 Live from the Computer History Museum in the heart of Silicon Valley, it's theCUBE. Covering OpenStack Silicon Valley 2016. Brought to you by Morantis. Now, here are your hosts, John Furrier and Lisa Martin. Hey, welcome back everyone. We are here live in Silicon Valley. This is theCUBE, Silicon Angles flagship program. We go out to the events and extract the signal from the noise. I'm John Furrier with my co-host Lisa Martin. Our next guest is Patrick Riley. He's an entrepreneur, formerly the founder of Kismatix, sold to Apprenda recently. Congratulations, we've had many conversations on theCUBE, good to see you again. It's nice to see you. So you're an entrepreneur now, back to doing entrepreneurial things. You start stuff up, you sell them, you get them passed off. Two OpenStacks ago, we had a great deep conversation with you and your guys. Just a small handful of founders and team digging into Kubernetes, which now is the rage. So you saw it early, you got a company out of it, sold it, are you surprised? Are you, by the growth? I mean, it's pretty spectacular. I'm not actually surprised. I mean, before that, I had a company called Orliatomics that we sold to Mesosphere. And when Google came to Mesosphere wanting to start the Kubernetes project, it made a lot of sense to me that that's going to be the winner. When we think about it, like, Mesos came out of a research project from someone that was an intern at Google who kind of saw the secret sauce and behind the curtain. And when you take those very same engineers at Google that built that with Borg and Omega and tell them, here's a green field, go make this thing, do it again, make it where everyone can use it, make Google's infrastructure for everyone else, of course it's going to be the winner. And so I felt like it was the right horse to bet on early, and that's why we... And the timing now is perfect because as the OpenStacks grow and the community's kind of solidified, they have survived, they've zigged and zagged, as we've been saying, but now they're on the doorstep of the growth. And the app growth is just phenomenal. So you're seeing Docker madness, Docker sanity, Docker madness, people call it madness and sanity, but Docker is kind of de facto standard at this point in the containers. But where does the Kubernetes fit in? Just give a quick 101 on where Kubernetes fits with Docker. Where is it? Is it a Lego block? Is it underneath? Is it on top? Is it overlay? Where does it all fit in? Well, it's really that orchestration layer. So now that I've containerized my application, I need to trust that those containers are going to scale up and scale down and keep pace with my customer's demand of my application. So Kubernetes is that fundamental building block that provides that orchestration. It's like scheduling. It's the scheduling, it's the access to cluster resources. Imagine if someone wanted to add GPU support to Kubernetes, they could. If someone had some sort of terrasorting they needed to do on their cluster, you could add those as resources. So it's just a pluggable building block that you can build a lot of things on. And I think that's why we're seeing a lot of the other projects out there like OpenShift and DS and Apprenda even, like choosing Kubernetes as a building block because why try to reinvent all of that technology? So OpenStack talks about being this open standard for managing applications, whether they're in virtual machines or on bare metal or in containers. Talk to us about Kubernetes containers from a couple of perspectives. One, complexity, and two, security. There's still information out there that maybe VMs might be more secure than containers. What's your thoughts on that? I think that's definitely true. The containerization, like there's still kernel key ring escalations, ways that you can break out of the container, especially containers that are being run as root. So VMs still have, they've been around longer. They have those security things fixed. We still have vulnerabilities in VMs, like you see the last time there was a problem with the floppy disk controller and someone figured out with that attack how to get past that. But traditionally VMs are more secure. And I think if you go back to the Marantis blog from a year or so ago, I wrote a guest blog post for them and I look at VMs as being an insurance policy for containers. I don't really have to pick containers or VMs. I can use both and get a lot out of both. And I think if we look at the demo that SAP did yesterday showing like Kubernetes and OpenStack and bare metal all together, that's really what most enterprises want. They still have a lot of legacy applications. And when we talk about some of these legacy applications, they were built on a particular type of hardware and have nuances for that hardware that you can't just containerize or even put in VMs. So we need to have something that'll work with everything. So in a legacy environment like that, we talk about complexity and some people say, oh, you know, the wave of OpenStack, OpenSource is Kubernetes, is containers. Where does the complexity go in an environment, in an enterprise that has the legacy applications and they've got to sort of bridge the need to manage those as well as cloud native applications. When we're looking at customers that are managing apps and VMs, bare metal and now containers, where does the complexity go? Is it spread out? Is Kubernetes going to help reduce complexity? Well, I look at it like, if I put my safe harbor air quotes on, like the future of cloud native compute foundation and Kubernetes is that we'd have that federated control plane where people can trust that this is the system that's going to orchestrate everything and make sure that your container services are running. So like back to the demo that was yesterday, showing off OpenStack running containerized for its core services on top of Kubernetes, and then being able to provide those other resources that people need in the enterprise. You want that single choke point. You want to be able to say like, this is where things start and are managed. And I think when you look at real world companies like a Bank of America's 1400 plus applications in production that they're running and some of these applications were written years ago for Solaris and hardly are ever touched and have 90 page PDFs that describe how to install them. Like they're non-trivial, right? The number of people in that organization that even have the expertise to know how to diagnose a problem with that application is probably a pretty small number. So I think these modern infrastructures having better instrumentation, having better metrics that you can provide, kind of help look at the overall cluster health. And I think that's what we'll see more and more, more dashboards, more ways to visualize what's going on. So there's a tweet out there on the SAP thing you mentioned, great talk by the guest. He says, SAP's OpenStack Infrastructures run on Kubernetes, separating the data and control plane. Your thoughts on that, is that a best practice? Is that where people are starting to look at the data and the control plane? Yeah, I think that separation is critical because I want to make sure that I can run my applications and store my data and they shouldn't necessarily have to be dependent on each other. I think when we see in Mezos and that community and Kubernetes, that community, everyone's trying to chase this holy grail of stateful storage and how to do it effectively and bigger companies. They have compliance and regulatory things that are guiding the way they do things that aren't moving nearly as fast as we can write code. So we need a way that we can kind of keep both working independently but working well together. So the question I want to ask you your thoughts on since you're the expert in Kubernetes, is a zillion Kubernetes orchestration products. You've got Kubernetes right now has the highest usage. You got Ansible, Mezos, Elastic Container Service, Docker Swarm, Hashacorp, Google Container Engine, OpenShift, Puppet, Chef, all the blah blah blah goes on and on and on. This is zillion orchestration products. Our customers confuse me and people always say what the hell is this Kubernetes thing and has that relate to what I'm using over here? How does this picture get cleared up? It seems, is that fragmented? Is that just use case driven? A different tool for the job? I mean, what's the, makes sense of that? I think it's really fragmented. I mean, at the end of the day, all we're really trying to do is run applications. And I think in this ecosystem, we forget that that's all our customers are trying to do. You know, if I'm Bloomberg and I'm responsible for running Bloomberg Terminal, all I really care about is that Bloomberg Terminal is up for my customers that are paying good money to access it. How I keep that up is I just want to pick the best of breed solution that's not going to crash at two in the morning and leave me guessing of how to get my application back and running. So this fragmentation I think is natural. Like we see, you know, different legacy providers picking Kubernetes as a building block. We see other people choosing other technologies, building their own, like in the HashiCorp example. I think it just boils down to whoever's going to do the most education, whoever's going to spend the time with the customers and educate them, like, this is how you do best to breed. So does fragmentation good or bad? I think it's good because it naturally kind of weeds out the, you know, the weaker solutions. And, you know, education. In terms of letting it rise to the top, if you will. Correct, yeah. And I think, you know, the companies that can afford to spend a lot, like you look at the Kelsey high towers of the world that get out there and like talk about education and show people real world examples that are driven from a terminal. Like he gets a lot of engagement. He gets people really excited. You see one of his talks and you go back and you want to do what Kelsey did. And we need that, you know. Well, Kelsey and Solomon had a little tiff. I don't know if you saw that on Twitter. Yes. Recently, what's that about? I mean, the container, is there a container war going on between? So it's not necessarily a container war. It's just, we don't want anyone to be a benevolent dictator. We don't want someone to say, you know, like you said earlier in our talk that, you know, Docker is the default standard for containerization. We also have rocket. We also have other things that are being created. We want to make sure that the community is being served with, you know, its needs. And it's not just one person saying, this is the way containers have to work or this is the way networking containers have to work. And when we see a company start to eat it's only ecosystem and, you know, all of a sudden they've acquired a company. Now they know how to do networking. How they've acquired a company. Now they know how to do virtualization. Like it's scary. So I think I can't. There's a red flag really from Kelsey saying, well, hold on, hold on a minute. That's my interpretation of it. And I think, you know, the flip side, and Luke who was just a guest, you know, he had some comments on that tweet steam as well. It's very difficult. Like if you're a VC funded startup, you're trying to, you know, provide value. You need to take control of your project. You need to help shape its technical direction. And it's really hard to have- Just for the folks watching, the reference we're talking about is a tweet storm between Solomon at Docker and Kelsey about, you know, the community's response to Docker. And just, it's kind of like an internal outside baseball thing, but important. Yeah, and, you know, if we go back to some of the earlier tweets that happened, one of the comments that I tried to make is like, let's maybe not do this on Twitter. Like we all have the ability to hop on a plane and go to the same place and like talk face to face. And it's better for us to talk, you know, directly and get to the kind of where we think the swim lanes should be defined. And it's not necessarily people have to stay in their swim lane. It's like, you know, know how to cross over from one swim lane to the other. Like we want to make sure that, like the customer is getting what they need. And the reality is like until we educate people on like how they use these things, they're not really going to- At the end of the day, Docker has made their bones in the community. So they really have to be focused on that aspect of it. And not be too aggressive. Exactly. And they have such an amazing community. I mean, you look at like the the turnouts at DockerCon, it's very impressive. I like your advice. I think it's, the approach is very pragmatic, which is essential. You talked about just, and I was looking at your screen, the amount of orchestration software out there. And you, I like your approach that, that fragmentation is actually an advantage for a couple of reasons. One, it gives customers choice. And I like that your view. And we've heard that articulated on this show the last couple of days is that customer centric approach. Not starting with, with, you know, let's put everything in containers. Let's really understand what we're doing and who is serving. So I think your, your pragmatic advice is very valuable for everyone that's here, those watching that aren't here. And really that choice can be an advantage and help customers make the right decision, leveraging the technology. And also especially in the spirit of community, help make it better. For sure. Where's the work areas right now? I mean, if you'd like to look at this and do a little audit, we're sitting here looking at the ecosystem, like a patient on the table, that's getting better, recovering people are working together, and it's, it's, it's healthy. And we give it a check. I would say, hmm, there's some areas we got to watch. What's your take on the ecosystem around? Kubernetes, Docker, OpenStack in general, the big picture, what are the hot spots that need to be really, I won't say it's on fire, that's on fire, but just in terms of areas that really need focus right now, to kind of go the next level. So like I already said, education's first and foremost to me, worrying that we can actually give people informed ways to use this. And I think with the various foundations that have popped up in the space, with CNCF, OCI, things like that, they need to be able to provide a lot of value. And, you know, we've had criticisms come out for some of those foundations as, you know, a bit pay-to-play, like you've got the big companies that can afford to become platinum sponsors, and they kind of have undue weight. We need to honor the people that- Which ones are those? I mean, just the IBMs of the world, and, you know, they can pay the money, no problem, you know, to join these things. We need to make sure that we honor the individual contributor as well. Like we look at the Kubernetes community, there's 800 plus, you know, community contributors to the project, and just because they don't happen to work at an Intel or a Google or an IBM, they should still have a voice. And that's something that CNCF's doing a great job to try to, you know, shepherd those people into the community, have an end user community for those people. And we just need to give everyone kind of a democratized weight. Yeah, and the thing is, is that, you know, Martin Casada yesterday's keynote was spectacular. He had the whole, you know, conversation around, look, it's a developer-led environment right now, and that really is individual, and they're from a big company, same thing, but it's a developer focus versus that. Just got a note here, apparently, people want to have a container debate. theCUBE is happy to host Solomon and Kelsey and have that container. We've certainly done it before in the OpenStack community. Randy Bias had the great API debate in the early days of theCUBE, so if you're watching, Solomon, Kelsey, happy to have that debate in the open, not on Twitter, but on theCUBE. I think that'd be very valuable. Patrick, thanks so much for coming on theCUBE. Really appreciate it. Good luck with everything, and can't wait to find out the next venture, what you're cooking up, and can you share things you're looking at right now, things you're kicking the tires with, things you're thinking about, top of mind? Really, my focus right now is to try to focus on the education, helping people pick, showing, contrasting the pros and cons of different solutions. I don't know what's next for me. Right now, I'm having fun going to meetups and talking to people and kind of helping make an approach. So you're chilling out right now. Haven't got that spark yet, or you're not going to tell us? We'll be on the start up. Maybe they'll... No comment. I'm sure they will. Patrick Riley, entrepreneur, founder, cosmetic, sole to Apprenda, formerly as last coming sole to May, so congratulations. Great to see you on theCUBE. CUBE alumni. We're back with more live coverage of Silicon Valley after this short break. I'm John Furrier with Lisa Martin. This is theCUBE.