Rising Interest in OpenStack for Workloads with Open Infrastructure Foundation’s Mark Collier

This is Textron tv. Hey guys, thanks for the throw. We're here with Mark Collier, his COO, for Open Infra Foundation, and we're talking about, well, to what degree, or people may be taking a second look at OpenStack in the week of the acquisition of VMware by Broadcom.

There's a lot to discuss. Mark, welcome to the show. Thank you.

Thank you. It's great to be here. I think just about everybody who evaluated VMware over the years has at least taken a look at OpenStack, and a lot of people have both OpenStack and VMware for that matter, but OpenStack has evolved over the years.

So what is the current level of maturity of this particular platform and how has it evolved recently? Yeah, great. Great question.

And I, I'm glad you point out that many people run OpenStack and VMware, um, because I think oftentimes there's kind of this overly simplified thought and that you pick one or you pick the other. And, you know, companies have a lot of different infrastructure, but in terms of OpenStack, you know, we just had our 29th on time release, uh, called caral. And every six months, like lock work, we re release a new version.

So it's very, very mature at this point. Um, billions of dollars of revenue our, our, uh, accounted for in sort of the overall OpenStack ecosystem every year. And billions of people actually just to use sort of the telco, uh, example of connect to mobile networks powered by OpenStack.

So there's so many different use cases today in terms of infrastructure and, you know, 29 releases later. It's, it's very, very solid. And we see just continued adoption and, and actually, you know, kind of this, this Broadcom move has really created an an another opportunity for people to kind of reevaluate, okay, if, if I looked at OpenStack before, maybe I'll look at it again, or if I ran some of my stuff on it today, maybe I'll move more of it to, to OpenStack.

So it's really created a ton of interest in just the last few months. It seems like the dilemma that people are facing is that the, unless you kind of structure your licensing fees, just so the cost of running VMware is higher for a lot of organizations, but the cost of migration is high as well. Mm-Hmm.

And they're trying to make an evaluation or a trade off between, well, what is the total cost of migrating versus, uh, continuing to license VMware? Has that gotten any easier? And, and 'cause the migration is more than just the VMs, right?

Right. Yeah, I think it's, it's a great point. Um, it is something that is, uh, there's no one, one size fits all.

And, and OpenStack isn't gonna just be, um, the right solution, a drop in replacement for everyone who's running VMware today. I think everyone is looking at it again, if they, if they weren't, uh, hadn't looked at it in a while. And so kind of making that choice for themselves and the calculation of the payback period or how, what's it gonna cost to migrate, um, is, you know, it's a little bit different for every organization.

But what, what what we see is, uh, if you look at companies where software is a, and technology in general is like a strategic part of their business, those are the people that it, it often makes the most sense for that, you know, really need to, uh, have their developers move more quickly and they need that kind of automation. And yeah, when you see your bill going up as much as like 10 x, then you really have to go, go run the numbers and do the math. Um, and so yeah, we, we have seen, um, just a lot of people, um, looking at that, those, those numbers more carefully.

And just one other point I'll, I'll mention, which is, um, uh, we're talking about sort of like OpenStack as an alternative to VMware, but the reality is, you know, VMware, uh, is one company's product, and OpenStack is an entire ecosystem. So when you look at, uh, your o options, if you decide to migrate to OpenStack or move more of your workloads to OpenStack, 'cause you've got Red Hat Mirantis, canonical, you've got all these public clouds V access, OVH, Rackspace, so you're not just sort of going from one vendor to another. You're going to a very rich ecosystem where you're, uh, much less likely to kind of, um, get locked in because you have all these different choices.

You have, you know, a competitive marketplace. And depending on your workloads, one of those particular products from one of those vendors might be a better fit than the other. They're not all, you know, identical.

They all have, you know, their own sort of strengths and weaknesses. Are some folks maybe looking at it more like this and they're saying, well, uh, we've got a number of workloads already running on VMware, and the cost can be difficult depending on the workload. So are they just kind of maybe drawing a line in the sand and saying, all right, all our new stuff is gonna run on this other platform and we will just go from there?

Yeah, I mean, I think that, uh, a lot of companies have been doing that for some time, but certainly, um, when they think about, okay, we've got these different platforms, we're not necessarily gonna shut one down 'cause it, you know, that that can have a lot of, uh, repercussions for the internal business units that are just kind of used to operating on it. But yeah, we're gonna put all of our new workloads on this, uh, more cost effective solution that has, you know, more competitive set of of vendors to, to help, help us, then yeah, they'll start to default to OpenStack. And, and again, just go back to this sort of competitive set of vendors, uh, question.

You know, I think, um, with VMware, uh, even in their ecosystem where they had ISPs and sort of like, um, uh, these, you know, consultants and this whole e they actually had a very strong ecosystem of third party companies that would help you buy and implement OpenStack. And they basically cut off a huge percentage of those and said, you can't even sell it anymore. So I, I think that one other thing I want to mention is what's happened here is not just about price, it's actually about uncertainty.

And companies really hate uncertainty. And so when you put this kind of uncertainty in the minds of your, of your customers about what's gonna happen next, whether it's price or something else, or they're gonna, you know, end of life something that I'm, that I'm expecting to, to run my business on, that really just causes people to think, um, beyond just the, the economic piece in terms of the price, One of the things that drives a lot of people in one platform or another, especially for the IT administrators, is there's a perception that one is a little smoother and easier to operate than the other. Um, has that improved in the, in the context of OpenStack?

'cause a lot of folks will say, well, yeah, the VMware user interface is a little smoother. Uh, there's certain capabilities that VMware makes available that are not always in OpenStack or probably will be at some point in the future if it's just a question of when. But, um, what is the actual IT administrator experience like these days?

Yeah, it's, that's a great question. I think, you know, going back to the sort of like 29 releases, like all of those improvements over the past, you know, 14 years really since Opens SEC started, they're driven by real world, uh, users and use cases and operators that, you know, when they have problems that that comes back into the upstream conversation when we, you know, meet twice a year to plan the next release. But it also very much comes through the commercial vendors, whether, like I said, the Red Hat or, or mortis or, or, or canonical, et cetera, um, who are, you know, uh, making money selling these products to their customers.

And so I, I guess just to kind of, uh, come back to this point about the ecosystem, they're not these, these are products, right, that are con powered by OpenStack, but they have more, uh, additional capabilities depending on which, which path you go. And, you know, the, we don't necessarily need every single feature to have feature parity between VMware and OpenStack for all the land in OpenStack proper, for there to be products that are very user friendly and, and easier for people to use. And, and you know, having said that, like for sure that it's not gonna always be the right, uh, use case.

If you're sort of like a, a company that's, you know, an auto sh shop or a small law firm or something and you've got Bmware running and it sort of meets your needs, you're not necessarily gonna go, uh, move over to OpenStack 'cause it's, it's gonna, there's gonna be some, some transition, uh, involved in that. But again, those companies, which is an increasing number of the companies in the world, right, for which software is strategic developing software is, is a big part of how the, the value is getting created. Those are the, the people that are gonna be, um, more of a fit.

Um, and, and in terms of OpenStack, when they take another look at it. To that end though, let's look to the future. Um, cannot walk down the street these days without somebody telling you about their great new AI thing.

And I wonder if the user interface of the future is gonna be the prompt. And, you know, it's not gonna be all about the, this wonderful GUI per se, as much as it will be an AI agent that is performing tasks on my behalf. And that will just smooth out the experience for everybody.

Yeah, it's, I, I'm glad you mentioned ai. Obviously everyone's talking about AI these days, but it, it's actually the, the other big driver of demand for OpenStack recently because, and, and for the open infrastructure foundation as a nonprofit, you know, we're helping to create this software to power infrastructure, and there's nothing driving infrastructure demand like ai right now you're seeing, talking more on the user interface side. But I, I completely agree.

I think that, um, we've all, if we've been around a while, lived through a lot of hype cycles, I think this particular one, there's actually a lot of substance to it in terms of it completely changing the paradigm in terms of the infrastructure, but also the, the user interfaces. I think that'll, you know, take longer to play out in the market. But I, I, I agree.

I think that, um, you know, in a lot of ways when you think about sort of legacy, uh, software and infrastructure like VMware, you know, it, I could see why there's, there's a, a, a phase in the market where the only thing that sort of makes sense if you're, um, looking at the spreadsheet is let's just raise prices. Let's just raise prices because the future is being built on, on other paradigms. And AI is a, is a great example of that.

And we're, we're seeing that, you know, there's a trillion dollars, according to Jensen from Nvidia. There's a currently a trillion dollars of infrastructure that's accelerated compute enabled, and another trillion will need to be built in the next four or five years. So anytime more infrastructure is being built, that means more infrastructure software.

And that is why we're seeing this kind of, uh, groundswell of interest in our foundation and OpenStack in particular, um, because people need to solve these problems at scale. What is the relationship between OpenStack and Kubernetes? And I'm asking, 'cause a lot of folks are struggling with, I got a monolithic app over here, I've got a brand new microservices app over here, and they're on different infrastructures, and that raises the total cost.

So, um, is that easier to do in the context of OpenStack? Yeah, so that's a great question and I feel like this, this question comes up, you know, pretty much every year for, for a decade. Um, and I think that it's extremely complimentary relationship.

You know, the vast majority of people who run OpenStack run it with Kubernetes. So they're, they're, they're very complimentary. And if you think about sort of the way I would define cloud native is, uh, this assumption that the cloud is there, like we can just assume it's there.

And if we're gonna build these applications and, and we're gonna, they're gonna need infrastructure, but we can just assume it's there. There's an API that provisions compute, storage and networking. And as sort of a developer, I don't have to think as much about the plumbing, but the plumbing is still there, right?

And so, um, that's very much the case with OpenStack. OpenStack is one of the ways you can get the cloud for you to run cloud native type tools on top, such as, such as Kubernetes. So we, we, we see that, um, if you look at the, the three of the, the biggest open source projects in the world, Linux, OpenStack and Kubernetes, they're very, very frequently run together.

Um, and there's a reason for that. It's because they solve different, uh, problems at different layers of the stack. They're all very open, they're all extremely diverse in terms of contributions for many, many companies.

They're not single vendor open source. Like there's a lot of open source now that's just kind of dominated by one vendor. But Linux, OpenStack and Kubernetes are, uh, great examples where there's a foundation behind them.

There's multi vendors, there's lots of choice and lots of ways you can get involved. And so we see them, uh, the three, uh, of course everyone just takes for granted. Linux is everywhere as they should, but it's very, very, very important building block.

And, uh, OpenStack is right in there in terms of how these tools get used together. So you have kind of the compute storage, networking, and you have sort of, you know, application orchestration and the lines are very blurry, right? And it's not just as simple as a layer cake.

We like to draw on the, on the architecture board. It's, it's much more involved in that when you're talking about these large scale distributed systems. But, um, you know, just to, to kind of give you another concrete example that ties it back to ai.

So, um, stack, HPC, which is one of the members of our foundation, they're, uh, working with the UK government to build the largest, fastest AI supercomputer in the uk. And it's powered by OpenStack, also uses Kubernetes, of course it uses Linux. Um, and that's, uh, that's just an example of how people put 'em together.

How active is the OpenStack community? And I asked this question because one of the things that a lot of folks like about VMware is there's a vendor there who provides support, and there's a sense of, uh, security that goes with that. But, um, how easy is it to get help within the OpenStack community when I have an issue?

Yeah, I mean, it's, it's extremely active. Like I said, it's, it's, uh, continues to be one of the largest, um, open source projects in the world, just measured by a, a lot, a number of different variables. But just if you look at, uh, code, contribution code commits every, every six months for these, you know, these releases that we do, it's still, um, very, very high on the charts.

So you have, you know, over 110,000 members of our foundation, and we have multiple open source projects, but OpenStack is by far the largest and most active. Uh, we continue to have events where people can get together, uh, both virtually and in person. Our next big event is our summit in Korea coming up in September.

So there's a lot of different, uh, venues, kind of both in the sort of pure community sense in terms of, um, mailing lists and, and IRC and places you can connect with developers and other operators. We also have work groups and other kind of, um, mechanisms to bring people together. And then of course, again, a lot of people do consume it through a vendor, and then they have those vendor relationships, and that's how they get the support.

And, and this ecosystem is stronger than ever. I've mentioned a bunch of the names, but, uh, that's, that's not a complete list. org website, there's a marketplace that shows like all the different vendors.

So there's just many different paths to get support. And there's, there's the community itself, which is extremely active, but, um, also, you know, through the commercial angles, that's oftentimes what people find meets their needs the best. I'm sure you've seen people make this migration.

Um, what did they do that kind of helped ensure their success? What did, what have you seen as a pattern that people who were doing this that others should follow? Yeah, great, great question.

And, and one of the first things we did when we saw this, this kind of like, um, overnight, um, uh, you know, our, our phones were, were blowing up with people calling and texting and, and contacting us about what was happening with this licensing change from Broadcom's acquisition. And we as a foundation, you know, our role is sort of a neutral body and, and a coordinator of this entire ecosystem as we reached out to our member companies, and we immediately surveyed them and said, okay, like, what are you seeing in terms of interest? And, you know, 82% of them that we talked to said they've already gotten inquiries.

This is as of like two or three months ago when we, when we first kind of pulled, um, pulled our, uh, member company. So 82% already had inbound inquiries from customers or potential customers saying, you know, help me migrate or help me look at migrating to OpenStack. And at, at this point when we did the survey, 61% had already helped a customer successfully migrate.

So I think that, um, that's just a sign of, of how quickly things are moving and how much investment is coming in. We also created a work group out, out after we surveyed everyone, and we started to, to bring people together to talk about what are the best practices. We're actually putting together a white paper that'll be out later this year, which answers a lot of the questions that you're asking in terms of what are the best practices.

So, you know, I probably can't rattle them all off, off the top of my head, um, live. Um, but, uh, but a big part of it is our job is to get the people that have those answers, encourage 'em to document 'em, encourage 'em to work together, even if they're competitors in the marketplace, they're collaborators in our ecosystem. And, and when it comes to OpenStack, so we'll be, we will be publishing a white paper later this year, including like, reference architectures and, and some of those details.

And, um, you know, in our first call that we set up to kind of, uh, bring all these people together, we had companies from four different continents actually participate. So this is a very much a global phenomenon. OpenStack's always been very global, but in terms of how quickly people, uh, I mean, there was not a company we talked to didn't want to take the survey, tell us, you know, what their plans were and hop on a call and then start to do meetings in person and meetings online, uh, to put these kind of best practices into play.

And, and, and including, I'll say, you know, uh, upstream contribution, right? I said, we've got 29 releases, but we're not done. We're gonna do, you know, 29 more.

So, uh, there will be continue to be improvements in the software upstream as well, which flows downstream to the products to kind of close the gaps. But, you know, the goal is never really to be just a one for one drop in replacement. I think that's, would be kind of aiming at the past or driving with the rear view mirror.

I think it's much more about, uh, helping people understand like, what are their workloads? Do you know, are they predominantly sort of like, uh, software development driven or is software strategic part of your business? And then just kind of put, put together a, a a migration plan.

And oftentimes, you know, companies will stand up a proof of concept or a small cluster, and they might do that in their data center if they're big, um, you know, on-prem folks. But even if you have an on-prem strategy, you can easily go to, uh, you know, public clouds like Veo, Rackspace, OVH and, and Deutsche Telecom and many others. And quickly with a few clicks, swipe a credit card and have an open stack environment, um, that you can, uh, you know, use to kind of kick the tires and test out.

So the beauty of like open source and, and being able to have like public cloud options for OpenStack is that you can start to experiment and sort of, uh, before you really, you know, pull the trigger and go, go all in. So that a lot of it is, comes down to kind of like documenting your, your, your application requirements and then just kind of experimenting and seeing, uh, what works. And you can do that very inexpensively with, you know, in your own data center with, you know, things like the public clouds out there that, that run.

OpenStack. All right folks, well, you heard it here. There's no silver bullet for anything as always, but if the one thing we have learned from all of this is well, maybe it's important to keep your options open.

Hey Mark, thanks for being on the show. Thank you very much. Alright, back to you guys in the studio.