 Live from Austin, Texas, it's theCUBE, covering KubeCon and CloudNativeCon 2017. Brought to you by Red Hat, the Linux Foundation, and theCUBE's ecosystem partners. Hello everyone, welcome back to theCUBE's exclusive coverage live in Austin, Texas, here for the three-day CloudNative, but now two days of KubeCon, Kubernetes conference, the second annual conference celebrating the evolution and growth of Kubernetes, I'm John Furrier, my co-student and our next guest, Steve Watt, Chief Architect of Emerging Technologies at Red Hat. Welcome back to theCUBE, good to see you. Thanks for having me, always a pleasure. So Red Hat making some good bets with Kubernetes, not a bad call. No, Kubernetes has done wonders for our OpenShift business here, absolutely. So how's this all playing out? Do we just talk before we came on camera here about just the pace of change? You've been at Red Hat five years, we interviewed when you were at HP during the big day-to-days. Boy, the world certainly has grown and changed. What has changed in your mind the most that people need to understand? You know, I think Kubernetes has been the single biggest driving force to shift all enterprise architecture from scale up to scale out, and I think that has just created a whole number of ripple effects across how applications are designed within the enterprise. I think that's the big one. So Steve, that whole shift from scale up to scale out has affected lots of parts of the stack, but storage is something you've been working on, something we've been keeping a close eye on, and it was one of the top items we wanted to kind of dig into this week. Maybe bring us inside a little bit, what's happening, what's Red Hat's role? Help explain. Absolutely, one of my favorite topics, right? It's kind of counter-intuitive. I work in the CT office, I run the Emerging Technologies team, which is sort of the team that does the experiments that help shape and inform our long-term strategy. And so you might think, well, storage is kind of old news, like how does that fit into this cloud native world? Why does Red Hat care about it so much for their platform? And I think if you look at the cloud native stack today, right? You have GKE, the new Amazon Kubernetes service, Azure, et cetera. These are all places where you can run your Kubernetes app, but just in that one place. Red Hat's platform perspective is a little different, right? We want you to be able to run your platform in the open hybrid cloud, whether that's in Google, in Azure, or on-premise, on open stack, or on bare metal, right? So you want to be able to run everywhere, but what's the biggest problem to achieving that application portability? It's data locking, right? So storage becomes cool again. We've got to solve this problem. Because you've got to store the data somewhere. Right. That's the storage devices. Right, exactly. In the new way, the new architecture. The new architectures, right? So the problem is you've got to be very careful that if you want to move ever, you should think upfront about your persistence platform so that it gives you the freedom to be able to move around. So Red Hat is investing heavily in trying to solve this problem. We've got a few exploratory prototypes that we're actually showing at this conference. And we work in both Kubernetes, building out the storage subsystem there, but also sort of in our products for container native storage. Yeah, Steve, walk us through a little bit because we've been talking about this in the Docker ecosystem for a bunch of years. Where are we? What's being worked on? What still needs to be sorted out? So, yeah, that's interesting. I think we're finally over the hump where everybody's asking who's solving the persistence problem for containers? They used to drive me crazy. That went on for about three years. I think people finally realize there are solutions. Kubernetes has always had them, actually. And so we've got past sort of the day one, like being able to dynamically provision, kind of like you would see with Cinder and OpenStack, be able to create storage. You've got a vibrant, huge storage ecosystem. At our Kubernetes face-to-face meetings, we have 50 people. They're like a mini conference. So we've got broad engagement from the entire storage ecosystem. And that's doing everything that you need sort of on the file level. But there is recent nascent work that we've done in Kubernetes for Service Broker is now the pattern to sort of provision object storage if you need it. And most importantly, we've just enabled block storage in Kubernetes in the 1.9 release that ships this week. And that is really interesting because it opens up the potential to run virtualization workloads on Kubernetes. Where's the action for the project with storage? I heard some hallway rumblings just around the Rook project. Yeah. Is that something that, I mean, what projects, if I'm interested in storage, where do I dive in? Where's the most action for moving the needle, continuing the innovation around storage? I think it's, you know, if you're a storage vendor, it's different if you're a storage consumer. So Rook is a project that's focused on providing a sort of an abstraction for software defined storage platforms to run inside Kubernetes. Gluster doesn't take that approach. We use sort of more of the pure Kubernetes approach, sort of get to the same place. But Rook is, you know, it's definitely an interesting project in that. Sort of an inception level project phase. Then, you know, for people that are wanting to consume storage, I think Kubernetes is the king of the pack. You know, I'm obviously a strong opinion on it amongst the other container orchestrators, but the amount of investment in allowing people to do, you know, more sophisticated, continually more sophisticated features, you know, snapshotting, you know, cloning, things like that. And then obviously, I'm sure you've heard a little bit about container storage interface, CSI. And that makes it a lot easier for storage vendors to build one adapter that works across, you know, decors, Cloud Foundry, Kubernetes, et cetera. What's the biggest surprise here for you? Because we've been looking, trying to read the tea leaves, obviously Kubernetes has cleared the runway, good standardization, seeing some commoditization, great adoption, although people can tailor it, a lot of different versions. Still early? Yeah. We have only two years old conference. I know. Three years it's been around. What's surprising you right now? What's jumping out at you? I think Amazon's announcement yesterday was very interesting. I think the fact that it's heartening to see that there's pure Kubernetes as a service being offered in Azure, Google, and Amazon. And I think that's quite interesting for a portability standpoint, right? And so I think to me that was big surprise. Amazon doesn't usually go the pure vanilla open source approach, and also their statements that they're going to contribute back to Kubernetes, I think is quite interesting as well. So to me that's the one thing that stood out. What's going on for the future too? You mentioned you have to set the roadmap. You guys have an agenda there, obviously have installed base with Linux, and now you got OpenShift doing really well. What are you guys looking at? What's on your radar? How do you see this thing unfolding? What's in your mind? Yeah, I think there's a couple of really interesting things, right? I think container orchestration is a legitimate disruption to virtualization, right? In that it solves the same problem of opportunity space, but in a fundamentally different manner that reshapes the market. I think the KubeVirt project is something that we're working on at Red Hat. It's another one of our sort of emerging technology focus areas. And when we enable block storage and it enables virtualization, what it gives us the opportunity to do in Kubernetes is have a single deployed platform that can serve both late adopters and early adopters. So the early adopters with pure container orchestration, but if you're wanting to have the same platform and do virtualization too on it, you can have sort of one investment, one shared experience to be able to do all of this. I think that's pretty cool. Steve, talk about the customers that are watching or will be hearing over the next few months and year around kind of how to architecturally package this and think about in their mind, whether it's a mental model or specifics, because there's always going to be that time-tested trade-off between performance, security, and so you have, obviously people have VMs that go in a way, but containerization, Google say, hey, we don't really care about VMs, we're a container company. There's always still going to be trade-offs. Speed, security. Yeah, so security factors in there, how should a practitioner think about getting their arms around this? You know, I think this is the tack that OpenShift takes, right? Which is that Kubernetes is a nascent project, despite the huge amount of interest and contributions that we have and on its maturity curve, as far as there's different things at tension, right? Like enterprise use cases versus public cloud use cases, right? And so we're very focused on our enterprise use cases and sort of enabling that inside OpenShift and bringing OpenShift up as a platform back to the sort of enterprise level that our customers would expect. Virtualization platforms are much further down the maturity curve, right? And so I think that's sort of our approach is that, like, Red Hat tries to meet our customers where they are. Some organizations have teams that are more advanced, some that are less advanced. And so we try to offer, you know, if you want to go virtualization, we've got OpenStack, we've got Rev, if you want to use this new school, Kubernetes-based container orchestration, and you've got teams understand it, and you grok microservices, then we've got a solution for that. Well, you know, the whole theme here is infrastructure is boring. Storage, as used to be called, Snorage back in the day, is pretty boring, but relevant. Most people look at, like, Lambda from Amazon and some of the serverless trends and certainly seeing it here with service measures and whatnot. The abstraction wave of infrastructure, it's almost eliminating storage from the mind of the developer, yet it's changing. How are you guys specifically riding that wave? Because, one, it's good for developers. Right. The velocity of developers increases, but the role of storage has changed. You mentioned block. People are like, oh, block and object's dead. I mean, storage has been dead for, what, 20 years now? Sleep's growing and growing, but now the role changes to the developer, abstracted away, and also more important for automation and some of the DevOps things. What specifically are you guys doing? So, I think you said the word role, that's really important, right? So, like, to an application developer, what you said is absolutely true, right? They want to use persistence platforms for storing their data in a cloud-native way, okay? However, the maturity curve is also important, right? Not every application developer team is fully microservice-based and understands all these architectural patterns. It's a journey, right? So, we want to basically give them multiple options along their journey. So, that's the one around the application persistence. So, if they're used to file storage or object storage, et cetera, like we have our container-native storage platform provides that for them from the application persistence level. But from an open-shift standpoint, right? Open-shift is our new platform, right? It's based on RHEL, but it's our new platform, our new surface area to build applications and, most notably, infrastructure services on this. So, just like with RHEL where we have, we created the opportunity to have a fertile ecosystem around it, we're doing the same with open-shift, which means that we've got to enable the companies that are providing those persistence platforms, those message queues, those NoSQL databases to run on open-shift, right? You want to run Cassandra on open-shift on-premise? What do you need underneath Cassandra? Block storage, direct attached block storage, which we're building in Kubernetes 110. Steve, any patterns you're seeing between kind of the customers that are being able to embrace really this new cloud-native world versus those that are having challenges with? Any advice you can give based on customer interactions and what you're seeing? Hmm, that's a good question. I think, you know, I just have to fall back on the fact that culture is a hard thing to change. It takes a long time. You know, institutions are persistent. And so, I think that for what we sort of say to our customers, our guidance on these topics, is that, you know, what we try and give you is choice. Depending on where you are on your journey, slowly move our customers forward through that journey, trying to give them a variety of different choices on that. I think, personally, like as with any new disruption, it usually has like 10x value, right? Like the one benefit of containers of virtual machines is you don't have to bring the operating system along every time you create a new container, right? You can much more densely pack a server with containers than you can virtual machines, right? Get more resource utilization. But it takes a long time for an application development team to fully get there, you know? And so, that's the thing, I think, is you just got to be judicious about the right tool at the right time. Yeah, but the other thing related to that is the pace of change, that's funny. I've talked to some of the people that created Kubernetes, the people running all this, and they're like, I can't keep up with all these projects. What are you finding internally in Red Hat as well as from your customers? Yeah, I think it's absolutely true. I was just remarking on that a minute ago, like with the, it's, you know, I'm walking around, you know, I heard this great quote, like why do you come to conferences? Do you come to conferences to learn or do you come to conferences to learn about what you need to learn, you know? And it's the latter for me, right? And the cloud-native ecosystem is exploding, you know? And so I think what we try and do at Red Hat is, especially our team, right? Our goal in emerging technologies is to look like 18 months down the road and pick the winners, like both from a like community vitality standpoint, but also like the right technology. And there's like this play through of choices that we need to wade through. And then what we tend to do is like, distill that down into our platform, that's like something our customers can rely on. And that's, you know, reliable and we've picked the right projects. But it's a, it's a big challenge. Like there's just so much happening, you know? And even in storage, it's becoming challenging. Steve Watt, the Chief Architect of Emerging Engineering at Red Hat, thanks for coming on theCUBE. Appreciate your perspective. It's an architectural game right now. A lot of people are putting these new architectures together. It's a cultural change. Congratulations on your success with OpenShift and everything else. Good job. Thank you very much. All right, more live coverage here on theCUBE after this short break. Thanks.