 Welcome back everyone, we're live in Silicon Valley. This is theCUBE's coverage of ONS 2014. That's Open Networking 2014. It's our flagship program. We go out to the events, distract the city from the noise. I'm John Furrier, the founder of SiliconANGLE. Join my co-host and lead analyst in networking in Wikibon. Stu Miniman, Stu, great guest here. We've got Dave Meyer, CTO, a chief scientist of Brocade, who is going to talk about the clear daylight ahead of us in Open Networking. Talk about code, Open Daylight. Dave, welcome to theCUBE. Thank you. Great coin here. I mean, how many of these are made, Stu? It was 150, Dave, you got to tell the quick story on this. Okay, here's the story. Everyone who committed code to Open Daylight to the hydrogen release, which is our first release, got one of these coins. On the front of it is Open Daylight Hydrogen. On the back, it says code. This is like a military challenge coin. And it came to be because one time I just said, well, you know, code is really the cone of the realm out here. And so this is the coin of the realm with code on it. Yeah, so I mean, John, we've been talking for a while. You know, if you look at standards, it's how many people did you put at it? How many people work on it? And the open source world, as Dave just said, it's code, you know, who contributes, who puts things into the code. And that's where the rubber meets the road. So, Dave, if I could start you off there. I mean, you've been involved in networking in a lot of different ways. You've been involved in ITF in the standards world. You're heavily involved in the open source Can you give us that high level, you know, how is networking really created today? And what's the change now versus even five years ago? That's a really good question. So over the last year of being inside a project like Open Daylight, a big open source project like this, it's really changed, it's fundamentally changed my world view on how networking is going to be done and how it is done. And really what it comes down to is collaborative effort in the open source community are transformative for networking. So not only is code important, but also the way that people build it, the way they collaborate to build it and the way that it gets deployed. So, and we're seeing that happen right now with things like OpenStack, Open Daylight, WhiteBox, Open Source, OSes and the like. Yeah, so, you know, from an open daylight perspective, can you just give our viewers who aren't familiar with it just kind of the top level, what is Open Daylight, who contributed to it and where, you know, so Hydrogen's out, Helium's next, following the noble gases, I guess. Yeah, yeah, so, yeah, that's kind of controversial. So Open Daylight is a project, an open source project that's designed to build an industry standard open source platform on top of which innovation can be done in the SDN space. It was founded by a group of what are called Platinum members in the project, including Cisco, IBM, NEC, Rokade and others. And its real goal is to provide a piece of standardized software on top of which people can build value added applications, they can bundle their own products with it and, or just open source it. So it's really about that and it's really about trying to build a platform for innovation. Okay, and you're the chairman of the technical committee, so can you give us, give us a little bit of view, you know, what's your involvement, you know, the code, you know, the first real round of kind of one dot overs of the codes now out there? You know, where are we with the project? Yeah, so the first release of Open Daylight was called Hydrogen, I think we've just alluded to that. We did a community kind of voting for that. And one of the things that, being chair of the technical steering committee, one of the things I did was I was really interested in having the project also play some new trails for openness. And so the first thing I did was I made the TSC calls open to all. So if you like, you can join the TSC calls, you can talk, you're not muted, you can read the minutes, you can get on the mailing list, it's completely open and transparent. And in fact, a lot of the, during the calls, a lot of the discussion is from non-TSC members and it's been very useful. So regarding where we are in the code base, while we just had the hydrogen release, that has been an experience, that was quite an experience. And what we're trying to do now is we're trying to build on what we have to not only do maintenance, stability and all the normal things that you do in software lifecycle, but also get new projects for the next release. As Steve said, it's gonna be called Hydrogen, helium, excuse me. And so we tried to do isotopes of hydrogen for minor releases, but then I found out that people don't know the periodic table that much. So it turned out to be confusing. So what the minor releases are gonna be like hydrogen one, hydrogen two, and then at some point we'll have helium. We're gonna try to get it on a six month cadence, which is kind of standard for a sort of project. You gotta be careful which isotopes you get, might end up with some poison gas there. Yeah, yeah, or it could be radioactive. But you know, there's another problem, because after helium is lithium, and that could be depressing. Well, you know, hey, we're a drug-free zone here. We're toxic-free here in the queue. Always getting the data. They've got to ask you though, relative to the geeks that are here, software defined networking, including about network virtualization. We had Rajiv on from Microsoft. You're seeing DevOps cloud really drive a lot of the pressure points into open source. How does open daylight efforts, and where is it now, and has it helped the enterprise data center game change? And how does that impact cloud? Take us down into the vision of, okay, open daylight, impact of the data center, and does that float the cloud boats in the harbor? Well, I think it's pretty easy that the initial use cases were in data centers you were pointing out. But what I'm kind of fond of saying is that, you know, like what's in hyperscale today is in enterprise tomorrow. So make sure you don't say that doesn't need to scale, because that's never going to be true. Everything needs to be scaled. And so that's part of what we're trying to do. We're trying to build platform that can be purpose for all of these kind of environments. So, open daylight is not a controller. It's a platform. So you can configure it in such a way that it would be appropriate for an enterprise, appropriate for a data center, appropriate for a service provider. So that means basically the enterprise can get onto a hood and tweak some things. Yeah, yeah. Not from the force fit. Here's the suit, wear it, whatever size you are. Yeah, it's a platform. It's a very dynamic, runtime configurable platform. So we're in that same thing with OpenStack. Two people love this idea of under the hood. I'll put this in, take that out. It's like a car engine. Absolutely, Dave, I'm wondering if you can talk to, I've heard you in some of your presentations talk about, it's not necessarily about what you build, but how you build it. You know, I've been talking about that this week a little bit. So I've been inside of this project for a year and it's, as I said, it really profoundly changed my worldview and what it really came down to was there's basically three things that are kind of embodied in the open source kind of development methodology that is really changing the landscape. The first one is that engineering systems are almost as important or maybe more important than the artifacts you're producing. So for example, if you look at the OpenStack community the tool chains that they use are so important and such an integral part and they get refined by the same people who are building artifacts, so engineering systems. The second thing is the culture of the organization or the open source project either way. And the third thing is of course you need the people and the processes that support that. So if you put those things together what you wind up with is the actual thing you build is less important and less a source of sustainable competitive advantage than the way you build it. And that's kind of a radical change because we're used to building artifacts as a source of competitive advantage. Yeah, actually, President, I heard James Hamilton with Amazon talks very similarly with how he designs things for scale and we actually talked about they build hyper optimized solutions at very large scale for Amazon. You've been to traveling a lot, I'm assuming in your role as kind of chief scientist at Bercadio talking to a lot of the big customers out there. Where do you see them in kind of really rocking this trend and making their decisions and moving their organizations forward to be able to take advantage of these new architectures? So let me answer that in two ways. I want to talk a little bit about the open source piece of it and then about what is SDN and what does that mean? So it's pretty clear that everybody's starting to go, hey, there's something going on here in open source. And not that open source is anything that new, but it's the network that we build and the ability to build these collaborations that is so powerful here. And that's, I think, what's starting to resonate with people in the open source world. And I think the other thing is that it's kind of amazing that in networking for so long all of our components, all of the network components were black boxes. You know, I couldn't see myself running a network where I couldn't look inside, like you said, look under the hood. I don't think I could sleep at night. And the reality is network ops nowadays can't sleep at night because they can't do that. So you have this net ops to DevOps thing going on and people are starting to get that. It's slow. I mean, because, you know, of course, there's cultural aspects to that. People need to, you know, be retrained or whatever. We heard one of the presentations out there talking about three kind of phases, organizing principle, operating system mindset and then deployment, right? And the kind of three phases, one figure out how you want to organize logical, physical resources, how to integrate it into some sort of operating system software model, which is networking and other stuff, and then deploy it. How complicated is that? And sounds simple, but just take us through and let's scale one thing. Where are we on that push button? Okay, lay it out, design it, integrate it, push some code like in a DevOps way and then I deploy some cloud services. So if everything as a service is going to be a model, what needs to happen under the hood, so to speak? Yeah, so, good question. In the case of the hyperscale sorority there, obviously they can't have humans in the loop, right? And that's one of the things I was saying during the whole hydrogen process was, hey, we need fewer humans. More code that writes code, but we also need humans. So, you know, more automation, more better, right? And that's actually, that's actually a trend, sort of one of these macro trend things that more automation, but not only more automation, but more intelligent automation. So, what you'll see happening is, and this is why I said, what's hyperscale today is enterprise tomorrow. People go, well, okay, that's for the hyperscales and I have an enterprise, I have news for you. Your enterprise is going to be doing tomorrow what the hyperscales are doing today, otherwise you're not gonna be able to compete. Simple as that, you know? And so, now if we go back to like, what is the SDN story around this? SDN opens all kinds of opportunities for doing automation and optimization, but the technology is still very young. And the architectures that you build with SDN are different than the ones you build today. So, there's time that's gonna be taken to get over that. In fact, you know, if you look at NFV, for example, NFV got a lot more rapid adoption because it's the same architecture, right? The SDN architecture is different. And that's another thing. Can you explain that a little? What do you mean kind of the same architecture? I mean, I've seen more solutions out there on NFV. I mean, Brocade with Viada is a good example of that. So, you're just saying it's software versus hardware, but... No, no, it's... So basically the architecture of the network today, you can think of it as in the old days or today, you have these embedded systems you call routers. And the routers have a control plane on them and they have a hoarding plane on them, right? And that basic architecture is what NFV is, only scale out. If you look at SDN, you have a separation of control and data of some extent. You have programmable hoarding planes, which you don't necessarily have in something like NFV. So NFV actually looks a lot more like the architectures that we have today than the SDN architecture does. And I think that helps people understand it a little bit more easily, because it's something closer to what they're used to and uptake it a little bit more easily. So I want to ask you about the hyperscale, because one of the things Dave Vellante and I always talk about is hyperscale is like the Facebook, people see the poster child, oh hyperscale, Facebook, Google, Amazon doing all that stuff on their own. And then on the enterprise side, there's a big mid-ranger, want to be hyperscale. So I agree with that. I think everyone will go hyperscale. I think that's the table stakes. But the reason why they couldn't do hyperscales, they didn't have the talent to build it. So if you look at Facebook, they didn't want to pay the premium, but they had the talent to build it. So now the enterprises need to get smarter. So tell us what the talent gap and how open source plays a big role in that. Because hey, if I have the brains, I don't have to pay the technology tax for general purpose software, which in essence is not really fits into the hyperscale model, do you agree? Yeah, so open source is key in this. And everybody, people always ask me why is that? And the reason why is because one of the most powerful things about the open source, the whole movement is the collaboration model. So for example, if you follow somebody like, I don't know, who's our favorite open source developer, Dave Tucker or Colin McNamara or any of these guys on Twitter, what they do is they build sort of cookbooks for people who are net ops, for how to get from there to be dev ops. And it's like, you could not buy that stuff, right? But these guys are giving it away because part of the thing is they're about helping that community grow. And that's a fundamentally different thing. So I think that's going to help bridge that gap. It's not going to be complete because the training gap is pretty significant. And I think that's what you're alluding to and that's going to take time. But that's a creek criteria. I mean, ultimately at the end of the day, that's why Facebook saved all zillions of dollars because they had guys that, hey, I just built them scratch. Right, no doubt. I'm just saying that in the enterprises that I encounter, training even for conventional networking stuff is pretty difficult. Yeah, I would say that our observation, again, going on all the different shows is the old school of training, here's some vendor gear and get certified. People don't want that. They want open source. Open source is a proven model for community, camaraderie, a tribe, if you will, but also career path. It's not just like I'm doing for the love of the game. I'm doing it because I want to be part of a group and it's a good career move. I will tell you right now, if you're an open source developer, you're sitting in a pretty good place right now. Yeah, yeah, because people want to, under the hood is a requirement. Future requirement for this new enterprise hyperscale for lack of better. There's actually also, there's skills, right? I mean, you know, there are unique individuals out there. So how many people could, how many people out there can build a, you know, enterprise class controller? It's like 30. You know, if you're one of those 30 people, you're in good shape. They're influencers. So Dave, I'm wondering if you can bring us through kind of the end of the story. So from a brocade standpoint, you know, what solutions are on the market today? What can we expect to see as, you know, Open Daylight gets rolled out from, you know, your group in the department side of brocade on the ethernet side of the house? Yeah, that's a good question. So brocade's committed to Open Daylight and committed to developing the controller and then applications on top of it. Obviously the ethernet fabric stuff is a kind of a low hanging fruit kind of thing because there's a lot of interesting things we can do there both in terms of Open Daylight and Open Stack and all of that stuff is getting worked right now. So I think what you'll see from brocade over time is more concerted effort to help foster development of Open Daylight and Open Stack and the community. One of the big things about this is, you know, if you're a developer in a company like, you know, a conventional developer, what will happen in a conventional company? What will happen is you work on product, right? You don't work on things that aren't product. If you're an open source developer, the influential open source developers sometimes work on things that are unrelated to their products. That's a whole mindset shift that we don't have right now. So you'll see brocade contributing to the community because that's what we're trying to grow and then we'll build products on top of that. Yeah, so I got one last question for you with your Open Daylight hat on. So when Open Daylight came out, kind of, you know, my take on it was the controller should be the easy part. It was, you know, will we have, you know, northbound and southbound APIs that are open so that anyone can create the applications on top of it which where the real value is and, you know, not have to worry about being locked into a specific platform. You know, I hear people are still concerned in the community, what can you tell them about what's happening in the community and what do we expect to see in the development of the project? I think, I mean, you know, there's a couple of things people are concerned about. You know, one is, you know, oh, it's just all a bunch of vendors trying to do, you know, trying to create some kind of barrier tantry for some, for some... Is it, you can tell us, Dave? It's not, it's not. And by the way, if it's successful as an open source, if it's successful as an open source project, there's no way that any vendor can stop it, right? That's one of the beautiful things about it, by the way. So that is all going on right now. And there are technology aspects to building a controller. It's not a turbulent matter, but it is, it is tough. And if you want to build something that's a platform like the open source controller is trying to be, or the open daylight controller is trying to be, it has challenging software features in it. So we're still working those out, right? And it's a, it's a fairly young project from that perspective. So I'm not really concerned about, you know, vendor domination now. Whether or not the controller is going to have standardized north and southbound interfaces. Well, if you have these plugable art and model driven architectures, they'll standardize by virtue of the fact that they're described by either the protocol on the southbound or in the northbound, you know, some kind of modeling language gang or whatever. So that's a standard. Dave, thanks for coming on theCUBE and sharing your perspective. Great, great data. Great, great shift in this market. You're doing some great work here. Tech athlete, as we say, we believe in open source. Reiterate everything you're saying, times a hundred, plus a hundred. But I want to talk about Brokeade because you work for Brokeade. Because, look at Brokeade's doing something. I want you to comment, what's Brokeade doing in all this? Get the plug in for Brokeade because, you know, I mean, you guys are doing some work there. Tell us what's the update, what's new, what do people know about, what people don't know about what's going on and share with the folks out there. Okay, so I'll talk about the software side of the house because that's kind of where I'm working right now. So we have this big effort in the NFV that's around the Viana, the Viana, what's going to be router, right? Virtual router. So again, this is a software component and it's kind of uniquely suited for NFV style applications both in terms of routing and appliances. So there's a lot of work going on there. And then on the other side of the house, on the open daylight side of the house, we're just ramping up a team right now. We just hired Benson Schisler and Tom Nadeau to help us build those teams. So I think in the very near term, you're going to see a lot of development and a lot of activity and open source around daylight, open daylight and around open stack coming from Brocade. Dave, thanks for coming on theCUBE. Dave Byer, tech athlete, Brocade, chief scientist, CTO, virtual routers, I mean virtualization has been a big enabler. The cloud is exploding, the applications are scaling. We have to scale, everything takes care of its Rasmus code, code writing code. That's the future. This is the future of networking here at Open Networking Summit 2014. Hashtag ONS 2014. This is theCUBE with John Furrier, Stu Miniman and Dave Vellante. We'll be right back after this short break.