>> Announcer: Live, from Santa Clara, California, it's The Cube. Covering Open Networking Summit 2017. Brought to you by The Linux Foundation. >> Welcome back, everybody. Jeff Frick here with The Cube. We are in Santa Clara, California at the Open Networking Summit 2017. I'm joined this whole show by my co-host, Scott Raynovich. Scott, great to see you. We're excited in this segment to get one of the keynote presenters to come down and spend some time with us on The Cube. So, Sandra Rivera, she's the Corporate VP and GM of Network Platform Groups at Intel. Welcome. >> Thank you. >> Jeff: And your keynote is all about >> 5G. >> Jeff: 5G is now. >> 5G is happening now. >> That is a powerful, declarative statement. >> Indeed, but it's true. >> Jeff: It's true. >> Yes. If you look at 5G being the true convergence of computing and communications, then you see that so much of the capabilities that we have had in the cloud and in the core of the network, really need to scale out to both the Edge and the Access network, to be ever closer to the end user or the end point. It could be a smart phone, it could be a laptop, it could be a tablet, or it could be some of the new devices that we see, drones and robots and connected cars. So this idea that we have to bring programmable, scalable, flexible computing closer to those end points is really the foundation upon which 5G is going to be built. All of that is really what we're driving with software defined networking and network functions virtualization. So 5G is indeed happening now. >> This is really a continuation of the theme from Mobile World Congress just a few weeks ago. Time flies. >> Is it a few weeks? I think it's a couple months. >> I don't know, I can't keep track anymore. >> 5G at Mobile Congress was all the rage. We were talking a lot about what 5G will enable. Connected cars and smart cities and smart factories and smart homes, as well as those immersive experiences that you'll have in your home, cloud gaming and 3D types of experiences and virtual reality or, actually what we're calling merged reality, the ability to put physical objects in the virtual world or virtual objects in your physical world. All of that requires way more bandwidth, very low latency, and much better responsiveness in that end point near the device or the user, which is what all the innovations in 5G from a radio perspective will enable, but of course the rest of the infrastructure has to support it as well. >> There was quite a bit of discussion at Mobile World Congress about 5G, and I think there was a lot of questions also being raised. Some of the larger carriers, such as Deutsche Telekom, I think maybe Orange, they were questioning the size of the investment that's necessary, and I think for some people it threw the timeline into question a little bit, as we know. As we were discussing prior to the show, the standard, we're looking at 2019, 2020 maybe for deployment? >> Sandra: Right. >> What's Intel's view on the deployment timeline? Does that matter to you? >> It matters a lot because we are investing now, and we're investing with a broad ecosystem of partners. If you look at it just from a pure radio perspective, yes indeed, the 3GPP spec for 5G doesn't really get nailed down until the end of '18. You'll start to see true compliant 5G devices introduced in 2019, and rampant scale in 2020. But the network infrastructure, that idea that you need this programmable, agile, composable infrastructure, really starts now, because you're not going to be able to have a light switch of, "Well, this is the network that I need to support all those devices and all those use cases." That composability of the network is anchored on having a programmable capability as opposed to a fixed function set of boxes or appliances, which is really how networks have been architected and built and deployed up until now. It embraces server volume economics, virtualization technologies and that pooling benefit that you get from sharing an underlying resource, as well as cloud architectures and cloud business models. The idea that you can pay as you go. You hear a lot about network slicing and that really is about composing the network for not too few or not too many resources that you need for that particular end use case. So all that is happening now. We are participating with Verizon in the 5G tech forum. We're working with KT and SKT as they get ready for the Winter Olympics. We're working with operators and telecom equipment manufacturers all over the world to prove out connected car and smart cities and smart factories types of use cases. I think that there's always some healthy skepticism about, are we over-investing or are we investing too early? But if you look at the amount of work that we have to get done in what is a relatively short window of time, we feel like we actually need to speed up. >> And 2019 is right around the corner, Scott. I can't believe we're already a third of the way through 2017. >> I have it marked on my calendar already. It's right here. 5G arrives. But tell us, the play for Intel is to be in the NFV Infrastructure for 5G, is that your play? >> Actually, Intel's strategy for 5G is end to end. Clearly we have modem technology that will go into client devices, yes smartphones, yes tablets, yes laptops, but also drones and robots and cars and any number of devices that haven't even yet been invented. We are in all of the infrastructure, from the access layer in terms of the base stations and a lot of the edge computing that is happening there, we're in the edge of the network which could be close to the enterprise, or close to the consumer, and we're in the core of the network which is where a lot of the switching and routing functions, the authentication functions, the security functions are done. Then, of course, we power most of the world's cloud infrastructure. So back into the cloud and the data center, that's Intel. It really is end to end. We have this broad view and this scalable architecture where it's a consistent silken architecture, a common tool chain, and a very broad access to ecosystem and developers to take you through that end to end portfolio of services and capabilities that you require. >> And at the end of the day, it's just eating up a lot of compute, right? >> Lots of compute. If it's a compute problem, Intel feels pretty comfortable that we have leadership there. Indeed. But we have some new announcements here. >> Okay, because you're here. Besides the keynote, you have announcements, too. >> We have some announcements around our data playing development kit, or DPDK. Intel invented DPDK in 2010. That was a set of libraries and optimized drivers for running high performance packet processing on general purpose CPUs. And of course, if you're in the network business it's all about moving the packets, so you need high performance packet processing. But the ability to have these optimized libraries for queue and buffer management, for flow classification, for quality of service, and run it on your standard server CPU, is a very powerful capability because you no longer need purpose built silken to run those functions. We invented DPDK, we contributed it into open source, it ran in an open source project called DPDK.org, but we announced on Monday of this week that that's moving to the Linux Foundation. We're broadening the community of developers, we are multi-architecture, we are very broad in terms of the developers that are contributing to DPDK and we think that this is a fundamental building block of networks that will be, again, built and deployed over time. >> So you'd already invented it, but you handed it over to Linux Foundation. >> We invented it and we contributed it to open source, actually some years ago, into a project called DPDK.org but the announcement was that it was now moving into a Linux Foundation hosted project, because that gives us a broader umbrella by which we can attract more developers and have greater contributions from a broad ecosystem. >> Right. And we saw AT&T just gave a bunch of stuff to the Linux Foundation. >> Sandra: That's right. >> Scott: Everybody's giving it to the Linux Foundation. >> That's right, it's a good place to be. I was curious. Tell me your take, from the Intel perspective on this show specifically, but also more just open source in general and the role that Linux Foundation plays in taking a project that was obviously of significant value, but enabling it to go places maybe that it wouldn't if it wasn't part of the Foundation. >> Indeed, yeah. So Intel is a big believer in open source, open standards, and a big enabler and investor in broad ecosystems. We're consistently the number one or the number two contributor to many of the projects that we participate in, including Linux, the actual Linux kernel. From networking projects perspective, we really do like the leadership that the Linux Foundation is demonstrating in coalescing the industry around some of the big problems and challenges, as well as opportunities that we face together. >> Yes, we're live. >> We're live, it's that stage. So, we do believe that having just a broader landing zone, if you will, for the work that we're contributing, and having that parallelization that comes from a community of developers tackling the same problems together as opposed to one at a time, or as opposed to doing the same thing in various places, is very, very powerful. So we're very happy to be part of many of these networking projects and, of course, we're a big supporter and partner to the Linux Foundation for many years. >> Okay. I guess we're a third of the way, or a quarter of the way through 2017, on our way to 2019, the launch of 5G. Just curious, Sandra, as you look at what you're working on in 2017, obviously the 5G Initiative and all the developments around that are very exciting, we really are excited about it for the IoT side. We don't really spend too much time on the handset side, per se, but obviously for IoT it's very exciting. But what are some of the other priorities you have for 2017 that you're working on if we catch up a year from now that you can report back on? >> We definitely are driving toward the commercialization of NFV and SDN. We have been through a period of time, of technical feasibility, a lot of early lab trials followed by field trials. But we are absolutely seeing now this much broader scale of commercial deployments and we're going to see that throughout 2017 and 2018. We think that, clearly 5G acts as an accelerant to a lot of that work. A lot of the foundational work that needs to be done in terms of network transformation and network virtualization, enables 5G, and then 5G creates a compelling event for us to go faster. So we're getting ready for some of the 2018 Olympics, types of demonstrations of early technologies on the path to 5G in 2019 and 2020. Network transformation, network virtualization is a fundamental piece of that. The other area that we're investing quite a bit in is data analytics. AI, machine learning, deep learning. One of the things that we know is once we have programmable computing in all parts of the network, in the entire spectrum, from the client, to the access, to the edge, the core, and the cloud, that you can actually collect and harness that data and turn it into business value, either upstream to the content providers or downstream to the consumers of the information or the data. We'll see much more of that really starting to come to fruition this year, not just in the big hyperscale cloud guys but a lot of ways that the enterprises can use data and turn that into business value. So we're pretty excited about everything that's happening on that front, as well. >> You're going to be a busy lady. >> Sandra: We're busy. >> All right. Well, Sandra, thanks for stopping by. I know for Mobile World Congress we could only get you on the phone so it was great to get to meet you in person. >> Sandra: I know, it's more fun this way. >> Absolutely, all right. She's Sandra Rivera, he's Scott Raynovich, I'm Jeff Frick, you're watching The Cube from Open Networking Summit 2017. We'll be back after this short break. Thanks for watching.