 Live from Los Angeles, it's theCUBE, covering Open Source Summit North America 2017, brought to you by the Linux Foundation and Red Hat. Hey, welcome back, everyone live here in Los Angeles. This is theCUBE's live coverage of Open Source Summit North America. I'm John Furrier, it's part of the Linux Foundation, here with Stu Miniman, co-host, and Alistair Mookie-Bomb, I guess it's bad to my seniors. The technical product marketing for Linux containers of Red Hat, welcome to theCUBE. Thanks for having me. Pleasure to be here. Thanks for coming on, I appreciate Red Hat has been, again, the gold standard when it comes to Open Source, this conference really is about Linux. You can't go any further than to look at the shining example of success that is Red Hat. From when I was growing up back in the day, when Open Source was radical, tier two, tier three, some would argue, alternative to the big boys who were the proprietary operating systems, now tier one, well-documented, don't need to recycle all that, but the fact is, REL is a tier one, supports multiple, seven, 10, what, how many years now, wait, support REL? Is it over 12, I think? Yeah, we're 15 years of Red Hat and Enterprise Linux at this point. Oh yeah, come on, John, remember when Red Hat Advanced Server came out and what was that, 2000, 2001, turned into, you know, REL eventually, but yeah, John, I was working for, you know, an infrastructure company and keeping up with kernel.org was a total nightmare and it needed some adult supervision and that's what Red Hat brought. Yeah, of course, too, and this is well-known, every bank, this is tier one, it's part of the operational infrastructure, so it's got to be stable. But now you've got all this growth going on, certainly we heard Zendlin talking about it on stage, stage, the executive director saying, look it, we're going to have, potentially by 2026, 400 million libraries in Open Source. So certainly the Open Source realm is growing. Sure. The operating system still has got to power all this application. Absolutely. And so you want the best of both worlds. You want the stability, foundational aspect of the operating system while still encouraging experimentation, failure, growth, iteration, so Agile and DevOps, Ethos is about Open Source, it is about trying it, same time, got to keep the lights on, they want downtime. What's your reaction? How do you guys look at that going forward? You want to enable more, but you don't want to break stuff. Yeah, I mean that's really kind of one of the hearts of most of our customers' problems, right, is if you put it in terms of spend 75, 80% of what people spend money on on IT right now, it's keeping the lights on. That's really long-term, not sustainable, right, for anybody involved. So one of the things that we need to do as an operating system, and as a, rather than just an operating system as a distribution, where customers come to us and not want just OS bits, but they also want tooling and application components. How do we draw that line between things that move a little bit faster and upstream that are popular and people want and need access to, at the same time providing that really long-term, stable system user space that really shouldn't change over a long period of time, because that's what provides that sort of application stability that we can ride out over a long period of time. Matt, in the keynote this morning, Jim gave a put out a lot of stats talking about 10,000 lines of code outed daily, 2,500 lines of code removed daily, 450 organizations contributing, so much going on in the space. What are they working on? What are some of the big issues because it's stability, we've added growth, sure there's cool things like Kubernetes and containers. I remember the hot t-shirt at the Red Hat Summit this year was, Linux is containers, containers are Linux, so we know a little bit about that story, so what sort of things is the community working on these days? Sure, so like you said, a lot of shiny objects, right? And even those objects, to be honest, they're not that shiny, right? You look at some of the original support for what's now Linux containers, talking to 2006, if you really want to draw the line, 2002, but there's a lot of things going on in new hardware enablement, right? It's not just new applications that are taking advantage of these different kinds of technologies, we've got new vendors coming out, Arm is about set to take off and add some new challenges and choices to the enterprise customers. We've got a lot of folks who are working in networking, right? The networking stack with NREL has changed dramatically over the past 10 years, right? And with OpenStack and things that are driving through the DPDK and into virtual functions and things along those lines, there's a lot of core stability and core change and things that we think of as stable over time. For instance, some of those new workloads, we spent a lot of time this last year hearing about edge computing, IoT, being something that's pretty important going forward. Linux looks like it's going to be a lot of these places, mobile, it's already all there. We talked this morning, 2017's the year Linux desktop just because there's so many devices now that are Linux. So, how does the workload impact that? Yeah, so everything these days is really starting to get to the point where almost everything's a distributed workload, right? We've really left the, definitely left the single system, single workload paradigm and even the kind of traditional up through the past few years, end tier, we have app web and database. That's really starting to get pushed out across multiple devices. Not only is it getting compute closer to the edge with some of the IoT devices, but simply looking at how we do reliability, stability. You mentioned DevOps, that whole, the ability to move that reliability layer away from relying on expensive components in hardware or expensive components in software. I really distribute that layer of knowledge at the application and use more replaceable, more commodity sorts of approaches. I like to get a spot on the operating systems. One of my degree and my undergraduate in computer science and back in the 80s, everything was here to build your own operating systems. Again, this is where systems come back. If you've got the cloud today, it's really the systems game. And all of us guys and gals from the old days are now in vogue again because the cloud is an operating system. Now you got subsystems, maybe it's just distributed a little bit more decentralized, but again, it's the same game, different era, if you will. So you're starting to see the emphasis on operating systems. The question is, Intel is a great example. Paul Moritz used to call Intel the hardened top where a lot of proprietary stuff underneath that process that no one really cares about because it just runs. I'm not saying that open source is proprietary, but at some point you have to harden the top. And you guys have that with rail, right? So okay, now you want to create more subsystems or new apps. How do you guys look at the future? Because it's a really dynamic environment. So DevOps is move fast and break stuff. It's basically the Facebook model. Now it's move fast and be reliable, which is now the ethos of ops guys. They want innovation, infrastructure is code, but they don't want it to break things. So you got to harden things. At the same time, you want to don't foreclose any opportunities. How do you guys look at that opportunity technically? What should customers look at as a navigation map to go down that journey path? Yeah, absolutely. So I mean, the most obvious way that we experiment in the open is with Fedora, right? That's where we start to play around with new technologies and some of these new concepts and figure out what is the appetite and how do we stabilize those and what's the maturity level of the code. But one of the other things that we do is, sometimes we get some flack for it, but we've got this idea of backcourting, right? So for example, rel74 just released about 30 days ago. It's a 310 based kernel. That's kind of old, except for the fact that we've added stuff that was in 4.12 and 4.8. We're looking at not only, how do we experiment upstream? How do we make sure that we're working in the right communities to understand what that innovation looks like, how we're going to bring that in and then bring it into Fedora, actually start to get some people using it. And once we actually realize that it's time to go ahead and stabilize this, we'll put some customer work on it. What gets your attention from an appetite standpoint? When do you guys start going, okay, we got to start really getting serious about this? I was just always going to be applied R&D going there. Big things happen. But at what point do you say, okay, we got to start getting down and dirty onto these new trends? And what specifically do you look at today and saying, okay, this is obviously on our radar, we're doubling down on it? Yeah, for us really, it comes down to our customers, right? They're the ones that we're really trying to drive when people start asking us, hey, we need to see some of these new features coming out. There's kind of a critical mass that we can see through the market. We can see through some of our partnerships. And for example, you know, we just added some NVMe over fabric enablement. That's brand new. There's not a lot of people out there that actually have devices available, let alone put them into production at this point in time. But we need to make sure for our customers that that ecosystem is stable, from a partner standpoint, an operating system standpoint, from a driver and device standpoint. So that way it's safe for customers to actually go ahead and use. NVMe over fabric, that would be probably service providers looking at that heavily, is it enterprises? We've got actually enterprises, we've got customer enterprises, and we've actually got some partners as well. Linus made a comment in the keynote this morning. He said, the concept of absolute security does not exist. There will be bugs. Can you speak to the development of Linux and kind of the communities? How do they look at security in there? Keeping up with those changes, making sure that you can move fast, yet be stable and handle those issues. How do we look at that from a Linux standpoint? Yeah, that's a really big challenge. And I think that one of the key things that happens from a community standpoint is we do have leadership that understands that community, that culture of security has to be baked in from the start. It's not something you can bolt on once the product has been released into the wild and someone else can fix that later. The community does a really good job of understanding that this has to be reviewed from the get-go. And there's a lot of things that we do as well that from working within the communities, trying to understand what really the threat levels are and what happens when a new threat comes down the pike. How do we close that gap faster? How do we react to it in a timely fashion? But also at the same time, preserving all the things we need to do the right way upstream. Okay, what sort of things is Red Hat working on in RHEL? We can share a little bit for the roadmap that your users are asking for that they're going to find interesting in the near future. So some of the things that we're finding that folks are asking for kind of goes back to the nature of distribution these days. People are asking for things that are a little more streamlined, a little more choice in what they can and can't use. And also looking for the operating system itself to be a little more automatable to make up a word. We've got this level of dev ops and automation that's going on, both for the kind of new CI CD systems as well as traditional fleets of virtual machines. So we're trying to work on how does the operating system act as a good citizen in a highly automated environment? And can we put facilities into the operating system that help folks who are trying to automate do that sort of natively rather than having the operating system fight it What areas of automation do you see as hot right now? Because that's a good thing. I'm a big fan of automation because yeah, maybe some manual labor will be automated away, but those are usually non differentiated labor roles which can be shifted to higher value opportunities. I'm sure you'd agree with that. But so what would the, or do you agree? And then what would be the areas that are being automated away? In terms of order of operations now? Sure, yeah, no, I absolutely agree. It's one of those things that my background is infrastructure and I had many, many failed attempts way back in the CVS and Canary servers to try to do these things in CF Engine in 2000 and I don't really want to admit how old. But I think that you look at the DevOps world, right? The dev side of the house has a pretty good idea of how to automate and how to build pipelines. I think that the operation side, we're kind of missing that pipeline view of infrastructure as code, right? We've got the state machine at the beginning, kind of the state machine at the end, but I would think where there's some interesting opportunities to get some new expertise is that how do I design and let an automated system create some sort of deployable artifact either by directly creating and pushing an image somewhere or some sort of other artifact that describes a higher order than just I need to configure these seven things and put an IP address on this box. How does the view of things, you got to have a good sense of the trees, not the forest, not the trees. Exactly, and it kind of goes back to that everything's a distributed system, right? We no longer have, oh, this is my web server. It's like, this is web server 77 of 82 that's part of this 200 node cluster that does whatever my line of business applications. Last to cloud operations. Matt, I'll give you the final word. Give a perspective of the folks watching. What are the cool things that you guys are bringing to the market in the future? What can they expect from you guys after the operations? This was a critical component we heard from the keynote today is powering a telescope in Arctic. So I mean, it's everywhere. It's done not only on Earth, but in space. Absolutely. So what's next? What's the exciting new things that people can expect? It's a different edge, right, John? What? Different edge there. Yeah. Going out to space. Space X, you got the telescope. It's an IoT device, technically. The rover has, you know, Linux on board too. It's literally on other planets at this point in time. Can we do better than Linux on other planets? That's a good question. I think the whole notion of pirate using the quote in the Martian movie. Exactly. Space pirate. Yeah. So I think, you know, part of what we're going to be seeing from the operating system is how do we drive choice? And how do we provide sort of a new way at looking at the way we distribute bits and pieces so that folks have more control over what they need, when they need it? Like, we've seen sort of an uptick in how many versions of operating systems. Like, we're back in that resurgence, like we had in the early 2000s, Atlantic's gone 99, say, of different operating systems of different sizes to do slightly different things. I think we're going to see a lot more shape around what some of that looks like. Hopefully, we'll see kind of that robot army of red hat enterprises, Linux systems that we can, you know, make that nicely differentiated space out of. I think that, you know, there's going to be a lot more in some of the networking space and that's coming out of it. And orchestration, your view on orchestration? Yes. Orchestrate. Personally, I think it's one of those right tool for the right job. I hate to say it depends, because that's that phrase that everyone hates to hear, even though it's usually true. And it comes down to, you mentioned containers, like what are we trying to orchestrate? And what's the right layer of abstraction? If we're trying to orchestrate applications, that's a different set of orchestration than perhaps trying to stand up and tear down data centers. But maybe not. And I think that from an operating system standpoint, those are all things that we can and should be able to enable, rather than, you know, resisting and causing friction for people who want to say, you know, we want Kubernetes, no, we want something else. Great. We're there to support all of those things because those are capabilities that we provide. Nathan. I see the inside the cube. Your Twitter handle is at Cleverbeard. You can see the Cleverbeard there. He's a technical product person, Linux and containers for Red Hat. Thanks for sharing your insight in the cube. We appreciate the support from Red Hat and continue to congratulations on the success of Open Source. You guys are the leader of the bellwether in Open Source and congratulations to you and your company. Appreciate it. Thanks. It's the live coverage with theCUBE. I'm John Furrier with Stu Miniman here in Los Angeles for exclusive coverage of the Linux Foundation Open Source Summit North America. We'll be right back with more live coverage, day one of two days of coverage after this short break.