 Today we have two guests from the QBird community, Fabian Deutsch, senior engineering manager at Red Hat and Ryan Halsey, senior software engineer at NVIDIA. Ryan, Fabian, it's great to have you both on the show. Nice to be here. Okay. Let's just start with some of the basics, which is, first of all, tell us, you know, what is this QBird project all about? QBird is an extension to Kubernetes with an API to define VMs and a runtime to run VMs. And that's an important part, right? So you're not able to only define and have the VMs run somewhere, but if you define the VMs using QBird, then the VMs are run on the Kubernetes cluster alongside the other pods, right, the regular workloads of your cluster. In fact, they are run as pods on the cluster as well, because QBird believes that a pod is the atomic computing unit of Kubernetes. QBird is a Kubernetes extension. It allows you to run virtual machines inside a Kubernetes cluster, just alongside any sort of any other workload that you'd have, like a pod or anything else, a virtual machine. It's just another workload that you can run inside the cluster. And so what's really exciting about that is, you know, we look at virtual machines today, they're the use case that'll never go away, just like containers. And it's really cool that we can actually take something like a virtual machine and we can apply it on top of an ecosystem like Kubernetes and actually run them alongside pods. So it's a cool use case. It allows people to take their traditional applications and actually run them alongside what people call the new age, the cloud-native workloads and sort of enable that smooth transition to the cloud. We kind of live in a cloud-driven or cloud-centric world. So from the kind of world that we are moving towards from the traditional IT, where does QBird fit into the larger cloud-native ecosystem? QBird is a reflection of the fact that VMs are probably not going away. They will at least last to be there to run Kubernetes itself. So if we look at the cloud-native ecosystem, Kubernetes is a central tool in that ecosystem. It allows you to, it's enabling many workflows. GitOps is a key word, empowering developers to bring their workloads to the cloud. But the requirement to run virtual machines for certain workloads doesn't go away. But we believe that containers are, and the container workflows are a valuable incremental step into where we can deliver applications more efficiently. By now giving users and also cluster administrators, actually whole organizations, the ability to run virtual machines on Kubernetes as well, we hope to reduce the headaches because there are fewer systems to care about. You've got one platform to run your compute workloads and that is what QBird helps with. So it helps you to worry less about other systems to run your production-ready virtual machines by focusing on Kubernetes alone. QBird and its role in the cloud-native ecosystem is really pretty interesting. Like I mentioned, it's one use case in terms of a workload. A pod is one and Kubernetes has many different ways to look at workloads. And what's really cool is that the virtual machine use case is one that is different than a container. And I think everyone would acknowledge that. And I think that there will probably always be someone that will want to run a virtual machine. And if you really love Kubernetes, I think it's cool that you can actually take what you love about Kubernetes and use it just like you would and still have the opportunity to run a virtual machine on. I think that's awesome. It's really empowering for developers. It's empowering for a lot of companies and vendors who want to actually still have that use case. And it's great because it really combines those two worlds and it does it seamlessly. And so it's really its role in the ecosystem is really empowering a lot of people to not only use the new container workloads, but also to use the traditional virtualization workloads. What are the other competing projects that you see? And what sets QBird apart from them? Then that will lead to the next question, which is more about incubation and moving further towards graduation phase. So let's talk about QBird and the projects that are there. So if we specifically look at the CNCF landscape, right, sandbox, incubation projects and the graduated projects, then there are a couple of projects that are virtualization related. So we've got, for example, G visor, the Cata project at Firecracker, for example, which in one way or the other are using KVM as a hypervisor technology or where it can be used. Sometimes they're different back ends. These technologies primarily focus or all of them have a slightly different focus, I would say, right. So if we look at Cata, for example, and G visor, but also Firecracker, it's more or less about isolating containers, right, adding another layer of isolation to existing container workloads. Important to understand is that in all of these three cases, working with a container and adopting container workflows is maintained. So a layer of isolation is added transparently when running containers. QBird, on the other hand, is specifically about running VMs, right. So the asset you're working with is not a container, but rather a traditional virtual machine. So a lot of stuff that we actually implemented in the QBird project and that we are adding to the ecosystem is about dealing with legacy virtual machines in the ecosystem. And now, getting back to the question, if so, if we look at the broader landscape, QBird is the only project at that maturity level that is really dealing with virtual machines by themselves, right, as an asset. So QBird, in sort of its role in the larger ecosystem, there are certainly different different projects out there that are tackling different use cases along with virtual machines and isolation, really, at the container level. And I think there's a lot of different ways you can view these projects like this. You've seen a few iterations, like Cata containers, Clear Containers, we've seen Cheat Advisor, Firecracker, all these things that exist in the ecosystem. And they're really sort of different ways that we can run workloads. And I think really the biggest differentiation between those projects and QBird is QBird is about taking what is really Kubernetes-native, the idea that you can take a workload and treat it just like any other pod that you would in Kubernetes. And you take the networking and you can detach networking the same way. Everything from live migration to all the things we like about our virtual machines, they're sort of applied in a Kubernetes way. And so really, the idea of QBird and where it differentiates itself, it's about looking at virtual machines and managing them through a Kubernetes way. And in a way that it integrates well with Kubernetes. And so it's not necessarily like, okay, we're going to isolate with workloads with this new technology or something like that. It's more like we're going to integrate and run virtual machines inside containers and we're going to integrate within the ecosystem. And we're going to provide all these different tools to make it very easy to do that. QBird has moved from the CNC of Sandbox to Incubation. The interesting thing with most of these CNC projects is that most of them are already being used in production. So when we look at the term Incubation, from that point of view, it doesn't like, hey, these projects actually are all in graduation stage. Your companies are running in production. So talk about what does this mean for the user community, for the developer community. And if you can also kind of also reflect on that when a project moves from one stage to another, is there some criteria that they have to fulfill? How do you know that, hey, you know what, this project is not ready for, you know, this stage? Yeah. So the CNCF really has taken the time to formalize the Incubation process, right? Just to look at what requirements do we put on a project to let it move up the ladder into Incubation? And I think it's all the good. So we went through that project. It's actually a couple of pages long, right? If you go to the relevant PR and the CNCF TOC, get a repository, I think there's a link to the document. It's all public, right? So you can review it for yourself. They ask about a lot of things, right? There are for sure stats about the project, right? How many adopters do you have? How many developers do we have? How many companies are interested in the project by contributing to it? But there are also questions about how's the governance model, right? And what kind of CICD do we have, right? Are we quality testing our code? How do we deal with security incidents? And so there's really, I would say, they're looking for the maturity of the project and not only from the code quality or from the adopters point of view, but a more holistic view. But I think that's great, right? Because as a consumer of the product, and now we get to the vendors and to the users, by seeing a project moving into the incubator, you're sure, right, that the CNCF has applied the requirements on the project and you can be sure, right, that there are certain processes in place that are required in order to really use that project and product, right? The things you want to rely on, that they're regular releases, that are well-defined releases, which is important also for vendors, right, to rebase potentially downstream products on this project. And all of this was reviewed during the incubation process. So the CNCF provides quite a few guidelines in terms of how projects go through the graduation process and, you know, Keyword has followed those guidelines and really has reached the point where a number of things like, you know, from the code coverage, the testing, the releases and the adoption, that it made sense for the community to continue on its journey up the ladder in graduation. And it's really exciting for the project. The project has really reached fairly wide adoption and solved a lot of use cases. And so it really makes sense for the product to really continue up into the graduation process to continue to expand its reach and to continue to grow its adoption. And I think, you know, just based on, you know, the quality and success the community has had, it's well-deserved and it's very exciting. And so that's why it's that Keyword has made this jump up to the incubation. Excellent. And as Kevin was touching upon, I also want to hear your thoughts. What does this mean for the user community or the developer community, oh, the maintainers have to like, first of all, because as we're discussing earlier, most of these projects are already used in protection, but it was more about, you know, giving that confidence, you know, there's a, you know, community around it, the code quality is also good. So if you can quickly reflect on that. Yeah, sure. So like, you know, the maintainers and, you know, the adopters of the project, I think it's good. It's like, you know, any project that, you know, as they go through different governance stages and in terms of their larger communities, it's, you know, it's important, you know, that they get the recognition that they deserve for the work that they've done. And, you know, this is the structure that CNCF puts in place in order for projects to really graduate and be recognized for their success in the community. And so it's important, you know, that, you know, that Keyword gets that recognition. And so it's important for the, you know, the maintainers, adopters, you know, that they feel like that, you know, that their work has really succeeded in propelling this community forward. And for anyone new that's, you know, is looking, you know, for a kind of solution like this, it's also a great stamp of approval that, you know, the community does have a lot of trust and puts a lot of trust into Keyword, and as well as you can see from their adopters and maintainers that they do as well. So it's good, you know, from many different perspectives, you know, from, you know, from how the far reach that the CNCF has that we can continue to get more people involved in the project, more people using it means, you know, more maintainers, more use cases, and the product just continues to improve and grow. What are the things that are new pipeline? What are the things, some features that you are working on? You're like, hey, these are the things we are excited about. These are the things that we are working on. I think there is, I think Keyword is at a good level, right? So if you want to get started running VMs today, Keyword is a good point to get started. I think also the getting started material is fairly welcoming and you can be successful in a couple of minutes. Also, thankfully to the integration like with MiniCube and to our operator, which makes it really easy to deploy Keyword. However, there are certain things that we want to address. There are some usability things. So Keyword tried to catch up with a lot of the legacy virtualization stuff that got invented over the last 20 years, right? So we rapidly added support for many virtualization features. And I think a round of usability focus is really helpful to make it ultimately easier for the consumer for VM, for a user to write them. There's also work we need to do in the scalability area, right? I think we're not that bad, but we speak literally about running thousands and tens of thousands of VMs on Kubernetes clusters. And we want to make that more efficient, right? We know that it's working today, but we always want to be careful about resources because they're not endless. Users, I think we need to pay some more attention to what is easy to get started, but to make all the features of Keyword discoverable, right? So I think paying some more attention to our user guide and also at conferences, right? Connecting better with our users. That's something that's for sure my list and make sure that we hear the voice of our users, of our adopters to see that we stay on the right path, right? Because we just started and it's great to see that we're useful to people, but we want to stay on that path. So listen to our users. There's many exciting features that the community has been developing and, you know, from stuff just from usability standpoint, from upgrades to all things in scalability and even the traditional virtual machine virtualization features that the community has been working really hard on, you know, from, you know, compute isolation in Yuma to device management, all the exciting stuff that goes around data retention and implementing things like live migration, things like isolating and security privileges for for the virtual machine and improving that area. Things like even save, restore, offline virtual machine migration. All things are very, very exciting and even scale and performance are also huge areas and huge focuses in the community. And, you know, there's always a question people always ask, you know, how many nodes can you scale to? And it's something that we've, you know, we've been working hard on, you know, Kubernetes is scales theoretically to 5,000 nodes and it's something that we've been working in the community to reaching and I think some people and we've seen this as high as over 1,000 nodes. And it's something that we continue to, you know, to strive and prove on to reach that 5,000 node number. So that's also very exciting. There's been a lot of work in metrics around performance. So you can actually measure how well a virtual machine can go and transition through the phases, which is very, very cool to actually see in real time and measure and then you can work on and, you know, improving different co-pasts and improving performance based on that information. So there's been a lot of things and, you know, the community's got a lot of exciting work and sort of the way that I would summarize it all is that, you know, there's, you know, anything with a new young platform that around virtualization, there's going to be many traditional ideas and things that have developed and the traditional virtualization software that are going to be needed to apply it onto things like Qvert and done in a Kubernetes way. And a lot of those features are have been developed and are continuing to be developed. And so there's, there's many, there's many, there's more to come and it's, it's pretty exciting. Fabian, Ryan, thank you so much for taking time out today. And of course, talk about not all the project, the whole move from sandbox to petition, but also what role is playing in the larger, you know, cloud native Kubernetes ecosystem and also give us a glimpse on what are the things that we should be excited about or looking forward to. So thank you for sharing all those insights and I would love to have you back on the show. We may not wait all the way for graduation, maybe before that when something new happens, but I really appreciate your time today. Thank you. Thank you. Thank you.