A 43-year Mainframe Odyssey, With Duane Baer
Duane Baer, CEO of Baer Consulting, got his first taste of the mainframe with the System/360 and never looked back—hear his account of the mainframe's evolution as he joins Reg Harbeck on TechTalk Enterprise
The following transcript has been edited for clarity:
Reg Harbeck:
Hi, I’m Reg Harbeck, and today I’m here with Duane Baer, and he’s the CEO of Baer Consulting. Dwayne, welcome.
Duane Baer: Good day, Reg.
Harbeck: So I understand you’ve been working on the mainframe for a while now. How did you end up on the mainframe?
Baer: Well, it was an interesting Yun activity, so right around the eighth grade, I tried to get as much science and math as I could, but I was really a little bit hesitant because at the time it was, your career was likely to be a teacher and that just wasn’t the track for me. So just kind of kept my head up and continued with math and a local college, North Dakota State School of Science, had sent out a flyer to the high schools around the area saying that they had acquired resources from what was then CDC, Control Data Corporation out of Minneapolis, Minnesota, that controlled data was willing to send the technology instructor over to the college at Wahpeton. And they started up a new curriculum called Data Processing Associate Degree. So I watched a little bit of that while I was early in high school and they had sent out a further notification that says, well, you can come to college camp for a week this summer and see if any of this program might be for you. So off to college, I went my sophomore year or summer and kind of took a few data-processing-like classes in a couple, three half-day sessions to fill up a week at camp, and essentially was hooked. I knew that mathematics was going to pay off, but certainly wasn’t going to be in the instructor program.
Harbeck: Cool. That’s very similar to my journey. I always thought I was going to be taking a math vector, if you will, in university. And then when I was in my last year of high school, they introduced computers and I was hooked as well. I’m going to guess that you and I were working different kind of computers. My computers were, first ones were Apple II, even though I became a mainframer. But with CDC would I be correct in guessing that you were dealing with something in 370 line of computers?
Baer: One more back.
Harbeck: 360.
Baer: Yep. Yeah, the college had acquired a real small 360 that they were actually used in the lab, and that was part of what CDC also brought to the program. So kind of academic mutuality there. So yeah, I found on my way through the System/360 and all the way up, so ended up finishing up that degree in North Dakota and kind of looked around. This may surprise you, but there wasn’t a single processor in all of North Dakota. Mainframe at the time, and you can already guess the decade, it’s the 1970s. So we’re looking back there when things were still kind of filling up. So anyway, I ended up out in Denver, Colorado, and had connected up with the banking industry, of which time Colorado in fact had banks. It was before all the branch banking and collection of what we see in finance today, but most every bank in town had a data processing mainframe computer running all of the financial applications and such.
So it ended up how I ended up in Colorado as early on operations at a bank and had a little bit of opportunity, if you will. So I never did attach to any of the business applications, and the track I took within college was to write an assembler rather than to go through the COBOL RPG language skillset learning. So always had a knack for kind of the system-level activity of it. And I was very, very blessed early on. A manager at the bank realized that I could have a lot of fun as a system programmer, so he offered me kind of a on-prem employed internship with the bank. So I kind of helped out all the assistant programmers every weekend as an operator. And so I afforded them to go home while some of their long sysgen runs were going on and I was able to pull ’em off and check ’em for ’em and they wouldn’t have to come back in and do it. So kind of earned a few kudos there and got transferred over to systems.
Harbeck: Cool. So basically by the late 70s you were a systems programmer.
Baer: Quite seasoned and scarred for it, if you will, because in the mid 70s, well, all through the 70s, because in that time or that era, IBM had what they call on-prem personnel and they called them PSR. So each client that had a mainframe had a designated IBM resource to kind of help ’em through all of the software challenges they might have with the operating system and such. And IBM also started as kind of a campaign to help customers become better at doing system diagnostics and repair as well. And they called it a red hat. And in those days, IBM had a massive, massive education system. So there was lots and lots of IBM training centers throughout the country, and they offered a significant quantity of various classes related to the IBM mainframe and its components. So in that era, it was a lot of challenge and had a lot of between hardware and software issues going on. It was, those that lived through it, we knew it as virtual storage constraint relief issues. So a lot of the CICS regions, early CICS regions struggled with memory and all of the memory leakage and that went through it. It was a struggle to, in those days to keep the mainframe up for much over a month or a month and a half at a time. And of course you now in today’s world, they don’t go down for years at a time. So quite a growth from then.
Harbeck: Now sometime during this 1970s to 80s journey, of course IBM went from 24 to 31 bits. That must have had a significant impact on the nature of your work.
Baer: Oh, it absolutely did. And that’s where a lot of this call it memory challenges and CICS, and again, the early days of IMS, because we really didn’t have, while IMS was DL 1, et cetera. Anyway, birth to come forward, we didn’t have any of the database that we do today until early 80s, I think is when a lot of that showed up. But yeah, to your point, there was a lot of challenge between the upgrades to the software and then the hardware would upgrade behind it. And it was a hop over each other for a good number of years. And then I’m not sure if you recall, but also in the late seventies, the red machine showed up, Gene Amdahl put out his machine. So there was a lot of that going on, and I did have the fortune of working for a client that did in fact have an Amdahl. Oh. So I got some experience on Amdahl machines as well. And again, that was kind of the whole, in those days, it was just, man, it was lots and lots and lots of information that was available to you if you just wanted to sit down and absorb it, if you will.
Harbeck: So it sounds like you took a lot of the IBM courses at that time.
Baer: We did. Had a very much, a large opportunity for being educated in that context. And the banks in those days were struggling to keep up. So I think the Colorado bank that I worked for was one of the initial owners of what is now Visa. Rocky Mountain Bank Card started in Denver, Colorado, which is now Visa organization. So that was kind of a fun point. And then kind of mutual to that is my first business partner, which is my wife, we started a consulting company in 1982.
Harbeck: Oh, okay.
Baer: And prior to that, she worked in the, what was called data entry section, where a lot of the key punch work, she worked at an administrative assistant there, but she had enough accounting skills as well that she was tagged to be the first ATM manager and balancer. So in the early days of that, the ATMs were connected to the mainframe to do authorization and cash disbursement. And so when we were having trouble with the mainframe availability, I would be in working on the mainframe to try to help get it back up. And once we did, then I’d go home and she’d have to tag her. She’d have to go in and try to catch up with all the ATM balancing. So had a mutual technology in the late 70s and the early days of our married life. But anyway, in 1982 we started a consulting company and have been doing that ever since. So we’ve had a real good opportunity in the growth, some years better than others.
Harbeck: So that’s basically 43 years.
Baer: Yep. It’s been going and still going. We got some more in this. So all doing well.
Harbeck: Now, I’m curious whether you were involved in upgrading to MVS XA before or after or both, before or after you started your consulting company?
Baer: It was right at that time. So I had had a little bit of exposure to the first scenario of the 31 bit machines in the late 70s. So the machines came out capable before the hardware or before the software actually became a GA. So we certainly went through the early stages of, there was a little bit of it in what they called an SP level first. So they went through SP 1, 2, 3, and then they started calling ‘em nicknames, so the XA and XA Enablement. So yeah, it was right at that time when a lot of the systems were bringing up hardware such that the memory extension was required for a lot of the banking applications.
Harbeck: So now, as you started your own consulting company, I’m going to guess that you found yourself doing more and more management and visionary things than you had to get other people to handle the technology. Was there some point in your career where you basically went let other people do the technology and stop focusing in it? Or have you’d always remained a bit focused on the technology?
Baer: Oh no, the kids these days won’t let me touch a keyboard anymore. They pushed me out of that ability quite a while ago. About all I can do now is talk about all the things that we used to coulda do on a keyboard. But no, to your point, the 80s was still very, very much, and which was our entry level into the consulting company, was a lot of physical people at a customer location, primarily because, one, the networks really didn’t support remote administration to the level that we needed remote connectivity. And a lot of organizations couldn’t quite handle the security concept of it. So very much the eighties was still send a person to a client for a period of time, finish a project initiative. So yeah, that’s when we expanded into additional staff and kind of working projects and multiple people. And to your point, a little bit more of the technical administration took more of my day than me actually doing technical work.
Harbeck: Now, did you at some point end up acquiring your own mainframe or just licensing access to it.
Baer: Nope, that’s a good segue. So in the late eighties, 87-ish, was about the time that IBM started having what they call business partners selling mainframes. And by early 90s, they had a pretty good network of what they called tier 1 business partners that could sell IBM new machines. So at that juncture, Baer Consulting then was engaged by several of the IBM tier 1s to do the actual perform to deploy the new machines. Oh, neat. So a lot of the tier 1s were, they did the front end, the sales and the engine and the relationships and tons and tons of contracts. And a lot of the mainframes in the 90s were shipped in climate-controlled trailers. So once they left Poughkeepsie, they didn’t stop until they found their destination to unload.
Harbeck: And now were you still in Denver area at that point?
Baer: Yep, Marilyn and I have, we moved to southwest of Denver up in the first ridge of the hills. We live at roughly 8,500 feet. We moved up there in 1978 and have been there since. So still live there.
Harbeck: It’s amazing how many of my favorite mainframe people live in the Denver to Boulder area. I’m just thinking of five different names just off the top of my head. It must be an absolutely lovely area because people go there, they just stay.
Baer: Yeah, it’s beautiful country. It really is. And the weather is very tolerable. We’re just up on that first sea ridge, so we get a little bit different weather than Denver proper, but just enough to really enjoy it. And it’s very, very comfortable even in the summer because it doesn’t get real hot at that level of the elevation.
Harbeck: Well, being someone who lives in Canada, I can appreciate the cooler weather. That said, obviously climate control has a very special meaning when you’re taking a mainframe to Denver. I’ve been through the Denver airport many, many times and I’m used to operating at high altitudes. And high altitude is definitely one of the first descriptions I would think of when I think of Denver. And obviously Poughkeepsie isn’t quite nearly as high altitude being in the Hudson River corridor. They’re very close to sea level and probably have a very different amount of humidity in other weather. What are some of the impacts that you saw that were related to that climate-controlled transport of the main.
Baer: Again, the whole issue there was to keep it out as you would expect going across country and different seasons if you will. But no, at that time they would put the machine in and as I recall, it would have to call it climatize in the actual protected environment, which was in humidity-controlled. And a lot of those machines in the mid 90s each had what they called a 400 cycle power. So the building had to be really, really prepped for different kinds of power and power sources. And then of course the air conditioning and the humidity control was massive. We were involved in a couple of data centers that used humidity control by actually running water in the channels under the floor to help with the humidity control. So they were very, very sensitive to environments in those days. They are a little bit today, but not near as much as it was radical in terms of data center control there.
Harbeck: Now in my imagination, I’m seeing a raised floor tile being picked off and it’s like, don’t step in the water. Is that anywhere close to—
Baer: It was actually ran water channels under the floor to, and again, they chose lots and lots of different ways in early data center designs to try to help with the stability of humidity control.
Harbeck: Now there’s I’m sure a lot of other factors that you’ve had to deal with in your role were, obviously you were involved in the installation of a lot of new mainframes, and I’m going to guess that that continued for quite some time. So you’ve seen some latest technology. What are some of the things that really jumped out to you in your experience of really interesting, shall we say, anecdotes of the experience of installing a new mainframe?
Baer: It’s a massive, massive amount of, call it quality control and significant amount of planning that even still goes into it today in the mainframe space, just because of the critical availability of them. And there was always the whole issue, the upgrade was what was called a push-pull. You spent, a lot of times you spent months getting ready for that one weekend because in a lot of environments, once you go, there’s no go back. And again, while those machines were down, so was the revenue stream of whatever institution it was providing service for. So it was very, very critical in those days to do adequate planning.
Harbeck: So I’m going to guess you’ve seen things fail in some interesting ways as part of getting them to succeed.
Baer: There was some long weeks some days, but yeah, and again, it was the, you say it as you kind of didn’t know any different. I mean, that’s what your job was, was to be ready and be supportive of others that were in the team. And again, it was the whole, took the hardware vendor, software vendor, and again, it was a whole infrastructure that knew how to support it.
Harbeck: Now, as a consulting company, have you ever thought of yourselves as either an outsourcer or a managed service provider?
Baer: So in the early 90s is when we had an opportunity to do a little bit of disaster recovery. So we were able to essentially acquire a real small mainframe that we actually set up a small data center in our offices, and we ran at least some DR exercises, if you will. And then after we suffered through Y2K and the world of post Y2K and kind of if you will, a little bit of individual company funding came back to start thinking about this again, we were able to open a tier 3 facility and we acquired some more mainframes at that context and we’re doing a DR service. And then I think we’re around 2014 or so, we opened our second tier 3 facility and we’re able to now do production and disaster recovery and the tier 3 facilities exceed the 250-mile requirement. But no, to your point, we own several mainframes, a lot of storage, and we now provide disaster recovery. And a lot of our customers are those that kind of, if you will, started clear back in the 1980s, where the death of the mainframe is us, we’re going to be off it in five or seven years. And so with a fourth or fifth renewal of that initiative be off in five years. Here we are 25 years later, they’re still on the mainframe.
Harbeck: I’m curious, have you seen any successful efforts to move off of mainframe?
Baer:
Yes. We’re clicking off, I’d say somewhere around 8 to 10%. So we do a couple a year that have actually made it and have actually shut it down. And it’s a little bit tongue in cheek, sad to say, but there are organizations that should never have been on the mainframe, but when they first started their compute requirement, there wasn’t really much else to be had. And of course, IBM’s smallest mainframe is now for the behemoths, right? So more and more companies are struggling to actually find a mainframe environment small enough for them to be able to operate on, hence the managed care that we’re providing.
And we also help ’em continue with their journey to rehost. As you know, a lot of companies have, call it lost applications in the mainframe space, not by choice. Healthcare is one of ’em. There was a primary healthcare provider that had an application on the mainframe and they no longer support the application distribution of it. So several clinics or healthcare organizations that we provided support for lost the primary application that supported the funding for the mainframe. And pretty soon they’re left with a big mainframe cost and not very many customers on it because the big one had to go off to a distributed application. So there’s some of that going on, even though there’s several that have just flat believe the mainframe is in fact gone or out of service and don’t allow any new development to be had on it, which is sad.
Harbeck: So that said, I think you probably being right in the middle of all this have a clearer perspective than many. What is the future of the mainframe, especially if you have anything to say about it from where you’re sitting?
Baer: I think it still is and has had a significant focus on all the threats of there’s no resources to support it, and it’s getting more and more difficult and it’s way too costly. So I think there’s some real good initiatives in play and some pretty big organizations are really focused on that. But one of the big things that we’ve seen is over the last 15, 20 years, there’s been, call it an acceptance of the mainframe support industry, if you will, that the system programmers and the system administration of the mainframe is in fact a user group deserving of some attention. And so it’s really gotten a fair amount of attention in the most recent operating system upgrades. So the tools are also taking benefit of learning. So the tools are learning how to repair themselves. More importantly, they’re learning how to offer up technical solutions and assistance to minimize the effort it takes to actually repair and continue to maintain the operating systems.
Upgrades to the operating systems from version to version upgrades of products, from version to release, so much more solid. They’ve just gotten the whole scripting and the discipline and the intellectual capital invested in the deployment sequences that they’re pretty much rock solid. And again, they’ve been requested to be, so is, I know IBM’s probably not allowed to say it, but the Z Series zero downtime. So there’s a lot more successful, call it upgrade maintenance, in place. And of course all that comes from teams and teams of people in various aspects of software development and deployment of upgrade packages. So that’s far more solid. And I also have met over the last couple, three years, just numerous younger people that have just found a phenomenal interest in the mainframe again. So it is so exciting to see the energy that they bring.
Harbeck: So I hear some optimism not just about the future of the mainframe, but of the future of the business interest that you started 43 years ago. How does the future of the mainframe look from the perspective of Baer Consulting?
Baer: Well, Baer Consulting, we’re very blessed to have been acquired in January by an organization called IT Service Alliance. And there are a couple of very, very intelligent guys that have an interest in being able to build on technology services across the platforms. They acquired another company earlier last year, which does more of a i series, P series type structure. And then they acquired us in January, which brings a Z series. So very, very energized with our organization, pretty much stayed intact, and they’re continuing to build out and invest in bear consulting as a Z series services organization. So I have a very much a high degree of confidence that it will be around for hopefully decades to come, because I just think the mainframe is in fact got a life of decades, not moments.
Harbeck: Oh yeah. I personally think in terms of it entirely being possible that a millennium for now, something that is based on today’s mainframe, will still be a critical part of our infrastructure. But that’s maybe just me. But I think if you take a look at the wheel, if you take a look at things, the great inventions of humanity, once you reach a stable state of functionality, it just sticks around and becomes standard. And so I’m also extremely optimistic about the future of the mainframe. That said, just before we finish up, any other closing thoughts that you want to share?
Baer: Again, I think to your point is the mainframe, as with other technologies, is they find a way to, if you will, survive and themselves find a way to create a better world for them. There’s lots of papers out there from the late 70s, early 80s saying, man, this is as dense as we can get. We can’t possibly do anything better, faster, more easily. And here we are 50 years later and we’re still making things smaller and faster. And more importantly is, I think the mainframe has done an excellent job in trying to be a reasonable steward to the environment. Just we use far less power than we used to. And just again, trying to be very, very cognizant of what we consume of the world around us. So I think it’s a very viable platform for the amount of users it can serve, for the footprint.
Harbeck: Well, thank you very much, Duane. This has been a fascinating and enjoyable conversation.
Baer: I thoroughly enjoyed it, I’m privileged to have received the invite.
Harbeck: So I’ll be back with another podcast next month. But in the meantime, check out the other content on Tech Channel. You can also subscribe to their week newsletters, webinars, eBooks, solutions directory and more on the subscription. I’m Reg Harbeck.