 Live from Boston, Massachusetts, it's theCUBE, covering OpenStack Summit 2017. Brought to you by the OpenStack Foundation, Red Hat, and additional ecosystem support. And we're back. I'm Stu Miniman, joining with my co-host for this week, John Troyer, and this is theCUBE's programming of OpenStack Summit 2017 in Boston. Happy to welcome back to the program a friend of the show, someone who's had on many times Lou Tucker, Vice President and CTO of Cisco Cloud Computing. Lou, welcome back. It's great to be here again. Yeah, so Lou, we've caught up with you many times at the OpenStack show. There's been the ebbs and the flow of the show. We know the only thing that's constant is that the rate of change is always going to increase, right? Absolutely is. So coming into the show, there's many people that were like, oh, well many of the big companies are pulling out or killing their OpenStack solutions. You had explained to us the Cisco, you know, inter-cloud solution a few years ago. And while, maybe you could explain to our audience what it means, what happened to that product, where OpenStack fits in your portfolio. I think like many companies, Cisco itself is going through transformation. As we are moving into the new world and cloud computing is continuing to come on very strong and be a real essential part of Cisco's strategy overall. We just found that the inter-cloud solution offering a kind of a public cloud for people was not the best use of our resources. Instead we're doubling down in terms of helping our customers build clouds, some private clouds and particularly seeing big uptakes in the telco space with NFV. And so that draws upon much more of Cisco's strength, really focusing on making our customers successful with cloud computing. As we continue to move up the stack and offer additional services from the cloud again, but that'll be a mixture of both private cloud and in public cloud, the kind of hybrid cloud movement that we're seeing everybody's going towards. Okay, and Cisco had made some acquisitions in this space. Maybe how do those fit with you? I think that you're seeing like with AppDynamics and others that we really are looking at being much more application focused in terms of helping our customers who are trying to deploy applications. So that fits in with this hybrid cloud strategy of ours. Applications can go across public cloud and private clouds. So we're seeing that and for example, a lot of those things are moving onto AWS and onto private clouds at the same time. I think that what we're seeing from an open stack perspective is this continued advance, particularly in terms of NFV and the telcos space. We saw that several years ago when we started some of the NFV working groups within the open stack foundation and we're seeing that comes to fruition now. And in fact, if you look at what's going on in China, for example, almost all of the big Chinese telecommunications companies now are adopting OpenStack as their foundational platform. Yeah, Lou, maybe you speak a little bit about where success is happening. Coming into the show, I'd heard a lot, you know, Europe and Asia and the telecoms especially. I think, you know, particularly I think because the primary public cloud market is in the US with Amazon, Google, Microsoft, that we're seeing broader adoption actually going on now outside of the US and OpenStack. So we're seeing that in Europe and Asia. More than, I think only 43% of the deployments of OpenStack are in the US. The rest are the rest of the world. So it truly is a global movement. In fact, that's why when we're hosting these summits we put them across the globe and we attract a global audience. Yeah, it was actually 74% of deployments were outside of the US. So I think it was, you know, users was the number you gave deployments was even higher outside of the US. So Lou, you talk with a lot of CIOs. One of our themes here is we're past the hype cycle. You know, OpenStack is real. It's in production with real use cases. A few years ago, I think you kind of referred to it. I think there was a lot of winner-take-all conversations in the cloud. You know, Amazon's going to win everything. There's more than one winner now. It's clear that the conversation, at least in the industry, has matured. As you talk with CIOs in their conference rooms, are they looking at real-world, multi-cloud and hybrid cloud scenarios? Absolutely. I think nobody's going to be on a single cloud. It just isn't practical anymore. So we're seeing that mixture and still the enterprise has got to spend $60 billion on cloud platforms over this next year. So in the past, I think that where some of the missteps have been, people tried to do it on their own too much. I mean, OpenStack, OpenSource, communities build these projects. And then though you really need curated additions and distributions and trusted companies to help you build that. Because you can't build up that expertise inside of every enterprise. It's just not an efficient way to do it. So I think that many times, the conversations that I have with CIOs have a lot to do with, they know they want to be on this journey. They know they're going to be some in public cloud. They know they have other requirements for what's transforming their data center. So I think we used to talk about software-defined infrastructure where software-defined data centers. And I think that what we're seeing with OpenStack, that is becoming in that layer. And then on top of that, we have all the newer technologies coming in, whether it be containers, or whether it be serverless computing and other kinds of things. So CIOs right now are going, please help. We need a way to go through this. They can either hire that talent, hire the experts, or they can talk to a lot of their traditional vendors to help them through that transition. All right, yeah, Lou, I remember the first couple of times we interviewed you, it was like, okay, how's Neutron doing? And all these side projects with kind of the big 10 to the core, you threw out there containers and serverless. A lot of that I see is really talking about kind of the application modernization, whereas OpenStacks tend to be at the infrastructure. What do you see? What's interesting to you, the future of the stack? I'd love to hear more about the service. I think that is a good way to characterize it, in fact. That OpenStack is really infrastructure as a service. The focus here is on infrastructure and abstracting the infrastructure so you can drive it through automation, through an API that lowers your costs, that improves your resiliency. And now, what is the best for the applications? In the traditional kind of virtualization world, all we did was recreate the infrastructure in a virtualized environment. That was great, but now you've got VMs, they act like servers and just like other machines. Instead, we saw it to most directly affect the application development lifecycle. We needed to move to silent cloud native computing, which is container-based using orchestration systems, such as Kubernetes. That's a much more application focus, less infrastructure. So we're seeing that OpenStack, I truly believe, is becoming the sort of standard and therefore less interesting over time because it's standardizing as being the infrastructure layer. And then on top of that, we're seeing all of the newer technologies, like you say, that are application-centered. Yeah, so I haven't heard serverless discuss. Is there a fit with that with OpenStack? I know how containers fit. Absolutely. I think, again, two or three years, we'll talk more and more about serverless. And one of the things that, as a computer scientist, I find fascinating about serverless is that we're finally talking about writing functions and putting them where they need to be applied, and now you're building a distributed application by connecting them. So we had SOA many years ago, service-oriented architectures. That's pretty to be too heavyweight. And now serverless is coming along as the complete opposite of that. It's ultralight. It's just right this business service I need. And what I find particularly interesting is it's happening the same time now as a lot of the open source projects that are in AI and machine learning and big data. So if you want to build an application, your application is probably going to span three or four different service providers, cloud providers, and your own data center. And so serverless is a way to allow you to connect all those things and start to build those applications so you can have an intelligent agent to whatever, reading your email and then scheduling something on your calendar, but then contacting another service to do some of the processing for you as well. As you talk with folks both at the infrastructure and application level, I'm not old infrastructure guys, too, right? It is sometimes a challenging conversation because they speak different languages. How can the infrastructure providers of the world speak effectively to the application developers with this whole new world of serverless? It's a great question because often they don't talk. And I think when they don't talk is where you have the real problems. Let me give you an example from like the network functionalization world. So that's where in the NFV case we have things such as mobile packet core, handling of cell phone calls that we virtualize now and is running on top of a cloud. Well, they need very good networking underneath it. So we need to be able to have the application ask the infrastructure through APIs, how can you improve me, give me guaranteed bandwidth. Make sure that I have connectivity between this point and this point. You want resiliency. So that's where none of these things exist in a vacuum and you can't run it on a single VM. You want the service to exist even though your infrastructure may have failures. So this is something that I think Netflix and Adrian talked about in time. And again, you really want application level resiliency. So again, these newer technologies allow us to do that. Whereas the infrastructure is mandate there is to get more and more efficient, more and more powerful, and then more and more automated so that it's driven from the application or from the management layer. We spend a lot of time talking about the hybrid cloud in this kind of on-premises versus public cloud or even multi-clouds. It's a huge network component. How's Cisco involved in kind of pulling some of these pieces together? That's why I get, we're very pleased to see the recognition come that it is a hybrid cloud world and therefore networking becomes something you really have to pay attention to. Just rely upon the public internet to just do everything magically for you is sometimes not the best kind of experience for your customers. So being able to have that kind of control over the networking as you span multiple providers is important. In the service provider space, that's what they've been doing forever. They've been delivering network services through their enterprise customers. And so now we're seeing that go on top of cloud such that AT&T is deploying many of these network, NFV applications across tens or hundreds of individual points of presence. And they have to work as a service. So the individual one can go down but you want the service to survive. A lot of talk about the edge at this show as people start to build maybe what used to be just a Linux kernel into a whole open stack distribution at the edge, IoT, as this hybrid cloud becomes an IoT cloud, are we there yet? Or is that still a few years ahead of us or are we there? I think what we're beginning to see, we actually saw Verizon today during one of the keynotes showing that a customer premise equipment can be a little instance of running a slimmed down version of open stack. And the reason you want that, as was mentioned, was so that you have a common way of managing that. So you can deploy applications through automation and a lot of the tools that we have around that all the way down to an edge device that's sitting in a retail store or sitting in a home environment. As we go, we are just beginning that journey. And I think that's why I tie that together with serverless, which to me is a data flow architecture, how one thing connects into another. You need all of those things working together as a distributed app and the extent to which that open stack starts to play a role there, I think will be important, as will containers. Yeah, Lou, the thing that flashes through my mind when I hear some of this, back to my early networking days, versioning and security. I mean, used to be you get your network, you get it all configured and then say don't breathe on this thing. Exactly right. Because you got duct tape and bailing wire and zip ties and everything like that. When I think about the edge, we've increased the surface area of attack and boy, we've already seen, oh that thing that was a five year old device out there is going to bring down the entire internet. So how do we tackle that? Yeah, so I think security is absolutely, one of the most paramount things that we have to solve. But what gives me hope is that the way we're doing it today is that one of the things that we're doing for example with open DNS is actually putting a lot of this in the cloud. With systems that, with assistive selling such as Maraki, we're doing managing those devices from the cloud. That means we can have a consistency of deployment out there. And the other thing is that we're moving this to software instead of hardware. So those devices that are out there are the five years old that you couldn't touch because you didn't want to have a truck roll to go out and replace them. That's, those days are over. Now we're actually, we can deploy several times a day new software, new containers out there. If we detect vulnerabilities, we can fix them very, very quickly. All right, Lou, I know open source is one of your passions. How's open source doing inside of Cisco and kind of the relationship with your customers in OpenStack? I think it's booming right now. Open source is continuing to advance. And I think that what we're finding is that, with for example a number of projects that are growing inside of OpenStack and in other areas is that there's more and more opportunities where we see where it's better for vendors to work together and to work independently. And so that I think is accelerating. And I know there's also discussions as well, where is open source going and how many software companies, Red Hat is a great example of successful software company based on open source, but where are the others? And I think that what we're seeing is what we're not counting is the Googles and the Facebooks and the Cisco's and others that are contributing heavily into open source, but we're not providing it as open source software. We're not providing it as a software company. We're providing services around it. We're providing solutions around it and leveraging the power of a community-developed software I think is the best way to do that. Okay, final question I have for you. You talked to a lot of CEOs talk about, what's the general mindset out there? How does technology drive business today? I find that it's coming at them so fast. And their biggest challenge is how do they keep up? How do they make choices? What do they commit to? What's going to be around in two years? The next shiny thing, is what scares the hell out of them right now because once they start embedding technologies into their enterprise, whatever stays there for a while, so they don't want to make the wrong choices. The way to approach it I think is to embrace change. It comes back to if you realize that you need to be transformative in your enterprise, you need to be able to quickly educate people so training is going to be important. CICD, continuous integration and deployment, everything allows you to make changes much more quickly. Those are the things that CICD needs to really start to focus on. All right, Lou Tucker, always a pleasure to catch up. We'll be back with much more coverage here. OpenStack 2017, Stu Miniman with John Troyer. You're watching theCUBE.