 The Cube presents KubeCon and CloudNativeCon Europe 2022, brought to you by Red Hat, the CloudNative Computing Foundation and its ecosystem partners. Welcome to Valencia, Spain. At KubeCon, CloudNativeCon Europe 2022, I'm your host Keith Townsend, along with Paul Gillan, senior editor, enterprise architecture for SiliconANGLE. Paul, we're going to talk to some amazing people this week. KubeCon, what, the energy here, what would you say about it? I'd say it's reminiscent of early year, early stage conferences I've seen with other technologies. There is a lot of startup activity here, there's a lot of money in the market, despite the sell-off in the stock market lately. A lot of anticipation that there could be big exits, there could be big things ahead for these companies. You don't see that when you go to the big established conferences. You see just anticipation here that I don't think you see. You'll see maybe in a couple of years. So it's fun to be here right now. I'm sure it'll be a very different experience in two or three years. So, welcome to our guest, KubeAlone Basam Tabar, the founder and CEO of UpBound. Welcome back. Thank you, yeah, pleasure to be on the show again. So Paul, tell us, we're in this phase of migrations and moving to CloudNativeStacks. Are we in another replatforming generation? I mean, the enterprise has done this time and time again, whether it's from Java to .NET or .NET to Java or from bare metal to VMs. But are we in another age of replatforming? You know, it's interesting that every company is now become a tech company. And every tech company needs to build a very modern, digital platform for them to actually run their business. And if they don't do that, then they'll probably be out of business. And it is interesting to think about how companies are platforming and replatforming. Like, you know, as you said, just a few years back, you know, we were all people using Cloud Foundry or using Heroku, you hear Heroku a lot. Or, you know, now it's CloudNative and Kubernetes. And it begs the question, you know, is this the end to your point? Is this, you know, do we have, you know, what makes us sure that this is the, you know, the last platform, the future proof platform that people are building? There's never a last platform, right? There's always something around the core. The question is, is Kubernetes Linux or is it Windows? That's a good question. It's more like Linux, I think. You know, you've heard this before, but people talk about Kubernetes as a platform, off platforms. You can use it to build other platforms if you know what you're doing. You can probably put, assemble a set of pieces around it and arrive at something that looks and can work for your business, but it requires a ton of talent. It requires a lot of people that actually can, you know, know how to put this thing together to work for your business. It is, there's not a lot of guidance. I think we were chatting earlier about the CSEF landscape and how there are all these different projects and companies around it, but they don't come together in meaningful ways. That actually, the enterprise itself has to figure out how to bring them together, right? And that's the combination of what they do there organically or not, is their platform, right? And that changes, it can change over time. Do you think they really do, they really want to put these things together? I mean, that's not what enterprises like to do. They want to find someone who's going to come in and turn key, do it all for them. Yeah, and if there was, this is the things. Like every week now you hear about another platform that says, this is the new Heroku, this is the new Cloud Foundry, this replaces ever, you know, some vendor has, and you can see them all around here, you know, companies that are basically selling platform solutions that do put them together. And the problem with it is that you typically outgrow these, like you're, it might solve 80% of the use cases you care about, but the other 20% are now represented. And so you end up outgrowing the platform itself, right? And the choice has been mostly around, you know, do you buy something off the shelf that solves 80% of your use cases? Or do you build something on your own? And then you have to spend all your resources actually going through and building all of it. And that's been the dilemma, you know, people who talk about this as a platform dilemma, but it's been, it's been the way for a long time. Like you, every, we go through this cycle every few years, and you know, people end up essentially oscillating between buying something off the, you know, that's off the shelf or building it, building it themselves. So what's the payoff? If I'm a CIO and I'm looking at the landscape, I don't need to understand, you know, I don't need to know what a pod is to know that looking at 200 plus projects in cloud native foundation and the bevy of co-located projects and conferences before the start of this, what's the payoff? Increasing the pace of innovation. I mean, that literally is, when we talk to customers, they all say roughly the same thing. They want something that works for their business. They want something that helps them take their, you know, line of business applications to production in a much quicker way, lets them innovate, lets them create higher engineers that can don't have to understand everything about every system, but can actually specialize and focus on the parts that they care about. But it's all in the context of, you know, people want to be able to innovate at a very high pace. Otherwise, they get disrupted. So I was at, you know, my favorite part of KubeCon in general is the hallway track and talking to people on the ground doing cool things. I was talking to a engineer who was able to take their Java stack, their .NET stack and start to create APIs between them, break them into microservices. Now teams are working across one another. Realizing that promise of innovation, but that was the end point, they're there. As companies are thinking about replatforming, where do we start? I mean, I'm looking at the CNCF, the map and this 200 plus projects, what do I start? You typically today start with Kubernetes and a lot of companies have now deployed Kubernetes to production as a container orchestrator. Whether they're going through a vendor or not, but now you're seeing all the things around it, whether it's CI CD or GitOps that they're looking at, or starting to build consoles around their platforms or looking at managing more than just containers. And that's a theme that we're seeing a lot now. People want to actually bring this modern stack to manage not just container workloads, but start looking at databases and cloud workloads and everything else that they're doing around it. Honestly, everybody's trying to do the same thing. They're trying to arrive at a single point of control, a single platform that can do it all, that they can centralize policies, centralize controls, to compliance, governance, cost controls, and then expose a self-service experience to the developers. Like they're all trying to build what we probably call an internal cloud platform. They talk about it in different ways, but almost everyone is trying to build some internal platform that sits on top of on-premises and on top of clouds, depending on their scenarios. You make an interesting point, which is that everyone here is, to some extent, trying to do the same thing. And there's fine points of granularity between how they're approaching it. As you walk around this floor, do you understand what all of these companies are doing? Not sure I understand all of them, but I do recognize a lot of them, yes. And in terms of your approach, you use the term control plane. What is distinctive about your approach? Very good question. So, we're trying to solve this problem as well. We're trying to help people build their own platforms. But let me, there's a lot to it, so let me actually step back and talk about the architecture of this. But if you were to look at any cloud platform, let's take the largest one, AWS. If you peek behind the scenes at AWS, it's basically a set of independent services, EC2, S3, databases, et cetera, that are essentially working on different parts of, like offer completely different pricing, different services, et cetera. They come together because they all integrate into a control plane. It's the thing that serves an API, it's the thing that gives it all a common feel. It's where you do access control, it's where you do billing, metering, cost control, policy, et cetera, right? And so our realization was, if the enterprises are platforming and replatforming, why shouldn't they build their platform in the same way that the cloud vendors build theirs? And so we started this project almost four years ago now, three and a half years, called Crossplane, which is essentially an open source control plane that can become the integration point for all services and essentially gives you a universal control plane for cloud. So you mentioned the idea of orchestrating or managing stuff other than containers. I think about companies that built amazing platforms, enterprise companies building amazing applications on AWS 10 years ago, and they're adopting the AWS control plane, and now I'm looking at Kubernetes. Is Kubernetes the way to multi-cloud, to be able to control those discrete services in the AWS or Google Cloud, Azure or Oracle Cloud, et cetera? We kind of have the tease of the part. So there are really two parts to Kubernetes, and everybody thinks of Kubernetes as a container orchestration platform, right? And there is a sense that people say, if I was to run Kubernetes on everywhere and can build everything on top of containers, that I get some kind of portability across clouds, that I can put things in containers and then they magically run in different environments. In reality, what we've seen is not everything fits in containers. It's not going to be, the world is not going to look like containers on the bottom, everything else is on top. Instead, what we're going to see is essentially a set of services that people are using across the different vendors. So if you look at, you could be at AWS shop primarily, but I bet you're using Confluent or Elastic or Databricks or Snowflake or Mongo or other services. I bet you're using things that are on-premises, right? And so when you look at that and you say, to build my platform as an enterprise, I have to consume services from multiple vendors, even if it's just one major cloud vendor, but I'm consuming services from others, how do I bring them together in meaningful ways so that I can build my platform on top of the collection of them and offer something that my developers can consume and self-service on? That's not just containers. What's interesting though is if you look at Kubernetes and look inside it, Kubernetes built a control plane that's actually quite useful and applicable outside of container scenarios. So this whole notion of CRDs and controllers, if you've heard that term, the ability, you know, like there are two parts to Kubernetes. There is a control plane and then there's the container workloads. And the control plane is generic. It could be used literally across, you can use it to manage things that are completely outside of container workloads. And that's what we did with cross-plane. We took the control plane of Kubernetes and then built bindings providers that connected to AWS, to Google, to Azure, to DigitalOcean, to all these different environments so you can bring the way of managing, you know, that style of managing that Kubernetes invented to more than just containers. You can now manage cloud services using the same approach that you are now using with Kubernetes and using the entire ecosystem of tooling around it. Enterprises have been under pressure to re-platform for a long time. It was first go to Unix, then to Linux, then to virtualize, then to move to the cloud. Now Kubernetes, do you think that this is the stack that enterprises can finally commit to? I think if you take the orientation of your deploying a control plane within your enterprise that is extensible, that enables you to actually connect it to all the things that are under your domain, that that actually can be a future proof way of doing a platform. And you know, if you look at the largest cloud platforms, AWS has been around for at least 15 years now. And they really haven't changed the architecture of AWS significantly. It's still a control plane, a set of control planes that are managing services. It's a legacy. They've added a lot of services. They've had a ton of diversity. They've added so many different things, but the architecture is still a hub and spoke that they built, right? And if the enterprise can take the same orientation, put a control plane, let it manage all the things that you know about today, arrive at a single point of control, have a single point where you can enforce policy, compliance, cost controls, et cetera, and then expose a self-service experience to your developers. That actually can become future proof. So we've heard this promise before, the cloud of clouds basically. Yes. To be able to manage everything. What we find is the devil's in the details. The being able to say, you know, a load balancer issuing a command to deploy a load balancer in AWS is different than it is in Azure, which is different than it is in GCP. How do enterprises know that we can talk to a single control plane to do that? I mean, that just seems extremely difficult to manage. Oh yeah, the approach is not, you're not trying to create a lowest common denominator between clouds. That's a really, really hard problem. And in fact, you get relegated to just using this, you know, really shallow features of each if you're going to do that. Like your example of load balancers. Load balancers look completely different between cloud vendors. The approach that we kind of advocate for is that you shouldn't think of them as, you shouldn't try to unify them in a way that makes them, you know, there's a global abstraction that says, oh, there's a load balancer and it somehow magically works across the different cloud vendors. I think that's a really, really hard thing to do as you pointed out. However, if you bring them all under a same control plane, as different as they are, you're able to now apply policies, you're able to set cost controls, you're able to expose a self-service experience on top of them, even if they are very different. And that's something that I think has, you know, been hard to do in the past. So, Basam, we'll love to dig deeper into this in future segments and I'm going to take a look at the product and project and see where you folks land in this conversation. From Valencia, Spain, I'm Keith Townsend along with Paul Gillan. And you're watching The Q, the leader in high tech coverage.