The Linux Foundation Projects
Skip to main content
Blog | I Am A Mainframer

I am a Mainframer: Vaughn Marshall

By | March 24, 2020

In today’s episode of the “I Am A Mainframer” podcast, Steven Dickens sits down with Vaughn Marshall, Sr Principal Product Manager at Broadcom Inc.  On this podcast, Vaugh discusses his journey with the mainframe, Zowe, and where they see the Mainframe going in the future.

Steven Dickens: Hello. My name’s Steven Dickens and you’re joining us for the I Am A Mainframer podcast. I’m joined today by Vaughn Marshall from Broadcom, who’s going to tell us a little bit about his journey on the platform and some of the cool projects he’s working on. So Vaughn, welcome to the show.

Vaughn Marshall: Thank you.

Steven Dickens: Vaughn, I always start by asking our guests on the show to tell us a little bit about kind of your personal journey. Just introduce yourself, give us a flavor of kind of how you’ve ended up on the platform and really get us orientated so the listeners can get started and start to connect with you and understand your journey.

Vaughn Marshall: Sure. So, I would say I have maybe a little bit of an atypical background in the terms of the mainframe platform. I’d say at the very beginning of my career, I was definitely working on mainframes, but it was pretty short-lived. This was probably in the early nineties and at that point in time, everybody was saying mainframes were on their way out. And certainly, it was, everybody was working on mainframes. That was the main thing. But what was becoming really, really prevalent or starting to was open system’s development. And so I actually ended up, through the company I was working with, working on open systems platform when we, we’re working on new technology. And so that was my introduction to the world of mainframe and I didn’t really end up back on the platform until much later.

Vaughn Marshall: Probably in I would say the 2010 timeframe. And that’s because some software that I had developed that was part of an application integration platform was acquired by what was CA Technologies now Broadcom. And ultimately that piece of software kind of made it in the CA Endevor product suite, which I am now the product manager for. So, a little bit different, but I certainly have always appreciated the mainframe as a platform that can process a lot of transactions. It’s very powerful and it’s just been a bit of a divergent journey for me to get back to the mainframe. But it’s somewhere where I started my career.

Steven Dickens: So maybe let’s pull on that thread a little Vaughn for analysis and maybe dig into what was, I suppose the reentry back into the mainframe space like. I’ve heard people describe that they were surprised about how vibrant and modern the platform was when they’d had time away. Was that your experience and maybe just give us your own perspective.

Vaughn Marshall: Yeah, absolutely. So, I wouldn’t say necessarily I was surprised by how vibrant it was, but when I was coming back into the world of mainframe, a lot of effort was being made to modernize the tools that were there for the mainframe. So for me personally, it was actually like a little bit like riding a bike. People who didn’t know me well didn’t know I had started out in mainframe, were very surprised to see that I could navigate a 3270 pretty quickly and efficiently and just thought I was a super fast learner. But definitely, once I got into the space, a lot of the work that I saw going on around me was modernizing the tools and the way people are interacting with the eye to essentially looking at the next generation of mainframe developers.

Vaughn Marshall: In terms of the platform itself, it’s always been something that seems to be pretty timely in terms of the technologies and spaces, sorry, the technologies that are being used on the platform. So things like pervasive encryption, container technologies, all these things that are considered pretty much on the cutting edge in the world. But distributed development are still things you encounter in the world of the mainframe. So I guess that was a little surprising. But definitely one thing that struck me was that the tools, at least when I started, were pretty much still 3270 oriented. But that’s changing really quickly.

Steven Dickens: And yeah, it’s interesting that you mentioned that obviously one of our projects under the open mainframe project is so easy. So maybe give us your personal perspective of how you see that 3270 to kind of modern interfaces today and maybe speak from the experience system of the Broadcom customers and kind of how you’re seeing them make that transition.

Vaughn Marshall: Yeah, there’s a huge amount of interest in Zowe and I certainly am part of that open mainframe project. So definitely got a great feel for that. You know, the way I look at it, it’s actually kind of interesting. It almost is a parallel to what we’ve seen happening in the world of applications, right? So if we look at applications, we’ve seen new technologies come along and really it’s almost like sedimentary rock. There’s just new layers coming on top of new layers and old ones don’t really go away. You just add new ways of interfacing. And if I look at the tools themselves, I’m seeing almost a similar sort of thing happen. Right?

Vaughn Marshall: So we’ve got the mainframe tools, which typically if you look at the things people are using on the mainframe, they’ve been there for many, many years. They have processes built up around them. They have all sorts of things that are embedded that make them hard to replace. And so just like when we were looking at modernizing our applications and we were starting to build web applications and now mobile applications where we add layers on top of everything, but it’s very painful to replace the core technologies that are under them. It’s the same thing that seems to be happening in terms of tools. So Zowe is definitely one of those things, right? People are changing the way that they interact with the mainframe. There’re new layers, whether it’s the Zowe desktop or Zowe CLI. There’re new ways of interacting with modern interfaces that kind of opened the mainframe up to that next generation of developers.

Steven Dickens: I think the way you’ve just described there with sedimentary rock, it’s a really good analogy that people and the listeners particularly will be able to get their head around of this layering, as you mentioned. Of technology on top of technology. I think I see a really interesting dynamic around people having more success with that approach than a rip and replace. I mean would that, would that be something you’d see from the Broadcom side?

Vaughn Marshall: Yeah, 100% it makes no sense. Rip and replace is basically reinventing the wheel over and over again. And I’m not saying that you can’t improve things, but it’s definitely a lot easier to stomach in terms of costs, in terms of risk if you’re not constantly restarting, right? If you’re taking the best of what’s there and maybe changing it to appeal to the next generation or so that you can achieve certain goals or integrations. If you can do that with minimal disruption, that’s the best way forward. That’s the best way forward in my opinion.

Steven Dickens: Yeah, I would tend to agree. So, I mean, let’s maybe do a shameless plug for the products that you’re responsible for within, from Broadcom and really just trying to understand when you talk about product manager for the mainframe application development business, kind of what that means. What the cool things you’re working on, maybe what would be of most interest to the listeners. We’ve got a pretty tight crowd of mainframe listeners here, so maybe just give us a view of kind of what’s the cool stuff you’re working on right now?

Vaughn Marshall: Certainly so shameless plug, right? So I’m working with CA Endevor and that is for those of you that aren’t familiar with it, it is a, we call it the original DevOps tool. So CA Endevor, actually, if you’ve seen it spelled out, it’s E-N-D-E-V-O-R, and not what people usually use when spelling the word Endevor as into strive forward. But the reason for that it’s an acronym, it’s not a word.

Steven Dickens: Oh, I didn’t know that. That’s news to me. I used to work for CA back in the late nineties and I didn’t know that. So you’re going to have to elaborate Vaughn. What does Endevor mean? I’m going to learn something here.

Vaughn Marshall: There you go. It stands for the environment for development and operations.

Steven Dickens: Environment for development and operations? Okay.

Vaughn Marshall: Yes, it sounds pretty close to that buzzword everybody’s saying these days, DevOps. So I like to say, oh yes, DevOps, like everything else, it was invented on the mainframe 30 years ago. But anyway, that’s so you know what we’re trying to do when we talk to customers is really figure out how they can take that investment in mainframe DevOps that they actually have and preserve it and make it relevant in the year 2020. It’s actually if you boil it down to what people have done for many many years and looked at the way people work with our tool Endevor, they’re doing a lot of the same practices that are done with modern DevOps. So for instance, I make a change. It automatically builds, you know, when I package things up, they start promoting through a defined life cycle, which to me is just a pipeline.

Vaughn Marshall: The thing that I find is, all of these things are happening, but when you start talking to DevOps center of excellence architects and next-generation developers, the view of what’s happening, the 3270 view becomes something that is a barrier for them to understanding that this is really all just the same thing. So the more that you can interface into the tools they’re more familiar with, the more that they understand that this is something that can fit quite easily into the world of DevOps. So that’s what we’re doing. And you mentioned Zowe, CA Endevor was one of the products that were first out the gate with the Zowe CLI plugin. So Zowe CLI or command-line interface is basically a framework for the mainframe that allows you to essentially interact with mainframe and mainframe products through a command-line interface.

Vaughn Marshall: Of course that’s another head-scratcher for us old guys. Cause I remember when we were all getting rid of our command-line interfaces, but they’re making a comeback these days. And of course the reason being, it’s much easier to script a command-line interface than it is a GUI. So basically when we start thinking about DevOps and things like that, we want to think about easy automation and something that runs universally in all sorts of different DevOps tools. So Zowe CLI has been one of the things we’ve done to modernize our interface into Endevor, making it scriptable in all sorts of different DevOps tools. And from there that means you can put basically pipeline views on top of what’s already happening and has been happening in Endevor and still preserve your investment. So that’s one thing we’ve been working on. I could go on.

Steven Dickens: It’s interesting that transition, as you mentioned, back to CLI. I remember as I say, I used to work at CA back in the late nineties and it was a move to GUI interfaces and everybody was going to move away from command-line and we seem to be seeing it hover in a sense. What would you put that down to and, I mean you obviously got that level of experience and see that dynamic firsthand. Maybe just you know, for some of our older listeners who’ve seen that transition and that trend or maybe for some of the new listeners and I know we’ve got a community that listens to us that would fit into that space. Maybe just give your perspective on that transition around the command-line.

Vaughn Marshall: So it’s a few things, for the most part, I would say number one is when you’re looking at a command-line interface, it’s something that you can script and make your own. So you can make new commands. It’s basically an API that you can also use interactively. So that scriptability and that natural fit into the world of automation is probably the number one reason. But it’s also more than that. If I look at people that are working on Linux or Unix technologies, they’re very comfortable in the world of command-line. And if you’re a developer, once you get good at something, you get very fast at it if you can keep your hands on the keyboard, right? You don’t have to go and reach over to a mouse and click something.

Vaughn Marshall: So you see the most productive developers and they’re just whizzing away on their keyboards. And so in that sense a command-line interface is also a very developer-friendly, advanced user interface. I won’t say that maybe they have the intuitiveness of a GUI, but once you know the command that you’re working with, you can issue it really quickly. And for that reason, developers really like command-line interfaces. Not to mention again its familiarity when they’re working on Unix systems through secure shell, it’s all command line. So it just becomes something that they’re very used to. So that next-generation developer is very, very used to using command lines and we’re starting to see a Renaissance or a comeback with command-line interfaces.

Steven Dickens: Yeah, I’d agree with that. I mean I see almost a bifurcation. You’ve almost got the low code, no code guys out there who sort of building composable objects into code and that’s happening at the business sort of line of business level layer. And then as you say, you’ve got to renaissance back to the command line for the sort of pure developer so it’s almost, I see those two streams. Would you agree? Would you sort of see the same dynamic?

Vaughn Marshall: Oh, absolutely. Absolutely. There are definitely people that like to get low level and the command line is probably in their wheelhouse.

Steven Dickens: Okay. So I mean let’s pivot maybe a little bit, start to look ahead. Where are you picking up any big trends, you’re picking up kind of feedback as you engage with the Broadcom client base around where they’re seeing the platform as they look ahead.

Vaughn Marshall: Yeah, absolutely. So the interesting thing is that is, like I said, when I started my career in mainframe, I was there for a bit. But the big thing that I had heard probably for 20 years was that the platform won’t be here for very long. And you know, I heard that drum get beat over and over and over. Mainframe’s going away, it’s a dinosaur, et cetera. But I’m actually starting to hear very, very different things now. Certainly, there’s probably been certain classes of companies that have got off the mainframe. Those are the guys that aren’t on the platform anymore. But if you look at companies where there’re massive amounts of transaction processing going on still, banks, insurers, governments, those types of things, a lot of them are starting to recommit to the mainframe. So that’s very interesting. And in that sense, in recommitting, they’re also looking at getting that next-generation onboard.

Vaughn Marshall: So what I see happening in terms of the future like you’ve said that the outset, the platform itself is pretty vibrant, there are lots of different things. But now what we’re facing is how do we get the next generation of developers involved in, happy to work in the platform and make it their career. And so what I’m predicting and seeing a would be a Renaissance in terms of the tools that are available for the mainframe. I think that type of thing is going to explode as well. You know, if we look at, we’re kind of in a transition period where you’ve got the constituency that that is very, very happy with their 3270 environments and don’t want to necessarily change that, but you’ve also got that next generation of developers that want to use all of the latest and greatest tools. So we’re kind of in that hybrid. So I think it’s going to be what’s going to really change in the platform, it’s going to be the tools that are in use.

Steven Dickens: Yeah. And I always ask this question of our guests, if you were to look into a crystal ball and predict the future sort of two, three, maybe four years out, what would you see for the mainframe as a platform in that sort of time horizon?

Vaughn Marshall: Yeah, that’s a good question. Well, I think if you walked around the shop, you’re going to probably see a very different crowd of people, a lot younger. I’m actually starting to see that at offices like Share. It’s definitely a lot more younger showing up. And the other thing that’ll be interesting too is whether the mainframe will stay kind of in this silo. I would like to hope that in the next little while there’ll be more of a, it’ll be integrated more into let’s say the regular development areas. So you’d have more full-stack developers where they think about their applications, not so much in terms of platforms in which parts of their applications run on which platforms, but as a holistic thing.

Vaughn Marshall: So you know, if I want to make a change to one of my front ends, I’m also equally comfortable and able through the tools to go into the backend and make those sorts of changes as well, rather than having to have these one section of the building where the mainframers live and then there’s everybody else. That’s what I’d like to see happen, and I think I’m starting to see that and I think it will be a bit to get there. But it’s certainly the more and more we train the next generation and make them comfortable in the platform and make the platform more like other platforms, the more that I see that happening.

Steven Dickens: Yeah, I’d agree. I’m starting to see that same blurring of the line. It’s more this is the developer and now it’s more independent of the platform.

Vaughn Marshall: Yep. And, and funny enough too, what I’m also seeing is it’s not just the newcomers coming in and learning from the old guard, but really also bringing new technologies to the platform. Zowe is a case in point right? That actually in our shop, Zowe CLI was something that was done by one of our teams of developers during one of their innovation sprints, and why they did that just because they wanted it to be able to use some of the tools that they were used to using in their careers. And so they enabled that through a building a CLY for themselves. It wasn’t a product or anything part of the open mainframe at that point in time. It was just that next generation of developers writing tools they wanted to use.

Steven Dickens: And that’s for me one of the really interesting trends seeing this platform open up the community we’ve built with the open mainframe project. Really just what was probably the poster child for a closed system sort of 10, 15 years ago has just opened up Linux come into the platform, Java and other sorts of restful APIs come into the platform, opening it up to different developer mindsets. You know, most recently with things like Zowe as you mentioned, and I think that the whole openness of the platform, it is very different from where it was just even a few years ago.

Vaughn Marshall: Yep. I would 100% agree and in fact, I think that having an open platform is critical. Not only that and having a community around that is critical to making things prevalent. It’s kind of the old beta versus VHS analogy, right. If you make something closed and proprietary then it’s not going to be as widely adopted. And certainly, I think the mainframe definitely needs to go more in that direction. It needs to be something that anybody can pick up and write code for. I’m not sure if IBM has plans around that, but they should, whether it’s emulators, I think emulated environments, things that make mainframe and mainframe technologies accessible will also make it more pervasive.

Steven Dickens: Fantastic. Vaughn, is there any of those sort of parting comments before we start to think about wrapping up?

Vaughn Marshall: No, I don’t really think I have any, I will say that for me, for the journey, it’s been actually a surprise ending up in, in a mainframe again, but a pleasant one. It’s one thing that I was reminded of as I joined the platform is the passion that the people who work there have for it. So you know, that’s unique when you go to events where you encounter mainframers they’re always a very tight community and that’s actually great. I like that a lot.

Steven Dickens: Yeah, I’ve picked that same experience while I was up at share and not this week, the week before and it’s always a tight-knit group. Very sort of keen to talk across vendors, across clients. You certainly feel part that kind of fraternity, that club if you will. And I think that’s, it’s got a different dynamic to it as a space.

Vaughn Marshall: Yep. Absolutely.

Steven Dickens: So Vaughn, that’s been a fantastic conversation. I think our listeners will have really enjoyed the show. Thank you for joining us.

Vaughn Marshall: My pleasure.

Steven Dickens: So that was Steven Dickens talking to Vaughn Marshall on the I Am A Mainframer podcast for the open mainframe project. If you liked today’s show, please click and subscribe and there are numerous other shows we can get you plugged into via a subscription. So please hit the subscribe button and come and join us on the I Am a Mainframer show for the open mainframe project.