 From Copenhagen, Denmark, it's theCUBE. Covering KubeCon and CloudNativeCon Europe 2018. Brought to you by the CloudNative Computing Foundation and its ecosystem partners. Okay, welcome back everyone. Live coverage here in theCUBE in Europe at Copenhagen, Denmark for KubeCon Europe 2018. This is theCUBE, it's supposed to have the CNC at the CloudNative Compute Foundation, part of Linux Foundation. I'm John Furry, co-host of theCUBE with Lauren Cooney, the founder of Spark Labs, new venture around open source and innovation. Our analysts here today with theCUBE and our two guests are Michael Hausenblass, who's the direct developer advocate at Red Hat. Diane Muller is the director of community development at Red Hat, talking about OpenShift, Red Hat, and just the rise and success of OpenShift has been really well documented here on theCUBE, but certainly in the industry everyone's taking notice. Great to see you again, welcome to theCUBE. Good to see you. Thank you. Wonderful to be here again. So first of all, a lot of big news going on. CoreOS is now part of Red Hat, so that's exciting. I haven't had a chance to talk to you guys about that yet here on theCUBE, but get great puzzle piece from the industry there for you guys. Congratulations. It's been a wonderful collaboration. Having the CoreOS team as part of the Red Hat and the OpenShift team, it's just a perfect fit. And the team from CoreOS, they've always been my favorite people. And Brandon Phillips and the team over there are just awesome and to have the expertise from Techconics, the operator framework, which you'll hear more about here at KubeCon EU this week, to have Kwee under the wings of Red Hat now, and Kwee is a registry with OpenShift or with any other Kubernetes thing. The stuff that they brought to the table and the expertise, and as well as the wonderful culture that they had is just, it was such a perfect fit with OpenShift. And you guys bring a lot to the table too, and I've been kind of critical of CoreOS in the past in a good way, because I love those guys. I had good chats with them over the years, but they were so pure open source guys, like Red Hat. Well, there's nothing wrong with being pure open source. No, I'm cool with that, but you guys have perfected the business model. You have great customers. So one of the things that they were always strong at was the open source piece, but when you start to monetize and you start to get into the commercialization, it's hard for a startup to be both pure open source and to monetize, you guys now have it together. Yeah, great fit. It's a wonderful thing. We at, on the OpenShift side, we have the OpenShift Commons, which is our open source community, and we've sort of flipped the model of community development, and that's at Red Hat. And one of the things is they've been really strong at CoreOS with their open source projects, whether it's at CD or, you know, a whole myriad of other things. Well, let's double down on that. I want to get your thoughts. What is this OpenShift Commons? Take a minute to talk about what you guys had. You had an event Monday. It was a word on the streets here in the hallways, it was very positive. Take a minute to explain what happened, what's going on with that program. So OpenShift Commons is the open source community around OpenShift Origin, but it also includes all the upstream projects that we collaborate with, with everybody from the Kubernetes world, from the Prometheus, all the CNCF project leads, all kinds of people from the upstream projects that are part of the OpenShift ecosystem, as well as all the service providers and partners who are doing wonderful things and all the hosts, like Google and, you know, there's just Microsoft Azure folks are in there. But it's, we've kind of flipped the model of community development on its head in the past, if you were a community manager, which is what I started out as, is you were trying to get people to contribute to your own code base. And here, because there's so much cross-community collaboration going on, we've got people working on Kubernetes, we've got Kubernetes people making commits to Origin, we work on the OCI Foundation, trying to get the container stuff all figured out. So you say flip the model, you mean there's now multiple project contributions going on? Yeah, there's multiple, we've got our fingers in lots of pies now and we have to, the collaboration has to be open and there has to be a lot of communication. So the OpenShift Commons is really about creating those peer-to-peer networks. We do a lot of stuff virtual. I host my own OpenShift Commons briefings twice a week and I could probably go to three or four days a week and do it because there's so much information. There's a firehose of new stuff, new features, new releases, the stuff Michael just did one on FAS. You did one before for the machine learning SIG on OpenShift on Caled. I want to just get your thoughts Michael on this because what came up yesterday in theCUBE was integration, glue layers are really important. So I can see the connection here. Having this Commons model allows people to kind of cross-pollinate one to talk about integration because if you've got Prometheus, I might use Kubeflow. So there's new things happening. What does this mean for the integration piece? Good for it or accelerating it? What's your thoughts? So I mainly work upstream, which means when it is Kubeflow and other projects and for me these kind of areas where you can bring together both the developers and the end users, which is super important for us to get the feedback to see where people really are struggling. And we hear a lot from those people that meet there what their pain points are. And that is the best way to essentially shape the agenda to say, well, maybe let's prioritize this over this other feature. And as you mentioned integration being one big part and functions of service being, could be considered as the visual basic of applications for cloud native computing. It can act as this kind of glue between different things there. And I'm super excited about Commons. That's for me a great place to actually meet these people. So the Commons is almost a cross pollination of folks that are actually using the code, building the code, and they see other projects that it makes sense to contribute to. And so it's an alignment where you allow for that cross pollination. It's a huge series of conversations. And one of the things that is really important to all of the projects is as Michael said, is getting that feedback from production deployments, people who are working on stuff. So we have, I think we're at around 375 organizational members. So there's... What percentage of end user organizations do you think? It's probably about 50-50. You can go to commons.openshift.org and look at the participants list. There's, I'm behind a little bit in getting everybody in there, but... So it's good healthy dose of end users. Good healthy dose of end users. There's some special interest groups. Our special interest groups are more around use cases. So they'll be, we just hosted a machine learning reception two nights ago. We had about 200 people in the room. I'd say 50% of them were from the Kubeflow community and the other 50% were users or people who are building frameworks for people to run on OpenShift. And so our goal, as always, is to make OpenShift the optimal, the best place to run your, in this case, machine learning workloads. And I think that's super critical because one of the things that I've been following a little bit and I have your blog entry in front of me is the operator framework and really what you're trying to do with that framework and how it's progressing and where it's going and really if you can talk a little bit about what you're doing there, I think that would be great for our viewers. So what I'm going to do is I'm going to make sure you get Brandon Phillips here on your Kubeflow sometime this week because I don't want to steal the thunder from his keynote tomorrow morning. Well, drop a couple hands. Share a little bit, come on. So the operator stuff that CoreOS and they brought it to the table. So it's really their baby. They had done a lot of work to make sure that they had first class access to be able to inject things into Kubernetes itself and make it run. And they're going to do a better technical talk on it and make things run. And so what they've done is they've opened it up and created an SDK for operators so other people can build more. And we think this is a tipping point for Kubernetes. So, and I really don't want to steal any thunder here. Or get in over my head is the other part of it too. I think Brandon is the right person to talk about that. We'll drag Brandon over here. I'm super excited about it, but let that be. Yeah, well let's talk about why you're super excited about it. Is there anything you can kind of tell us in terms of what? It enables people to run any kind of workload in Kubernetes in a reliable, automated fashion. So you bring the experience that human operators have into software. So you automate that bit, which makes it even more suitable to run your enterprise application that so far might not have been the best place to run. That's great. And yeah, I'm also looking forward to Brandon explaining the details here. So yeah, I think it's great hearing about that. And we talk a lot about how it's great for users. It's great operators, developers, how they're building things out and things along those lines. But one of the things that we are not hearing a ton about here and we want to hear more about is security. Right. Security is increasingly important. You know, we're hearing bits and pieces, but nothing's really kind of coming together here. And what are your thoughts on that? Security, I was recently, when I logged about it and people on Twitter said, well, is that really true that, you know, Kubernetes is secure by default? It's like, well, all the pieces are there. You need to be aware of it. You need to know what you're doing. But it is there, right? It might not, all the defaults might not be as you would expect it, but you can enable it and you can. And I think we did a lot of innovations there as well. Be it, you know, RBAC and the security context and so on. And actually, Liz Rice, myself, we're working on a Kubernetes security cookbook for Riley that will come out later this year. We're trying to document these best practices because it is early days and it's quite a range of things from building container images in a secure way to access control and so on. So there's a lot of stuff there. What are some of the end user feedback sessions or feedback data that you're getting from these sessions? What are some of the things you guys are hearing? What's the patterns? What's the things that are boiling up to the top? Yeah, there's so many. I mean, this conference is one of those ones where it's a cornucopia of talks and trying to, I've just wrote a little blog post code, The Hitchhiker's Guide to KubeCon. It's on blog.openship.com. And because you could spend all of your time here in a different track and never leave it like a security one or an operations one. It's a lot of great content. I think the Istio stuff is probably the hottest thing I'm hearing people going to. There was a great deep dive training session hands on on Monday here that got incredible feedback, IBM, and Google did that one. We had a lot of customer talks and hands on training sessions on Monday. Here there are pretty much, there's a great talk coming up this afternoon on KubeControllers, that magic. I think that's at 11.45-ish. There are a lot of the stuff around service mesh and service brokers is really kind of the hot thing that people are looking for to get implemented. And we've got a lot of people from Red Hat working on that. There's, oh man, there's at CD updates. There's a bazillion things. It's exploding big time here. No doubt about it. The number one thing that I'm seeing last couple of months being answered with customers and also here is that given that Kubernetes is now the defective standard of container orchestration people are much more willing to go all in. A lot of folks were on the fence for a couple of years going like which one is going to make it. Now it's kind of like this is a gift. You can, you know, just as Linux is everywhere on servers, that's the same with Kubernetes and people are now happy to really invest. They're like, okay, let's do it. Well, we're hearing too, just stepping back and looking at the big picture is we see the trend, kind of hearing and connecting the dots is the number of nodes is going to expand significantly. I mean, CERN was on stage yesterday and we heard theirs and still small. Not a lot of huge, not a lot of scale. So we think that the scale question is coming quickly. Well, I think it already came. All right, in the machine learning reception that we had tonight, one of the gentlemen, Willem Buchwalter from Microsoft and Diane Fedema from Red Hat. And a whole lot of people are talking about how do we get, because machine learning workloads have such huge work, you know, GPU and Google has their GPU requirements to get to scale to run these things that people are already pushing the envelope on Kubernetes. Jeremy Eater from Red Hat has done some incredible performance management work and there's on the CNCF blog, they've posted all of that. To get the optimal performance and to get the scale is now, I think, one of the next big things and there's a lot of talks on that. And Istio, that's Istio's kind of big service mesh opportunity there is to bring that to the next level. To the next level, you know, there's going to be a lot of things that people are going to experience trying to get the most out of their clusters. But also, I think we're still at the edge of that. I mean, someone said something about getting to 2,500 nodes and I'm like thinking that's just the beginning, baby. Yeah, it's going to be a couple zeros. I got to ask you guys, got to put you both on the spot here because this is what we do on theCUBE. You guys are great supporters of theCUBE, we appreciate that, but we've had many conversations over the years with OpenShift. Going back to OpenStacks, I don't know what year it was, maybe 2012 or I forget what year it was. Now, the success of OpenShift was really interesting. You guys took this to a whole other level. What's the reaction as you look back now on where you were with OpenShift and where you are today? Yeah. Do you pinch yourself and saying, damn, or what's your view on what you're doing? Red Hat made a big bet on Kubernetes three years ago, three and a half years ago when people thought we were crazy. They hadn't seen it, they didn't understand what Google was trying to open source and some of the engineers inside of Red Hat, Clayton Coleman, Matt Hicks, a lot of great people saw what was coming, reached out, worked with Google and the rest of us were like, well, what about Ruby and Rails and MongoDB and doing all this stuff and we've invested so much in gears and cartridges and then once they explained it and once Google really open sourced the whole thing, making that bet as a company and pivoting on that dime and making version 3.0 of OpenShift and OpenShift Origin as a Kubernetes-based platform as a service and then switching over to being a container platform. That was a huge thing and if you had talked to me back then, three years ago, it was kind of like, is this the right way to go? But then, you know, okay. It's an important history document at that point because I remember we talked about it and one of the things, you guys made a good bet and people were scratching their head at the time, big time, but also you got to give credit to the community because the leaders in the community recognized the importance of Kubernetes early on. We've been in those conversations and said, hey, you know, you can't screw this up because it was an opportunity. People saw the vision and saw it as a great opportunity. I think as much as the technical bits as an engineer, the API being written and go and so on, I really think the community, that is what really makes a difference. If you compare it with others, they're also successful, but here with CNCF, all the projects, all the people coming together, and I love the community. It's a case study of how to execute, in my opinion. You guys did a great job in your role and the people didn't get in the way and try to mess it up. Great, smart people understood it, shepherded it through, let it grow. And it really is kudos to the Kubernetes community and the CNCF for incubating all of this wonderful cross-community collaboration. They do a great job with their ambassadors program. The Kubernetes community does amazing stuff around their SIGs and making sure that projects get correctly incubated. They're not afraid to re-jig the processes. They've just done a wonderful thing changing the way that new projects come into the Kubernetes. And I think that willingness to learn, learn from mistakes, to evolve is something that's really kind of unique to the whole new way of thinking about open source now and that's the change that we've seen. Open source, open movements, always have a defining moment. The OSI model, remember, that stack never got fully standardized but it stopped at a really important point. GCP IP, IP became really important. It created the interoperability world, Cisco as we know and others. This is that kind of moment where there's going to be a massive wealth creation, value creation opportunity because you have people getting behind something as a de facto standard and then there's a lot of edge work around it that can be innovated on. I think to me, this is going to be one of those moments we look back on. And I think it's that willingness to adjust the processes, to work with the community and that Kubernetes, the ethos that's around this project, we've learned from a lot of other foundations mistakes. We, you know, not that they're better or worse but we've learned that you could see the way we're bringing in new projects and adding them on. We took a step back as a community and said, okay, this is, we're getting too many too soon, too fast and maybe, you know, this is not quite the right way to go and rather than doing the big tent umbrella approach, we've actually started doing some really rethinking of our processes and the governing board and the TOC of the CNCF have done an awesome job getting that done. When you got lighting in a bottle, you stop and you package it up and you run with it. Congratulations, Red Hat Summit next week will be there. Thank you. Looking forward to, you know, going deep on this. Well, the OpenShift Commons gathering is the day before Red Hat Summit. We've completely sold out, so, you know, sorry, there's a wait list. We've gone from being our first one, I think we had 150 people come. There's over 700 people now coming to the gathering one and 25 customers with production deployments speaking and this is the day before Red Hat Summit and I lost count of how many OpenShift stories are being told at Red Hat Summit. It's going to be a crazy jet laggy week next week, so. Congratulations, you guys got a spring in your step. Well done. OpenShift going to the next level. Certainly the industry with Kubernetes. Service measures is to do a lot of great coverage here in the cube. Here in Europe for KubeCon 2018 in Copenhagen, Denmark. I'm John Furrier, Lauren Cooney, the founder of Spark Labs. I'm with theCUBE. We'll be back with more live coverage. Stay with us, day two here at KubeCon. We'll be right back.