 OpenStack Summit 2016 brought to you by the OpenStack Foundation and headline sponsors Red Hat and Cisco. Now here are your hosts, Stu Miniman. Hi, welcome back to theCUBE. I'm Stu Miniman joined with Brian Gracely. It's day one of three of our coverage of OpenStack Summit 2016. I've got back on the program Chris Wright, who's with Red Hat and first time person on theCUBE. It's Vijay Venagupal, who's the director of product management with Cisco NFV. Thank you for joining us. All right, so first of all, you know, we throw around these terms, you know, NFV all the times, you know, those of us for the networking background, we know it's network function virtualization, but maybe Vijay, could you tell us a little bit about, you know, your job, what you do at Cisco and what is this NFV thing that, you know, to a non networking person? Vijay Venagupal and I lead product management for NFV infrastructure inside of Cisco's NFV business unit. It's a newly formed business unit inside of Cisco, specifically focused on enabling NFV for customers. To answer your question, what is NFV? It's really about the transformation of the service provider infrastructure from the way it's built today to a clarification of that infrastructure. It's about virtualizing the infrastructure and allowing services to run as virtual, you know, virtualized functions on a general purpose compute infrastructure. But it really means much more than that because it's about the transformation of operations. It's about the transformation of services. So it's really a complete transformation of service provider infrastructure and operations, and that's what NFV enables. Yeah, I mean, an oversimplification, I heard a couple of years ago at the Open Networking Summit was, if you think about like kind of the cable companies, they've got like apps and services that they could just push down to the consumer and it's not, you know, having to put different devices and everything, it's software, it goes in there and they deliver kind of that end to end service. So it's a little bit oversimplified. Chris, maybe you can help, you know, why are we sitting here talking to Cisco and Red Hat? You know, how does that work? How do you put together, you know, the kind of the peanut butter and chocolate that makes this NFV solution? Well, so we're coming at this from the point of view of providing infrastructure for virtualizing these network functions. What that means for us is there's open source expertise, there's specific projects, Linux, OpenStack, some of the layers in between. We're working with Cisco, they're bringing to this picture huge depth in network experience. So this is a great partnership where our experience in open source, our ability to build up these fundamental platforms working together with Cisco to bring these solutions, these networking solutions to customers, that's really what the goal is here. Yeah, you know, Jonathan Bryce during the keynote said, hey, you know, Telco slash NFV is now a very distinct use case, very distinct market space. What's going on in the market that that all of a sudden is up on the keynote at OpenStack, what's changing for them from a networking perspective that that's this critical, that we're having AT&T in Verizon, and Swisscom all talked about as key customers, what's the big change, what's the shift? You want me to take that? So what happened is that when the large service providers like AT&T and Verizon started looking at the way they run their networks and the way they run their operations, and they compare that with the way the IT industry has evolved over the last 10 to 15 years, they find that their speed of operations and their agility for delivering services is far lower compared to what the enterprises and IT companies are doing. They look to the likes of the big web giants like Amazon and Google, and see how they've transformed their network operations by using a lot of software and using a lot of virtualization. And the benefits of that have started becoming very abundantly clear because everything is available at the click of a button, services can be provisioned through a GUI, there's no manual operations or CLI involved, which is how network operators used to manage their networks. So the telcos have been looking at what the rest of the IT industry has done, and suddenly there is a necessity that they want to transform themselves. And OpenStack has become central to that transformation because the first thing that you need to do when you start to virtualize your network, you need a very strong and scalable virtualization infrastructure. And being the telecom industry is very open to adopting open source and very open to adopting standard-based technology. And so OpenStack de facto became that platform of choice of virtualization. So over the last three years, we've seen a tremendous amount of increase of adoption of OpenStack. It's almost become a de facto topic of condensation because NFV and OpenStack have actually become inseparable right now. I'd even go one step further and say the service providers are seeing an economic reality that doesn't really map out well into the future. And that is explosive data growth, consumption through their networks. And as Vijay was saying, how are we operating these networks? And look at our kind of administrator-to-box ratio, and it's fundamentally different from the enterprise or IT workloads. And some of those environments are providing services over the top of these networks. So they're competing with a whole new class of services and looking at kind of a, I like to call it an existential crisis. I mean, this is real business urgency for these guys. And so I think it's interesting when you contrast, compare and contrast with the enterprise world where virtualization is well understood. So here, moving to cloud is kind of a refinement, it's an evolution of that path on the service provider side. This is the beginning of a journey from hardware-centric to virtualization, and it's driven by a real sense of urgency, which is why it's showing up at the top of the line for all of these keynote sessions where you see, this is really important for these businesses. So earlier when we talked to you, Chris, we talked about kind of the reliability and everything. But what about scale? I would think that really kind of the telcos and the way that they build this really speak to the maturity of OpenStack and how it can scale. Well, this is the early days of applying OpenStack to the NFV problem domain. We were seeing a lot of rough edges. So those days were identifying all of the areas where we needed to improve the platform. And that's what we were talking about earlier with NOVA and Neutron and some of the core services in OpenStack. Today, we're seeing that this is actually functional. We're working together to bring this, our joint offering to customers. They're deploying it not just in POCs, but in production and moving actual traffic, consumer facing traffic with this technology with OpenStack. I think over the last three years, definitely there's been a smoothing of the edges. Because three years back, it was even difficult to get, to get a pod up reliably. You mentioned scale. And scale is a big problem for many of these customers. How do you run it with large deployments? There was an announcement earlier today where Verizon announced that the largest, the largest OpenStack deployment working together with Red Hat. So it's an example of the kind of scale that customers are looking for. So I'd say the last three years has seen a significant stabilization and improvement of OpenStack to address these telco kind of requirements. But we still have a long way to go. I think it's still the tip of the iceberg in terms of what's possible. There's a next generation set of capabilities required. And that's where we're working together. And as Chris said, we're bringing in the deep expertise of networking and network operations. And Red Hat brings in their deep open source and OpenStack experience. And what we're looking to do is bring these two symbiotic capabilities together to make that infrastructure for the next gen SP. VJ, could you just explain to our audience, when they're working with Cisco and Red Hat, what are they actually getting? What are the components to make it up? Kind of hardware, software, services component of that. So what Cisco has packaged together is we've announced a solution that we call the Cisco NFE infrastructure that I lead inside of Cisco. And the NFE infrastructure, think of it as a pre-integrated hardware and software solution for NFE infrastructure. So it includes the necessary compute, the networking, the storage, pre-packaged into a pod along with OpenStack. And that's where we partner with Red Hat. We run a rel Linux KVM on the compute servers. We use Red Hat Ceph for our storage virtualization. And we use Red Hat's Liberty release, the OSP8 release, as the OpenStack platform, which serves as the VIM for the NFE stack. And all of this together, what Cisco is doing is that we are pre-packaging all of this into that integrated pod. And we are working together with Red Hat to provide a single interface to the customer. Because the technology is so new and it's so complex, many vendors are finding it difficult, many customers are finding it difficult having to work with multiple different vendors. And the problem becomes who do you call when things go wrong? How do you troubleshoot such a complex environment? So what Cisco is saying is that we're going to step forward as a single throat to choke for the customer. So they're a single vendor that they have to hold responsible when things go wrong. And we work in the backend with Red Hat to ensure that we provide that SLA for the telco customer. Because SLA is how they live and die. They live and die by the SLA they offer to their customers. So we guarantee that SLA, but working in the backend with Red Hat to make sure the issues are resolved for the customer. So the value to the customer is that we've taken away the complexity of OpenStack. We've made it available in a pre-integrated and pre-packaged manner. And the whole thing is available as an integrated single, call it a single window support system. Very good. But we heard AT&T talk quite a bit today about their journey, Verizon probably has a similar journey in terms of scale. But one of the things that was interesting to me that it kind of went along with this idea that OpenStack plays a role, there's lots of other challenges to deal with, whether it's scale or whatever. What are you seeing the conversation change as you talk to carriers? I mean, that conversation today, they were talking about writing a lot of their own software, which maybe didn't happen five years ago, 10 years ago when they were building telephones switches. How does the conversation change? Who's in the room? Who's making up that audience these days? For my point of view, this is a conversation that's happening in the upstream communities. There's some activity still in the standards bodies. And then some in this sort of emerging open source space where we're trying to figure out how to apply standards to open source and how to open source can sort of feed back into the standards definition loop. And so a lot of these conversations are happening in open public forum. But you go back in time and you had a situation where really a standards focus industry, the customers would work together with vendors in the Sanders communities and then point at that standard and say to their vendor, give me one of those. And we're in a new place now where we're trying to work together, defining what it is we're building and then being able to deliver that as integrated solution to the end customer. And one of the big changes in the way the industry is responding to this is in the partnership and forums like OPNFE. Because OPNFE starts becoming very critical because in the absence of standards, we need a reference system that all, because every vendor is trying to build an NFE platform or NFE infrastructure. So OPNFE as a forum for arriving at a common reference definition of the NFE stack is critical. And again, Cisco and Red Hat are working very closely as partners in OPNFE too. And everything changes when you go from a 100, 500 page standard document to a bunch of GitHub repositories that you can actually go do things and validate if things are standards compliant. And no, it's unbelievable to think of that. How much things have changed even over the last five or 10 years? Yeah, it's a challenging transition for the operators that don't necessarily have the expertise, the developers, in-house to really engage in those communities. And so part of what we're doing is vendors where we provide these developer resources, we work together in upstream communities like OPNFE to understand the actual requirements. How do we get, how do we translate this kind of foreign language of telco speak into something that's consumable by the open stack community? Because that's who's going to try to actually develop software to meet some of these functional requirements. Good. Yeah, and we talked to, we were talking to, Rick Cash earlier about just Red Hat's view of should there be carrier open stacks, should there be, and I think what he was saying is the same thing you were. You have to work upstream, you have to sort of work in the areas, understand where scale is an issue, where redundancies an issue, and then bring that back into core so people aren't getting variation A, variation B, variation C, and you get these crazy forking scenarios. If you look at, what is one of the goals? One of the goals is leveraging open source to build this next generation network, architecture, this next generation network. If by leveraging open source we're creating a set of parallel forks, we've lost a significant amount of the developer efficiency that we gained by going to the open source development model. So it's absolutely critical to be upstream first and helping the upstream community understand why these problems matter to them. Because initially it might look like crazy talk. Why do I care about your weird esoteric NFE use case? And what we're seeing is they're not necessarily that, the use cases aren't necessarily that strange or unique. The language to describe them might be special to the telco industry, but if you think about a network function whose job in life is to efficiently process packets. And the way we enable that is by carefully placing that application workload on a piece of hardware. An HPC workload could benefit from that same application type placement. So we're looking for ways that we can communicate the requirements in ways that are sort of broadly accessible to the community. VJ, can you explain to our audience how big is this opportunity? How wide of an impact will it have in the marketplace? We estimate this is a $20 billion opportunity worldwide across the entire NFE portfolio. Including the hardware, OpenStack, the VNF, the orchestration system. So it's a $20 billion opportunity over the next five years. It's massive. So if you walked around and asked anybody here and you say, first impression of Red Hat, it's an open source company. First impression of Cisco, maybe not an open source company. Help us just, is that an impression you're trying to change? I mean, obviously you've got an entire business unit now focused on NFV where open source is core. Is that something Cisco's trying to get out in the marketplace? We've actively, Cisco's been actively investing in software a lot more over the last many, many years. It's at least five years that we've been focusing a lot more on software. Our open source strategy has evolved a lot over the last five years. We are probably the prime drivers of open daylight. So we saw the coming of the open source movement in networking. And so we were, David Ward as the chairman of the Open Daylight Foundation. And we have been very, very active participants in both open daylight and open stack. We have a huge team in Cisco under Lutaka's organization, which is actually actively investing in upstream development of open source. So Cisco has been investing very heavily in software and very heavily in open source over the past few years. So it is a transformation because we do see where the market is moving to and we're responding to what we're seeing in the market. So last question I have for people watching that haven't made it to Austin. How do they find out more about the solution? How do they hear from some of their peers that are leveraging these kind of technologies? Vijay, start with you. So the best starting place is to go to the Cisco and Red Hat product websites. Both of us have product websites that point to the NFE infrastructure product page. In Cisco it's called Cisco.com slash go slash NFVI. So that's the URL. And as always you can reach out to the Cisco accounting contact for customers and open the community mailing list for reaching out to us. Yeah, similarly for us, we have a telco specific portion of our redhat.com website and then we highlight customer success stories on redhat.com. So you can go there and learn about telco specific solutions and customer success stories. All right, well, Chris and Vijay, thank you so much for joining us here. We'll be back with lots more coverage here from OpenStack 2016 in Austin. Thanks for watching theCUBE. It's always fun to come back.