 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 here live in Austin, Texas for KubeCon and CloudNativeCon. I'm John Furrier, the co-founder of SiliconANGLE Media with my co-host Stu Miniman, our next is Nigel Polden who's the author of the Kubernetes book, also container, guru, trainer, been in the business for a long time in the community. Great to have you on for our intro. Thank you. Stu, keynote, let's get down to it. What was the big highlights? Yeah, well, first of all John, we officially entered KubeCon days here. So CloudNativeCon was yesterday, we've got two more days of KubeCon. Kelsey Hightower, we had him on theCUBE yesterday, phenomenal speaker, everybody's looking forward to him, usually the lines to talk to him. Made sure that there was the standing ovation before and after his, very demo heavy. I mean, this group loves it. There were a lot of great pithy lines, arguments over which is the best language, which is the best way to do things, knocking on things like Gamble. So it was definitely a fun, geeky discussion. I'm a big Game of Thrones fan, so love to see that Season 7 delivered on Kubernetes. What was the summary of the keynote? What was it take? So I think from my perspective, the summary was Kubernetes is boring, which translates to us generally as in it's maturing. It's something that you might want to be able to trust in your production environment if you're an enterprise. I mean, look, as a technology guy, we always think, we like to know the details, the weeds, and we like to play with YAML and stuff like that. But at the end of the day, businesses don't and developers tend not to want to. They don't want a smooth pipeline. And that's boring, and so boring is good. Yeah, and I do want to poke it a little bit. And, Agile, I definitely want your opinion on this, because there's certain technologies we say, oh, right, it's reached that boring phase, but it's just kind of steady state. Kubernetes is at like 1.9. Coming into the show, it was like, how complex it is? Oh my God, there's all these things above and below. Hen gave a really nice keynote, showing kind of the layer cake there. I think maybe the Kubernetes layer might be, it's stable enough and used and people can use it, but this ecosystem by no means is it boring and there's lots of things to bake out. What are you seeing? Totally, and it's that definition of boring, really. So I would say boring would translate into usable, but you're right, in no way is it boring in any sense. In fact, it's exciting and it's dangerous as well. Yeah, and the other way. So I'll give you an example, right? So Kubernetes is massively successful. I think we all grok that at the moment, okay? But it's almost potentially going to be a victim of its own success. So I was at one of the mini summits that was going on before KubeCon and CloudNativeCon started. And it was about networking and there was a bunch of guys here from big carriers and they really want to take the simple networking model that Kubernetes currently has and make it fit their needs, which would make it really complex. Dare I say, almost open-stack neutron. And I think there's so many people here at this conference right now that want to take Kubernetes and use it for their own purposes. And as successful as it is and as much uptake as it's got, there is a potential danger there. I think that it explodes out of control and I don't want to knock open-stack but becomes difficult and not what we want it to be. And that's dangerous for me. And I know you're big up a great point here because something we've been looking at is every time we abstract or make this new design model, it's, oh well, we want to make sure that the developer doesn't have to worry about that infrastructure. Clayton from Red Hat, we had him on theCUBE and he talked about in the keynote, right? Boring means when I write my code, I don't have to think about the infrastructure but networking and storage, networking, some of the basic pieces are done but there's a lot of activity in that space. And storage, we're still arguing over what container native storage should be, what cloud native storage should be. So it still is my definition. It's not boring yet. That's the direction and I like it. Kind of was where we talked about invisible infrastructure. What are you seeing? I mean, you've got a heavy background on that side too. So I think I quite like the space that networking is at within Kubernetes. It's simple and that works for me, right? Storage is certainly, it's still playing catch up there and I think a lot of decisions still need to be made. The futures, in my opinion, is still not clear there but I think a lot of games have got to be played to say, now how far do we take networking and how far do we take storage and things like that so that it in the one sense doesn't balloon out of control but on the other side, you do want it to meet more use cases than just the very basic use cases. So I mean, that plays back to my idea that that danger aspect of Kubernetes, it seems to have won in the orchestration space at the moment, but I think the road ahead, there's still loads of potholes and there's tight bends and there's cliff edges and things that we still could fall off and that's exciting. Nigel, your dangerous comment reminds me some of the early days of VMware. People that would get in there, they'd do some really cool things, they'd write it up, share it with the community, absolutely, it feels like that, it's almost even bigger. Yeah, like the top layer that interfaces with the developers and things like that, that's getting pretty stable but underneath, I mean, that is a happening place underneath right now and I imagine it's going to be a happening place for quite a few years. What about service meshes and also pluggable architectures because that seems to be the answer to the dangerous question. Oh, don't worry about it, carriers and whatnot, you can just build pluggable architectures, no one's going to get hurt. Yeah. Ready for prime time, what's your thoughts? So I think service mesh is almost certainly, in my opinion, the hot topic of the conference so far. I like this idea of it getting boring and stuff and that's good for the project but if there's one takeaway, if it's something that you're not quite clued up on at the moment, go away and look into service mesh, I've got to do a lot of that myself to be perfectly honest. But this whole idea of running side car containers and what have you inside of your pods alongside your applications to look at your ingress, traffic, your incoming traffic, your outgoing traffic, it's all cool and it can add so much functionality and make it so much more usable to a lot of users but at the same time, there's that, I don't know, right? Look, I'm a little bit old fashioned, I remember the days of deploying agents on servers and we would have server builds that had agent upon agent upon agent and we had this backlash in the industry of like, you're not bringing your product in vendor X, Y or Z, okay? If it deploys an agent, we're going fully agentless here, we're sick of managing all these different agents in our stack and I wonder again, playing to the danger topic here, that like, are we going to end up having loads of these side car containers in our pods that are effectively the modern day agents that we then have to manage and consume resources? It's playing the side car dynamic, it's important. I think take a minute to explain the dynamic because containersization, been around for a while, Google and everyone else knows but Docker really put it on the map, now the commoditization of the containers with Kubernetes, what's this side car thing about? Quick, take a minute to explain for the folks. Right, so in the Kubernetes world, I guess the atomic unit of deployment, the equivalent of a VM from the VM world space, right, would be the pod, which is effectively a container, right? But within that pod, you run your application container. And I think for most people, right, you run one container inside of that pod, it's your application, right? What we're starting to see now is, and Kubernetes has always had this ability to run multiple containers inside of a pod. Most people don't do it, and it seems that a lot of the external projects and a lot of the third-party vendors are starting to pick up on this and say, all right, well, let's run another container inside of that pod. It's not your actual application, and we call it a side car container. And it adds functionality and what have you, but it also potentially eats through resources, makes your deployments maybe more complicated. I mean, it's always a trade-off, isn't it? You get additional functionality, but it's never for free. Yeah, it's overhead. All right, talk about the customer guys. What we saw in Keynote, saw HBO on stage. How are customers using Kubernetes? Because I'm trying to put my finger on it. I love orchestration, I know what that does and I understand the benefits, but how are actually people using it today? So I think it's a little bit like the whole container thing, right? The early adopters are the Netflixers and the HBOs, and the people like that, that have got large engineering teams, that have got a lot of developers on staff, that are really just comfortable going and taking these new technologies and rolling them themselves. And I've got this appetite for danger again within their organization almost, right? The risk-taking organizations, right? They're all over the containers and the Kubernetes. The more traditional enterprises, I think are still kicking the tires, are still throwing out the occasional new project within the organization. So let's test the waters with this new feature that we want to add to our main product door, or we've got something new. Let's try containers and Kubernetes. There's certainly, at least the ones I speak to, right? Certainly not at the phase where they're taking their legacy apps. And HBO is using it for like traffic, identifying ingress, you mentioned that earlier. I mean, basic stuff, not a lot of heavy lifting, or is it? Well, I think the HBO, I mean, how much they ran this season seven of Game of Thrones on Kubernetes. I mean, I'm sure there was some non-Kubernetes stuff in there as well, but it seemed like from the presentation, pretty much, well, a lot of that stack was running containers and Kubernetes. And let's be fair, when it comes to HBO, Game of Thrones is like their, it's their killer product at the end of the day, isn't it? So they've taken a risk there with that. But again, you know, HBO are probably a risk. They have a lot of online viewers, by the way, on that too, for HBO. An insane number. But I would say, compared to a traditional enterprise, they're a risk-taking organization. They live in the cloud. They like living on the edge. They're willing to take risks with new technologies to push the product forward. All right, so I want to give you guys the thoughts on a tweet I saw out there. Think of Kubernetes as the kernel for modern distributed systems. It's not about zero ops. It's about power tools to unlock developer productivity. Craig Mclucky from Heptio mentioned that on stage. Really kind of rallying around Kubernetes. Thoughts on that quote. What does that mean? Yeah, so I mean, John, there was for a while people saying, you know, how do we deprecate or even go to kind of no ops? And absolutely, it's, you know, many of the keynotes talked about who's deploying them and who's running them. We're not talking about eliminating ops. Even when I can have a voice assistant help roll things out, you know, there's still absolutely major piece of who needs to run this. But, you know, the right things are the right part of the organization. Yeah, I think instead of using the word kernel, maybe use the word Linux, you know, looking at Kubernetes as the Linux of the cloud. And that's not, that's not my term. I've heard other people say it, right? But it's open source for a start like Linux is. It's got a great thriving community of people contributing to it. You can fork it, you can do whatever you want with it. But if you're going to deploy a cloud native application right now, then Kubernetes is that substrate. You've just got to look at what came out of re-invent. So AWS is now offering a native Kubernetes hosted service. Obviously Google does it, Azure does it with Microsoft. They're all picking up on this like realizing that people deploying cloud native apps, they're going to be deploying it on Kubernetes. That's about Red Hat. I just saw Gabe Mondoi tweet from the keynote, Stu. Red Hat's contribution to hardening Kubernetes cannot be overstated. CC OpenShift. And we had Brian Gracely on yesterday. I mean, OpenShift, what a bet. Microsoft betting heavily on Kubernetes. Google obviously sees this as an opportunity. Multi-cloud fantasies out there somewhere. But that's what customers are kind of asking for, not yet intangible product. But this is interesting. You got Red Hat, the king of the enterprise. Open source, no debate about that. Microsoft and Google, old guard with Microsoft, and then new guard in Google. Really, if they don't throw a line at the main cloud trend with Kubernetes, they could be left in the dust. So I see a lot of things to play. How is the Red Hat and the Kubernetes investment paying off? How do you guys see that, playing a good strategic move, headroom to it with comments and color commentary on that? Well, I think if you compare Red Hat to Microsoft, if you don't mind me doing that, right? Microsoft had a cash cow in Windows in the past. And I think it quickly realized that the cash cow was not going to live forever and they invested heavily in Azure. Red Hat live a lot, I guess, as well, off support contracts and things like that, the Red Hat Enterprise Linux. How long of a tail that has, I'm not sure. So certainly they're doing at least, they're looking in the right direction at least by investing heavily in Kubernetes. If they want to go in and be the enterprises trusted Kubernetes partner, I think they've got a great story. They've contributed a ton to it. They're already in the door at most enterprises. And I think you couple those two things together if the enterprise is going to adopt Kubernetes at some point. I'm not saying they've got the best story, but they've got a pretty decent story. All right, in the last minute, I want to ask both of you guys this question because it's been kind of on my mind, I've been thinking about it. Maybe I'm over-stretching here, but three-day conference, one day to cloud native, two days to Kubernetes. KubeCon, why? More important, we're growing community. Cloud native I think would be probably stronger sessions. Is it because there's more emphasis on the Kubernetes? Kubernetes is the core. Kubernetes is what started the CNCF. All the other projects really build off of it. So, I think it's pretty safe for it. It needs more attention. Kubernetes, I mean, while there's, I love Kelsey's line this morning. He looked out at the audience. He says, I think everyone that's running Kubernetes in the globe is here. So, there's jokes about how many people are actually running in production. Yeah, they're probably here. So, look, there's still so many people that are getting the Kubernetes 101. The whole cloud native, all of these other projects are all building off of it. I think it's really straightforward on there. We've even heard, do we call it the CNCF? Do we rename it to something that's a little bit more Kubernetes focused because cloud native gets talked about some. There's service mesh, absolutely Nigel was the buzz coming into the show. I hear those sessions are overflowing here. We didn't even get to talk about there's like another alternative to Istio that's there, but yeah. And Lou Tucker, by the way, affirmed that same thread yesterday about this service mesh. Nigel, final word for you on this segment. How big order of magnitude is important as Kubernetes? I mean, given you've seen talk about aging over the old days and all the ways have come, there's been kind of incremental improvement. Balls have been moved down the field here and there in some big chunk yardage, if you will, these are football analogy. How big is, because I've seen Kubernetes go from here to here, really move the needle on the community, it's galvanized. How important is Kubernetes from an order of magnitude? What do we look back a few years from now? What are we going to be saying? Hey, remember KubeCon in 2017? How important is Kubernetes? Well, can I say, I think it's really early days, okay? And I like the analogy that it is the Linux of the cloud or of cloud native, okay? But I think there's danger in that as well because the world is changing so fast now that, I mean, Linux has lived for a very long time, okay? Will Kubernetes live that long or will it be replaced by something else? It probably will be. But I do feel these are early days and I think it has got a long stretch ahead, a long stretch as in like good four or five years and within two to three years, you know, just about every organization in my opinion is going to have some Kubernetes in it. And the beginning signs of maturity is coming. Stack wars, stew, all the vendors really trying to figure out strategically it's like a 3D chess match right now. Open source is kind of like the arbiter of this, really good stuff. I think it's going to be super important. Thanks for the commentary. Kicking off day two of Kube exclusive coverage here at KubeCon. Cloud NativeCon was yesterday, two days at KubeCon. We're back with more live coverage from theCUBE. I'm John Furrier with Stu Miniman and Nigel Poulton after this short break.