IBM Destination Z - Group home

Peaceful, Easy Mainframe

By Destination Z posted Mon December 23, 2019 03:38 PM


Being eagle-eyed goes with being a mainframer—so much so, that the SHARE mascot for MVS is an eagle. Of course, these days, wearing bifocals or progressive lenses also often goes with the job, which is fine, since it’s not about fine print but about refined thinking and attention. That’s one reason the mainframe soars with the eagles rather than being sore with the turkeys: reliable technology we can take for granted to handle the details.

However, we’ve spent too long silently soaring above the Frey. It’s time to acknowledge the billion points of enlightenment that make the mainframe the true star of business computing, while discovering the important ways that the mainframe is also down-to-earth, as it keeps silently slipping into the future.

To rediscover and proclaim the “easy” in IEASYS, you may wish to make the three-hour drive from Winslow, Arizona to Phoenix for SHARE, where I’m giving the zNextGen keynote on this topic on Monday, March 11 at 3:15 p.m., room 106C. We’ll be having some fun discussing all the foundational and touchstone insights that emerged on the mainframe and still have their definitive manifestation there.

Preparing for this discussion, and more importantly for the discussions you’ll have with non-mainframe colleagues, friends and family before and after SHARE, here are some reasons why we on the mainframe don’t need to let the squeaky wheels drive us crazy, because we’re already standing on solid ground.

Costs and Benefits 

Have you ever noticed that the majority of unsuccessful efforts to move off the mainframe rarely have a provable cost/benefit case that takes into account all the costs of the mainframe, all the costs of the alternate platform, all the costs of moving, all the benefits from staying on the mainframe and how to achieve as many or more benefits from moving to the new platform? Instead, cost/benefit analyses are often based on some combination of anti-legacy attitudes and aspirational sales pitches by a charismatic sales person who will say what it takes to get their commission and get out of Dodge before the dust settles.

We mainframers have become cynically comfortable with the knowledge that such migrations fail, and annoyed that management doesn’t run the numbers more thoroughly in order to figure out how they’ll even come close to 100 percent of what the mainframe is doing at any price.

But there’s a more obvious principle at work here that no one seems to acknowledge: It’s not good enough to replace an older system with an equivalent newer system. In fact, to make it worthwhile to replace an established, functional system, the alternative has to be quite a few times better in every relevant way for it to be justifiable. But every study I’ve seen says that the mainframe wins these comparisons at any scale that makes significant use of the mainframe.


It’s hard to replace something good with anything equivalent or only slightly better because of the effort and costs necessary. Even if the incumbent were noticeably worse it would still be hard to replace just because it got there first and got its foot in the door. Remember VHS versus Beta or Windows versus OS/2? No matter how much better the latter may have been, the former got there first, and therefore sunk deep roots in many different ways.

Likewise for those organizations that already rely on mainframes, and for the entire world economy, the mainframe is demonstrably better, and it got there first to set the rules for how to do things properly. Often the only way other platforms have had any hope of differentiating themselves is by following fewer rules, allowing for the formation of a “cowboy” culture among those who work with them. In a world where privacy and security are only getting more critically important, that makes it even less feasible to move off of an established, trusted incumbent.


And speaking of security, let me first say that I’m very proud to be able to work with all three z/OS external security managers (ACF2, RACF and Top Secret)—each of which is an excellent part of a comprehensive mainframe security context—and you must have at least one to be running a properly secure z/OS mainframe.

With the full range of security behaviors from PKI and LDAP through policy-based and role-based access controls to pervasive encryption and beyond, the mainframe is a powerhouse of security. Then there’s the fact that its architecture is so different from the consumer electronics machines that no viruses or such malware written for those platforms have any functionality on the mainframe.

Of course, as my friends Big Endian Smalls and Soldier of Fortran will point out, that doesn’t mean your mainframe is automatically hack-proof. You still have to practice a culture of proactive security. Fortunately, that’s also something we do on the mainframe, and I expect the investments necessary to build on that to continue to grow.


Five and a half decades after IBM announced System/360, IBM and other vendors and mainframe-using organizations have put together a platform (and ecosystem) of such proven quality that its reputation should be unassailable—if people were just to look at the facts.

It really is time to start taking a jaundiced look at the people who have been assaulting everything good and established in the world just because they have some new snake oil to offer in its place. When you start making a list of every way you rely on your mainframe environment and take it for granted, encourage your colleagues to do so and maybe even get a place where people can add to it, you may find that it keeps growing indefinitely. It certainly should, as long as someone in upper management doesn’t shut it down because they want to pad their resume with a project to move off of a mainframe.

Optimized Configuration 

While we’re at it, who says proprietary is a nasty word? Yes, we’ve all seen that open source can be a good thing in some contexts—heck, we even have a bunch of it available on the mainframe now. But that doesn’t change the fact that the hardware, OS and other key systems and applications on the mainframe are written and optimized to work together according to an extremely functional architecture. Any mainframer worth their salt wouldn’t look down on such an architecturally-optimized approach for maximizing performance and value, particularly when the mainframe is capable of running 100 percent busy all of the time with no significant degradation.

But Wait, There’s More! 

Whether in the Southwestern deserts around Phoenix or in the electronic wasteland of exploding infrastructure with insufficient staffing and receding security, on the mainframe we can still put the “calm” in “commerce” as we take it easy and let everyone know that we have a platform that works, and isn’t going away.

And that is a peaceful, easy feeling indeed, knowing the mainframe doesn’t let us down.

1 view