Pete Wright:
Hello, everybody, and welcome to Trilio Insights on TruStory FM. I’m Pete Wright. Virtual machines have long provided IT teams with the important flexibility and isolation for running workloads. But as container adoption accelerates, how can organizations integrate their VM environments into modern infrastructure? This week on the show, I have the distinct pleasure to explore KubeVirt with our own Kevin Jackson. Our hope is that you’ll come away with a 360 degree view of KubeVirt and how it might fit into your container strategy. The future of virtualization is coming into focus this week, on Trilio Insights.
Kevin Jackson, how are you?

Kevin Jackson:
I’m good. How are you, Pete?

Pete Wright:
I’m very good. I’ve been studying up on this conversation. I have been studying up because I think it is fascinating. It’s a fascinating exploration of technology. When we talk about integration VMs into our Kubernetes environments, I hope we help some people today. I hope we help some people see the light.

Kevin Jackson:
Oh, I hope so, too. I’m excited about this emerging technology.

Pete Wright:
Yes, it’s fascinating. Let’s explain, first of all, what is KubeVirt and how does it work?

Kevin Jackson:
Yeah. KubeVirt is basically virtual machine management through the ways, the wonderful ways of managing stuff through in Kubernetes. So just how you can manage applications that are based on containers, it’s the natural world of Kubernetes. Red Hat got a little bit curious, looking back at the history, around about 2016, they got curious in wondering can we manage virtual machines in the same way, as we manage any other Kubernetes thing, and they did.

Pete Wright:
There’s a certain recursiveness that delights me about that exploration. It is a snake eating its tail, in some respects, it feels like. Or nesting dolls.

Kevin Jackson:
It is. Look, we’re encroaching into territory where we know that virtual machines are the bedrock of the infrastructure. Obviously, you’ve got bare metal and then on top of that. To make the best use of bare metal, you’re going to carve the resources out using virtual machines. That typically is what runs things like Kubernetes. People do run bare metal Kubernetes, but most would run it in a virtual environment and virtual machines.
You’ve got this world of Kubernetes, and then somebody comes along and adds this extra feature called KubeVirt. Or in the Red Hat world, OpenShift Virt. Suddenly, you’ve got the ability to run virtual machines. If you think about the thing, it’s like turtles all the way down.

Pete Wright:
Yes, exactly.

Kevin Jackson:
Then, you can run Kubernetes inside of those virtual machines and I don’t know where it’s going to stop.

Pete Wright:
In so far as we’ve talked about the beautiful metaphors for how this thing works, give us some concrete use cases. Why would I want to consider this, if I’m an infrastructure manager? How’s it going to help me?

Kevin Jackson:
The appeal is around the Kubernetes. Right off the bat, the world we’re in is a Kubernetes focused world. What’s getting all the action is obviously, like I said before, applications based on containers. I guess, there’s almost a barrier to entry in the world of Kubernetes. For people who are born into that world, that barrier didn’t even exist. But for others, the vat majority of us others, let’s be really honest, still run virtualization. But know there’s a world out there, which is highly orchestratable, and getting all the richness and all the good features of Kubernetes. They want to be able to have that cookie and eat it as well.
They’re left with a couple of choices. “How do I get there? I’ve got my applications running in virtual machines. I’ve got this alien world of Kubernetes and containers.”

Pete Wright:
Simply, it’s a legacy question, right?

Kevin Jackson:
Yeah. One thing off the bat is running virtual machines in Kubernetes. To take advantage of the Kubernetes ecosystem is running virtual machines in Kubernetes is a nice little stepping stone to that. I think it’s a very valid stepping stone because I think not everything should run in a container. I don’t know how controversial that is to our listeners, but I just don’t think that everything should be containerized. Virtual machines exist. There’s far more virtual machines than containers.

Pete Wright:
Right.

Kevin Jackson:
I don’t know, given the fact that containers are the tiniest slices of a computer, maybe it all equalizes.

Pete Wright:
Yeah.

Kevin Jackson:
But ultimately, the world does not all run containers. There’s big databases out there, there’s big systems that run perfectly well and are suitable-

Pete Wright:
In a VM.

Kevin Jackson:
For VMs. But, I would like to use the orchestration features, I’d like to use the capabilities of Kubernetes. Suddenly, I’ve got this little bit of consistency for when I do get to the containerized world. Or if I’m already there, I’ve got this consistency of manning virtual machines and my containers.

Pete Wright:
It sounds very much like the benefits that you’re getting out of this structure are resiliency and scalability, right? You have a system that can bend under pressure, even though you’re running VMs right alongside all your containers.

Kevin Jackson:
Yeah. Yeah, yeah. Yeah, exactly. Yeah, in terms of features, you get the framework of Kubernetes and that comes with things like scalability. However, I guess there’s a world of it’s not all a panacea. We get to a stage where yes, we can run virtual machines inside of Kubernetes, but then we have to have the conversation around the fact that … We’ve discussed, or certainly touched on the why, and you get the added benefits through a Kubernetes ecosystem of running virtual machines on there, but virtual machines also run equally as well on the platforms that they were designed to run on.

Pete Wright:
Right.

Kevin Jackson:
You know what I mean?

Pete Wright:
Just because we can do a thing does not necessarily mean we should.

Kevin Jackson:
No, exactly. We would not be able to have these kind of conversations if we would just come on and say, “Oh yeah, you should just run everything now under KubeVirt.” If you’ve got a virtual machine that’s running on, I don’t know, you’ve got a KVM hypervisor, you’ve got oVirt, you’ve got, I don’t know, OpenStack, whatever. You might be thinking, “Why do I have to change?” Our listeners do need to understand you don’t have to any of these things, you’ve got choice. You’ve got to migrate for the right reasons. But I think the right reasons is to have consistency of your future estate. There’s consistency of managing both containers and virtual machines. Again, you don’t have to run everything in a container. Suddenly, you’ve got this nice little landscape which gives you the benefits of a virtual machine, and all the good stuff that virtual machine runs, and all the good stuff about scaling out your containers in a very Kubernetes, auto-scaling, auto-healing kind of way.

Pete Wright:
Yeah, and the auto-deployments. We touched on that last time, with Rudolpho. There are clearly benefits, for sure. Spiderman’s dilemma, “With great power comes great responsibility.”
Let’s talk a little bit about KubeVirt in industry. What are our other options?

Kevin Jackson:
Yeah. Yeah, obviously if we look at the landscape that this now operates in, obviously there’s a driver. You want to go there with your virtual machines because you want to run Kubernetes, let’s be really honest. I think in terms of the choices where things can go, if you think about the interesting landscape that’s come about recently with the Broadcom VMware acquisition, it’s caused a little bit of noise in the industry.

Pete Wright:
Sure.

Kevin Jackson:
You can’t avoid reading the news articles to see yet another grievance as to price hikes, and all that normal stuff that comes with acquisition. People obviously are a little bit touchy about where can I run my VMware workloads, for example. In that context, what are the choices? KubeVirt obviously operates in a world next to OpenStack. It obviously runs next to the world of VMware itself anyway, and any of the hypervised Nutanix, and stuff like that. There’s a lot of options for people to run virtual machines.
It’s worth noting the fact that KubeVirt is relatively mature, in comparison. It’s KVM based, like a lot of hypervisors that are out there. A lot of the OpenStack based distributions are based on KVM, other hypervisors exist. It is relatively mature. Even the fact that it’s been in curiosity mode, in development for quite a while. In fact, looking at the history of KubeVirt, it was, as I say, curiosity got the better of some Red Hatters in 2016. It became a project of CNCF in 2019. Then, it got a level of maturity being an incubated project in 2022, 2023, ’22, somewhere around that kind of time I think. But we got version 1.0 last year.
You contrast that to say, OpenStack, again which, arguably in this context, a management of a hypervisor. It’s not a hypervisor itself. That’s been around for 14 years.

Pete Wright:
Yeah.

Kevin Jackson:
There is a level of sanity into this conversation. It’s not the panacea, but the acceleration of adoption, or the acceleration of a certain interest. Certainly, what we’ve seen is quite immense.
In fact, last year we were having conversations, early last year we were having conversations around say, OpenShift virtualization. Which sparked us to do development, being able to back up OpenShift virtual machines using our Trilio of Kubernetes workflow. That was back then. I think now, we’re getting two or three conversations a week around KubeVirt based environments.

Pete Wright:
A question on this discussion around why not use OpenStack. How much of this is a discussion around managing your budget of the unknown, budget or uncertainty in deployments?

Kevin Jackson:
Yeah. I think because of the relative immaturity of KubeVirt versus the maturity of OpenStack, I think it just depends on which angle you’re coming at. If you’re a Kubernetes person and you know that there’s a desire to enhance the environment with this operator called KubeVirt, so this extra additional capability, then that angle is different from somebody who is an infrastructure person. An infrastructure person will probably think that the Kubernetes environment is complex to manage. But from a Kubernetes angle, a Kubernetes person, they might think the OpenStack environment is bloated and complex.

Pete Wright:
Yeah.

Kevin Jackson:
Depending on your entry, I think you get two different answers.

Pete Wright:
When you look at the landscape overall, what major players are really investing in this? Whose really pushing? Who are our role model infrastructure players who are pushing KubeVirt?

Kevin Jackson:
Yeah. The biggest by far, not even the elephant in the room, is Red Hat.

Pete Wright:
Yeah.

Kevin Jackson:
They developed the thing. They have it as part of a certified operator and a certified offering for their OpenShift environments. We certainly see a hell of a lot of them in our conversations, and quite rightly so. In the upstream world, when it was donated to the CNCF Foundation. Out in the big, bad world of just upstream KubeVirt, you’ve got service providers that are using this. You’ve got people like a Spectro Cloud. You’ve got people like a few others, maybe Platform9. People who enhance, they have offerings of Kubernetes and then enhance this with virtualization. But they can hide a lot of complexity through their own user interfaces as well, so they can get a seamless experience of say running virtual machines. But the backend and what they know is a Kubernetes based world. They can offer both to their customers.

Pete Wright:
Sure.

Kevin Jackson:
But yeah, by far is easily Red Hat.

Pete Wright:
Red Hat, for sure. Okay. You dropped the word sanity earlier, and now there is an elephant in the room we have to talk about. What are the factors we should be considering if we have this opportunity to potentially discover if KubeVirt is right for us? Why would you move?

Kevin Jackson:
Yeah. Hey, why wouldn’t you?

Pete Wright:
Well, that’s really the thing because we’re all susceptible to new shiny. There are great new infrastructure plays we can make to extend our operation, let’s seriously look at it.

Kevin Jackson:
Yeah. I think if you are comfortable with Kubernetes … Let’s talk about, I guess, the requirements. I think it’s an easier way of looking at it. I think if you’re comfortable with Kubernetes and you have an additional requirement for running virtual machines, for whatever reason, your applications just don’t work well in containers or your managing something, to bring the sanity word in there, “Why am I managing this VMware environment or an OpenStack environment, when do you know what, I’ve got 90% of my staff know Kubernetes.” In that panacea, it’s a good place to go.
We’re on KubeVirt 1.0, with a ton of solid features. It’s still an incubating project. It still needs to get to that maturity to progress to the next stage. But even if you look at the CNCF or the KubeVirt pages, you will see the fact that their incubating projects are heavily used in the industry, they just need a little bit more time to mature and just to develop. It’s under active development.

Pete Wright:
Yeah.

Kevin Jackson:
But, the real elephant in the room is, “Well, will it do everything I need it to do?” If you have got an OpenStack environment and you’re thinking, “Oh, do you know what, I can migrate all my virtual machines to run in KubeVirt,” you have to question why you’ve got OpenStack in the first place and what that’s providing you. OpenStack provides you with a ton of capabilities that currently can’t or don’t exist in the KubeVirt world, as it stands today.

Pete Wright:
Yeah.

Kevin Jackson:
Again, it’s still a maturing project. It’ll allow you to run your virtual machines, you can attach your desk, you can attach your networks. You can do all your virtual machine stuff. But, there’s a reason why these other things exist. You have to be grownup and adult to have this justification of moving your workloads into there. You could have basically all the right reasons. We know the world of IT, people move stuff for the wrong reasons.

Pete Wright:
Yeah.

Kevin Jackson:
There’s a wave, there’s a bubble happening. Just because somebody says, “This is the way you should do things.” Unfortunately, people like me get caught up in that kind of hype.

Pete Wright:
I didn’t mean to call you out with the new shiny comment earlier, but maybe that’s where we all are.

Kevin Jackson:
[inaudible 00:17:44], okay.

Pete Wright:
Yeah.

Kevin Jackson:
Just addressing that.

Pete Wright:
Yeah.

Kevin Jackson:
But yeah, the point I’m making, look it’s really awesome technology. Honestly, I think for a lot of use cases that go beyond just basics, you will get to a place which you will be very happy managing Kubernetes based virtual machines. But I think the, as I say, the entry point is the fact that you’re a Kubernetes person, you’re a Kubernetes administrator that’s comfortable with the networking elements and don’t see Kubernetes as complicated. From the outside in, Kubernetes looks complicated. But to a Kubernetes person looking at OpenStack, or VMware or stuff like that, honestly, they’re alien environments in that context.

Pete Wright:
It’s a walk in a park, yeah. All right, so Kevin, tell us what is Trilio doing in the KubeVirt space?

Kevin Jackson:
Yeah. Yeah, good question. I’d say quite a while ago, we were first to market with an early release of being able to backup OpenShift virtualization VMs using Trilio for Kubernetes. So the certified operator for OpenShift. That was a request for a large environment, going back, way back now. What year are we in? Anyway, it feels like a long time ago. We’ve obviously matured since then. Kubernetes and KubeVirt itself has matured. OpenShift itself has matured.
For us, from a Trilio point of view, we treat a Trilio virtual machine pretty much like anything else in Kubernetes. That’s a little bit of the magic. You’re not having to switch gears to think, “Oh no, I’ve got a virtual machine, so I now need to go and put my virtual machine hat on to understand how to backup this virtual machine.” To us, and in fact the construct of a pod, which is the outside element and then this container inside, you’ve got a pod and then you’ve got a virtual machine inside. It makes logical sense that Trilio just treats this as just a resource that sits inside a pod. It’s got extra resources, like a system volume. It’s got a resource like a network. For us, it was very easy to fit into our workflow.
We have a good ground in basics, I would say a long time ago, and where we’re up to now, you can do things like fast DR failover. We’ve got a feature called Continuous Restore, works extremely well for containers, with storage attached. With virtual machines, they immediately have storage attached.

Pete Wright:
Wow.

Kevin Jackson:
That’s the difference between the container structures, to process the runs in memory optionally has disk, you can’t get away from the fact that a virtual machine has a disk.

Pete Wright:
Yeah, it needs resources. Right.

Kevin Jackson:
Yeah. Immediately, Trilio is perfect for backing this up. There’s a disk there, we grab the disk, the information, all the Kubernetes, metadata around this. But we extended our Continuous Restore functionality to virtual machines as well. You’ve got an environment, say in data center one, to run all your virtual machines. You’ve got an environment, data center two. And, we’re able to backup those virtual machines. But not only can we backup those virtual machines, we can almost, I like to phrase it as pre-stage those virtual machines, in a different data center. You’ve got this almost site recovery manager, to coin a phrase from VMware, kind of like a world where you’ve got a scenario that, if your OpenShift cluster or Kubernetes cluster has failed in data center one, your virtual machines are ready to go, including the disks are already on the right storage in the right data center, for you to have fast failover.
Yeah, that was new for the latest releases that we’re on. We’re on version 4.0 of Trilio for Kubernetes now. That’s baked into that, just out of the box. You get that free.

Pete Wright:
Fascinating.

Kevin Jackson:
An additional feature for all our customers.

Pete Wright:
Yeah, right. Right. Thrown in, you’re just throwing it in.

Kevin Jackson:
Absolutely. We love our customers.

Pete Wright:
So generous.

Kevin Jackson:
Yeah.

Pete Wright:
Well, it’s fascinating. I think, if we’re going to leave folks with anything, any sort of homework, where would you suggest they go to learn more?

Kevin Jackson:
Well, there’s a few websites. KubeVirt.io is probably the main ones. We’ve got some blogs either in the pipeline or about to go on trilio.io, around backing up and managing virtual machines. We’ve got some videos on YouTube. If you look at the Trilio channel, you’ll find that we show OpenShift integration. We show the console, you can backup your virtual machines and recover your virtual machines directly from the OpenShift console, for example. They’re good resources to go to.

Pete Wright:
All right. We will put … Just swipe up in your show notes, everybody. We’ll put direct links to those resources in the show notes so you can catch them there.
Thanks for the tour, Kevin. I appreciate it.

Kevin Jackson:
Yeah, no. Thank you very much for having me on again. Yeah, I always like chatting to you. I hope our listeners are happy and are now ready to explore Kubernetes with KubeVirt.

Pete Wright:
Likewise, water’s warm. We certainly appreciate all of you downloading and listening to this show. Thanks for your time and your attention. Again, encourage you to head over to those show notes, click those links. Learn more about all the resources we’ve talked about on the show today, on behalf of Kevin Jackson. I’m Pete Wright. We’ll catch up with you next time on Trilio Insights.