Reg Harbeck’s interview of Connor Krukosky

There’s nothing more mainframe than a good assembler.

Being a good disassembler, in contrast, might be called “meta-mainframe” – to be able to dig into the details and pull out the patterns like Little Jack Horner pulling out a plum.

Now, a good mainframer may be tempted to dis disassembling, or even to suggest that those who claim to have done it may have been dissembling. We kind of like to “build forwards” in the world of top-end business computing, and while reading a good core dump may give a semblance of turning memory into actions, we are generally prone to comply with compilers instead.

Be that as it may, whether being thankful for the contents of memory or making sure the chips are up, digging into the details is a pastime reserved for those who genuinely like to understand how everything connects – and disconnects.

Which is exactly what Connor Krukosky had to do upon taking delivery of his used IBM z890 mainframe.

Fortunately, because it was no longer owned by IBM, he had the right to ship and dip his chips as he saw fit, which was important in order to get it to fit onto affordable transportation and then squeeze into his parents’ basement.

In addition to an inquiring mind and an interest in exacting detail, there was another requirement for Connor to survive the quest: patience.

So, he carefully took apart his mainframe and tracked the process with sufficient precision to reassemble it once he’d squeezed everything under the floorboards. Talk about the opposite of a raised floor!

This is where kudos must go to the good folks at IBM who designed the mainframe to survive just about anything. Who’d have guessed that would include being taken apart, relocated and put back together by a teenager? But Connor did exactly that, carefully putting each piece back together and getting everything so exactly in place that the mainframe actually worked when he turned it on.

Well, except for the software, which he didn’t have yet…

Because, of course, this is one of the biggest challenges of running a mainframe environment: The hardware, with all its excellence in leading-edge functionality and resilience, is just the beginning.

After all, when you first acquire a personal computer, tablet or smart phone, you just take for granted that the operating system and productivity apps are included, or at the very least easy to download and install.

But when you get a hold of a second-hand mainframe – particularly for less than the cost of a single copy of MS-Office – getting needed software is up to you. So Connor had to get a hold of an operating system.

Now, if he had acquired a big old system from the 60s or 70s (or something that works like it), he might have been inclined to install an open-source version of a legacy mainframe OS such as MVS 3.8. But Connor, being from the post-penguin (i.e., after the invention of Linux) era, automatically thought of putting together a Linux environment instead, to ensure he was running with something that didn’t violate any rules and because he was familiar with Linux.

Ah, software licensing. Yes, that’s another important reason that a teenager, who can just afford the power bill for his mainframe with assistance from his parents, might be dissuaded from assembling a complete z/OS environment to run on it. How many hardware, software and other innovations that have become part of the mainframe were about handling software licensing? If you think you’ve counted them all, you’ve likely missed more.

But not so for Linux – right? Well, sort of, but the folks at SUSE did get their hopes up when Connor reached out to acquire a trial copy of their Linux for the IBM mainframe. Within a month, he was receiving emails from them asking how his trial was going, hoping to add his environment to their revenue stream. Fortunately, he was able to get together with them at SHARE and sort everything out.

And so, up came the basement mainframe running Linux, and ready to connect with the universe. Now Connor just had to assemble a team of connectors – which wasn’t quite as easy as it might seem. After all, mainframers have generally learned during their careers not to “play with” the mainframe – even one just sitting there ready to telnet or ftp onto.

Fortunately, he was able to assemble a different sort of team made of experienced mainframers; his mentors … but more about that next time.