By Jon William Toigo
It occurred to me a couple of years ago, when IBM began touting strong quarterly growth numbers both in numbers of MIPS and in revenues derived from mainframe hardware and software sales, that the burgeoning narrative about a “mainframe renaissance” might be short-lived. If I had written what I was thinking at the time, I would probably sound a lot smarter today. Heck, maybe I could be introduced at my next talk as a respected industry analyst.
Instead, I took the occasion to indulge in a bit of guilty pleasure. It was fun watching the smoke billowing from the ears of those industry analysts who had been proclaiming the end of the mainframe as an article of faith for nearly three decades. Something evil in me delighted in seeing their discomfort as they fumbled to find words to reconcile the striking differences between their analytical prognostications and realities on the ground.
I rarely indulge in such schadenfreude, and feel a little petty when I do—except when the victim is some analyst idiot who has probably never set foot inside a data center and yet who feels completely reasonable telling anyone silly enough to listen to him or her about the final days of tape, the burgeoning data explosion and storage gap, the rise of cloud computing and end of the data center, or (and especially) the impending demise of mainframe computing. That kind of willful ignorance of reality grinds my gears, sometimes making me so sick when I hear it that I find myself walking out of a meeting room where an analyst has been granted a block of time to spew.
“No, I didn’t walk out on your talk,” I might say when cornered by the speaker after the meeting, “I had an important tweet to read.” As a cool columnist, I smile at my wit, my demonstration of savvy in “new media technology.” I rather like being the only person I know who uses a 3270 emulator as a Twitter console.
The sad truth is that, in the rarified world of IT industry analysts, egos simply won’t allow a contrarian view to be considered. The analysts who had proverbial egg on their face for at least three quarters of strong mainframe earnings growth never stopped looking for a sign—any sign—that their initial views would ultimately be proved.
They got it last month when IBM reported that growth had slowed. One of the major trade press pubs, which is usually loath to talk about mainframes in the same sentence as computing, did a round-up of analyst commentary following the IBM quarterly earnings call. All of the leading brand IT analysts had their knives sharpened and ready to go. Doomsday for mainframes was back in vogue.
However, they overlooked the fact that “slower growth” meant there was only a 15 percent increase in sales over the previous quarter, which in my book is still pretty good—especially when you consider that this “slowdown” comes during a quarter that’s traditionally the slowest for every company in the tech sector.
The analysts also failed to mention that slower growth translated only to a 34 percent increase in MIPS, which again strikes me as a pretty impressive number. Companies are actually processing more data on mainframes than at any time in history; this, during a time when most companies are finding it difficult to sell just about anything.
I was reminded of these facts recently when CA Technologies called to brief me about the latest developments in their Mainframe 2.0 portfolio and roadmap. The CA folks hardly talk about the analysts, as they’re too busy creating the real “new technology” and delivering it in sensible ways. Mainframe Software Manager (MSM), sort of an InstallShield for Big Iron apps, is an example; this tool is available at no extra charge to customers on active maintenance for any of CA Technologies’ z/OS products.
CA Technologies has published the OASYS standards-based Application Programming Interface (API) for MSM for use by other vendors, even competitors—an initiative that would make life a lot easier for mainframers who use products from different sources. Sadly, there have been few takers.
If I were an analyst, this is the kind of information I would cite to question the efficacy of mainframes going forward. Not some goofy take on IBM quarterly earnings.