>> Announcer: Live from San Francisco, it's theCube. Covering DevNet Create 2017, brought to you by Cisco. >> Hey welcome back everyone. Live here in San Francisco, exclusive coverage with theCube at Cisco's inaugural DevNet Create event. I'm John Furrier with my co-host Peter Burris. We're breaking down the new foray into the open source world with a big presence. Cisco expanding their DevNet core developer classic program and creating an open source model with collaboration, 90% of that activity is non-Cisco, really a good formula. And to help us this down is Krish Subramanian, Principal Analyst at Rishidot Research, formerly of Red Hat, formerly of a start up that was recently sold. Can't talk about it because it's not released yet. Friend of theCube, Cube alumni, part of the Clouderati, going way back. Krish, we've seen a lot of the waves of how cloud has evolved from the early days. I remember when EngineYard was a startup, Haruku was a couple guys, we were having our meetups. >> And the AWS was still like people who weren't able to make money. >> They were poo-pooing the hell out of it. It was EC2 and S3 with a couple different, I mean RightScale did everything back then, so think about the changes. And now Cisco here with the formula, they have the right formula, I got to give them props for that, doing it right. They're not trying to come in and do a land grab and sort of, "Ahh, we're Cisco", throwing their elbows around. Really doing it right, your thoughts? >> Yeah, definitely, come back to what some other legacy companies tried to do. Cisco didn't try to jump in and say, "Hey, we are going to run public cloud, compete with Amazon", and sort of take them down. They sort of waited for right moment, they initially started with the InterCloud, which will go much further, but when IoT came into picture, they were there right for that and they were there taking advantage of that. And with the increasing focus on developers, they are going right to capture the minds of developers. Especially for IoT, that is critical for Cisco to go-- >> Well, I'm really glad you're on with Peter. We have two analysts here who know the industry up and down, from every dimension. Of course, I'll add my color, but I want to ask both of you guys a couple questions. One, do you think Cisco's making the right moves by coming out and really focusing on their core competency, which is the network? They also bought AppDynamics, so that is a big purchase. So, you got apps meets infrastructure, programmable infrastructure, which means infrastructure as code. You really can't have infrastructure as code unless Cisco gets behind it, they're the leader. So, with IoT looming, this seems like a good move for Cisco. What do you think? >> Yeah, definitely, they are going in the right direction, so it's really like IoT's still in the early stages and we have to wait and see how it is going to evolve, but Cisco is very persistent. Especially I like the AppDynamics acquisition because they are clearly telling the world that we understand that applications are the future and developers need the right tools if they are to develop their apps on Cisco infrastructure. And with the emphasis on programmability, Cisco is taking right steps towards capturing developer attention and I hope with successful events like this, they will be able to get there. >> Peter, I want to go to you for a second because we just found out, in talking to Suzy, I did not know this, but in your previous life, when you ran research at META Group, folks may not know what that was, it was a big research firm at the time, you did some really similar work around the infrastructure developer. >> Yeah >> Okay, and our comment was, "What is old is now new". I got a degree in operating systems and computer science and that seems to be the model. What is this notion of an infrastructure developer? It was mentioned in the keynote today. Does that exist in this new scenario? Do you see it being viable? It seems like the messaging is tight. What is your reaction to this notion? You've done a lot of work on that. >> Well, as a way of answering the question John, and I'll play off of something you just said, when we talk about the degree with which this is relevant to Cisco, here's what I say. Everybody's always looking for what is it that's different from today, relative to yesterday? And there's a lot of things that are different. One of the most important ones is that yesterday's computing industry emphasized a priority set of models about how you do things. So, if you thought about the network, the network had a modeled structure. You sat there and you designed a network to be as relevant to as many things as possible. Same with the database. You sat there and you designed the database to be as relevant to whatever notion of applications. When we start talking about the new world, now what we're discovering is the data is going to force a reconfiguration. That's what big data is. In many respects, it's non-structured, non-modeled data, but we still want to do analytics. Same thing with the network. We want the network to evolve and emerge, have emerging characteristics that allow us to do things that we never really anticipated when we first put this stuff down. And so, the thing that an infrastructure developer, at least as we conceived it, and we were way ahead and probably wrong for that reason, but the way we conceive it is someone has to take some degree of responsibility for starting to characterize, fill that gap, characterize the services in the infrastructure that need to be made available to application developers in a way that makes coherent and consistent sense so that an application written to an infrastructure, in fact, may become a service to another application at some point in time in the future, because they make consistent assumptions about where they operate within that margin between the application and the infrastructure. >> John: Does that environment exist today, in your opinion? >> It does in certain places. It does in certain places. I think the whole notion of containers is making, in Kubernetes for example, is making some very powerful presumptions about how applications are going to interact with each other in the future. Now, we had SOA, but we also talked about Conway's law, it just never happened because the structure of the organizations that were using SOA just guaranteed you end up with monolithic, crap applications anyway. >> Explain Conway's law real quick for people who didn't-- >> Yeah, Conway's law is, it's been mentioned in theCube a couple times, basically, it's a suggestion that the structure of the application is a reflection of the structure of the organization that created it. And so, if you have a silo-based application development organization that's looking at the application for the finance group, or the marketing group, you are going to get a structured, siloed-oriented application, no matter what underlying technology you use. And that's been that way forever. >> And so, Krish, I want to get your thoughts because let's take that to the next level. So, one of the benefits of cloud was horizontally scalable model. That really kind of, to me, was the big ah-ha moment around software. And with DevOps, which is now called cloud native, which is the same thing, infrastructures code was, hey, I'm not not an infrastructure person. I just want it to be available for me and help me configure it out and programmable, as Suzy was saying. Okay, so if you take what Peter's saying about data, you've lived through the infrastructure as a service, platform as a service, SAS wars or evolution, however you want to look at it. And, now you see that kind of coalescing into SAS and infrastructure and PAS kind of folding away and kind of becoming less of a contentious conversation. But, now that same thing's happening with data, we believe. I mean I think, maybe he may disagree, but now data's now the new data layer. What's your thoughts on that? Because now, if you inject data into what was the old cloud stack, new things are really possible. >> Yeah, the thing is, data brings in a new dimensionality to what we are seeing right now. Everything from infrastructure to application, everything requires a mindset change in terms of seeing them as services. So, even if it is a physical hardware you are dealing with, you have to make it more service-like by putting an API in front of it. So, it's changing the way how we consume these services. But, data is the one that is bringing business value to customers. When you make data easily, sort of like, inter operate with the services, let's say call it, for lack of a better term, a services ocean kind of IT model you have in your enterprise. So, when you offer to bring data into it, it offers you a lot of opportunities which didn't exist in the past. It opens up new avenues in which you could manipulate data, make sense out of it and probably get more value than what you were getting in the past. >> What's interesting, if you bring micro services, if you think about Docker and Kubernetes, as you were saying, and you bring data now into the equation and the notion of microservices, you can apply all that microservices knowledge to data. That's what you were saying, from what I hear. Or concepts of-- >> Sort of like you will bring data close to take, earlier as Peter pointed out, data was in silos, representative of the organizational structure. So, by taking a more services approach and spreading the services across these siloed, PAS, siloed organization, you are bringing the entire organization into one single umbrella, sharing the data and thereby benefiting much more than what they were getting in the past. >> So John, in the opening, one of the things we talked about, and I'll repeat it here because he's probably going to see it and I'd love to hear your comments on it, is that we went to hardware-defined networking. And then we went to software-defined networking. And, Wikibon's working on a proposition and I'm sure we'll find reasons why it's not going to play out, but again, I'd like to hear your position, is what I'll call data-defined infrastructure. So, we were on theCube last week at Informatica and we heard a lot about the role that metadata's going to play in discovery of data resources and whatnot. I can imagine adding metadata when we start talking about dependencies and time and location and things that are relevant to how a network or how an infrastructure might configure itself to serve the data, becoming a feature of the programmability of the underlying infrastructure so that we end up, in five years, we do talk about data-defined infrastructure. Just as today, we're talking about software-defined infrastructure, where the infrastructure, literally, responds to the needs of the data because that, ultimately, is the most flexible way of think about this. What do you think? >> Yeah, I fully agree with you. In fact, data brings in a new dimensionality to the equation where applications, it's a morph based on what is there in the data. So, on-the-fly, the infrastructure needs to be modified. So, data sort of brings in a new way of doing infrastructure from what we have done in the past. I fully agree with the role of data in that and how, through the application, that influences how we deal with infrastructure. It does change completely. >> All right, so I got to ask you guys a question. Journeys, is journey to DevOps, journey to digital transformation, certainly has a lot of cloud, has a lot of open source involved with it. We're seeing the Ford CEO get fired, he hasn't been on the job for four years, right? So, you guys both work with end users and advise them, so what's your advise to CXOs where, hey the clock now is, I thought four years was short. It really should be seven to 10 on the transformation scale, but people are getting axed in their third year, so they got to show results. How does an executive make all this stuff happen in such a short time? Or should they just reset expectations? >> When the executive comes in, he, or she, not only should look at their core business, they should also think that they are a technology business and change the mindset completely. That mindset change needs a push from the top that's going to accelerate the change down the lane and I think the executive should think that they are becoming a CEO, or CXO, of a technology company, rather than a manufacturing company or a automobile company kind of thing. >> I think that's true, but look, we haven't studied what happened at Ford in detail because I'm sure there's some subtleties in there that we just don't fully understand, but on the surface, it sounds like he might have gotten a little bit of a raw deal, just from the pure standpoint of-- >> Well the stock was down 39%, so my guess is total Wall Street hatchet job, but -- >> Peter: Exactly. >> We don't know a lot of the politics, but Val Bercovici, who was on earlier, who has a lot of experience in organizations that net app since 97, or late 90s, brought an interesting point, you were saying earlier. Tesla creates a car that's a service. And so, to me, I hate to use the cliche, "Everything as a service", but essentially, that's what software's going to. So, where you make up a day, that's why I'm kind of poking at the data thing because I think you're on to som-- >> But it's the end of the day, Tesla still has to have a shop that bends metal, there's still some car manufacturing things that have to happen. And, in many respects, whether the old CEO is saying, well the value proposition is, someday this autonomous vehicle is going to happen, but right now, we still got to build cars that can compete in the world market. There's a lot of subtleties here. There are-- >> Yeah, but Tesla does upgrade with software over the network. >> For an 80 to $100,000 price point and there's about four billion people that are going to buy cars in the next five years that may, or may not, be able to buy a 80,000 to $100,000 car. So anyway, coming back to your core point, I think what it really means is that if you're in a situation where you don't have visibility in a how, some of these new, digital approaches are going to create value for your business, you're doomed. So, I think the first thing you got to do is you got to be very explicit. This is how digital technology's going to create value for my business, that's number one. And, be able to articulate that to, virtually, anybody that's capable of understanding it, including Wall Street. But, to do that, you have to step back and say, and what is it about that digital technology that's going to create value for my business. And the thing that's going to do it, or not, is the data. >> And the asset configuration around, the work around the assets. >> Especially the asset configuration, as it's defined by the data. And, increasingly, there's an economics terms, what we're going to see happen over the next 10 years is the asset specificities are going to go down dramatically. In other words, the ability to which, or the degree to which an asset can only be configured to a specific purpose. Software's going to change that dynamic dramatically. And that, in many respects, is one of the fundamental, underlying things that's going on here. But, at the end of the day, you have to say, what role is data going to play in my business? How am I going to articulate that role by saying that I'm going to incorporate digital in this way? And then, put in place a plan that demonstrates that you're competent about some of these things. And, if your shareholders don't like it, they're not going to like it from anybody, not just you. >> Krish, I want to get your thoughts on the Cloud Native Compute Foundation. Why it's so successful. Why, in your opinion, do you think, there just booming with vendors, a lot of cash infusion, a lot of activity, projects went from one, three, 10. We had Dan on earlier, a lot of growth in the cloud native. And then, also, Kubernetes as a, kind of as an emerging, really interesting dynamic, vis-a-vis multicloud. So why cloud native is so popular and the impact of Kubernetes. >> Cloud native is popular because of late, developers are understanding that the role we are building applications is not going to work in cloud. When containers came into picture, that really made it easy for developers to develop cloud native apps. It got them to take advantage of the more distributed nature of the underlying infrastructure. So, the containers are the main reason why cloud native has become the household term, even in the enterprises. That could be one of the reason why Cloud Native Foundation is popular. Because they came at the right time to host all these development projects and evangelize with the developers and take steps in that. As far as Kubernetes is concerned, it worked at Google's CE. If it can work at Google's CE and then solve Google's problem, it should be able to help-- >> If it's good for Google, it's good for me. That's their strategy. >> And also, people are slowly realizing that as more and more enterprises go to cloud, they are realizing that going with a single cloud provider may not solve all their problems because different cloud providers have different set of services. So, they want to take advantage of all that. But, they want a single pane of glass to manage everything. Kubernetes is this general to be that at the cloud-- >> Krish, thanks for coming on. Peter, thanks for the comments, I'll just wrap up the analyst segment by saying, in my opinion, I think Cisco's making a good move here because, to your point about Google and Kubernetes is, and that's one of many examples of great software being contributed to open source. And open source, for all the times I've been involved with it since I was in college, is this more great software coming to the table now than ever before and that's creating great innovation. So, combined with the cloud and cloud native and Kubernetes, a perfect storm of innovation is coming. And it's coming, not from vendors, it's coming from open source. And, so the smart vendors are putting their toe in the water and really figuring it out. And again, the-- >> Peter: It is coming from vendor support though. >> Well the vendors are smart by putting their people in open source as a proxy for contribution. That's the open source model. That, to me, is the new R&D. It's a new innovation strategy, coupled with some proprietary R&D. Not saying they should be going all open source. >> I agree with it completely. In fact, I would even go one step further and say open source is completely disrupting the traditional enterprise software in modern business. Think about someone like Capital One putting critical software as open source and disrupting all the vendors in the space, so it's-- >> Well, let's continue the conversation in studio or tomorrow. Again, open source is horizontally scaling as well. Great stuff, great projects. More exclusive coverage from the inaugural event for Cisco's DevNet Create after this short break. (up-tempo music) >> Hi, I'm April Mitchell and I'm the senior director of strategy--