20250828 The Opportunity for OpenShift Infrastructure

1:15PM Sep 17, 2025

Speakers:

Rob Hirschfeld

Shane Gibson

Keywords:

OpenShift

Kubernetes

virtualization

VMware

Red Hat

cluster manager

enterprise adoption

AI readiness

infrastructure abstraction

cost reduction

migration path

strategic benefits

IT strategy

RackN

cloud 2030.

Hello, I'm Rob Hirschfeld, CEO and co founder of RackN, and your host for the cloud 2030 podcast. In this episode, we talk about something that is deep in rackns wheelhouse, which is the adoption and the challenges for adoption of OpenShift virtualization. In this discussion, we really dig into what the hurdles are, and ultimately get to what the strategic benefits and positioning are around this migration from virtualization focused infrastructure to Container managed, focused infrastructure, so with VMs, but the we're in a generational switch to using Kubernetes as the infrastructure manager, infrastructure abstraction layer. But that doesn't necessarily mean we're just going to dump everything into Kubernetes and be done with it. There's a lot more challenges to that, and this is something that my company, RackN, is specifically trying to help people with. We're at the crux of this, and you will hear quite a bit of positioning here in the discussion around RackN, and what RackN should do. If that troubles you, I'm sorry. I think it's very easy to substitute in your mission, your drive to think about how we're treating the industry more generally, and how the C suite in your organization should be thinking about it. That's ultimately what we're discussing here, and I know you'll enjoy the discussion.

It's becoming very clear to me in our engagements in enterprise that OpenShift is the strategic replacement for VMware and OpenShift virtualization, people are going to migrate our fortune. 500 accounts are looking at OpenShift virtualization as the migration path, rather than new, new tent, the way it's the way it's shaking out, and it's worth a recap, is there's VMware, which is very, you know, going up dramatically in price. Nutanix, which is the same price as VMware and OpenShift virtualization, which, if you do it the way Red Hat wants, it is also just as expensive. They have licensed approaches that aren't as expensive, like a fifth of the price. And so one of the conflicts that we see as Red Hat is in promoting the most expensive approach to our customers, and some of some of them are rejecting that, and they're looking for the price reduction. And we can talk through, you know, we'll, we'll talk through that, but

really quickly, what's the basis for the price reduction? What? What is it that you don't get?

It's actually the key. It's a key question. So if you buy open, if you license open shift virtualization, then, then you can, you can buy a license for that, and it's a fifth of the price of the VMware. But Red Hat is, is promoting people doing this using their their advanced cluster manager to handle cluster life cycle. And so when cluster manager on a per core basis is very expensive because it's designed for Kubernetes development teams. And this is, this is actually going to lay into some of the other dilemma that I want to talk through, yeah, if, if you're a Kubernetes developer, and you're delivering applications using Kubernetes open show, and I'm just going to say open shift, because this is, in a lot of ways, an open shift conversation. We're in the Kubernetes conversation because for our customers, they want the license. They want the supported license. Well, an open shift is a whole vertical stack, right? The problem with, yeah, sure, there's no,

well, you get this, you've got, you've got a support, you've got the whole support backing backing

up. So you need the storage pieces. You need the V motion stuff. You need the sand, right? You Need to Know Your gateways, right? It is a a ridiculously sprawling landscape of technologies. And one of the things that OpenShift has done successfully is pick those technologies and support them, right, and then integrate them into it, into a thing, and then, because this is the thing that really scares enterprises about this, this product, especially compared to VMware, is, you know, you know, you go in and you're like, Well, I don't want to use this I don't want to use this gateway. I want to use that gateway. And now the whole suite that you have is out of support, and you're on your own for it, because you've you, you know, changing one service inside of the Kubernetes thing actually changes the behavior of the whole cluster. So it has to be an integrated tool.

Delivery. So quick question, is anybody delivering the result that you would get by, you know, swapping out some aspect of open of open shift, is anybody delivering? I'm going to say bespoke configurations, probably with some sort of automation to kind of bring it all in into place, but then providing ongoing support for it, because it sounds like it's being driven Well, I know, I know a company that I need to introduce, I need to introduce you to.

So go ahead. I am, I am, literally, my team just spent an hour talking about part of this. What is the because this is a pretty direction. It's not the alternative, actually, it's, how do you re How do you what do enterprises need to do to get over the step function to start doing this work? It's right, there's a degree of what, what hardware do I buy? How do I configure it? How do I install it? How do I manage it? There's also a All right, well, what is my v San look like? What does my NSX look like? What is my VM admin system look like? So, and one of the challenges is it's not, it's it still is more Kubernetes E and you still have to understand the Kubernetes constructs, because network segmentation in OpenShift and Kubernetes in general, for VMs is tied to pods and namespaces. And so when you you have to be able to, if you want network segmentation for a group of VMs, you have to give them up on, I think it's a pod, I don't know it's a namespace. You have to give them a namespace, and then they happily segment traffic inside of the namespace. But you have to

be, you know, be able to managing the name

spaces and and there aren't the tooling isn't quite there yet. It'll, it'll get there.

It's quick question, Are most of the enterprise customers, you know, kind of faced with this. Are they dealing with well established, fairly up to date applications in their IT estate, or is there a lot of legacy stuff that has to be maintained inside? What's the kind of a balance act there? Do you have a sense of that?

I don't have a clear sense of it. What what I have, and I think this is actually true even for the for the Kubernetes workloads, is that there's a ton of stuff running on VMware right now in VMs, and a fair bit of it is actually even in Kubernetes itself. And I think the real secret is not that people are going to run Kubernetes bare metal or OpenShift on bare metal. What they want to run is a whole bunch of VMs and then keep the people who are doing Kubernetes and OpenShift today on VMs, want to keep doing OpenShift and Kubernetes in VMs, there is no there's actually a disadvantage to trying to migrate to use bare metal as your primary Kubernetes platform. This is something that is not yet in not yet understood in market, although it's, it's sort of clear, right? Kubernetes, writ large, if you're writing applications, you're going to run it in VMs. We don't want to run it on metal. It adds complexity and reduces portability and a whole bunch of problems. So the thing that we haven't been able to have a voice in yet is this idea that, hey, I'm going to use, and I did some work on this, we have to re promote it. I'm going to use Cluster Manager to manage my VM workloads, because my app, my app practices, and this is, this is where the market just isn't there yet. And I'm trying to figure out how to educate the market and how to engage with the market to have these conversations. If I'm building applications, I'm going to have hundreds of clusters. It's the way people use Kubernetes. If I'm doing a virtualization system or an AI system on bare the both are in bare metal. I I'm not going to have hundreds of clusters. I'm going to have you can I'm going to have one big cluster, or I'm gonna have an AI training cluster with 64 machines in it, right? 50 to 100 machines in it, and then prod, and then it's gonna be and it's gonna be stable in production. And there's not gonna be a dev test, prod, motion, there's not gonna be a whole bunch of dynasism where I'm constantly turning those machines over. They're running virtual they're running one workload, AI training, right? Running, you know, virtualization as a workload, and so the benefit of cluster manager for that use case is actually pretty minimal. Yeah, even though they're they're selling it really hard. You're not turning over clusters like like you do in a in an application where you're constantly updating the code, you're you're just running the VMs, and so we haven't seen in market. I mean, I did some interesting stuff. I think it got some a little bit of attention, but we're not engaging. And here's, this is part of the business problem that I'm trying to work through is how much, how up stack do we need? How much expertise and how much consulting and how much training do I need to do in market to help these enterprises over the hump on adopting this platform

is the well the what constitutes, what constitutes, in your mind, an appropriate adoption. Does it mean, almost, you know, I don't mean to be, I don't mean to be deprecating this or, or, you know, derogatory. But is it a lift and shift mindset, or is it a transformation? Is it Is this an opportunity to change the nature of what you've got in place now, so that you, you've kind of, I won't say, future proofed yourself, but you've at least given yourself kind of more range of motion, more latitude in in the way you've you've done this. And if you do this on an application by application basis, do you come up with the same the same configurations to the same constellation show up? Yeah, you're asking

a key the there is a degree of lift and shift to this, but it's not 100% the target isn't 100% replacement. The target, the target is and and I don't see any practical use cases where people repurpose gear, which is an important thing to think through. So if you're, if you have, if you bought VMware clusters, you're running VMware clusters, and you're not going to transition that gear into an open shift cluster or a Nutanix cluster. I mean, basically you have to see a new gear.

Yeah, you have to think about this as, you know, a a basis. And it may be a multi step process, but it's a process by which you're also, you're not just and doing an end end of life on, you know, your old tomcat web servers and, you know, and stuff like that. What you're doing is also doing kind of a serious end of life on parts of your infrastructure to replace it, correct?

Because you have, you're going to align your exiting the VMware licenses with your infrastructure, right for that, and to an extent, I think what we're seeing right now is people are, since they don't know where they're going, they don't want to buy more VMware, but they're going to renew the VMware licenses they have the window. Okay?

I see this somewhat differently, and maybe it's just me to me, this is a golden opportunity to show your value in the marketplace and the business value of following the path that you're going to lay out as the educational component of educating the market to push the engagements that you want, because even if they're not going to build it on bare metal, you're the trusted advisory partner that helps them lower cost, increase their value, give them the latitude to future proof and and, most importantly, set up for AI.

Yeah, the AI, the AI pieces, is definitely part of this, right? This, we're, we're pursuing both paths in parallel. There's, there's an AI build path, and there's an open shift build path. The thing that you're saying resonates with what the dilemma that that I'm going I'm figuring out is, so far we've we've approached this sort of like we're going to help you install and operate the platform. But the more I see. It, the more we have to build expertise to help you be a manager of the, you know, we to go up higher in the stack.

You're managing, the you're you're managing the full estate, is what it comes down to, and that's that means you're, you you're not necessarily responsible for, you know the, you know, kind of the migration or the or the the renovation of all the legacy apps and so forth. But it probably behooves you to find of the partners who do, and that's why I want to introduce you to some people.

And I would love to find partners who want to do that work, we do not have. The challenge for us is, if it's necessary for enterprises to do the adoption,

both of you, both of you, should take a look at the company that will be making a lot more noise in the next months, called App factor, A, P, P, F, A, C, T, O, R, and I'm going to introduce you to the CEO. And then the the guy on the West Coast that's doing a lot of the kind of business arrangements.

Nice. Yeah, I Oh, nice knowledge graphs in the back, yeah, and, and, by the way, Rob, the reason that you haven't heard back from OpenShift is only because, no, no, we, I told you, we would make that connection for you, travel schedules and stuff like that. So it'll probably happen like after our launch towards the end of September, but it's definitely coming your way. Yeah, no, absolutely, but I will look into this app factor thing. But the one point that I would make is dual streaming, the approach to the market might not be in your best interest. I understand why you're doing it, but I would say, Go up. Go up. Up stack number one, and it would be probably the best time to do that would be as soon as possible in whatever campaign you're going to do to engage with the market. Because here, you know, come September one, every enterprise goes through their budgetary cycle for 2026 right? And if you can save them money, save them time or make the process more fluid for them with their existing platform team or their ops team, you are ahead of the game.

This is, this is this is the question that we're trying to figure out. I mean, we're not, we're not doing both the right now we're putting chips in the OpenShift virtualization bucket more than the AI budget bucket, and I can explain why, but it's in part because that's customers are pulling us there very hard at the moment. But the thing that so we're putting together an OpenShift virtual VMware to OpenShift migration offer as because I want to test the market and get calls. The thing, the thing that we're trying to figure out is in that offer, right? What's scaring these companies? Right? We know the companies are looking at this strategically, what we're trying to figure out is who to engage with at the companies to get a call about, yes, I want to accelerate or I want to de risk the this migration path. Because that's ultimately what the message is, is like, de risk here, migration to OpenShift virtualization, that's the, that's the offer is, is literally de risk, right? Or, except, usually we put it in terms of move faster, right? But it's a de risking move faster. You have to get off of this platform. We're gonna, we're gonna help you move faster. What, what I keep stumbling over, is who to reach with that offer. Okay, so this is just, we're literally going to the executive suite, and they're going to have to pull us in.

Yeah, yeah, you need to go to the executive seat. And the hook might be if you're planning, and it may sound weird to you, if you're planning on going down the AI road, you need to consider this very carefully. First, it's a it's a must step in your AI journey.

Yeah, that's part of it. The other is. Well, your budget, you know, looking backward, your budget for both continual maintenance of all the legacy crap that you've got, and this is your opportunity to kind of, if not, do a complete clean sweep right now, it is the, it's, it's a phased approach to getting, getting yourself to, you know, kind of future proofing,

yeah. But you can, you can, instead of, with all due respect to rich, instead of using words like future proofing, you know they're going to go down the AI path one way or the other.

I do. The thing that I'll tell you from our engagements is there is narrow, well, there's there's just no AI in it. The people that we're talking to, the people making this decision, are not at all involved in making AI decisions. There they are. And this is what this so, this is why. This is why the OpenShift but, but there's a but, the reason why OpenShift is better positioned, or the migration is the target for the migration is the OpenShift developers in these orgs are coming in and saying, and the executives are hearing this too, and it does AI, and it does virtualization, it's the one. It's the One Ring to

the that that altogether I understand what, what Joanne saying and and legitimate. It's legitimate. It's scary as crap for a lot of folks, and they don't want to hear about it. They don't want to deal with it. The place where you may find that the introduction of AI, let's make it machine learning and AI actually works is as the tooling and as the means by which to move. It's the

Yeah. It's

the dilemma, yeah.

The dilemma that I've seen with this, and we've been testing it, is, if I talk about open shift bare metal, or Kubernetes bare metal. Market interest is near zero. Here zero, right? There is, there is. And they're very, they're they're literally different. Go to markets. There's a open shift virtualization path which is still open shift bare metal, and there's an open shift AI path, which is the bare metal, but, but that's

why the bare metal, but that's why we're saying, I think why both Joanne and I are saying, If you remain solely in the the kind of the infrastructure, the lowest, the lowest levels, and do not say it's going to be driven, it's going to be defined. You are going to be defining it more from top down, and that that is what's going to dictate a lot of what has to happen underneath there in lies the the hook for for AI and a number of other aspects, which is these, this is the way to make those transitions happen and and, you know, It's, it's overkill, but the entire stack is moving kind of as as a stack, instead of, you know, kind of disjointed and piecemeal, that's, that's the, that's the kind of high order story. This is, again, why I think you need to talk to the app factor folks,

I actually love what so the thing that you're saying is, you're this is, this is a generational change in it, in that we're unifying, we're reunifying all of this disjointed it under A, you know, a shared platform. I actually am working on a piece about this, okay,

yeah, well, yeah, it's not a shared platform, if not a shared platform, at least a share a it's a common platform. It's a common word, it's a common platform, and it's a common methodology that you're using,

and everybody is making the case about data and data quality and all of this stuff around data governance, storage, whatever, all of that that case, nobody is making the case about the fact that if you are going to go down the AI path that your infrastructure needs. To change. So that's your innovative approach to building a solid. You know? Why? Why you need the solidity? No pun intended, underneath the AI. And this needs to be one of the one of the cornerstones of it. And the other part is that you can now, if the AI decision, which is a business value decision, comes from the top down, that there's an alignment from top to bottom to infrastructure, not just about the data. Because here's the little known fact, and I'm sure Rich will back me up on this in the connector, literally between the cable and the computers that are used for AI, there is a latency effect that comes in right then and there. So you need to shift the stack. You need to have the commonality you're doing this in preparation for AI, whether AI terrifies you or not, you're eventually going to go down that road in some format. In order to get the best performance out of your data, quality spend and the best performance out of AI, you need to do this infrastructure shift, not only because of the VMware debacle, but also to set yourself up for the future. Very solid business case by by saying that

this is, this is what, where you're going, is this is an aligned IT strategy, yes, yes, using, using upright, using OpenShift infrastructure as a common platform, right? And this is, you know, RackN needs to emerge as an expert in this open shift infrastructure,

right? And, and one of the things that you're offering up here potentially is, you know, it's probably not a good use of the term to call it bespoke. But your, your purpose in all of this is to make for it to be aligned. Everybody does not get aligned in the same way everybody does not, you know, have boxes that all have the same logos on them, top to bottom.

They shouldn't. That's bad policy.

They should. That's, that's bad policy.

So sorry for interrupting you rich, but I just occurs to me, and I don't want to lose it out of my brain. All right, when you pour concrete, you need to put the rebar in it.

Uh huh, that's right. That's why we call the product rebar.

Rebar.

Exactly. It's it's time to use that promo gizmo, Gadget, whatever. Yeah. When you're building the stack, it needs to be supported. That support is the foundations.

The foundations make a huge difference. No, this is, this is the thing that, that is, you know, our customers understand this, convincing our prospects or the market, that of this is, has been the challenge I really like Where y'all took this, because what you're describing is at the CIO level and aligned strategy based on right, you know, you know, ultimately on a Kubernetes infrastructure platform, the place I was working on an article around this with Mike Manny. And some of this goes back, actually, to the vision of Kubernetes at Google, right? What Google is Borg for, and they actually have something below board called the hive mind, or the it's not, it's the back brain or the lizard brain. Yeah, right. Kubernetes there, or the pre precursor to Kubernetes, runs the infrastructure. It's their infrastructure abstraction, and then they have some bare metal infrastructure that feeds it. But they haven't aligned, aligned strategy around how they use, you know, container and container, and they use the fvms running, and they do all this stuff, but they have an aligned strategy around the platform that runs their infrastructure and so,

and that's why they get the kind of and that's the reason they get such an incredible utilization factors on everything that they do, from, you know, the the bare Metal on up their networks. I mean, it's just, you know, it defies anybody, anybody else's numbers. I mean, when you start to look at, you know, everybody else, you know, if they can, if they can, kind of creep up to 60, 70% they're, you know, they're kind of, you know, just. General utilization, you know, their efficiency. That's a, you know, that's a big win. And Google's running almost consistently, you know, kind of bumping around 90. And it's like, holy crap. So, yeah, you're, I think this

is, I like, this is, this is the aligned message, yeah.

Two, two visuals that will go next to it. One, the cost of doing nothing. Two, the cost of just doing the VM renewal license.

Yeah. Basically, it's like saying, Look, if you had the, if you had the towers and the straight, you know, the stacked piece, piece parts, and you only move one part of it and not the rest of it, you basically lost a certain amount of balance. There's only so far you can shift

it without, I don't think you I even have to warn people about that. I think the at the at the executive level, the goal is the Align strategy. And if you're telling people, I have analyzed, it's where you start with the Align strategy.

And it doesn't all happen at once. It's, it's for the same reason that people first efforts into Cloud were, you know, so disastrous weather. It is not a lift and shift in its entirety. It is, there is an alignment, and it is a phased there's a, there are, there's a, there's a, there's a motion involved in it. There's a transition. And you we have that's, I think, I think that's the right kind of message.

Wow, it's really fascinating how much we can get in these conversations. In this case, there's a considerable amount of crossover, like I said in the intro to what RackN should do and how we specifically are trying to help the industry. And I think it's also important to think through what it takes to make sweeping it changes. Where do those changes start? Who do you have to help? What knowledge do you have to build to make that happen? Ultimately, that is what we love discussing in these sessions. And if you're still listening, then you enjoy learning about too. I don't think I've heard any other conversations that have such a pragmatic and sweeping focus to them as what we're doing in this podcast. And I hope you are enjoying it. Thank you for listening to the cloud 2030 podcast. It is sponsored by RackN, where we are really working to build a community of people who are using and thinking about infrastructure differently, because that's what RackN does. We write software that helps put operators back in control of distributed infrastructure, really thinking about how things should be run and building software that makes that possible. If this is interesting to you, please try out the software. We would love to get your opinion and hear how you think this could transform infrastructure more broadly, or just keep enjoying the podcast and coming to the discussions and laying out your thoughts and how you see the future unfolding. It's all part of building a better infrastructure operations community. Thank you.