IBM Destination Z - Group home

Ten Thousand Mainframes Today: A Million Tomorrow

By Destination Z posted Mon December 23, 2019 03:41 PM


It’s unofficial! I’ve heard it from everyone who’s heard it from someone else who heard about it! Oh, except for those who have heard a different number.

Yes, validating how many IBM System/360-descended mainframes there are in the world is either impossible, obvious or confidentially certain, depending on whether or not who you talk to has an NDA so they can’t tell you, works for IBM so they won’t tell you or has a pretty good idea and wants to tell you.

Guess which group I’m in.

Actually, it’s rather freeing to not know for certain. It’s a bit like a special mainframe version of Heisenberg’s Uncertainty Principle. The good news is, I know I’m in the right order of magnitude, because there are a few thousand z/OS shops in the world, slightly fewer z/VSE shops, an important handful of z/TPF shops and a growing number of z/Linux shops. I’m sure there must even be some z/VM shops that don’t have any of the above operating systems. Not to mention many pathological configurations of maximally recursive z/VMs running under other z/VMs, all hiding in IBM’s development and testing facilities.

If the average mainframe shop has about 1.7 mainframes, the normal mainframe shop has 1.1 mainframes and the biggest mainframe shops have a double-digit (or in a few cases triple-digit) number of mainframes, that doesn’t account for all the “undeclared” mainframes IBM has, just waiting to become the worldwide cloud.

Ah, to heck with it: I’m saying ten thousand, and you either can’t or won’t disprove me. Trust me: I’ve tried to get someone (anyone?) to do so.

But it ain’t a million…yet. Unless IBM has a terracotta army of buried mainframes just waiting to rise up and face the challenge of the demise of every other production platform. But I don’t see that happening…

What I do see happening is the discovery of the citius, altius and fortius of computing by everyone who has outgrown, or never had, a bias against a legacy of excellence. And I’m really hoping that IBM is tooled up and ready to start mass production with serious momentum when the right time forces them to new heights.

Stepping Toward Excellence 

In my opinion, excellence is as close to truth as Heisenberg’s Uncertainty Principle will normally bend to allow. And the truth will out.

But wait: there’s UNIX, Linux, Windows, Android, Apple and something else that’s on the tip of my tongue. And you can Internet, JAVA, Hadoop and Dante who knows what else with them. Why would you bother moving to some legacy system instead?

Yes, I said “moving to some legacy system.” Did you catch that? I just Jiu Jitsued every argument ever made against legacy computing (thanks for the paradigm, Holden!). I used their most cynical and dismissive argument against them—because everything is now legacy, so it’s time to choose what works. After all, the mainframe was in its teens or younger when these other systems emerged, and over three decades later they still haven’t caught up with what the mainframe was doing before it was a decade old.

I know—bad management decisions, marketing, in-flight magazines, etc. It’s like the evil twin of the adoption curve: the disadoption curve. Once something good has established a space, the parasites dig in. But they can only exist to the capacity of the context they’re sucking. So even the commodity consumer devices that want you to think they’re production computers need something better to exist in order to fake being as good. If all we had were cheap consumer toys, no one would believe quality computing existed.

But, of course, it does. That’s why your bank account, insurance, taxes and every other system of record keeps working. And that’s why every other platform needs that to keep happening. But eventually, people are going to start asking, “if that’s really happening, why isn’t it happening for me and my organization?”

The tipping point will come when a critical mass of people stop unconsciously inserting “except for the mainframe” every time they ask, “what other computing platform can give me what my current one isn’t living up to?”

What’s Stopping Us? 

So what’s stopping us from already being there? Oh yeah—we are. To quote Pogo, “We have met the enemy, and he is us.” Every one of us who have kept the world economy humming with a smidgeon of excellence and a dollop of attitude. We learned the most important word in our vocabulary when we were two years old, and we’ve found more and more convoluted ways to employ it ever since: “No.”

So, IBM, I know you’re still smarting from accidentally making a bunch of mainframe documentation unavailable online and a few other “oops” moments, but listen up: This is your density! Let it also be your destiny. The world is waiting for a platform that actually works, and they’ve heard rumors of this glorious paradigm every time they took something for granted that ran on a computer that was practically invisible because it worked.

This was your warning shot. My next salvo will be at SHARE in St. Louis, where I’ll be giving the zNextGen keynote presentation on this subject on Monday, August 13 at 3:15pm in room 242. We’ll discuss the geometric implications of this impending tipping point, as our supercooled potential to be the platform of reference for all serious business computing crystallizes.

See you there!