 Live from San Francisco, it's theCUBE. Covering IBM Think 2019, brought to you by IBM. Welcome back to theCUBE here in Moscone North at IBM Think 2019. I'm Stu Minim, my co-host for this segment is Dave Vellante. We have four days of wall-to-wall coverage of this big show. Happy to welcome back to the program. Jason McGee, who is an IBM Fellow and he's the Vice President and CTO of Cloud Platform at IBM. Jason, great to see you. Great to have you, be here. All right, so Jason, we spoke with you at KubeCon. As we were saying, it's a slightly different audience. It's a little bit bigger here. Not as many hoodies and jeans and t-shirts. A little bit more of a business crowd, but we're still talking about clouds. So, let's talk about your role here at the show. What's going to keep you busy all week? Yeah, so obviously cloud is a huge part of what's going on, I think. Talking a lot about both public and private, about hybrid and some of our multi-cloud management capabilities, my role as the leader of Cloud Platform, I'm talking a lot about platform as a service and Kubernetes and containers and Istio and kind of all the new technologies that people are using to help build the next generation of applications. All right, so we've had a few interviews today, already talked about some of the multi-cloud pieces. We had Dan Berg on earlier, talked about Kubernetes, so first you're going to help correct the things that he got wrong. Yeah, anything Dan got wrong off it. And service meshes have been a really hot conversation of the last year or so, Istio, Envoy and the like. Talk to us about where IBM fits into this discussion of service meshes. Yeah, so I think we've been on this kind of journey as an industry over the last few years to build a new app platform. And service meshes kind of fit the part of the problem, which is how does everything talk to each other and how do I actually control that and get visibility into it. IBM has had a founding role in that project. My team at IBM and Google got together with the guys at Lyft to create Istio. What I'm most excited about I think in 2019 is that technology is really transitioning into something people are using in production and their applications. It's becoming more of kind of the default stack that people are using, really helping them do security and visibility and control over their applications. Yeah, one thing that I heard just from the community and wonder if you can tell me is Istio itself, the governance model is still not fully into CNCF. I heard a little bit of hesitancy on some, Envoy of course out there and the like. So where are we, what needs to happen to kind of move that forward? Yeah, you're right. So we're not there quite yet. We're pushing hard to make that happen. Certainly from an IBM perspective, we absolutely believe that CNCF is the right home for Istio. As you mentioned, some of the pieces like Envoy are there already. CNCF has done such a tremendous job over the last 18 months really rallying all the core technologies that make up this new cloud native platform that we're building. And so Istio is out there, it's 1.0, it's mature, people are using it, that last step needs to happen to get it into the community. So I have to ask you, things move so fast in this world. Do you ever go back to the open stack days and that was going to change the world and then Docker and containers and then Kubernetes and Istio, I can't help but thinking, okay, this isn't the end of the line. Jason, what's the underlying trend here that's going on in the coding world? Yeah, sure, I'll put it in maybe in my own lens, given my history, I'm an old web sphere app server guy, that's the first half of my career I built that. And I think the fundamental problem we're solving is actually exactly the same. It's like, how do you build a platform that lets app developers focus on building their apps and not focus on all the plumbing and the infrastructure for running those apps. And we did that 20 years ago in Java with app servers and we're doing it now with cloud and we're doing it on top of containers. Things like Istio, while they're important in their own right, they're really actually more important because they're just part of this bigger puzzle that we're putting together. And I think for the average software developer, they shouldn't really have to care about what part is Istio, what part is Kubernetes and which part is Knative, like all of that needs to come together into a single platform that they can use to build their apps and run them securely, right? And I think Istio is just recognizing that next piece. I think we've all agreed on containers and Kubernetes, we all talk about it all the time and Istio is that next layer, like how do I secure and control things, right? Yeah, so you teed it up nicely because we want app developers just to be able to worry about the application. So you mentioned Knative, the whole serverless trend is one where, the idea, of course, is I shouldn't have to worry about the infrastructure layer, it should just be taking care of me. We've talked about it for Paz for a number of years, there's various ways to do it. So at KubeCon, and we've been looking for about the last year now, where does Kubernetes and serverless, how do they fit together? And Knative looks to be a piece as to how to bridge some of those worlds, so where are we and what's IBM doing there? So I think you rightly say that they should fit together, or like they're all part of this continuum of how developers build apps. And if you look at serverless applications, you know, there's the serverless dimension, I'm personally not a big serverless terminology fan, I think they're more about event-oriented computing and how do you have a good model for event-oriented systems. Today, with Kubernetes and Istio, I think we've built the base platform, I think with Knative, what we're doing is bringing serverless and also just kind of 12-factor applications into the fold in a more formal way. And when we get all those pieces together and we integrate them, I think then developers are really unleashed to just build their application in whatever way makes the most sense for what they're doing. And some things like serverless, event-oriented, it's going to be easier and some problems, straight containers will be an easier way to do it. Yeah. So you say you don't like serverless, you like event-better, a function? Yeah, a function. So explain that to the audience. Like why should we care and why is that different? How is that different? Yeah, I think for a couple of things. First off, the idea of serverless applies much more broadly than just what we think of as kind of function-based programming. You know, like any system that does a good job of managing and masking the infrastructure below me, you could consider a serverless system, right? So when you just say serverless, it's kind of like second hand for functions, I'd rather we just kind of say functions because that's actually a different programming model where you kind of trigger off of events and you write a functional piece of code and the system takes care of those details. You could argue that Cod Foundry is a serverless system in the sense that you just, as a developer anyway, you just CF push your code and it just runs and it scales and it does whatever you need, right? So part of my mission, part of what I look at a lot is how do we bring all these things together in a way that is easy for the developer to stay focused? Istio's a great example. You know, one of the things we're announcing this week is managed Istio support as part of our Kubernetes service. What does that really mean? It means the developer can use the capability of Istio without worrying about how do I install and run Istio, which they don't really care about. They just really care about how they get value out of its capabilities. Yeah, that's one of the things that having watched all these Kubernetes Istio and the like is how many companies really need to understand how to build this and run this because can I just get it delivered to me as a service? And therefore that whole, what do I want out of cloud? I want a simple model to be able to consume not necessarily, I want to build the stuff that's important to me and not the rest of the system. Yeah, and I think if you look at the industry, there's really, I think kind of two dominant consumption models that have actually emerged for people really using these things. There's public cloud platforms who are delivering things as a service. And then there's kind of platform software stacks like OpenShift, like IBM Cloud Private, which take all of these pieces and bring them together. And I think for most developers, they'll consume in one of those two ways because they don't really want the task of how to assemble all these pieces together. To go back to the serverless piece, like one distinction I heard made is, okay, if I can really scale it down to zero if I don't need to manage it, then that can be serverless, but there's alternatives coming out there like what Knative has to, if I want to run this in my own environment, it's not serverless because I do need to, it might be functions, but I need to manage this environment. The infrastructure is my responsibility, not some service provider. Right, and I think if you look at serverless, to me, personally, I always think of it in kind of two scenarios. There's serverless as a programming model and a technology and serverless as a business model. As a consumption model for payment. I think the programming model part is applicable in lots of cases, including private clouds and in cluster. The business model part is, I think, frankly, a unique to public cloud thing that says I can just pay for the milliseconds of CPU compute that I'm using and nothing more. That's a good thing for consumers. It's a good thing for the consumer. It's actually a good thing for cloud providers too because it gives us a way to reuse our infrastructure in creative ways, right? But I think first and foremost, we have to get more adoption of it as a programming model that developers use to build their applications and do it combined with other things because I think most realistic apps aren't going to all be serverless or all be Kubernetes, they're going to be some mix. Yeah, right, it's like everything else. It's what percentage of the applications will this take over? We had this discussion with virtualization. We've been having this discussion with cloud and serverless, of course, is pretty early in that environment. Knative, did I hear there's some announcements this week that IBM's making? Yeah, so Knative obviously as a project is kind of much earlier in its maturation than something like Istio is, but we're making that available as part of our public and private clouds as well. Really, so people can get started with the ideas of Knative, they can have an easy way to get that environment stood up and they can start building those applications. And so that's now something that we're kind of bringing out as we work in the community to actually mature the project itself. Excellent, one of the things everybody's of course keeping an eye on, I saw Arvind Krishnan talking about the cloud strategy is how Red Hat fits into all of this. So we know you can't talk about kind of post acquisition, but Red Hat's involved in Knative, they're involved in a lot of the service ventures. 8 million developers, that's got to be exciting for you. Yeah, it is. And obviously, like, look, we've been partners for many years. You know, on the open source side of things, we've worked closely with Red Hat for a long time. We actually view the world in very similar ways. You know, like you said, we're working on Knative together, we've been working on OpenWist together, we obviously work in Kubernetes together. So personally, I'm pretty excited about them coming into IBM, assuming that acquisition goes through. They, you know, they fit into our strategy really well, and I think they'll just kind of enhance what we've all been working to build. Right. All right, Jason, what else, what should be looking, you talk about the maturity of these solutions, give us some guideposts for the people watching the industry that we should be looking at as 2019 rolls through. Yeah, so I think there's a couple things. You know, I think this unified application platform notion that we've been kind of touching on here, I think will really come into its own in 2019, and I would really love to see people kind of embrace that idea that we don't need three container stacks, we're not trying to build these different things. You know, one of the things I'm kind of excited about with Knative is by bringing serverless and 12-factor into Kubernetes, it allows each of those frameworks to be kind of the best they can be at their part of the problem space, and not solve unrelated problems. You know, I looked at the kind of serverless versus Kube camps. You know, the purists in both think all problems will be solved in their camp, which means they try to solve all problems, like how do I do stateful systems in serverless, and how do I bring in storage and solve all these things that maybe containers is better at? So I think this unification that I see happening will allow us to have really high efficiency, 12-factor and serverless in the context of Kube, and we'll change how people are able to use these platforms. I think 2019 is really about adoption of all of this stuff. You know, we still are really early, frankly, in the kind of container adoption landscape, and I think most people in the broader industry are just kind of getting in their feet wet. They all agree with it, they're all trying, but they're just starting, and you know, there's a lot of interesting work to do together. Jason, are there any, anything that are holding people back? Anything that you, you know, what do you see as some of the things that might help accelerate some of this adoption? Yeah, I think one of the things that's holding people back is just the diversity of options that exists in the cognitive space. I mean, you guys have all probably seen like the CNCF landscape chart. I've never seen so many icons on something, that's really frightening for the average enterprise to look at a picture like that and go like, which of these things are going to be useful, which are going to exist in a year, like how do I bet, make bets on these things? So I think that's actually helped people back a lot. I think the kind of agreement around Kubernetes that happened in the last 18 months or so was really liberating for a lot of people that helped them kind of move forward there. I think if we can all agree on a few more pieces around Istio or around Knative, like it'll really help kind of unlock people and get them trying. Actually doing it, I don't think is anything more than picking a project and starting. I think a lot of enterprises overanalyze everything and they just need to pick something and go and learn and they'll find. So pick some narrow use case. Yeah, pick an app, pick a use case and go do it. And you'll learn and you'll figure out how it works for you and then you do the second and the fourth and the 10th and before you know it, you're on your way. That's what we did at IBM ourselves and now we're running our whole entire public lot on top of Kubernetes. All right, Jason, any learnings from that kind of experience that you'd share to users as they look forward? Yeah, we had a lot of learnings from using Kube. One is we could run a heck of a lot more diverse workloads than we thought when we started. We're running databases, we're running data warehouses, we're running machine learning, we're running blockchain, we're running every kind of application you didn't think could ever work on containers on containers. So one of the lessons was it's much more flexible than you think it is, right? The other thing is you really have to rethink everything like the way you do compliance, the way you do security, the way you monitor the system, like all of those things need to change because the underlying kind of container system enables you to solve them in such a powerful way. And so if you go into it just thinking, oh, I'm just going to change this one part of how I do apps and the rest won't change. I think you'll find in a year that you're changing the whole operating model around your environment. Well, Jason, rethink everything. We're here at IBM Think 2019. Thanks as always for catching up with us on everything going on. For Dave Vellante, I'm Stu Miniman. Got three more days of live coverage here from us going north. If you're here, stop by and say hi or reach out to us on the interwebs. Thanks so much for watching theCUBE.