 the whole story around because the pro is born from interactions with very fast scaling startups and very large enterprises. Hi, this is your host, Martin Parthia. And welcome to another episode here for Let's Talk. And today we have with us once again, Lukas Tintli, CEO of Love Lab. Lukas, good to see you after a long time. Yeah, hi Swapno. Good seeing you again. Thanks for having me on today. It's my pleasure to of course host you again. And today we are going to talk about Vcluster Pro. Talk a bit about the pro in the name Vcluster. There's so many people excited about our open source project Vcluster. You know, there's over 40 million virtual clusters out there. There's some major companies building on top of our solution. There were people speaking at KubeCon about how they were rolling out hundreds and thousands of virtual cluster. A lot of people save a ton of infrastructure costs with virtual clusters, but there are certain things that you cannot do in Vcluster in the, in the, you know, part itself that you're launching to start up a virtual cluster. You need something external. You need kind of a control plane. You need a CRD. You need a management UI, right? You may want to add like audit logging and SSO centralized for all of your Vclusters. So we are thinking, okay, this is something that needs to live outside of Vcluster, but it needs to be connected with Vcluster. And then there are certain things in the Vcluster core itself in the distribution, the Vcluster distro where people just are demanding more security, guardrails, more performance, optimized features and very, very enterprise focused features. So we were thinking about creating a commercial offering around it. And yeah, we're announcing that right now with Vcluster Pro. Can you also talk about when we look at this pro version? Because sometimes what happened a lot of, you know, when you do come in the enterprise editions, you are working with a company because they had the pain point that you had this and solves. So can you elaborate a bit on if there is any design partner or customer for Vcluster Pro? The whole story around Vcluster Pro is born from interactions with very fast scaling startups and very large enterprises. Most of the folks that we engaged with in the pilot phase were either like, you know, Global Fortune 500 companies or really, really quickly growing startups. One company in particular that was an amazing design partner for us was CoreWeave. CoreWeave is a cloud provider that is focused on GPUs. They're really shipping the next wave of, you know, Nvidia GPUs for AI workloads to a massive amount of people at this point. They're based in New York City and they have, you know, and they build an incredible product to power these AI workloads. A lot of these startups building on AI right now are running on top of CoreWeave. And, you know, you may have seen like OpenAI's engineering blog. Everything they do is based on Kubernetes, right? So a lot of these AI companies are banking heavily on Kubernetes. They need a lot of Kubernetes power. So CoreWeave was looking to partner with a company that can help them scale Kubernetes beyond the limits of what you can do with physical clusters. And then virtual clusters was a perfect solution. So they were a very early design partner for us and we're super excited that we, you know, that they give us the permission to also talk about this case. And yeah, we also have a joint KubeCon talk that will discuss how in detail they're using Vcluster and Vcluster Pro under the hood to deliver GPUs to their customers with Kubernetes. Pro will be serving a much larger, you know, customer base there. Can you talk about how and where does Vcluster Pro fit in a company's cloud native or infrastructure stack? Yeah, absolutely. So if you're running more than just a couple of virtual clusters, then you probably want to look into Vcluster Pro. Most people that we engage with today, they're thinking in the numbers of like 100 plus virtual clusters. And then really the management and the security controls around so many virtual clusters is a big focus as well as the, you know, resource consumption and performance of these virtual clusters. One very exciting thing we're doing in Vcluster Pro, for example, is we're pulling CoreDNS which was running alongside each of your Vclusters. We pull it inside of Vcluster and ship it directly with Vcluster in a security hardened container image. That means we can, instead of launching two pods, we just launch one. So that's obviously a lot more efficient. It stands up quicker. You have less pieces to manage, right? And essentially we make this part of our promise to you that we're kind of obviously updating this, looking at, you know, any patch releases, feature releases coming out and really pushing out CoreDNS joined with Vcluster rather than as a separate DNS solution for each Vcluster. And then on top of that, there's a lot of like management functionality that people are profiting from. So if you are setting up authentication for each one of your virtual clusters individually, that becomes very tedious when you have, you know, 100 plus virtual clusters and different teams and users using them. So with Vcluster Pro, you launch a management UI and you have a management CLI that will essentially allow your engineers to access these Vclusters via their SSO. You have audit logging for any interactions that they do inside their virtual clusters and a lot more. And one feature in particular that's important for the security perspective is, you know, typically when you launch a Vcluster, the Vcluster pod itself, alongside with the applications that you launch in the Vcluster, typically run in the same namespace, in the same Kubernetes cluster. But we launched a feature in Vcluster Pro called isolated control plane that allows you to run the Vcluster pod, the control plane of the Vcluster in one Kubernetes cluster and then the tenant workloads of your engineers that are working in that Vcluster in another Kubernetes cluster. That means when you have hundreds of Kubernetes, hundreds of virtual clusters, you can have one cluster that manages all of your virtual clusters control planes and nobody can touch it and launch any other workloads in this cluster, which makes it much more stable, much more secure. If you're a platform team, you can now guarantee uptime of these virtual clusters. You can upgrade them. You can have like rolling upgrade windows, right? There's a lot of things you can do with centralizing these control planes and then your workloads, they run in separate clusters so they can't reach the control plane unless through the regular API calls, right? And that's super valuable from a security perspective as well. How easy or difficult it is to work between, you know, the community version, which is Vcluster open source project and Vcluster Pro. We made the focus of this initial release to make it as easy as possible for people to get started with Vcluster Pro and explore it. So if you already have the Vcluster CLI installed, well, you're ready to go. The only thing you need to run is the command Vcluster Pro start and it's going to fire up Vcluster Pro in your cluster, which is a centralized control plane for all your virtual clusters and then essentially it immediately detects all these virtual clusters and then you can import them into Pro. That means you upgrade them technically, right? Because these Vclusters, they run their separate core DNS, right? And they don't have these features enabled like as their control plane. There's a couple of other security and performance features that are not available in the open source. If you are opening up the dashboard or you're running the command Vcluster import with one command, you can essentially turn a Vcluster into a Pro distro Vcluster and it couldn't be easier than that. And we also have the functionality to revert things back and give you a smooth path back to the open source. If Vcluster Pro is just something you were curious about and just wanted to look into that, but we're trying to keep them as compatible as possible and keep the experience as much the same as possible. So you have a very, very smooth transition. You're essentially running Vcluster Pro start and then you're running Vcluster create and now you're creating a Pro Vcluster instead of an open source one. Very, very smooth transition. Same CLI, same dox link, right? Obviously we show you which of the pages focus on Pro and which focus on the open source piece of the project, but we really want to make it a very integrated experience to not confuse things between the commercial Vcluster Pro and the open source version that's available. Talk a bit about the benefits and use cases for Vcluster Pro, not just Vcluster. Yeah, so I think anyone who wants to enable self-service for virtual clusters and who wants to run virtual clusters at scale in a very secure way, they're probably going to look into Vcluster Pro for CoreV, for example, right? They obviously, they're cloud provider, right? So security is of utmost importance for these guys, right? If you're using Vcluster for internal use cases, obviously security is a priority as well, but they're dealing with other people's workloads, right? With three, five different clusters, right? Running alongside each other on the same host cluster and they need to isolate them, right? They need to make sure that obviously customers can cross boundaries, right? And they're locked into the Vcluster because the Pro makes it incredibly easy and also it helps you with the CRD, right? You have a control on a CRD. Vcluster open source is just a help chart. So there's not a lot of lifecycle management. We all know that upgrading help charts, there's a lot of things that get you wrong when you do that. If you have a controller in a CRD, then you have a Vcluster object in Kubernetes now and that object is managed by our software that makes sure you can smoothly upgrade them. You can rollback changes, right? You can see which Vcluster is behind and it's a security patch, right? You can version and create templates for these individual Vclusters to say like, hey, this is our standard Vcluster with Argo CD batteries loaded, right? And if someone in the company needs that, boom, just instantiate the template and Argo is already running in there with the version and with the security constraints that you put in place for that setup. And that makes it really great for self-service. So if you're looking to make Vcluster available internally, that's a really great solution. If you're like CoreWeave in your cloud provider and you're literally spinning up Vclusters for other people or launch your Argo application side of it and then give Argo to your customers, that's another really great use case. So we typically see the internal use case, multi-tenancy for Kubernetes for internal use to reduce cost, right? Not to have so many clusters around, right? That's a lot of the reasons why people switch over to Vclusters instead of Vclusters. And then we see the production use case where companies like CoreWeave want to achieve multi-tenancy to host their own software or in this case, even cloud infrastructure that needs to be isolated on the Kubernetes level for their different customers. If you look at this space, the fact is that cloud native Kubernetes ecosystem, it's very like overall, maybe if you look at the CNC of LandScape, so many logos of their complexity is there and it is going to stay there. That is the fact that complexity is not going to go away. What we have to do is to make it easier for customers to deal with this complexity. There are a lot of players who offer managed Kubernetes solutions. They are known as managed Kubernetes providers. Even Google, with the JK Enterprise, they are trying to make it easier to kind of manage fleet of clusters. Do you compete with folks like Google or just talk a bit about, of course, this is a market where there will be a mix of solutions for folks who pick and choose whichever work for them. They will even mix two different vendors. But what I want to understand is that how do you differentiate yourself from some of these players and on what level you compete with them? We're glad that we found a niche with virtual clusters that only we are serving at this point. There are a couple of smaller companies that also have integrated our open source vcluster, but they're essentially relying purely on what's available in the open source. All these innovations that we put in vcluster pro, for example, in the past, this is a project that's been going on for like over nine months internally. There's so much innovation going in this as well that if you're looking into vcluster, it's the obvious choice to go with the makers of vcluster, right, with the inventors of the technology, with the ones that have the deep knowledge and really the deep experience on how to run these things for security related use cases or for large scale or really that enterprise environment and people trust us, right? They saw our open source, they saw the quality we put into vcluster and they know that vcluster pro and anything we add on top on the commercial side has that same level of sophistication and reliability that they can trust. I think with regards to competition, obviously there's offerings like EKS and GKE that make it very easy for you to create a Kubernetes cluster and we are complementary to them. You cannot run a vcluster without a real cluster, right? So ultimately, you've got to have an EKS cluster and then you launch five or 10 or 100 vclusters on top of it, right? So it's not like you won't need EKS, you will still need the nodes, the VMs that power that cluster, right? We're all going to pay the AWS tax ultimately, right? That's no way to do this without a cloud provider like CoreWeave or like AWS. And the great thing is, and you can see that with us partnering with CoreWeave, we're not really competitive to them. We actually drive more people towards cloud and towards spinning up more Kubernetes workloads because we make these clusters more cost-effective and because we make these clusters easier to scale and manage, right? Because managing 500 EKS clusters is really, really hard. Managing three of them and then having dynamic, ephemeral and non-ephemeral virtual clusters running on top of it is so much easier and that's why a lot of people go down that route. Of course, you folks keep working on new projects. A lot of things that you cannot talk officially at this point of camera, CubeCon is coming up. So we will hear a lot of things there, but if you can just get a glimpse, what are the things that loft is working on these days? Yeah, we launched another open source project, believe it or not, VCluster was a hit, right? We started it two and a half years ago and now we launched another project about a half a year ago called DevPod and that keeps us busy as well. It's like in the early stages of the hype cycle. It's another very exciting thing to work on. It's super, super early, right? Like we're very far away from anything like commercialization of the project or anything like that. It's really more like an exploration, but it's the massive fraction that was trending on Hacker News. That was like the day we launched it and it became trending on Hacker News. 60,000 people started using it day one, right? That was amazing to see and we got so much positive response in terms of like, oh, we've been waiting on like a GitHub code spaces, open source alternative, right? We've been starting this little Python project on the side to kind of emulate. It's great that you have this project out there and you open source it. I think there's a lot of really interesting stuff going on in the DevPod space as well. And I'm sure like 2024, we're going to have some really exciting announcements in that direction as well. But obviously for KubeCon, we are heavily focusing on the cluster because that's very dear to our heart and to our Kubernetes community. And it has so many exciting folks out there who are like just, you know, have so many good use cases for because there's so much untapped potential still here. We've just recently learned people are using it for internal workshops and e-learning and that's even things that we haven't even, you know, like we've been focusing so much on like multi-tenancy internal platforms, production multi-tenancy, right? That we haven't even tapped into like learning experience yet. So I think on both of these, you know, I guess we're a multi-product company now but we definitely drive both of these projects forward and I think there's a lot more exciting, you know, stuff to come. Lukas, thank you so much for taking time out today and walk us through, you know, the whole evolution of B cluster Pro, how it came to exist and the importance of customers that they play in this, the importance of the role that customers play in this space in kind of emergence of some new product, which actually end up helping everybody, you know. So, you know, rising tide lift everybody. That's what the saying goes. So thank you for all those insights and as usual, I look forward to chat with you again soon at KubeCon. Thank you. Absolutely. Thank you.