(cheerful music) >> Voiceover: Live, from Los Angeles, it's theCUBE covering Open Source Summit North America 2017, brought to you by the Linux Foundation and Red Hat. >> Okay, welcome back here when we're here live with theCUBE coverage of Linux Foundation Open Source Summit North America in Los Angeles, I'm John Furrier, Stu Miniman, our next guest is Arpit Joshipura, General Manager of Networking the Linux Foundation. Welcome back to theCUBE, great to see you. >> Thank you, nice to be here again. >> Always good to talk networking, as Stu and I always say networking is probably the most active audience in our community, because at the end of the day, everything rolls downhill to networking when the people complain. It's like "where the hell's my WiFi, "where's the patent latency," networking SDN was supposed to solve all that. Stu, we're still talking about networking. When are we going to fix the network? It's always in the network, but important. In all seriousness, a lot of action continues and innovation to networking. >> Absolutely. >> What's the update? >> Update is very exciting. So first of all, I can confidently say that open source networking, not just networking, but open source networking is now mainstream. And it's mainstream in the telcos, in the carriers, service providers, it's getting there in the enterprise. And Linux Foundation is really proud to host eight of the top 10 projects that are in open source networking. ONAP, ODL, OPNFV, Fido, you know, the list goes on. And we're really excited about each of these projects, so good momentum. >> We've been seeing and talking about it too, we all, joking aside, the intro there, but in all seriousness we've been saying, we get better the network, it's finally happening. Has it been a maturization of the network itself, has it been industry force and what have been the forces of innovations been? OpenStack has done some great work, they're not getting a lot of love these days with some people, but still we've seen a lot of production workflows at OpenStack, OpenStack's still there, rocking and rolling. New projects are onboarding, you see the telcos getting business models around digital. What's the drivers? Why is network mainstream now? >> I think it's a very simple answer to that, and that is before 5G and IoT hit the market, network better be automated. It's a very simple requirement. And the reason is very self-explanatory, right? You can't have an IoT device on the call on hold while you get your service up (laughs). So, it's IoT, right? And it is the same thing on 5G, a lot of new use cases around cars or around low latency apps. You need automation, and in order to have automation, a carrier or a solution provider goes through a simple journey. Am I virtualized? Yes or no? Am I using the building blocks of SDN and NFV? Yes or no? And the third, which is now reality, which is, am I using open source to do it? Yes, and I'm going to do it. And that's the driver right? I mean it's all- >> Automation, when you started throwing out a lot of TLAs, you talk about SDN and NFV, we've got a four-letter acronym that we need to talk about. The Open Network Automation Platform. Why don't you bring your audience up to speed, what that is, the news that you have this week. >> Absolutely, so ONAP was launched earlier in 2017. It's a combination of two open source projects, ECOMP and Open-O, and we wanted to bring the community together versus sort of fragmented, and because our end users are asking for a harmonized solution. So we brought it together. It was launched earlier this year as we talked about, but the most significant thing is it has received tremendous support from the member community. So at OSS today, we just announced that Vodafone has joined as a platinum member. They will be on our board, and as you know Vodafone is one of the top providers. So if you add up all the subscribers that are being influenced by ONAP, they come to 55%. So out of the 4.5 billion subscribers that exist, more than 55% will be influenced by ONAP and the work that happens. That includes China Mobile, China Telecom, China Unicom, all of the China, Bell Canada, AT&T obviously who sort of was the founding member, Orange, Reliance Jio from India. So we've got, Comcast joined earlier in the quarter, so we've got cable companies, carriers, all joining. And to be very honest, I'll probably just give you the list of all the networking vendors that are participating here, and I've list Amdocs, Cisco, Ericsson, GigaSpaces, Hua Wei, IBM, Intel, Nokia, Tech Mahindra, VMware, ZTE, Juniper, you know, you name it. >> Arpit, I mean the long story short is-- >> Just cause they're involved does that mean they're actually working-- >> They're active. Active. >> we're not going to be critical on this. >> But come on, even Cisco's involved in the open source stuff, right? >> They've very active. >> We've had lots of guests on from Cisco, Lulu Tucker's been on many many times. We know the open source there, but it used to be, networking was very proprietary. Now, it wasn't SDNs going to totally change everything, it's lots of different pieces, lots of different projects. It kind of felt like the river slowly wearing down the mountain as to this transition from proprietary to open source. >> I think what happened is if you just look at four years back, it was proprietary. Not because people liked it, that was the only game in town. When the open source industry, especially in the networking, and this is a hundred year old industry, telecom right? When it came in in the desegregated manner, hardware and software separated, control plane separated from data plane, all of that happened, and what happened suddenly was each components started becoming mature. So they're production-ready components, and what ONAP and what Linux Foundation is intending to do this year is trying to bring all the components into a system solution. So that it's easy to deploy, and all you have to do is point, click a service, everything below it will all be automated and integrated. >> Well the telcos are under a lot of pressure. I mean this has been a decade run, over-the-top they've been struggling with that from years ago, decade ago or more. But now they're getting their act together. We're seeing some signs, even VMworld. Stu, Pat Gelsinger said 5G's the next big kahuna in networking the next 20 years, you can validate it. This is going to be a 20 year changeover, so as the Linux Foundation, which essentially is the organic growth engine for this community, what do you guys see in that 20 years? Cause I see 5G's going to create all these connection points. IoT is going to be massive. That's going to increase the surface area for potential attacks. We're seeing a networking paradigm that's moving from old guards Cisco, Juniper, and some of the names you mentioned. They got to make some changes. How are they adjusting? What's going on so the next 20 years we don't have more conflict and more identity politics. >> I'll tell you one thing, I come from a vendor community, right? So I really appreciate the work they're doing. Part of the reason you would have seen in the past a vendor dragging their feet is because of fragmentation in the community. You as a vendor do not know where to put your resources, people, and where you put your money. What we're doing at the Linux Foundation is starting to harmonize all that. And once you do that and you have enough of a scale and enough of a community, there is no shortage of people and developers that the vendors are contributing to. >> John: What's some of the proof points that you can share? >> Okay, so ONAP, from start to now, about 1100 Wiki members already. That means 1100 unique developers are joining the project. Over 50 members. We ran out of VMs, I mean it's like that has not happened in any project for over five years. We had to fire up people more. So you can see that... And this is not just, these are competitors, but if you step back and look at it, they're competitors from an end user perspective, but they're solving the common problem in which they don't get any money. They don't make any money. These are things that absolutely need to happen. The plumbing, the infrastructure, the orchestration, the control layer, the data plane layer, all of that need to just happen, it should just work. And let them differentiate on top. We are actively seeing almost everybody participating significantly. >> Stu, let's hear your thoughts on this. You guys are both, I view you guys both as experts and influencers in this networking ecosystem, so I got to ask you both a question. CNCF has gotten a lot of traction with funding, sponsorships are off the charts, you're seeing massive tractions, Stu, where you also see that KubeCon Cloud Native, but you have native clouds, I call them native clouds, in Amazon and then soon-to-be enterprises that want to run software-defined networking. So the question is do you see the same kind of support going for your group as CNCF's getting? Is it just fashionable at this point, CNCF? Why isn't the networking getting as much love at least from a sponsorship standpoint. >> Let's define love. So if you define love as the 2017 ONS, which is our largest networking summit, we grew that 10%, everything was off the charts. The feedback, the content-- >> John: The attendance growth or sponsorships? >> Attendance, sponsorships, CFPs were 5x oversubscribed. Call for papers, for submissions, 5x oversubscribed. So we had a hard time picking the best of the best. ONS 2018 is going to be here in LA, we've already started getting requests on, you know, so we're the same boat. >> So you feel good. >> We feel good. >> Not about this, like you're winning. >> No, but I tell you-- >> There'll be positive numbers we know from the hype scale horses, Stu, answer your question and then maybe you guys can comment. So is it a matter of that there's more buzz in positioning involved in the hype side of CNCF now, and there's just meat and potatoes being done in the networking world, Stu? Cause you and I both know, if no one has nothing to say, they've got to kind of market themselves. >> So John, think back to five years ago, how much hype and buzz there was around SDN. John, you and I interviewed like Martin Casado, he just bought for $1.4 billion, all these startups, lots of VC investment, so I think we're further down the maturity curve. Now networking's always-- >> John: People going to work, they're doing their job. >> It's real, it's in production-- >> It's funny-- >> It's not parb, I always say when you move from PowerPoint to production, real things happen. >> I always say, if there's going to be sizzle, I better see some steak on the grill, so what's happening is steak is cooking right now. >> And John, so one of the things we say, networking, no offense to all my friends in networking, networking is never sexy. >> Oh, come on Stu, networking is totally sexy. >> I always say it's cool again. >> Networking has never lost its edge. >> It absolutely is majorly important, but Arpit, take us in, you know, Kubernetes is hot, containers get a lot of buzz and everything. Networking, critical piece of making sure that this works, feels like, I think back to the virtualization days, it took us 10 years to kind of solve those things that that abstraction layer broke. It feels like networking is further ahead than it was, it's moving faster, we understand it's not something that's just kind of oh we'll let the networking guys get to it eventually. Networking and security, which often has that networking tie are front and center now. >> Very good point, and I think what you have to also sort of step back and look at is what are the problems that need to be solved from an end user perspective? So the hardest networking problems at the data plane control layers, check. Next big problem that remain to be solved was orchestration, data analytics, and things like that. Check, solve, with ONAP. Now the next problems that need to be solved are containerization of enterprise app, which is where Kubernetes and... and then how does containerization work with networking? That's all the C&I, the interfaces. I would say next year, you will start to see the interworking and the blend of these "hot projects" where they can all come together. >> Stu, you were there in 2010, I looked right in the camera and said to Dave Vellante, storage is not as sexy. And Dave called it snoreage, cause snoreage is boring. (Stu laughs) >> And at that time, the storage industry went on a run. And we well-documented that. Sexy is, networking is sexy. And I think that we-- >> I call it cool. >> And I just tweeted, 25g is a good indicator of a 20 year run, and networking is the big kahuna as Pat Gelsinger said in IoT, so I think, Stu, I think it's going to be very apparent, sexy. I just don't see a lot of amplifications, so you don't see a lot of people marketing the sizzle. I think, being done I would agree, but Stu, there's more buzz and hype on the CNCF side than networking. >> That's fair. I think it is always as you said, it's the initial phase of any project that gets a lot of clicks and a lot of interest, and people want to know about it. A lot of the buzz is around, just awareness. The classic marketing cycle, and I think we're past that. It was therefore ONAP in January, we're past that. >> Alright, so here's the question, final question. So the steak is coming off the grill in our metaphor here, what are people-- what is that product, what's happening, what is the big deliverable right now from a networking standpoint that people can bet on and know that they can cross the bridge into the future with it. >> You will see a visible difference, you as in an end user, an enterprise, or a residential consumer. You will see a significant difference in terms of how you get services. It's as simple as that. Why? Because it's all automated. Network on-demand, disaster recovery, video conference services. Why did over-the-top players, why were they so successful? If you need a Gmail ID, you go in, you get one. It's right there. Try getting a T1 line five years ago. That would be six weeks, six months. So with the automation in place, the models are converging. >> So provisionings are automatically happening-- >> Provisionings, service, and then the thing that you will not see but you will see in the services impact, is the closed loop automation that has all the analytics built in. Huge, huge. I mean, network is the richest source, and by the way, I'll come back next year and I'll tell you why we are cool again. Because all of a sudden, it's like oh my god look at that data and the analytics that the network is giving me. What can I do with it? You can do AI, you can do machine learning, you can do all these things. >> Well, we're looking forward to it, the eye of the storm is kind of happening now I think in networking, Stu and I always have debates about this, cause we see a lot of great action. Question is, let's see the proof points, you guys are doing some good work. Thanks for sharing, Arpit, really appreciate, General Manager of Networking at Linux Foundation. It's theCUBE, more live coverage from Los Angeles, the Open Source Summit North America. I'm John Furrier, Stu Miniman, be back with more live coverage after this short break. (techno music)

(upbeat electronic music) >> Narrator: Live, from Santa Clara, California, it's theCube. Covering Open Networking Summit 2017. Brought to you by The Linux Foundation. >> Woman: Sure. Um, so, yeah, as you were saying, OpenDaylight really kind of kicked things off from a open source networking standpoint. I mean, there were certainly other open source controllers earlier, in sort of the market life cycle, but they kind of never really made their way out of the universities. OpenDaylight was the first that really had a lot of commercial participation and uptake, kind of in the real world, so to speak. Um, so with that, I think there was a lot of learning that happened, both on the vendor's side, with regard to open source, as well as on the user side. Um, and as the OpenDaylight platform matured and started coming to fruition, we started seeing a lot of other projects sort of both below at the platform layer as well as further up the stack. So at this point, and we've been talking about this quite a bit here at ONS, um, we've been talking a lot about the whole open networking stack that has sort of come to fruition now. You know, really low level stuff, DPDK was just announced today. Fido, which is sort of big data for networking. Then all the way up the stack to ONAP, which was just announced last month. ONAP is a bringing together of the ECOMP Project that was started by AT&T and then they brought it to The Linux Foundation and Open-O, which actually sort of germinated within The Linux Foundation with a lot of input from, um, a number of small vendors, as well as major carriers, particularly in Asia. So, um, bringing those things together at the orchestration layer, and so now we've got this sort of whole stack. Some of it, a lot of it is Linux Foundation projects, some of it is other projects with other open source foundations. All of which we work with very collaboratively across all those different projects. >> Man: Right, right. >> But at this point, we're really kind of looking at how do we enable people to consume this a little bit more easily from the user side? And then also from the developer side. There are a lot of developers who are involved in multiple different projects. Which of course means that they're spread very thin across all those projects. And we're looking at how do we make it a more feasible and scalable activity for them? >> Right. >> So for example, you know, OpenDaylight is upstream of a lot of other projects. There are a lot of other projects that have a lot of dependencies on OpenDaylight. So how do we streamline the release train in such a way that, you know, everybody gets what they need at the time that they need it, so they can do their releases on a timely basis and so forth and so on. And that just, you know, that makes things a lot easier from a developer standpoint. That also sort of naturally increases the, improves the integration points between those projects which is, of course, better for users. >> Man: Right. >> Um, so those are a lot of the things that we have in motion sort of across the Linux Foundation, um, and I think that the other thing that we've really seen over the last year come to fruition is a lot of the early adopters of OpenDaylight in particular have now spent enough time working with the open source community, either through their vendors or increasingly directly themselves, that they kind of get this open source thing, and they understand kind of what the processes are and why we do things they way they do. >> Right. >> And so they're willing to take a much more active role. AT&T is a prime example of that. They were working on ECOMP themselves internally, and they, very quickly, came to the realization that in order to scale it as quickly as they needed to, I mean, they were putting tens of thousands of their developers through specialized boot camps, right? >> Man: Right, right. >> The networking people to become networking developers. But at the same point, you just can't push people through the system that fast enough, nor can you hire enough people that fast enough. And so that's why this has decided to bring it to the open source community. >> Man: It seems like there's kind of an acceleration of carving out some piece of what was proprietary and putting it out to continue the development in an open source world. >> Any "why", you kind of answered the question just now in terms of there's not enough people. But more interestingly, you talked about some open source stuff just never gets going. What are some of the real secrets that make an open source project run? >> Yeah. >> Versus those that don't, or you know, die on the vine. >> Yeah. Um, there are a lot of different components, of course, like with anything. Some of it is technical, right? Do you have the right architecture? Is it one that can scale? Is it extensible? Are the right kinds of people involved in the project? Is the project being informed by the right kinds of people? So if you go and build something that nobody needs, either because you don't have the right people involved, or because you're not open to that feedback, it's going to die on the vine. So, you know, a successful project really has to have a strong community around it. And it's a-- >> Jeff: Chicken and egg. >> Chicken and egg thing, right? How do you get a strong community? Well, you have the right processes in place, but you also make sure that you have the right people involved so that they can build the right kind of thing. And that they have the skills to do it effectively. >> Right. And then the other interesting trend we're seeing is, The Linux Foundation is becoming kind of the hub where you put these things, um, to grow, and as you said, really to cross-pollinate with the other open source projects that have all these interdependencies. >> And that seems to be an accelerating trend as well, as least from the outside looking in. >> Lisa: Yeah, no, it absolutely is. And I think we learned a lot with, with OpenDaylight and also with OpenStack. You know, when OpenStack started, and OpenStack of course is even older than OpenDaylight, but when OpenStack started, I think there was all kinds of euphoria in the industry because open source was relatively new to infrastructure, and infrastructure people, it was like, "Oh, I can build everything "that I ever wanted to build now!" Um, and so there was this sort of irrational exuberance about feature proliferation. In some ways, kind of at the expense of platform stability initially. And at a certain point, the users, again, started getting involved and said, "That's great. We need the thing to actually work. "At scale, in real world environments. "Please focus on that." And you know, that's the real beauty and strength of open source, is when you have users who care, and see the possibility of a project, they can be actively involved and actively influence where the focus of the project is going to be. And that's how you get to something that's going to be useful to people quickly. >> Thank you. >> Well, it'd be great to hear a little bit more about how you-- on these, I'm always kind of mystified as an analyst or a journalist or whatever, when you see these things. The press release comes out, "ONAP is the new thing", right? There's a new thing every week. How do you ensure the success? How do you get the momentum behind it? I imagine there's a lot of stuff that's been happening behind the scenes for ONAP. >> Lisa: Yep. We try not to keep it too behind the scenes. It has always been part of open source culture and what's proven to be a best practice is openness and transparency of not just the code itself but the processes around it. >> Scott: Mhm. >> Um, if people feel like they understand what's going on, that things aren't being hidden from them, that they can have a voice. >> Scott: Right. >> They're much more actively willing to participate. So that's really kind of the key to building any kind of community. >> And how do you work with a big carrier, like, I mean, the fascinating part about this for me is for our viewers who don't know what ONAP and ONOS and ODL are, it's basically all this carrier software that's becoming open source and they're just putting it out there, saying, "It's no longer our family jewels. "Everybody can use it." I mean, that's a big leap for an AT&T, you know? Tell us how you work with AT&T or Verizon or some of these big, gigantic organizations. Like, they just hand you a thumb drive? (laughter) How do you get the intellectual property? How's that process start? >> In the case of AT&T, they reached out to The Linux Foundation and said, "We want and need to do this. "Help us do it. We don't know how this works. "Help us, teach us." But it's very much a, you know, a big part of the role of The Linux Foundation in all of this project proliferation and so forth is teaching people how to do open source effectively. Because, again, it's not just about throwing coders at a problem, 'cause you can do that inside your own organization as well. It's understanding how to do that in a collaborative manner, how to carve off what parts to open source, 'cause AT&T's ECOMP platform, not all of it has been open source. Some portion of it, the stuff that's really important and proprietary and is considered the crown jewels, that has stayed internal, but they've shared a reasonable, fairly large percentage of the base platform with the open source community. And learning to draw that line is an art. And figuring out what is commodity and really could and should be shared with the rest of the world so we're not all reinventing the same wheel. >> Scott: Right. >> But rather than having ten developers here doing that and ten developers here and duh duh duh dah, we can put 30 developers, all working together, to get the same thing more quickly. That shifted mindset can take a little bit of time, a bit of education, and that's kind of part of what The Linux Foundation brings to that process of onboarding new open source projects. >> Jeff: Right. And then on the other end, I always think of Randy Bias. He's one of our favorite guests, Especially with OpenStack, and he knows a couple OpenStack Silicon Valleys ago, where he was somewhat critical on the other end, saying we also have to kind of reign things in, and you have all this risks of stuff going all over the place, and how do you kind of have some organization at the top end because of successful growth can drive a bunch of different agendas and things can get forked. It's not a simple thing to manage. >> No, and we've tried different models and different approaches within different projects and we've learned a lot from that. OpenDaylight was very much a, you know, you guys figure it out, hands-off kind of model. Other open source projects have been very top-down, from their governant structure to everything else. Others, like Open-O are kind of in-between where they did specifically set up an architecture committee that was composed of the leading members of the project because, again, Open-O in particular is touching the business layer of these carriers. So they really need that architecture to be meeting their specifications. >> Right, right. >> Sort of a lower layer, so it's a little bit less critical. There are lots of different models and sort of a gradation of top-down versus bottom-up and, you know, let a thousand flowers bloom. (chuckling) There are pluses and minuses to all of them. I think that we've been sort of learning as we go through all of these different projects what works. And different--sometimes it's worth shifting the model and starting out one way and shifting as you go along as the project matures, too. >> Jeff: But the net-net, which I think, you said at the beginning, is that big companies are now really learning how to operate effectively in this world, in this open source paradigm. It's matured way, way, way beyond what, we used to always joke, years ago, is a free puppy, you know? (laughing) >> You know, I mean, I think Tokus understand now that it is, yes, it's a free puppy. You still have to do lots of work. I think that understanding is sort of starting to trickle into the enterprise. I still have, every time I do a briefing, people will ask me to tell them about my product, and I say, "I don't have a product. I can't sell you anything." I help bring together a bunch of building blocks that you and your vendors can put together. But I don't have a product. And that, you know, that's a major mind shift for, especially, enterprise IT, where they're used to buying things off the shelf. >> Right. >> So larger enterprises, um, are starting again. They tend to take their cues from the carriers as things get proven out in the carrier world. And so we're starting to see that the same level of understanding and also, drivers in large, especially very distributed types of organizations, where they have 50, a hundred, hundreds of different sites around the world that they need to have a centralized few of in some fashion. And the only way they can get there is with SDN and they have a very strong preference, very clear preference for open source. >> Scott: How big is The Linux Foundation now? >> Lisa: By what metric? >> Uh, people, I guess. >> Lisa: Oh, people. Um... We're a few hundred, no more. But it's not just--we're not the ones doing all the work, right? We organize things. We help things happen. We help teach people. We provide the infrastructure. >> It seems to be growing very fast, like new projects are being added and merged. >> Lisa: But again, it's vendors and it's users. >> Very grassroots. >> Yeah. We help provide the ground, the legal framework, and the technical test facilities and things like that, and kind of the organizational guide rails. But we're here to help, we're not the ones doing the work. >> Right, right. Alright, Lisa, so I'll give you the last word before we sign off here. As you look forward to 2017, what are some of your top priorities for this next year? >> Lisa: Yeah, so, several things. First order is really enabling our users to really be successful with the projects that they already have in hand. In many cases, they're well through the phase of proof of concept and all the way onto production, and we just want to make sure that they're continuing to get everything they want out of the project and supporting them and supporting their vendors. And really building out the commercial ecosystem around it, so that they have a strong base of support. So that's one thing. Certainly on the OpenDaylight side, with some of the newer projects, it's really about figuring out what are the best practices that we can implement for this project, for this project, and for this project in order to make sure that they're successful. And a lot of that, again, is that whole harmonization effort that we have going on. >> Right, right. Alright, Lisa Caywood. She knows all about bringing open source to the enterprise, and thanks for taking a few minutes out of your day. >> Thank you very much for having me. >> Absolutely. I'm Jeff Frick, he's Scott Raynovich. You're watching theCube from Open Networking Summit 2017 in Santa Clara, California. We'll be back after the short break. Thanks for watching. (electronic music)