. But, again, the form follows the function as needed by the users, and rarely is a purely arbitrary aesthetic with no function. After all, computers are all about function.
Even so, we in the world of high-end business computing are a sort of pocket of production and productivity that is even more functionally intensive from a fiscal perspective. And if there’s no accounting for taste, we have always preferred accounting. Indeed, the original IT departments often started out as part of accounting. Many of the early users of business computers couldn’t even imagine using computers for anything but calculations. Thank heavens for lateral thinkers such as Rear Admiral Grace Hopper who took risks beyond calculation.
None of which changes the fact that “pretty” is not a typical way of describing a mainframe. And our lovely 3270 “green screen” terminal interfaces are even less attractive—though, as I’m fond of saying, I’ll take a green screen over a blue screen any day.
But if aesthetic innovation is not very evident in the mainframe world, many other kinds are, including those which appeal to the functional sensibilities of its power users. I recently asked a group of new and experienced mainframers on social media what kinds of innovations they’d like to see on the mainframe. Every answer was about helping them perform their jobs more effectively and efficiently.

Of course, if form and function are key aspects of our platform, so is security, which manifests itself in a culture of deeply-ingrained secrecy and caution.
NDA Requirements
You could say that there are two kinds of people in the IBM Z ecosystem: those who don’t know what’s really going on, and those who are under Non-Disclosure Agreement (NDA). This helps protect everyone’s investments in the system of record that runs the world economy, but it can hamper certain types of innovation that rely on shared insights.
Part of that comes from our history. When the world of business computing was sorting itself out in the 1970s, and IBM was pulling ahead of everyone else, the authorities got worried that they might get too powerful, and imposed some strict limitations on IBM. 
This had a number of results, one of which being the software industry, since IBM’s software was no longer bundled with their hardware, so you could spend some of your money on competitors’ solutions instead. That increased innovation through competition.
Another result, however, was a great reticence on the part of IBM to publicly pre-announce anything that might be construed as anti-competitive, as it could result in potential customers waiting for the IBM solution rather than buying a competitor’s offering that was already on the market. So, in order for established IBM customers to be privy to upcoming enhancements, it became standard for them to sign an NDA saying they wouldn’t let the word get out in a way that could harm competition.
Fortunately, there were still other mechanisms to promote innovation on the mainframe, one of the most important being the SHARE requirements process, whereby IBM customers would get together to list and lobby for enhancements to the platform—hardware, OS, middleware, etc.—that met their business needs, and IBM would carefully take account of these in deciding future directions. It has served us all very well, but it’s also always been about enterprise-class business needs, though admittedly including ease of use for the technologists who work for those organizations that have mainframes.
Entry Prize
This brings us to the question I’ve been leading up to: Are there new types of innovation we need on the IBM Z platform? After all, it’s all about that business, and particularly big business, so just how much account can be taken of individual consumer preferences?
The world of physics has striven to find a grand unified theory that consistently explains how things work from the tiniest scale to the largest, but so far in vain. In my opinion, that’s likely at least in part because our focus on such different scales of existence requires very different—perhaps even mutually exclusive—approaches and perspectives.
I mention that as an analogy between consumer computing and enterprise-scale business computing.  And yet… as your colleagues who use z/VM and CMS will tell you, the original PC was under CP on the mainframe. Before hobbyist consumer electronics invaded the office, mainframe shops already had people running VM guests as if they had the whole mainframe to themselves. Having a personal slice of the mainframe has been part of what we do since our first decade. And not only has it been about business, it has also allowed for a fair amount of personalization.
Today, with the cloud increasingly being expected to have the qualities of service that are only consistently available on the IBM Z platform, a great re-integration has begun. IBM has significantly assisted this with their latest hardware and OS announcements, allowing vast numbers of dynamically instantiated environments to effervesce, even under z/OS using containers.
In other words, we just got more room to improve—a lot more. So, while I love all of the great innovations that keep bubbling up through our ecosystem such as Zowe and containers and pervasive encryption and Cinderella footprints, the time is now to look to even more unexpected innovations.
Potential IBM Z Innovations
What? You want me to speculate on what they will be? OK, but keep in mind that this is just an example in one tiny direction, and nothing like what we’ll actually see when the new innovations start arriving faster and faster and the IBM Z platform retakes the world of business computing by storm.

That said, to illustrate my thesis, let me combine my above thoughts into what I see as a potential tipping point of reemergence: consumer IBM Z services with dynamically adjustable qualities of service.
After all, there’s no other platform that has the capacity to measure, plan and adjust qualities of service with such exactitude as the IBM Z platform. Why not slice it very finely, slap an adjustable potentiometer on it, and let people buy cloud services where they get to implicitly negotiate the price and deliverable using a sliding scale of their choosing? Once people realize they’re getting a scientifically-provable cost-benefit service, all the other pretenders won’t have a footprint to stand on.

But what do I know? I’m just waiting for the happy laughter among all my mainframe colleagues when the rest of the world discovers the IBM Z platform and begins a growing chorus of, “That’s funny…”