In today’s episode of the “I Am A Mainframer” podcast, Steven Dickens sits down with Robert Garrett zOS/Mainframe Operating Systems Software Consultant at zResource. On this podcast, Robert discusses his journey with the mainframe, the changes he has seen over the year, and where he sees the Mainframe going in the future.
Steven Dickens: Hello, and welcome. My name’s Steven Dickens, and you’re listening to the “I Am A Mainframer” podcast, brought to you by the Open Mainframe Project and Linux Foundation collaborative project, put in place to advocate and promote open source on the mainframe platform. I’m joined today by a good friend of the platform, Robert Garrett, currently an independent contractor. Robert, welcome to the show.
Robert Garrett: Thank you. Glad to be here.
Steven Dickens: So, reading a lot about the bio… But I think if we could just get started, Robert, could you just give the listeners a brief introduction to your career? I think there’s going to be many highlights and things to talk about here, but just get us started. Give us a view of your time on the platform, and then we’ll able to get from there.
Robert Garrett: Certainly. Yeah, I got into this business, really, by accident. I went to college and got a degree in mathematics and physics, never giving a thought to what I would do with that after I graduated. I had also taken some courses in computer science and programming. That was my second minor, and just fell into it by accident, but through a series of events, some of which are a little bit humorous, I wound up being a system programmer back in Lubbock, Texas in the late 1970s. But I’ve loved it ever since; I’ve always been the guy that can’t resist tearing something apart to see what makes it work. And so, this has been tailor-made for me, and there have been times when I almost feel guilty about being paid this much for something I enjoy doing so much, but I’ve tried to keep that a secret.
Steven Dickens: I love that. So, you must’ve seen the platform change massively over the years; just give me some perspective of how you got in, what those first few months and years looked like on the platform, and really what you experienced whilst getting onto the Mainframe for the first time.
Robert Garrett: Oh, definitely. And I have seen a lot of changes. The very first platform I worked on as an operator, which is where I started, was on good old DOS/VS Release 34 on a 370/145, and I believe it had maybe a whopping 768 kilobytes of memory on the whole machine. If you look at that now, how even your toaster has memory measured in gigabytes, so that’s been eye-opening for me. My wife tends to take all this stuff for granted. Every time I discover something that I can do, like when she’s driving down the road and I’m using the Wi-Fi system on my car to connect to somebody’s mainframe platform and do work for them while we’re watching the scenery go by. Some of those things just at this point blow me away because I remember how it used to be. She looks at me just matter of factly and says, “So what’s the big deal?” But it has been a big change over the years, and I think that will continue to be true.
Steven Dickens: So when I was looking at the byways that came through, you spent a lot of time on CICS. Can you talk to me a little bit about that journey you’ve been on, particularly with that part of the platform, and how you’ve seen that evolve over the years?
Robert Garrett: Oh, yes. Well, in the beginning, when I first started working with it, it was back in, I believe on CICS 1.1.2, which was two or three releases before the command level interface in the product was actually introduced. At that time, it was mostly COBOL; there was some PL/1 around and Assembler. Those were your three options, and it was all green screen interactive type transaction related programming. We did quite a few things with it, even back in the seventies where I worked with Lubbock. But you fast-forward to now look at it; it bears very little resemblance to the original although the underpaintings in the ancestry are still evident. Now CICS is mostly driven by the web. It’s a backend server in a lot of cases. It supports Java, it supports REXX, it supports Node.js. A lot of traffic comes in over MQ, and you look at the massive transaction volumes. (Editor’s Note: Worldwide, CICS systems process more than 1.6 million transactions every second of every day.) Robert Garrett: The last real job I had, which was working for one of the largest financial services firms in the world, every morning when the bell rang, and the market would open, it was a frightening thing to watch. We had, I forget exactly how many MIPS on the floor, but it was a lot, and our transaction volumes would go from nothing to 60-70-80,000 transactions per second in the space of one clock cycle. It was a scary thing to watch. But the amount of volume that gets processed through CICS and the reliability with which it processes all that stuff, it’s still just amazing. There’s nothing else in the world like it.
Steven Dickens: Give us… We may have some new listeners, maybe to the mainframe experts; what would that be supporting? You mentioned a large financial customer; obviously, you probably can’t say the name, but give us a flavor of what those transactions are doing for that bank?
Robert Garrett: In the financial market, what usually happens is that the markets are only open for a defined number of hours per day. But what happens in overnight trading is that people will get on, and they will queue these trades up, and they say, “I want to buy X shares of Y stock at the opening in the morning.” And so you’ll have this tremendous backlog of work that’s queued up to hit. Then when the bell finally rings in the mornings and the markets do open, all those transactions that have been queued up worldwide overnight come flooding in.
Robert Garrett: It was a real challenge with that particular customer because, if you’re familiar at all with Mainframe technology, PR/SM, logical partitioning, and all that sort of thing, there is a configuration that you can set where it will automatically provide a capacity on demand. We could not use that function because it did not react quickly enough to meet that market demand. It was literally going from idle to many thousands of transactions per second in the space of one clock cycle. It was a frightening thing to watch. That was where that volume came from. If you contrast that to inventory or manufacturing where their workload tends to be more even; the defining thing about this company was the spikiness of their volume; they had to size everything about their capacity just to get through the first 15 minutes of every day.
Steven Dickens: Wow. And how do you see that to the world change? I know with CICS we see new graphical interfaces, we see things like Zowe, most recently, that’s come through the project. What’s been your experience maybe over that last three to four years? Where do you see the platform go?
Robert Garrett: Well, there’s a big push still towards automation, 24 by seven operation, and zero downtime. That is a challenge for any mainframe environment. The way that CICS and other parts of the platform are addressing that of course is with parallel sysplex and dynamic configuration. Things that you used to have to take systems down for, you can now do dynamically. You’re able to shift workload around as you need to with things like sysplex optimized workload management. Everything is geared towards continuous operation. And also the mainframe talent pool continues to evaporate as people who are really experienced at this, they all look like me: they’re all old, and they’re retiring. So companies are concerned about where the next generation of support coming from. And I think that’s driving towards things like Zowe and other tools that attempt to simplify what it takes to maintain this environment.
Steven Dickens: And how are you seeing that transition? Personally, my journey on the platform over the last 10 years, I’ve seen a lot more younger college kids coming into the platform and that volume of younger talent, is that what you’re seeing?
Robert Garrett: I’m seeing some of it. There are … If you look at least here in the States, there aren’t that many colleges or universities that have any sort of mainframe curriculum at all. In fact, there’s my alma mater (I work with them providing free system programmer support in exchange for having access to their systems so I can work on development and such things) but there’s probably less than a dozen, maybe even less than a half a dozen universities in the whole country that have any sort of mainframe program. This is really upside down from the way it was when I went to school, back when dinosaurs roamed the earth. The way curriculum was chosen back then was that the business community pretty much dictated what colleges taught because of their needs. It’s really upside down now because it seems like businesses are adapting what they use to fit what colleges are producing. That’s what’s causing the drive towards Java and open systems, I believe. At least that’s my opinion anyway.
Steven Dickens: Yeah, it’s interesting, maybe questioned some of the numbers, but I think directionally, you’re not far off. So, interesting to get your perspective on where you see the platform going more generally; what are you starting to see grow? What are the big trends that you’re seeing going forward?
Robert Garrett: The most current version of z/OS has introduced support for containers. I expect that that trend will continue, that’s become very popular. I’ve just now started working with containers myself on Docker, and I’m still learning about a lot of it. But I’m quite surprised at how quickly you can spin something up and get it running. I haven’t really quite wrapped my mind around all of it yet, I’m still learning it, but that’s pretty impressive.
Steven Dickens: What have you thought about zCX, that’s new technology to the platform, but the container extension piece of … Have you got your hands on with that and started to have it played?
Robert Garrett: I haven’t played with it yet; it’s on my list of things to do and WT where I’m working now, they are on 2.4, so it’s available to me. It’s just hasn’t been something that I’ve gotten around to yet. But it is something that’s on my list of things to do, I do plan to explore that.
Steven Dickens: So are you going to be taking some of that experience you’ve had just in the Linux space over onto z/OS, is that the plan?
Robert Garrett: It cross-pollinates everywhere. I’m kind of a glutton for punishment; I do this for a living as I’ve been doing for 40 plus years now, and as if that’s not enough, I also have my own personal Windows domain network here at home that I brought up and support. That grew up really just out of frustration. I got tired of having to change my email address every time my ISP got bought or sold. So years ago I decided I was going to bring up my own email server so I wouldn’t have to do that anymore, and that ballooned into multiple redundant servers and stuff here. I guess I just can’t get enough of it.
Robert Garrett: It’s funny the way that you start to see parallels between the different architecture and the way things are done, and the way things are implemented and accomplished, the different sorts of things we all have to do in the IT world. It can be a little bit confusing sometimes, too, because I get my environments crossed up from time to time. I’ll be trying to do something on the Windows network that really comes from Linux or the mainframe, or vice versa. It’s hard to keep it all straight sometimes. But one thing that continues is that I do enjoy tinkering and playing with it all. I don’t think I’ll ever get tired of that.
Steven Dickens: Well, we need to get you to follow in the likes of Connor Krukosky and those guys who get me on the mainframe in the house by the sound of it, Robert.
Robert Garrett: Yeah. I’ve certainly got enough room out here to do it. My wife and I, we live out West of Fort Worth about 45 minutes to an hour out, on 13 and a half acres in a rural setting. So I’ve got the property out here to put something up if I can just get enough power out here to run it. That would be fun.
Steven Dickens: Yeah, you wouldn’t be the first. I think I saw somebody online recently talking about one of the college kids had just bought a platform and put it into his basement, and sort of got it stood up, so that’s at least two in the wild now I know that people have got in a home setting. So maybe that’s something to start to tinker with.
Robert Garrett: The challenge is that software is changing so fast, even with the CICS team. I’m very fortunate to be involved in their early support program for the product. For customers, there’s something that they call the design partnership on the customer side. But when I retired from my previous company, I had to be disconnected from that. But I’ve re-joined now that I have my own company, as a business partner, but it’s the same program where you get to be involved in setting the future direction of the product. One of the things they’ve been moving towards is more frequent releases, and I guess buying into the whole agile concept of making smaller changes, but making them more frequently.
Robert Garrett: So that’s really one thing that I’ve observed over the years. Stuff just keeps coming out faster, and you have to be able to move with it, set up your environment and your processes so that you can adapt. Even on the beta programs working with CICS, it’s very common to get a new beta every month or even sooner. If you’re going to be involved in that, you really need to have your installation processes fine-tuned so that you actually have time to work with the product and explore it, instead of spending several weeks trying to get it installed and configured.
Steven Dickens: Yeah, for sure. So, we’ve obviously got many younger listeners who listen to the show; you’ve got a fantastic track record and pedigree on the platform. What would you be saying to your younger self? So say you can go back to 25-year-old Robert, but you’ve got your experience and perspective now; what would you be saying to your younger self as you were embarking on your career Robert? What would the advice be?
Robert Garrett: The very first mentor I had, working with systems, was back when I was still an operator in Lubbock, Texas, beginning to get curious about the thing that system programmers did. We had a guy who himself was an outside consultant. His name was Donn Thornton. He was out of California, had an ego the size of Texas, but he had the skills to back it up. This guy was really good. So I started pumping him for information and trying to get him to explain things to me. He was really great about it. He would always give me just enough to get me started and enable me to go dig things out on my own, but he never would spoon-feed me about anything.
Robert Garrett: I was talking to him one day, and he told me, he said, “You know there are only two rules for system programmers. Number one is: Always make sure that you have a way to put it back exactly like it was before you touched it. Rule number two is: Go have fun.” And I have followed both of those my entire career. It’s a blast; I love exploring things. I love having these insights that allow me to take new concepts, new tools, and new technologies and figure out how they can be applied. There’s a feature that’s been in CICS for several years now called System Events; they’ve renamed it; it’s now called System Policies, and when I first learned about it, I was a little bit unsure of its use and what it could be done with it, but then I had this insight of, “Oh, these are really just exit points that you don’t have to write any code for, but that you can configure to do things.”
Robert Garrett: And once I figured that out, I started seeing all kinds of things I could do with them, and now I’m surprised that they haven’t taken off any more than they have. But that’s the thing: always be creative, look for different ways to apply the things that you run across, and try not to pay attention to all the naysayers that keep predicting that the mainframe is dead. They’ve been saying that for at least the last 20 plus years now, and the old boy, he’s still kicking around pretty good for somebody that’s dead. I don’t think that’ll ever happen. I think it’ll change. I think the role, how it’s used, and the place it has in business will continue to change. But the fact is there is no other platform that can handle the volume of work with the level of reliability that you get on the mainframe, and that’s just the truth.
Steven Dickens: So what I think “comes across there” is just your intellectual curiosity, Robert, and that, just passion for learning, and I think that’s really solid advice. You sound like you’ve had that level of curiosity for the last 30 plus years, and I think if I could give advice to my younger self, it would be to foster that curiosity. Would that be a fair statement, do you think?
Robert Garrett: Oh, definitely. My wife is always complaining that she never gets anything new because I always repair the old stuff, and I’ll tell her, “Honey, what you don’t realize is, what’s going on here are actually two personality defects. One is that I’m too dumb to know I shouldn’t take something apart. And then two is once I get it spread out all over the floor, I’m too stubborn to quit on it until I make it work.” And that’s really all that’s going on.
Steven Dickens: I love that. I’ve got a vision of you taking apart home appliances now, Robert. So that maybe that’s one for our listeners. As you start to think going forward, where do you see the platform? If I was to give you a crystal ball, where would you see the platform three to five years ahead? Where do you think we’re going?
Robert Garrett: Wow, but that’s a tough one to predict. A lot of that depends on software and reliability. That is the biggest advantage I believe that the mainframe holds over everything else, if you look at, and I don’t mean to be picking on anybody in particular but this is just one example; if you look at the news from the last, what is it, three or four now major upgrades that have been released by Microsoft for Windows 10, they’ve all had major problems, and it’s gotten to the point where people are afraid to even put one on because they don’t know what it’s going to break, and not without reason. And as long as… and it’s a challenge, it’s a challenge to meet the demands of agile deployment for more software, quicker, without sacrificing reliability. A lot of people seem to think that agile is a code word for you don’t have to test, which isn’t the case.
Robert Garrett: But I think as long as we can not sacrifice the advantage that the mainframe has in terms of reliability, as we pursue being able to be more responsive to change, that’s going to be the challenge. But if we can keep that going, I believe there’s always going to be a place for the mainframe; I think its role will change. I know of some companies that are already regretting their decision for having outsourced and trying to offshore a lot of their talent. They’re now actively bringing work back in-house, and they’re reinvesting in the platform. So I think that will continue to happen.
Robert Garrett: A lot of people want to try to run down the mainframe and say, “Well, it can’t possibly be good because, look at how old this is, this is an old thing.” But I mean, in an environment like we have today, if you have a mobile phone that’s even, I don’t know, one or two years old, it’s already hopelessly obsolete. But people don’t stop to think through, in a world where technology ages that fast, how does something manage to become as old as the mainframe is and still be viable? You don’t accomplish that by getting it wrong. So I think that people are going to begin to realize, as some companies already have, the advantage the platform enjoys. I hope that IBM as a company does not sacrifice the attributes and qualities of that platform that have always set it apart from everything else in their quest to try to make it adapt to popular culture, and the other things, the other technologies that seem to be emerging. They’re already ahead.
Steven Dickens: Well, it’s interesting. I know you’re a classic car guy from your bio, and the way I look at the mainframe platform is the same way I’d look at a Porsche 911. So maybe go with me on my analogy as a fellow car not; both the mainframe and the Porsche 911 were both launched in the 1960s, both in the same year. We’ve seen with both of those platforms that they get refreshed every couple of years; there’ll be a new model that comes out engines faster, knot 60 times faster. It will lack the Nurburgring faster than the previous version, keeps the core of what it means to be a Porsche 911, but nobody would look at the latest 2020 model of a Porsche 911 and go, “That’s a vintage car.” And I think they should do the same when they look at the latest, ed 15 and say, “That’s a modern computer.” In the same way, you’d look at it as a Porsche 911.” So I wondered whether you’d agree with me on that analogy, Robert?
Robert Garrett: Oh, definitely. If you compare the current Z series machines to the 360, and the 370, even the 30xx machines, there’s no comparison. They’re light years faster, less expensive, takes less power to run, more flexible, and all of those things, yet they still have the basic DNA that has been common to that platform all along. When you look at, and I’ll pick on Java here for a minute, how frequently are people required to go back in and revisit their applications because some function or some library or something else in the language has been deprecated and no longer supported, that happens fairly often. But I have code today that is still running at the object code, load model level that’s been running unchanged for the last 10 or 15 years. I’ve never had to touch it (or even recompile it). Part of that I’ll take credit for because it’s the way I design code, but part of that is inherent in the platform. They don’t sacrifice the old just to put in something new. And that’s a big advantage.
Steven Dickens: I think that’s how we’ve got to educate and position the platform that it’s got the cutting edge design, it’s got all the latest bells and whistles if you will, that you would need, but it’s also got that you described it as DNA and I’d like that description, that DNA that enables it to make the best of what is a heritage we need to look after and respect, but also merge it with the new. As we look to wrap up today, Robert, are there any other parting comments or things you want to share with the listeners? I think we’ve gone on a fantastic journey together over these last few moments, but is there anything else you’d like to add this as we start to think about wrapping up?
Robert Garrett: Gosh, I could go on and on as you’ve probably figured out by now, but I would just like to encourage people to try to take an objective look at the actual capabilities of the platform, what it can do today, combined with its track history of what it has been doing, and it continues to do for businesses since the sixties, and try not to fall into the trap of chasing the latest, new bright shiny object just because it’s new. This stuff works and it’s solid. And most businesses, they’re not in business to explore and play with the software; to them, it’s a tool to run their business, where it can support and pursue the objectives that the business actually has. And I don’t think there’s anything better for doing that.
Steven Dickens: Well, Robert, I think that’s been a fantastic way to summarize up. Thank you very much for your time. Ladies and gentlemen, my name’s Steven Dickens. You’ve been listening to the “I Am A Mainframer” podcast, brought to you by the open mainframe project. Please click and subscribe. We’d love a five-star review if you could take a moment, and we’ll welcome you to the next show, which will hopefully be up soon. Thanks so much again.
Robert Garrett: Thank you.
