 Welcome to Amsterdam, and KubeCon, CloudNativeCon 2023, join John Furrier, Savannah Peterson, Rob Stretche, and UPscot, as the Kube covers the largest conference on Kubernetes, CloudNative, and open source technologies together with developers, engineers, and IT leaders from around the globe. Live coverage of KubeCon, CloudNativeCon 2023 is made possible by the support of Red Hat, the CNCF, and its ecosystem partners. Hello, and welcome to KubeCon Europe. I'm John Furrier, Rob Stretche, my co-host here, got a great guest here. Wall-to-wall coverage. We've got multiple days at CloudNativeCon, KubeCon, 2023, Europe, got a great guest here. Cube alumni, Merly, third, Umali, who's back here, formerly CEO of Portworx, now with Pure Storage. Great to see you. It's a passage, right up ahead. Every KubeCon, you come on and give us the update. Well, there's a lot going on. Every KubeCon has been something exciting. Well, at this time, we have a lot of exciting stuff, but I'd love to kind of talk a little bit about what we're seeing in the industry and what's happening with some of our customers, John and Rob. I think it's a great time in the industry, post-COVID actually, there's been a lot going on. So let me kind of talk about a few trends that I think are particularly interesting. The first one is actually is sort of the headline that forms in my head is DevOps is dead, long-lived the platform, right? And in fact, it speaks to not so much the demise of DevOps, but the success of DevOps to the point now where DevOps was a cultural phenomenon about 10 years ago, right? So, you know, developers and ops coming together and forming a joint alliance, but now that that has been so successful that it's resulted in kind of what we see as the emergence of platform engineering. So it's been formalized. So DevOps now has a budget, it has a team, and it has a set of tools that is, you know, called platform engineering. Essentially, the platform engineering team is there to provide a service of self-service to developers. And the old SRE role has been kind of mitigated and is now part of infrastructure, right? It's been so successful again that now SREs are there just to perform, you know, upgrades of network compute storage. And developers help themselves using the platform, which is usually a cloud-native platform, right? So platform engineering is usually cloud-native and it's anchored in Kubernetes, and it has to do with sort of this, you know, some sort of a Kubernetes distro, some sort of a security specific to Kubernetes. And the third part is something that is data on Kubernetes. So these are the three ingredients of platform engineering. Simar, I got to ask you, because we've been talking on theCUBE, Rob, you've been on many times with us on this. Platform engineering, we love, by the way, we think it's awesome. I'm sorry, it's captain obvious, right? I mean, I'm sorry. It is so obvious. But it's always been that, okay, SRE, Google kind of big iron. It's involved. There's been a lot of debate around what DevOps is. Oh, we do DevOps or you're the department. What's been the mainstream adoption of where it's settled in from a functional standpoint? Did Kubernetes bring that IT replacement kind of thing? Because we need to look at IT infrastructure. What is platform engineering now? How would you compare to us? There's a lot of debates. Not everyone could be Google. So that was one. That was obvious. Where has it standardized in the minds of the enterprise? It's actually very, very specific and very clear in my head. There is kind of the old infrastructure teams and there's a lot of on-prem infrastructure and there's a lot of cloud infrastructure and all of that, the infrastructure itself is being cloudized, so that is pretty straightforward. Developers have multiplied. What's changed is that apps have grown and so now where there used to be a hundred developers there's thousands of developers. The thing that's changed is, remember the old middleware, the old concept of middleware, the app server, all of that has now been replaced with this platform engineering. Platform engineering is the new middleware. It is now where people can now, the difference is the old middleware was always on call and was kind of still ticketing-based, now it's self-service-based. So platform engineering replaced the old middleware with a self-service model for developers. I think that what's really interesting about that is that you had a lot of developers who said, don't make me an SRE or that maybe they started out with junior developers in that role and I think we're platform engineering to exactly what you're saying from the customers I've been talking to is that they're really looking at that replacing IT, per se. Now it's become platform engineering and they're getting a new set of skills as well. Is that what you're seeing is that they're trying to understand that? Absolutely. So you know the platform engineering group is really anchored in two types of sets of technologies, cloud-native technologies and the other one is modern databases and modern data services. So Rob, the thing I'd say is the Kubernetes is really kind of three areas that we are seeing are where people require more and more expertise and are building part of the platform. One is the actual Kubernetes destroy itself, whether it's OpenChift or GKE or EKS and so on. The second one is security, whether some specific platforms, Prisma Cloud is obviously a good example. SysTake is a good example and the third one is data on Kubernetes of which Portworx is a great example where essentially people want to manage their storage resources, backup, DR and databases and data services underneath the auspices of Kubernetes. So that's kind of the key areas and then modern data services which of course are things like Postgres, Redis, Cassandra, Kafka, even streaming services like Spark, all of them being offered as a service by the platform team to the developers. That's what we're seeing. So one of the themes in the hallways here, obviously the AI which I'm going to get to later in the interview, but is which is going to be fun to hear your response to that because it's emerging and everyone's talking about it from the hallway. But then the hallway talk that we're hearing is the edge with Kubernetes. Maturity is here, we're seeing maturity and it's getting better. Automation is there, management is getting better. Edge is wide open. The Kubernetes at the edge, what's the state in your opinion? What's your view on the Kubernetes at the edge? As platform engineering teams look at the edge, it's programmable. You can put devices, you've got compute storage, data at the edge, AI is going to be there. What's the edge Kubernetes state of the union current situation? I think there's a lot of things about cloud native that enables the edge. One of them is the very concept of microservices. By having microservices now, everything doesn't need to be really one big, heavy monolithic application. You can kind of structure them as one app can be four to five microservices and you can run different microservices at the edge versus the core, depending on sort of the capability and depending on what you need run at it. So one is microservices are allowing a distributed mechanism. The second thing is just what you talked about, John, is that there are thinner things that you can run a slimmer version of Kubernetes. You can run slimmer storage versus the core. And the third thing is the availability of high bandwidth, right? So 5G services now enable you to kind of run a lot more stuff at the edge and then stream it back to the core. So the combination of distributed computing, lighter weight kind of protocols, and third thing is the availability of high bandwidth is enabling this. So we're seeing a lot of our customers in the retail area. Are they actually deploying or are they kicking the tires? They're deploying. And particularly food services. Food service vertical is one that we're seeing smart stores kind of emerge. The other one, of course, has always been one has been kind of anything that uses 5G, so distributed 5G services as well. And then finally, the old IoT, it still exists. Maybe less of a buzz. Now it's the reality of IoT. Those are the three areas. We had Audi on earlier there. The car is the ultimate edge. Yeah, and I think what's interesting is there's a lot of complexity that comes with Kubernetes. And I think you mentioned that early on. I think, what are you seeing that you're helping customers? Because one of the things just talking to some of the people in the hall and after the keynote this morning was, where do I get started? And then how do I operate it in day two? And that speaks to that whole platform engineering. Great question. So everybody wants platform engineering. Everybody has platform engineering. But not a lot of people have a lot of Kubernetes expertise. It's a pretty tough thing. Partly it's complex. Partly it's just availability of skills. So what has changed now? The other trend I would say in our industry is as a service versions of all of these platforms. So while I say platform engineering, platform engineering is available in two flavors. Typically it's been available in the past five, ten years as a software stack. But now it's available as a service stack. For example, now Portworx itself has offered now backup as a service. So we offer a best service under Kubernetes control. The other thing we're doing is offer something called PDS, Portworx Data Services. We're offering about 11 data services that are essentially a simple one-click kind of We host the control plane, right? So all you have to do is take your containers, plug in our PDS service into it. So platform engineering operating a platform is now as simple as collating all these services as opposed to actually kind of running them yourself. So you're saying if I hear you correctly, platform engineering teams are running the platforms where the Kubernetes piece is better managed because it's kind of heavy lifting that you don't really want to deal with because it's operationally impacting the engineers who have to come over here and work on Kubernetes. Yeah, what they're doing is essentially they are operating these as a collection of services, plug-and-play services. So it's no big surprise, right? Again, another kept an obvious thing. What's happening in the cloud is happening on prem as well, right? So people are taking that cloud operating model and trying to emulate it. And what's a good way to do it? Take your platform and turn it into a bunch of services and essentially offer plug-able services. So I got to ask you the question around, obviously Portworx has great success acquired by Pure Storage in 2020, continuing the container. Now Kubernetes containers, marriage made in heaven, a lot of virtualization migrating into off-beer metal into cloud native. So a lot of growth opportunities for sure. So good business opportunity, check, check, check. The AI impact now is upon us. In the past six months, it's been really kind of mind blowing. Everyone saw chat GPT, they go, wow, it's horizontally. Everyone loves it because it appeals to a horizontal use cases. People, prompting, prompt engineering, prompt tuning. We see this opportunity potentially coming in and providing auto tuning, self-healing, these are topics we've been talking about in the queue for a decade. Now with AI, this is opportunity, what's your view? What do you see? Because it's now just forming as a DevOps opportunity. Where do you see that with Kubernetes, DevOps, security? Where's the bullseye? Where's the frame? How would you frame Kubernetes and containers with an AI impact? Look, so first let me start by saying, we're clearly riding the hype curve of chat GPT and GPT. So with all that being said, let me actually start by saying what I think it's not, right? What I think AI is not is a replacement for a human, right? So it's not going to replace humans, at least not for a long time. However, what it can be is a strong aid to a human. In my view, I don't think GPT or AI takes a human and turns them into a super human. I think that's too high a standard to meet. I think it turns them into an Uber human. What does that mean? I think you can use chat GPT in a couple of different ways. One of them is to supplement a human in the following way. One way that we're seeing it being used is to either confirm or deny a hypothesis. So if you have a hypothesis, let's say you start out and say, I believe that containers are replacing VMs and therefore you don't need VM anymore. You can use chat GPT to kind of go out and collect data to say, can you confirm or deny this hypothesis? So one is it's a great way for you to check and test some hypothesis. The other way is if you believe and you have some information that you believe to be true, you can use chat GPT to go collect information to help you support that hypothesis. So what you really need to kind of, there needs to be a human at the end of that to look at that data. Now the other place where I do think there's going to be huge strides in the way we are ourselves using at Pure and Portworx is in very, very detailed, vertical slices where you know the data to be true because it's your data. When it's your data and it's a very, very specific set of tightly controlled, tightly defined use cases, then you can come in and turn GPT to provide very, very, very specific, useful things. A great example for that is test cases, right? If you want to generate test cases for a new capability, you can point it to all of the existing tests, the suites that you have, BVTs, smoke tests, system tests, and then it can generate, chat GPT or GPT with its APIs can generate very, very specific new test cases. And those are going to be good test cases because they're all based off of very accurate data. So it's kind of the old ego, right? Garbage in garbage out. Well, yeah, here you have sort of the ability. So to me, as an aid to a human with some human judgment being applied and the second is in very, very specific vertical use cases is where you could use it. That's sort of my view. Rob, what's your take real quick? Yeah, I think it's dead on with that. I think it's the quality of the data. And I think in that, it's how you have to manage that data. And I think this is where things like Port Works helps you really get there. And I think that's the critical thing that's been missing from a lot of these is how do you manage all of this data? And I think it seems like that's the route that you guys are going down as well is to help them manage all of that with the things like Kafka integration and helping people get to those streams. Yeah, no, I should point out one thing, right? Let's not forget ML ops and not just worry. You know, there's AI and there's ML. ML, by the way, is AI without the mystery, right? And actually, we've had machine learning. We have something called autopilot which learns how your storage is being distributed by the application and we can now apply some automatic rules learning from the machine. So ML ops is way ahead and it's actually much more reliable and it's something that we've been using for a while inside of our product with our autopilot solutions. And that's what Andy Jassy said when Amazon launched, they've been using ML for years. This is where the pivot is. As the new wave comes in, you're just going to extend out. So final question we're running out of time real quick is what's going on with Portworx? How's this KubeCon going for you? What are some of the meetings you're having? Give a plug for the company and the opportunity to ask you. Yeah, look, I mean, we are having a great deal of success with the emergence of platform engineering. Now there's kind of a budget, there's a center in the organization where these decisions are being made and being driven. So with that, what we're doing is really driving a couple of things. One of them is we're providing everything as a service. Everything Portworx is now available as a service. So we're offering backup as a service, we're offering Portworx data services with 11 data services so on. So that's one big theme that's happening. The other thing that is happening in the industry, frankly, we're beginning to see a VM takeout by containers. I don't know how much you're kind of seeing. There's a lot of migration. Over a quarter of our deployments are on bare metal. And essentially, if you think about it, it's very, very straightforward. Kubernetes sprays the application to the next free node with the free CPU available. And Portworx does the same with the storage. So combination of something like an open ship with the Portworx is basically, Portworx is like V-SAN for Kubernetes, right? And then what Kubernetes itself does is provide that virtualization. And so folks are saying, I don't need to pay a V-Tax anymore. Now, we're deployed in a lot of VM applications, but essentially, with the help of Kubevert, you don't even need to kind of stand up a complete VMware shop. So there's a lot of opportunity. And also the enablement of the apps. People want to write the modern apps, AI native. I'm really going to leave it there. Thanks for coming on. Yeah, thank you. Great update. Portworx, a big success story. Now part of pure storage. Again, storage computer that's going to enable more opportunities. This is the Kube. Here, KubeCon, stay with us for more live coverage. The leader in tech coverage. We'll be right back.