Skip to main content

The Triumphant Return of Centralized Systems

Greg Szabo on his long mainframe career and still loving what he does. (There's also a quick lesson on Canadian geography.)

This transcript is edited for clarity.

Reg Harbeck: Hi, I’m Reg Harbeck, and today I’m here with my colleague Greg Szabo. Now Greg isn’t just any mainframe colleagueI think he may be the one mainframe colleague I still know from the earliest in my career. I was still a young junior systems programmer when I first met Greg and he was already quite established in his career, and he keeps getting even more established. But Greg, let’s turn it over to you: How on earth did you end up working on the mainframe?

Greg Szabo: Oh, that’s a great question, Reg. It goes back to when I was 15 or 16 years old, and that was around the 1966-67 timeframe, and I was interested in computing. There was a computer science club at my school with a great math teacher who headed it all up and I was exceptional in math, and I just migrated to it. I started going to the board of education once a week, putting my programs together and running ’em through the smaller computer system they had there. It was an IBM—I can’t remember what it was, a 1430 or something—it was a very small disk system with punch cards and printed output, and we’d get the privilege once a week—I think it was Tuesday night for about 3-4 hours where we’d go in there and meet till from 7 till 10, and I started coding then. And by the way, we also had a computer science club. Took a trip to Waterloo once a year, and when you went to Waterloo, you had to code up everything with job control language. So at around 16 or 17, I read my first JCL manual and I’m still training from that today. So, quite interesting. So if you think about that, I was around—I think it was grade 11. So that would make me 16 years old and I’m 73 now, so that gives me how many years of experience? I think some people are going to have a hard time beating that.

Reg: OK, it’s impressive. But I have to say that at SHARE, there probably were a double-digit number of people with your level of experience, which really struck me because there’s also a whole lot of people who are in their 20s, so we’ve got a whole wide range. But okay, now here’s the thing: This is quite rare for me, but we’re both speaking Canadian, and so I’m going to do a little bit of translating here. And the first one is Waterloo. Now Waterloo is Canada’s MIT. The University of Waterloo is Canada’s great technical and science university. It’s in Waterloo, Ontario, probably just under 2 hours drive from Toronto, and just a little bit north and east—sorry, north and west of Toronto. Every Canadian kid growing up who wants to go to university in Canada, wants to go to a great technical and science and engineering university, aspires to go to Waterloo. So that said, you, I’m going to guess, grew up in Ontario, but probably a little bit of a drive from Waterloo.

Greg: That is correct. I was living in Windsor and the high school was Kennedy Collegiate, and we would make that trip once a year. I actually attended Waterloo. I didn’t graduate from there, but my first 2 years I was in the advanced math class with computer science and I worked there. I studied there extensively. I actually graduated from University of Windsor in an honors degree, computer science and math. But there’s a little bit more history I wanted to share with you before we get to university.

Reg: Okay now remember that then, because I just want to add one more thought for our listeners, and that is Waterloo happens to be the university where Martin Timmerman worked, and he is a past president of SHARE. He recently retired, but he was actually a University of Waterloo employee and they used to have an IBM mainframe. They moved off that mainframe in recent years, but when he came to volunteer at SHARE, they were actually a mainframe organization. So now back to you, Greg: Tell me what else you wanted to share.

Greg: So my passion for computing was quite extensive. When I was 18, IBM had a contest, a computing program coding contest. So at 18 I wrote a compiler—

Reg: Whoa.

Greg: Using FORTRAN, and I compiled PL/1 code into machine language at that young age and won the contest. It paid my first year of university, and that first year of university I was an Ontario scholar and I ended up at Waterloo. So there’s a little bit of my history. When I went to Waterloo, you got to imagine we were using punch cards and printed output, and turnaround time was 2-3 hours. So you had to know everything perfect or you’d never get your assignments done. So I learned how to read PDFs—or they weren’t PDFs but the books that you read—the IBM manuals or the books you bought from the bookstore. I read the JCL manual cover to cover and memorized it. I did the same with Assembly, and I developed a skillset of being able to study manuals really, really completely. I read the Principles of Operations twice when it was 500 pages while I was in university. It was different than today. Today they just do a Google search and they get a quick solution. But my skillset has lended me to be able to expand my knowledge base as time went on because now I’m a very diverse instructor across a multitude of disciplines. And as I pointed out to you a little earlier, I’ve taught around the world for IBM training partners, and now I’m working with ProTech. So I’ll give you a little bit more history as we’re going along. So my passion from it grew at a young age and it’s more than, to me, it’s never been work, it’s been a passion. I love the creativity, I love the understanding of how things work, and that’s what I teach my students is I teach them why, and I give them a different paradigm view on the whole system. And we can get into more discussions on that as we get into advanced subject matters. When I graduated—in my fourth year I was actually a tutor for engineers who were studying at University of Windsor, and that paid my college fee as I went through. So it’s been very lucrative for me as I’ve got along. Eventually I took a little sojourn and I became an application developer on Philips mag ledger system for about 3-4 years. That was the only compute shop when I went out looking for a job in Windsor, Ontario, and I got the job.

Reg: Let’s put this again: When I talk to my colleagues in the U.S. and they talk about well-known place names in the U.S., I’m assuming everybody knows where they are. But not everybody knows where these various place names in Canada are, and one of the unique things about Windsor, if I recall correctly, is it’s actually south of its twin city across the river. That’s a very well known American city.

Greg: Yes, we all know that city, Detroit—exactly, right across from Detroit.

Reg: But south of Detroit, which is—

Greg: Really cool, absolutely correct, Reg. And I demonstrate that to my students in class and I kind of freak them all out as I show that to ’em, right? So when we get a little bit of a break, sometimes I get into talking about something to break up the educational fire hose I’m feeding ’em from.

Reg: So back to the fire hose. Keep going.

Greg: Yeah. So when my sister, who was working at Chrysler in Windsor—because Windsor was an automobile center for Canada at Oshawa. Chrysler was there, Ford was there, GM was there. I got a job at Chrysler, and she was making more money than I was at this small company. So I got in and I wrote an exam, they hired me and I worked 6 months as a COBOL programmer and then they advanced me into being a system programmer. And this was just at the time that Lee Iacocca was taking over and Chrysler almost went under—and there was, out of three of us, only one of ’em was kept, and I was the one they chose. It wasn’t based upon seniority. So I was a young buck and they still went with me. At that time I understudied a wonderful guy whose name was George Kramer—he worked at IBM. I convinced Chrysler to hire him so that I could understudy him. He retired with the company and I understudied him for 3 1/2 to 4 years. And I then took on a role as a lead system programmer at Petro Canada. Chrysler was in trouble a little bit. The oil patch was booming in ’81. I went out for five different interviews and had five job offers, and I took the job offer that offered the greatest challenge there was—they didn’t have a compute system in there. They didn’t have a computer in there. They had no standards, no naming conventions, no RACF, nothing. They were rolling in the computer. I think I arrived March 1 and April 1, the computer arrived. I was the lead system programmer who designed everything from the ground up. I played the world series of system programming. I did everything. I was the lead guy who ensured that there was always a system to recover with. Again, I got into performance tuning. I got into system tuning. It was a JES3 system. I had to have it available for the application people by July 1. I worked so hard in that period of time that they brought a cot into my room. I worked 110 hours a week. I worked so much that they gave me November and December off. Now when I went to work for them, they made it very lucrative for me. They gave me a 40% boost in salary. They bought my house from me in Windsor, Ontario—bought it from me because I couldn’t sell it. They gave me a $25,000 interest-free mortgage for a new house. They flew my wife out and moved me from Windsor, Ontario, to Calgary. And they put me up at the Westin for 3 months, gave me a 1-month signing bonus, and at the age of 30, I had 5 weeks of holiday a year.

Reg: Wow.

Greg: And there I became manager of tech support. I have 11 guys reporting to me. When we finally settled down and then Storage Technology came knocking, they heard about my reputation and Storage Technology came knocking at my door and made me an offer I couldn’t refuse. They made me western district system engineering manager for pre- and post sales at StorageTek. This was when they were coming out with the robotic library, which was a huge success for them. And I took on that role and eventually, I actually made it to club a few times, and in bonuses for solid state disk attached to the mainframe, I made more money than any other person in the entire corporation. But I was kind of a nice guy; I shared it with all my guys who reported to me. Even though I was a guy closing the business, I wasn’t greedy. It was like, you’re not going to believe this—$150 for, I think it was 12 megs of memory if you sold it in solid state disk. And then what happened was Memorex Telex—or it used to be Memorex—was into selling peripheral devices and had geared up with Fujitsu and to sell our competitive robotic library to StorageTek. And they hired me on as national system engineering manager. So I would be in charge of all system engineers, pre- and post-sales, for Memorex Telex. That stint with Storage Technology lasted about 4 1/2, 5 years. And then I moved forward to work with Memorex, who took over Telex, and I became an international product driver and lived in Italy for 6 months while we introduced the Fujitsu library to the world. And as a matter of fact, through my efforts, the new company that started up in Calgary called New Era landed this contract for writing the software on the mainframe to make the Fujitsu library work. And the idea of caching of robotic most used cartridges in the library reduced the library size to about 5% of all the cartridges that were in the tape library, because after you wrote to a tape there was very low likelihood of using it. So we purged them in the evening and introduced scratch tapes during the day. That whole algorithm and approach to interfacing with the mainframe was developed by New Era Software in Calgary—

Reg: I’m learning stuff right now from you here. The thing is, in 1989 after getting my computer science degree in Calgary, I went up to Edmonton and I moved back to Calgary in 1989. And so this is roughly when you’re talking about—it’s like around the 1988 Olympics, that sort of era, isn’t it?

Greg: That is correct. I took a chance, remortgaged my house and invested in New Era, which ended up being a home run for me because that was sold for $23 million to Interlink. I don’t know if you remember Interlink, but they created a protocol stack for the mainframe z/OS—or it wasn’t the z/OS operating system, but the OS/390 operating system at that time.

Reg: Was it OS/390 or was it still MVS?

Greg: It might’ve been MVS. It might’ve been MVS/ESA going into OS/390. So they and all my friends who worked in Calgary got hired by New Era. It was a fairly substantial company. I don’t know if they had 20, 30, or 40 people working for them, but the creativity in Calgary was incredible. There’s something called Hardware Configuration Manager that’s a GUI-based way of doing a hardware configuration definition that was put out by Tom McDonald who came out of Calgary as well and bought by IBM and is now part of the z/OS system and software that they sell. So at that time, Memorex Telex got in a little bit of trouble—and by the way, I was rookie salesman of the year. I became a salesman and selling robotic libraries and peripheral equipment. I was a rookie salesman of the year with Memorex Telex—and then they got a little bit into financial trouble. That stint that I was an international product driver, that gave me a lot of influence to bring the software development home to a team of talented individuals in Canada, in the Canadian city of Calgary. I don’t know how well the American people are going to feel about that, but it was my hometown and my friends.

Reg: Well, Calgary is the most American city in Canada. There’s more Americans per capita in Calgary than any other city in Canada.

Greg: I didn’t know that. Oh that’s fascinating, so thank you for that.

Reg: The oil industry is a big part of that. I mean, the connection between Alberta and Texas is really deep.

Greg: I lived in Calgary for 20 years and I then started, until Memorex Telex got into trouble, but I had done significant performance analysis for Memorex Telex and Storage Technology with all the chief information officers at all the big companies in Calgary—NOVA [Chemicals Corporation], city of Calgary, you’re well aware of that—and a number of others. Actually I was at the city of Winnipeg as well, and when I left Memorex Telex at that time, New Era had hit the home run and I was in good shape financially. So I stepped out on my own and began my own consulting company and my own software development company. And that’s when I formed my company, Expanded System Services out of Calgary, and we worked at the city. That’s where I met you.

Reg: Yes. That’s around the time that we met. So I was a junior system programmer. I had started as a CICS system programmer working in the Alberta government in the late 1980s and then moved down to Calgary at the beginning of the 1990s and did a half a year stint as a COBOL programmer myself, and then went back into systems. And so I was doing system programming at The City of Calgary when I met you.

Greg: Yeah, exactly. And some of your people, I’m sure you remember, came to work for my company because I formed a cooperative, like a law firm, who we all shared in the cost of software development equally and we tried to build a few products ourselves. I’m sure you remember Harvey Desereaux—

Reg: That sure rings a bell.

Greg: And you remember my manager, Peter Chang.

Reg: Oh yeah.

Greg: Peter Chang came to work for me. He was doing education through me with IBM. There’s a great story there.

Reg: He was my manager for a little while.

Greg: Yeah, I believe it. So Peter joined me and the guy who I worked for at NOVA joined my company as well. And all these were very successful just around Y2K, I built my company to about 20 people and we were supporting all those people who eventually, when they decided to migrate off the mainframe, they came to us to support it. Because there’s a multitude of mainframes in Calgary—more than any place I believe in all of Canada in terms of density of companies that were there had migrated over. Dome Petroleum, [which no longer exists], we had Gulf Oil who actually built TLMS or TMS—they were actually the ones that developed it. I knew the guy who did it. His name was Mickey Way, out of Gulf Petroleum out of Calgary. Now that went through a lot of morphs before CA took it over. I’m just going to say there’s a long history behind it. Now Mickey was a good friend of mine as well, and the person who headed up New Era Calgary was a good close friend of mine, Ben Dully, who I did triathlons with. He was a very brilliant guy. He was so brilliant that he bought back New Era from Interlink when they got into trouble because IBM wiped them off the face of the earth when they put their own TCP/IP protocol stack into z/OS and gave it away for free. And he bought it back for $3 million and within a year he sold it for $32 million to Beta Systems. And that product is called Harbor. I was part owner in both those deals. So it was very lucrative and successful for me. It gave me the ability to continue to focus on building my company. We did that together and I am very pleased to say that everybody who came to work for me was making over six figures a year and having a blast and just enjoying working with me in the positive, creative environment we were in. Now myself, personally at that time, I was working on our product called Adm. I don’t want to get too much detail into Adm. It was relatively successful. We developed it totally through sales into large organizations. It modified print streams and injected metadata into it and allowed us to take that print stream and organize it according to postal codes. And we put the bar codes on it for envelope stuffing. We created versions of it that allowed you to look at it through a PDF interactively instead of having to have it mailed. So we built this very sophisticated system through customer input, and today it still was used by about five or six large corporations, one of those being Desjardins, and maybe I’ll have to explain who Desjardins is—

Reg: Yes, good old CCPEDQ. That was the first mainframe customer that I worked with when I left city hall in Calgary and went to work for a large mainframe software company. They flew me out to Montréal to meet with Caisse Populaire Desjardins CCPEDQ, and talked to them about mainframe security and Db2. But back to you.

Greg: So they bought our product—we beat out IBM, I’m sure they’re not very happy with us—but we demonstrated our software product written in Java on a desktop when the contract and the bid went out and they were blown away. And the portability up to the mainframe being Java-based was easy as pie. So we had a database behind it for referencing documents and pages on the printed output stream. The printed output stream became the data component of the database where the index itself was in Db2. So we built this product, were relatively successful with it. We are still collecting maintenance revenue on it today. And we have Lloyd’s of London, we had GM using it, to name a few, as well as the biggest user was Desjardins. So that is still in effect today. It works and helps with their mailing out of their statements, both for their banking and for their credit cards. So that was pretty cool. The product never really took off. After significant investment, it sort of stalemated. Content Manager on Demand and customers’ desire to go to a well-established, large vendor kind of put us to the sideline. I’m not complaining. It was a great lesson. I drove all the marketing, I organized the software. I wasn’t the person doing the coding on it because I was still working as a full-time consultant at the time and doing this in my off hours. I’m a real workaholic and I have a passion for this stuff. I live and breathe this. To me, a weekend isn’t a weekend, it’s more time to get on and play some more. It’s always been playing, right?

Reg: Well, what do they say: if you find something you love, you never work a day in your life.

Greg: That’s right. I haven’t really worked a day in my life and that’s why I’m still doing it today. And as time went on, I got hired at NOVA as a system programmer and I implemented their first web server on z/OS, on their OS/390 1.3 as we were going to 2.4.

Reg: So this would’ve be late 1990s at this point if it was OS/390, because z/OS, that basically saw the light of day around the time of the turn of the millennia.

Greg: That’s correct. That is correct. And at that point, the organization was built to around 20 people, and Y2K was very good to us because all my guys were gainfully employed. I had instructors, I had consultants out working at the various companies throughout western Canada. We were teaching internationally for IBM. As a matter of fact, I was teaching so much for IBM that eventually I moved to Toronto. Now this was 20 years ago. So as they were hiring me relentlessly, they were shipping me around the world. I was doing things and making education work where it hadn’t worked before.

Reg: So you’re based in Toronto at this point, moved from Calgary to Toronto, but you’re working actually around the world—and quite a few different countries as I understand.

Greg: Yes. At that point in time, IBM was a little bit more lax with terms of farming individuals out to other IBM organizations in other countries. And I was doing leading edge stuff where other people were afraid to touch it. They were just stepping into UNIX System Services and TCP/IP and they had trouble with some of the courseware, and I created courses that worked. Take for example, they had a diagnostic class on TCP/IP that they never had run before. So I took the courseware and I took a P500, which is a desktop base, z/OS, had a card in the motherboard. It was an MCA—remember the old MCA connections on the motherboards? They had a motherboard that you could plug into the back plane of the PC that had the z/OS or the OS/390 machine language instruction set on it. It was a desktop system, and I implemented OS/390 on it and set up the diagnostic class, put VMware up on my portable powerful laptop, took a router along—actually emulated a router under VMware on my desktop and created a little network of routers and Linux images, plus going out to the z/OS box—and had a portable TCP/IP diagnostic class and lab that I could take with me. And that’s exactly what I did. I built it, I made it work, I shipped it out to a number of IBM customers and had a real success story with it.

Reg: I can’t believe that we’ve been talking for 26 minutes and we’ve still got 20 years left before we get to today. I hate to put the turbo charge on you here, but we don’t want to run too, too long here. But this is great!

Greg: You have to appreciate—you know we’re talking 57 years of efforts here. So it’s not light weight, right? This mainframe had the 60th anniversary. Think about it when it’s 3 years old, I came onto the scene, right?

Reg: Yeah.

Greg: So I mentioned the stint at Chrysler, just to let you know, when I was a sysprog there, we were going from VS1 to MVS when I stepped in. That’s when I came to system programming. Years before they were on MVT and they had gone to VS1 and I came in. They jumped over VS2 and they went to MVS and that’s when I stepped into my system programming career. So I’ve gone through XA, ESA OS/390 and z/OS as well. And by the way, the disciplines I’ve taught in my career, it’s easier for me to say what I do not teach in terms of the mainframe. Whether it’s Rexx coding, COBOL coding, Assembly coding, workload manager tuning, these are all the things I teach.

Reg: What don’t you teach?

Greg: What I don’t teach—there’s just very few areas. I teach WebSphere as well, I just want to throw that in. I teach storage classes, security classes, performance classes, diagnostic classes, maintenance and implementation classes for server pack. The only thing I don’t teach is MQ—WebSphere it’s called now—and CICS, IMS and Db2. Those are the four areas I do not teach.

Reg: Okay. That’s so funny because IMS and Db2 and Datacom and ADABAS and IDMS and tableBASE and [Model 204]. Those are the things I can’t really teach. I can speak at a high level about databases, but I’ve never been up to my elbows in databases. So I know better than to pretend. But it sounds like we have an interesting overlap there.

Greg: So those are the areas I do not teach. And you know ADABAS was at the city, right? And the Natural language—

Reg: And also at the government. I mean both places I worked in my beginning career, they had ADABAS.

Greg: So with my educational journey, in order to make a living at it, you had to diversify, and I did. I was very good at diversification because of my skillset of being able to study manuals. So in about 10 to 12, 14 days, I could cover a 500-page PDF manual today and absorb it enough to be able to instruct it because of my background and skills. To that end, eventually, as you know, IBM has kind of stepped away from the educational business a bit and they farmed it out to training partners. I’ve trained for all the training partners. I’m still doing occasional consultant stints on the side. And recently with the Vitality Program—

Reg: At Broadcom.

Greg: I’ve been working extensively with Broadcom, who have kept me busy quite a bit for the last 5 years. And I also—to a large U.S. organization, I don’t want to mention it—I do a 10-week Assembly class because their system depends entirely on Assembly. So I do that twice a year and with the Broadcom Vitality classes of 5 and 7 weeks, I’m pretty booked up, but I’m still loving it. I love the new generation. I love everything about the mainframe. And part of my efforts in instructing is to open the door to how significant the z/OS operating system is, the zSeries hardware and the Z architecture. Creatively, it’s in front of everyone because I teach how to implement Linux on z/OS. So I’m pretty extensive in Linux and have taught Linux classes. I teach how to implement zed/VM on z/OS.

Reg: Yes. He says zed—I was about to chastise you. Here we are, two Canadians, and we keep saying Z just because we’re so used to working with our American colleagues. But we’re allowed to say zed, too.

Greg: You notice there’s not—I try to watch my Canadian side of things. I don’t say eh—I don’t do that very much. So in terms of what I’m doing today, as I host this series of new people coming into the mainframe world, I try to turn ’em onto how exciting an operating system it is. I strongly believe being a mainframer is more than being a mainframer. It’s being a disciplined technical professional who works under the highest guidelines of care and diligence before implementation, testing to the nth degree, going through change control. It’s more than just the hardware and the software. It is a professional organization of engineers who comprehend things at a lower level consistently. Focusing on high availability and performance excellence in terms of maximizing resource utilization while achieving the response time goals of the environment. It sounds like a commercial, but it’s true. We have led the pack in terms of cloud computing. When you think of the 3270 emulator, which is fairly powerful unto itself and does a lot of things, it’s going out to a Z box anywhere in the world, isn’t it?

Reg: Yeah. Well I know I’ve met system programmers who work way remote—

Greg: Yep, and the concept of dynamically being able to create a [VM] image of a z/OS, running z/OS underneath the covers, can be done in literally hours. And the levels of security that exist there—and the level of AI. Everybody’s talking AI—woo, AI. Well, when I talk to the students about comparing other operating systems to z/OS, there’s no other operating system that allows you to set by transaction level the response time goal and measuring real time against it and dynamically make changes using artificial intelligence through the workload manager, gathering data from RMF, talking to the system resource manager and modifying things like dispatching priorities, page fencing, which is giving you more real storage, and I/O priority queuing. There’s no other operating system that does that in real time. You compare it to Linux, Windows, I don’t care where you go—that is AI, artificial intelligence, and we’ve been doing it for 10-15 years. I just don’t think we word it properly. So the concept of where we are in terms of comparisons, because of that discipline that we develop and the leading edge architecture that we have, when we go to other shops in the distributed world we have no problem. I quite frankly had no difficulty in picking up Linux and teaching it. But if you try to take a distributed person and bring ’em to the mainframe world, it’s another thing. It’ll take him years before he develops that sense of professionalism and attitude and hard work that we have.

Reg: I often thought some day in the far distant future, being a computer professional might actually be being a professional like a doctor or a lawyer or an engineer or an accountant. But if they’re ever going to form people as professionals in computing, the only platform where you can do that is the mainframe, because we have that level of professional behavior that is baked so into the system and the culture that you can’t escape it, and it’s necessary. Where you can get away without doing that, you start with your own home computer and move into a job without ever learning professionalism in other platforms. I think that’s a challenge that we really need to rise to.

Greg: I agree. Couple other points I wanted to make. You know around Y2K, everybody went distributed. Remember the city was actually moving away from it. We ran the mainframe for 7-8 years behind the covers for them. I’d like to talk about this a little bit because I think this is an important point to make.

Reg: At some point we need to have another one of these conversations, but I’ll give you a couple more minutes, but we’re going way over time here. But go ahead.

Greg: Okay, couple more minutes. Last couple. It’s very interesting what I’ve observed over the years, because I’ve been around a long time. So this tendency to go distributed was embraced a lot in the Oil Patch, because now you didn’t have to wait for your applications to be done. You just went out and bought the box and bought the software and did it yourself.

Reg: Now let me just add in for our listeners: The Oil Patch refers to on the one hand, all the oil companies in Alberta and especially in Calgary, but also then there are twin oil companies down in Texas and such places.

Greg: So what they had to do, they wanted to be able to break out financially the resources dedicated to it. And when you bought a little computer and bought the software dedicated to that project, it was much easier than when you’re on a centralized system, super server like what we like to refer to as a mainframe. And so there was a real strong tendency to go that way. What’s interesting is to observe what went on. These companies that bought these distributed systems, everybody went in different directions. Some bought IBM AIX, some bought Sun Solaris, some bought HP. They had flavors of Linux; some bought Windows servers. Eventually, the engineers who were supposed to be doing engineering type things were supporting the computing system more than they were doing their work because they had to put maintenance on, there was backups to be taken. And it’s quite interesting because at times these little department systems were angry at the central group supporting the mainframe when their system broke and said, why didn’t you have a backup for me when the mainframe people or the centralized group of people weren’t even involved? True stories. And to that end, what I saw happen was, as time went on and they gave up and pushed it back to the central organization. We reduced the number of vendors; we went to maybe one particular vendor, maybe IBM AIX. We bought blades, which centralized the hardware, and they implemented VMware, which allowed the images to ship between the hardware seamlessly, put it into a single frame, not unlike address spaces or VM under z/OS. Then what they did is they took storage and they didn’t do just a bunch of disk on the servers, they created a storage area network and centralized all the storage. As a matter of fact, that centralized storage can also support the z/OS system and the zSeries box using the DS8000. And they recentralized again. They recentralized again! They cut down on the vendors, they merged the hardware together, they centralized on storage, on networking through Cisco routers and Brocade routers in those days, and IBM was in the routing business for a while as well. And now they went back to a central system because a central system reduces the skillset that you have to have. You can do more maintenance with less. You apply it once and you’ve got it under control. It allows a rollout of operating systems easier. You have a single vendor to communicate with when you have problems and no finger pointing. It just makes better sense. So really what I’d like to conclude this with, is the people who never left the mainframe were really 30 years in front of the people who did, because they’re all back in a centralized system where they started, and maybe they would’ve saved millions and millions and millions of dollars had they never left. So there’s my wrap up, okay?

Reg: Oh yes. Well I so emphatically agree with you that if I were to even give a hint of my emotionally positive response, it would sound like a revival meeting.

Greg: It is!

Reg: I guess that’s true. So that said, just before we kind of finish up, first of all, is there any other thoughts you had? And then also what does the future of the mainframe look like if you have anything to say about it?

Greg: Well, I love what’s going on now in the mainframe. Container management, AI for predictive analysis of failures, pervasive encryption, these are the things that are going to keep it at the forefront. I think the challenge has always been is to get the message out more and with this new generation that’s coming up, I’m actually having an opportunity to evangelize with them on the merits of it. And I enjoy that part of the business, I enjoy continuing to learn. I think it will continue to evolve. I think Zowe is a great idea. I think what’s missing a little bit is the merging of the legacy way of things with the new world of things. As you know, I’m working on some software now to try to bridge that gap. I want to play an important role as we go through the future and try to bridge that gap and continue to play an important role in this hardware that I’ve had such a passion for all my life. I want to continue to be creative just like I was when I was 18, continue to build exciting things, and continue to partake in the revitalization of this platform. I was told—and you too were told 20 years ago—that the mainframe was dead. We’ve heard that cliche a million times and genuinely, it’s nice to see it’s morphing and it’s changing. I think we have to be careful of the abstraction and simplification, because one of the beauties was is that we always understood what was under the covers, and we could fix the problems accordingly. Now we’re introducing another layer of abstraction that could itself have problems. We have to be sensitive to never losing the underlying skills of comprehension that we have because that gives us the skills to just fix things and not just do things. And that’s the way I’ve always felt with this architecture. All the manuals and the lower layers have always been available to us. We have been professionals that don’t just do it, we fix it and we understand how and why it works. That is what differentiates us. We aren’t just like the Nike symbol, “Just Do It.” No, no, no. We don’t just do it, we do it right. We do it with high availability, with the professionalism and control and communication that’s required to support a large scale infrastructure to the point of where we have banks like the ABC Bank of China, who have 40 million customers online in a given instant. This is the architecture that does it. We’re dealing with rocket ships, not bicycles and cars. We are the Ferrari mechanics who are going to win the race. That’s who we are. We have to appreciate ourselves. I would love to see—and it was talked about many years ago—that we become certified engineers with the little silver ring which we don’t have today. There was talk about it many times, but now more than ever, I think it’s appropriate for us as we delve into this new world. The new technologies have never scared me. I’m using Chat-GPT in my classes. I have my students create quizzes for themselves using Chat-GPT to augment their education. So I’m not shunning anything; I’m embracing it. The only thing I am cautious of is, and what I’m looking for is a Chat-GPT that is not going to give you hallucinations, because every time you want to use it, whether it’s checking out JCL or not, it’s led me down an inappropriate path.

Reg: But you know why? Provenance of data, and IBM is doing that with their AI on the mainframe. They’re literally ensuring that you’re using quality data so you don’t get that weird stuff that gets into your data and makes it behave hallucinogenically.

Greg: Yes, yes. So one of the things that I’m working on is to have not quite—it’s available today, but people aren’t implementing it—is to have all the IBM documentation available under Adobe. There’s a feature that they give you an index and you can search for a single word without going to the internet and get every manual that hits it. And I’m going to simplify that process. And that’s another one of the things that I’m sort of working on and developing.

Reg: Greg, we could go on for a week. I am acutely aware of how much we didn’t talk about that we could have, even though this is as long as I’ve gone with one of these. But it’s been absolutely worth it. I’m so thankful to have had the opportunity to introduce you to everybody else, given that you were like—seriously, I think of all the mainframers I know that are still out there today, you’re probably the one I’ve known longest. So it’s a real pleasure to be able to introduce you to everybody. Thank you so much. It’s an honor to have been able to finally do one of these TechTalks with you.

Greg: Well, I want to thank you for taking the time to interview me and encourage me to do this because I’ve been very quiet because of my travel. I haven’t really exposed myself too much, worrying about my family, you know, so I do appreciate it. I [feel] almost like I’m coming out now to share my visions with the world, and you are the person who’s leading the charge on that. So I appreciate your efforts, Reg. Thank you.

Reg: It’s a pleasure. So I’ll be back with another podcast next month. But in the meantime, check out the other content on TechChannel. You can also subscribe to their weekly newsletters, webinars, e-books, Solutions Directory, and more on the subscription page. I’m Reg Harbeck.