Open Mainframe Project Opens Mainframe Infrastructure to Open Source

Hey everyone, welcome back here to Techstrong TV. My next guest is an old friend of mine. Him and I used to do the Open Mainframe podcast together.

His name's John Murdick. John is the Director of Program Management for the Linux Foundation, but also, John, still Executive Director of Open Mainframe Project or no? I help, I definitely help work and lead with this community, along with a number of other communities here in various different vertical spaces.

But yeah, I've been 11 years strong working with the mainframe ecosystem. Good for you, man. God bless you.

Yeah. John, beyond 11 years of working with the Mainframe, you're also an author. Yes.

A bit of an open source expert. Yes. I don't want to tell your story.

It's not going to over very well, but we'll do that for the shameless plug here. Okay. Available at great online book sellers and everything.

Yeah, my expertise, but also just a lot of my passion, just working with open source communities. They've been really one of the really joys of my career overall. I've got to meet some really interesting people doing some really interesting things, and the idea of collaborating like this is not first nature to everybody.

And there's a lot of nuances to really doing it well, and so I just love to partner with these projects, these communities, these ecosystems, and sort of help them achieve what they're looking to do, is build great technology. They're the stars of the show here. They're the ones that know these spaces inside and out.

We come in, and we help them figure out how you do this in open source. And that's really, myself along with the larger Linux Foundation, our value proposition. Very cool.

John, let's talk Open Mainframe Project. Yep. I mentioned it now three or four times.

I take it for granted because you and I did the podcast for, hell, probably more than a year, maybe two. Oh, yeah. And so I'm intimately familiar with it, but not everyone out here is.

No. Where they are, I think everyone out here knows the Linux Foundation, certainly. But for people who aren't familiar with the Open Mainframe Project, how do you describe it to them?

So the Open Mainframe Project, the aim there is a collaboration space for open source on the mainframe. It started around Linux on Z specifically. If you remember again, about 10, 11 years ago, IBM did a big announcement around the LinuxONE hardware, along with a number of commitments they had there.

Open Mainframe Project was one of those. Since then, and it's really interesting because while the project started then, the concept of open source goes back decades, and in fact, as we look at open source in general, we trace it back to the mainframe ecosystem. It's the same joke than ever.

You talk with technologists all over the place, and every single technology we have traces back to the mainframers. And so does open source with Share in the 1950s was effectively preceded all of the work that open source was. Getting together, collaborating, sharing code, sharing insights.

It became more a name in that decades later, but that's really where we're at. So the concept has always been there. Where this has really given is a neutral place to actually help do this.

And it started in the Linux space, but it's grown into z/OS and COBOL over the years. And really now is the focal point for anybody that's looking to do cross-vendor, cross-industry development. They're coming here to say, "Hey, this is the place that we do it," based upon the reputation the Linux Foundation has of hosting efforts like this in just about every horizontal and vertical market out there.

Energy, motion picture, networking, you name it. And cloud tags, AI, all of those things, all of those are hosted here as well because of the expertise. And that's the same rationale where the Open Mainframe Project comes in.

So yeah, so we have a lot of great collaboration. But I want to emphasize, it's not meant as a competitor to the Share. Right.

As a matter of fact, you guys put on a big thing at Share usually too, right? Yeah. We do a number of tracks at Share.

We do a community data tech exchange. We're doing stuff with GS Forge event, which was the previous GS UK event. So, we participate in a lot of things, and you're exactly right, this isn't a competition with them.

It's a niche audience, and I think there's a lot of different needs everyone serves. And we see that in a lot of different industries, where this, again, our strength is how to do open source code collaboration and open source standards and other sort of collaboration. That's the sweet spot here, and that's really the focus that we really build a lot of work on.

I think another important thing is, and you and I have had this conversation before. Right? We're in what I call the foundational era of open source, right?

Where most of the most successful open source projects are managed by foundations like the Linux Foundation. And the Linux Foundation, I forgot, there's like 40 daughter foundations or organizations underneath it, right? More than that now.

Yeah. But yeah. Covering all of these different areas.

But one of the beauties of it is it's a safe space for companies that would otherwise be competitors to collaborate. Mm-hmm. And that's an interesting thing, right?

It's kind of like a United Nations sort of thing, right? And whether you're a fan of the UN or not, we haven't had a World War since we had the UN, right? Sure.

That's one way of looking at it. But that being said, the Linux Foundation and the Open Mainframe Project flowing from it, serves as a place where otherwise competitive companies can collaborate and work together, that rising tide lifts all boats. And from that common open source foundation, people can soar as high as their talent, skills, investments will let them.

That's people and companies. Mm-hmm. So it's a great model.

But John, I want to turn specifically to the Open Frame Project, right? Because this is something I learned firsthand doing this with you, which is the Open Frame Project itself is a baby Linux Foundation in many ways, in that it has many different projects which it manages. Mm-hmm.

You didn't mention Zowe, I noticed, but- Yeah ... a lot of people, they just know Zowe and think that's what the Open Mainframe Project. That's the only thing that, much like Linux Foundation and Linux, or people may think that Kubernetes is the CNCF, when they have 220 or whatever it is- Right ...

projects in CNCF. Zowe, Open Zowe, whatever you want to, is just one- Mm-hmm ... of, what is it, 20 something projects- Yeah ...

within Open Mainframe? Tons of different efforts in there, and I think what's actually interesting about Zowe, and a lot of people don't realize is not only is it kind of, I would say, an anchor project here in Open Mainframe Project, it's probably one of the largest projects in open source in general. CNCF, you're probably familiar with the Velocity reports that they do from time to time, and Zowe is always in the top echelon, if not very much towards the top in Velocity commits, things like that, not just across Linux Foundation, but across open source in general.

So it's a very successful project. Absolutely. Excellent.

All right. Enough background, John. We got to jump into what we want to talk about today, and that's the Open Mainframe Project's new mainframe infrastructure.

Mm-hmm. Talk to us about that. So I've been with this project for 11 years.

The same question I've got every, well, probably almost every day feels like over the last 11 years is, hey, it's great that you have a project that's supporting open source in this mainframe. How can I actually go use this hardware to write open source? And that's been a constant challenge.

I mean, that's not something new to us. Groups before and everyone that's been trying to do this have had that same challenge. And honestly, it's just because really running a mainframe environment, it's not like going to Micro Center or Best Buy or something like that and going buying server hardware and throwing it up in a rack somewhere and calling it good.

There's a lot of specialty work that goes into setting these up, maintaining them. They're designed for enterprise deployments and you know from the enterprise world, deploying enterprise software is not three clicks in an install wizard. It's a lot of work.

Even with AI. Even with AI. Yeah.

" Uh-huh. So it's hard to do. They're expensive machines and we often get feedback of, well, how come this customer has it?

Could they use some of theirs or whatever? These machines often aren't even connected to the Internet. They're completely air-gapped.

So it's been a huge challenge over the years, and there's always been some bespoke ways of solving it. We've had Viacom Infinity, who's a part of Converge Technologies now. They've provided some resources here and there.

We've had other groups, Marius has provided some resources here and there. But it's never really been-- It's always been a knock on the right door and find out who you know, and then you can maybe get some access. And even then it's a little bit spotty.

It isn't always consistent. IBM, for the Linux side, had the LinuxONE Community Cloud, which was really good, but again, that was Linux only, and the vast majority of workloads on Mainframe are z/OS. So that didn't help that side of the house.

So this is something we've been working on probably at least for all these 11 years, I would say very actively, at least for the last five or six years, of how can we have infrastructure available that any open source project can come apply for that they can use to basically help develop support open source for z/OS and z Linux. And it's been an interesting journey through all of this. We've had a lot of great partners along the way.

Broadcom partnered with a lot of hardware and expertise. We've had a number of our other members that have pledged to bring software and other tools to this. And IBM has come to the table here with a lot of background infrastructure and things of that nature.

And so now we're excited to say, hey, when somebody comes with a question of where can I get this mainframe to try out, use my open source or use my project or use it as a part of a CI for my builds, we have it for you. And that's really been the long goal that we've been on these years, and it's a huge accomplishment for this community as a whole because this has just been the struggle point for decades as open source has tried to support this. You know, John, I'm thinking back, I think it was maybe towards the end of the run of the podcast- Mm-hmm ...

that the announcement was made that this mainframe had been donated. Mm-hmm. And then, so that's got to be what, two years ago?

More than two years ago, I think. Three? It's been a while.

Yeah, it's been a while. Right. And it's taken this long because, okay, now you got a mainframe.

Where do I keep it? In my closet? Literally in my closet?

No, it doesn't fit in your closet. Yeah. No, my wife would not like it if it was in our closet.

No, probably right. We'd have to move all her clothes out. So you had to have it housed, and it had to be set up, and then it had to be made accessible, but it had to be secure because this is a little different than your average mainframe access, right?

Yeah. Where it's open to the public, so to speak. Mm-hmm.

And setting up a mainframe's not trivial to begin with. Right. But setting up one for this sort of mission, really, I don't know if there was a manual to cover for this one.

No, not even on the IBM side. I think we were very much blazing the trail on a lot of different fronts. I think honestly, that's what's taken so long, is all of those things of just the logistical and operational side, then also licensing and things like that.

There's a ton of commercial software that you have-- z/OS is a commercial operating system. It's proprietary software. There's a lot of licensing that needs to go into place just to give people access to it.

And it's not even necessarily a cost question, it's just the way that the license for those things are structured to make that happen. So there's just so much work in just figuring all of this out, how it's done, because again there's never been the concept of how do we make something publicly accessible to someone who is not an employee of a given organization, right? And that has just been a huge turning point on so many fronts.

Again, the licensing, the legal aspect, or like you said, the operational, the infrastructure, all of those things. How do we make it accessible? So it's, yeah, it's taken every day of these last several years to figure all of this out and figure out the way to go about it.

But we're at the end of the journey here, and we're having projects that are up and going on this hardware, and it's super exciting to see. Hey, the good news is, think how easy it'll be when you have to set up the second mainframe. That's true.

Everything now is very trivial from here. And that's actually why it's great to have good partners because now that we kind of have the entire runway built of how to do this, now hardware ringing in is trivial. That's not a hard thing to do, and we have partners that we can work with very quickly on that.

And we have a lot of flexibility in what we do. Most open source projects are going to say, yeah, I need a mainframe. They're not going to know how many LPARs do I need, and all of these things.

The concept of LPARs they're not going to have any idea about. " They have a standard thing they pick from, right? But we also have the same flexibility that, hey, if your project is needing to use specialized items or have specialized access, we have ways to accommodate that as well because we have so much of the expertise from our member companies that are coming here that can help work through those requirements.

But John, as big as the mainframe is, and the load that it can handle, there's still a limit, right, to how many different projects, uses, what have you, users. How are you guys determining who gets access and how much access and so forth? I think the big thing that we're looking for is open source projects that are in an active area of wanting to support the mainframe.

So not sort of the, hey, I want to have this spot, but I'm kind of like, I don't know, mainframe resource squatting, if you would, on it. You're in a position where you can start executing on it. And we've had a number of projects that have started to come to us with that sort of understanding of like, hey, we're at the point now we're ready to do it.

So I think that's a major aspect to it. Clearly, we're looking for projects that are looking to grow the ecosystem and add value to the ecosystem. So being conscious and there's always a lot of work.

There's a lot of groups out there that are working to migrate workloads off the mainframe. We're not looking for projects in that realm. We're looking for projects that are making the mainframe a much more attractive platform, so ones that kind of continue to help grow the ecosystem.

Outside of that, I think the flexibility of what it looks like is pretty broad, and we've had some really broad conversations with vendors and open source projects wanting to do specific different things. Again, this can't be used for a commercial. So if you're a vendor that's wanting to build a product, you can't come apply to us for resources.

Go to IBM's ISP program. That's what that's for. We're just strictly for open source projects and open source collaboration.

But otherwise, in terms of capacity and things like that, we've got tons of it. And mainframes, not only are they designed for very high capacity, but most of them are designed for a very transactional workload that's very consistent over time. So you have just a huge volume of transactions always coming through.

CI workloads and testing workloads, they're lumpy all over the place. So we have a lot of flexibility in there that we're able to punch a lot into a single infrastructure and even branch out to others. So we really don't have a cap of number of projects that we can support.

It's sort of more of we want to make sure that if a project wants to use it, that they're in a position where they can use it fairly quickly. Not just, hey, I want to get it and then six months, nine months, 12 months later, maybe I'll get around to something. Sure.

John, where can people go to find out more and maybe if they have a use or they're thinking about it? Yeah. So our TAC is the one that oversees which projects and which groups get access.

They have a whole application process that they can go through here. I'll drop a link here in the chat that maybe you can put in the show notes or whatever that include there. It's really simple.

Really, there's a form you go through and put all your information, which you're looking to use it for, all of that. We triage it, we get it in front of our TAC to kind of take a look at it. A lot of times they have some questions and things like that of what it looks like.

And then from there, the TAC approves, and then we sort of pass off to the backend teams that work on getting this up and going for you. So it's very, very simple. And again, I highly encourage any projects out there that have had requests around, hey, I wish I could-- Do you support mainframe?

You have interest in it, things like that. Go check this out. I guess the big thing I would advise is making sure that there are some resources that sort of a bit of understanding of how to support it.

But we also have a large community of folks that are happy to partner with your projects and work with you. So that's another really good option there. Good stuff.

Hey, John, I know what a big deal this was for you and all at May and all the folks at The Open Mainframe Project. Congratulations on getting this spun up now and fully functional. You guys do great work.

Keep it up and come back and visit us. I miss you. I know.

It's been too long. It's great. I think the big things to this is just all of the entire mainframe ecosystem, IBM, Broadcom, just a bunch of folks that have came together to make this happen.

And I think everyone had the intention of we want to make this work, and now it's kind of come to fruition, and I think this is just a huge taking off point for this ecosystem of really being able to embrace open source at a large scale. But yeah. We'll have stuff coming down the road, Alan.

We'll have to chat more. All righty. John Mertic, Director of Program Management, Linux Foundation, key player in The Open Mainframe Project.

Their new mainframe infrastructure is available. You're watching techstrong TV. We'll be back in a moment.