 from Seattle, Washington. It's theCUBE, covering KubeCon and CloudNativeCon North America 2018. Brought to you by Red Hat, the CloudNative Computing Foundation and its ecosystem partners. Okay, welcome back everyone, we're here live with KubeCoverage here in Seattle for KubeCon and CloudNativeCon. I'm John Furrier, Stu Miniman and this is Jason McGee who's an IBM Fellow, CTO of IBM's Cloud Platform, Kube alumni, great to see you. Welcome back. Great to be here. Jump right in, you got a talk coming up, you got a show here that's doubling in size, the community has clearly resonates around Kubernetes. Yeah, absolutely. Which is goodness for the industry we covered that last year, how people start to snap in and getting it, getting it together, seeing visibility into value points where people can coexist and create value. But we're now going to the next level. Cloud's certainly been validated, hybrid Cloud on-premises and public Cloud working. Customers are seeing it, uptake is there. Where's the big thread now that's being worked on? Because as going to the next level, it's an app market, we've also got some systems in there, where do you see this coming together? I know you're giving a talk on this. Yeah, I mean, I think at the end of the day people are trying to run applications. That's what this game has always been about. They have applications they're trying to build and run that run their business. And I think as a community, this group of people here has been working together to build that platform, right? And I think it's been actually incredible to watch the last couple of years. Everyone rally around Kubernetes and containers. That agreement amongst everyone happened so much faster than I thought it would. I was pretty confident two or three years ago that Kube was the right path forward, but that everyone came there has been pretty amazing. I think what's happening now is, well, what about stateless 12-factor apps? What about functions? What about the rest of the stack? And how do we all come together as a community to find that going forward? Talk about the role of functions and as compute, storage and networking we call the holy trinity of IT. Those things have changed with cloud, but specifically compute. I mean, you know, I say spin up a server in 10 seconds. Well, I need now milliseconds. So you see functions in Amazon with Lambda, these things are changing the game now with containers and functions, a dynamic is evolving pretty interestingly. How do you see that evolving and the impact of that piece? Because compute certainly is goodness to a lot of things. Sure. I mean, I think functions is interesting because there's kind of two angles on it. There's like functions as a business model and functions as an architecture, right? And I think the architecture part, the programming model part is quite interesting. There are certain styles of applications, mostly event-oriented applications, where that is a really natural way to solve a problem, right? And I think what platforms are all about is having the platform be rich enough that for a diversity of workloads that you're running, it's easy to consume the platform. And so us all agreeing on functions as a programming model and getting out of the platform and integrated with Kubernetes and integrated with Istio, I think we'll enable people to build apps much more quickly. And you see that's good signs right now. Good signals, all green light. K-native project is a great example of something new, right? Yeah, Jason, I wonder if we could pull on that thread a little bit there because the holy grail has always been, I just want to worry about my application and all that storage and networking stuff should just work. You know, when we went to virtualization, it helped to a level, but that was just an abstraction. What's the same and what's different about when we go to something like functions compared to what we've been doing in the past? Well, I think there's a couple of things. I mean, first, I think IT is under this kind of, we're trying to flip the model. Like for my whole 20 plus year career, IT has been mostly about infrastructure and we started infrastructure and we build our way up to apps. And what I think we've been trying to do with Kubernetes and with K-native is flip it and start at apps and move our way down. Now, Kube was a good step in that journey, but it's still pretty raw, you know, it's still, you know, compute, you still have storage abstractions, you still have networking abstractions. What you want is for certain workloads to not worry about any of that and functions and also 12 factor systems like Cloud Foundry, both play a role in if you fit within a paradigm, we can get rid of all of that for you. And that's what developers want. Now, it doesn't work for everything, right? Like not every application follows the rules. And I think, you know, Cloud Foundry has a particular opinionated view of 12 factor stateless apps. Functions have a opinionated view of event-oriented apps. We need those abstractions and we need them to be done consistently with the rest of the platform so you can kind of mix and match as you see fit, right? Istio's gotten hot too, so service matches are coming in. I know there's been some debate around how much does Kubernetes become or stay in core last year was a big competition around, you know, kind of core and let things fill in around it. Your thoughts on this trend and how people are thinking about it and what's being actually implemented. So my view is I think the community has done a good job on letting different projects fill in their role but us all agreeing on the stack, right? I mean, Container D and Kubernetes and Istio and Knative, Prometheus, all these things are kind of slotting into their place. And I think in general, we've done a good job of avoiding like one mega system design, right? And I think CNCF has done a good job of letting a few, you know, competitors play with each other in the community and like make each other better. Well, Jason, you bring up such a great point there because one of the things when we reach this size and there's so many people here, there's the obvious comparison to, is this open stack? Sure. And you just brought up one of the biggest things that I've seen is before it was like, okay, well, how many different pieces are in the core and I've got the big tent and all these things but it all needed to live together as opposed to here, I've got all of these components and in many ways we're trying to decompose Kubernetes and we've got all these various pieces and they're not all dependent on each other and we don't all have to agree, there can be from observability for, you know, management, there's so many different ways that I could take the pieces and put them together. So, you know, we'd love your viewpoint as to, you know, what we're getting right now and you know, how do we not duplicate some of the sins of the past? Yeah, I mean, first off, it's always something a community as vibrant as this has to keep their eye on is like, is it all getting out of control? So far, I think they've done it, we've all done a good job because we've been very application oriented and we've also been very focused on real usage, like most of the technologies we're talking about here, people are really using in production, at scale, like there's somebody who has real learning behind that and I think it's driven good decision making. I think one of the maybe unsung things about Kubernetes is the extensibility model that's built into Kubernetes. The loose coupling that's built into this community has been incredibly powerful because it's allowed new things like Istio is a great example, like, you know, we with Google and Lyft and others built Istio, we built it in this completely native experience inside of Kubernetes without changing anything about Kubernetes, like we were able to insert it into the system in a very natural way and I think that allows us to experiment and figure out where we need to go without it becoming this big mess, right? Scale's great and that's a key value of the cloud. Security's number one, what's your view on security? How's that going? What are end users experiencing? Obviously it was a security issue recently in Kubernetes. Seemed to work from the recovery standpoint when things were automated pretty quickly but security's a concern, it's the top of mind. You got the security container boundary there, the boundary within containers, you got a roll of VMs, kind of a new dimension. How do you view the security piece of Kubernetes? So I think it's letting us solve those problems in completely different ways. The Holy Grail for a long time has been, get to standardized systems and I think with containers, we're as close as we've ever been, I wouldn't say we're there, but we're awfully close to having a model where we've got clean separation between the application layer and the system, we can plug in security, we can do image enforcement, we can do scanning, we can do firewalling and network stuff in very different ways. I mean even Istio, like Istio at the end of the day, a lot of what people are interested in with Istio is the security idea. They're like I can do encrypted communications between microservices and that's all kind of done for me in the infrastructure underneath. So I think security is important, I think we're making it easier for developers to be successful building secure systems with platforms like we're talking about here because we're able to solve them in new ways. We've got IBM Think coming up, theCUBE will be there February, I think 15th. 12th of the 15th. 12th of the 15th in San Francisco. Yeah, first time. What are you guys going to be talking about at IBM Think for folks that are going where people might want to sign up, plug for theCUBE and IBM Think there for a quick second. What's going to be there? What's the focus within IBM? You guys got a lot of customers. What's their resonance to Kubernetes? How are they thinking about it? How are they consuming it? Can you share a little bit of what's coming up? Yeah, I mean look, at IBM we're focused on helping customers make that journey to cloud and we're very pragmatic. We understand the complexity of the environments they have. They're building awesome new cloud native stuff. They have a bunch of existing middleware workloads. So we're going to talk a lot about how we help you get there and how you handle the diversity of workloads. We're going to talk a lot about technology or about Kubernetes. We're going to do some fun stuff. We're going to do an awesome, we have a session that's all drones, flying drones demo of how Kubernetes works. Like all live, maybe somebody will get hurt, I'm not sure, but we're going to do some awesome tech demos. If you're sure it's right, they're going big time on that one. It's a joy, one thing, we've heard a little bit of discussion about IoT here but not a lot about AI when it comes here. I'm wondering if you might be able to help connect the dots for us. Yeah, I mean, so I'd say two things. AI is its own domain, right? I think the intersection with AI and the conference like this is Kubernetes is the platform for AI too. I mean, at IBM we run all of Watson on Kubernetes, right? We run all of our machine learning and deep learning systems on Kubernetes. So it is becoming the platform for AI developers as well to be able to be successful, taking advantage of all the compute, resource, you know, custom hardware and stuff that's available in clouds. So I think there's a strong intersection of this being the platform for those workloads. So on the cloud native stuff, we know we've been covering you guys for a long time. You had Softflare as an acquisition but even before Softflare you had BlueMix. BlueMix was developing a lot of cloud native technologies. How is the result in years of investment around BlueMix changing or evolving with the rise of Kubernetes and the rise of these new sets of microservices because you got operations impact, you got developer impact, you got the simplicity model you were just talking about. How was IBM bringing that to bear? Can you share some inside commentary on what's happening? So, you know, we've over the last two plus years we've been building out the platform I've been describing to you in our cloud. You know, we made a decision that Kubernetes was the foundation both for the existing apps to modernize and for new things. And then we've been taking our serverless platform, our cloud foundry investment, our DevOps tools and bringing them all together. I mean, my goal is to build that new platform. You know, as an old WebSeer guy from 20 years ago I saw the value in the industry rallying around a common platform for apps. I think we can do that again. I think we've made so much progress. And on IBM we're trying to drive that forward both in our products and in these community interactions. Talk about that dynamic. You mentioned before we came on camera here about how I was saying it's a systems world now. People who have a mindset seem to resonate well with cloud. You mentioned the app server days, those glory days. Is there a renaissance of those two dimensions going on? Just share your thoughts on that. Interesting insight. Yeah, I mean, I think it's interesting. Cloud is absolutely a systems kind of problem. It's, you know, how do you bring hardware and networking, abstractions around compute, all these pieces together and do it in a way that's composable. I think that's like the really interesting part of cloud is like you have a hundred things that all on their own have to have solid capability and then you'd be able to mix and match them and you can't do that unless you take a systems view, right? Like that security is the same. The user experience is the same. APIs are the same. And it's been actually really challenging to do that in the context of open source because every open source project has its own viewpoints on how you do authentication, authorization, users and like getting all this stuff to work together is hard. And so I do think we have a little bit of a resurgence of people who understand how to build complete end-to-end systems. And then once you enable that you have some horizontally scalable capabilities. You got data and kind of virtual specialization. You can specialize and you can have some common base. So now at the top above that is the app server kind of vibe that you went through. That's kind of happening now. You see that, which I mean. Absolutely. And we see it for our clients and ourselves like all of IBM cloud we've moved to run on the same platform. Like we run all of our services on Kubernetes. And so we've kind of used the platform ourselves to prove out that can handle this diverse set of workloads. This is really disruptive. I think that's a great angle. Jason, thanks for sharing on theCUBE. We really want to get that out. Cloud is disrupting IT, open source communities and the developer market with horizontal scale and new kinds of application environments. Certainly exciting. Thanks for having us here at KubeCon. Three days of wall-to-wall coverage. I'm John Forrest, do minimum day one. Stay with us for more interviews after this short break.