 Live from San Francisco, California, it's the Cube at VMworld 2014, brought to you by VMware, Cisco, EMC, HP, and Nutanix. Now here are your hosts, John Furrier and Dave Vellante. Welcome back. We're live in San Francisco, California, VMworld 2014. This is the Cube where we extract the signal from the noise. I'm John Furrier, Dave Vellante. Our fifth year here at VMworld broadcasting wall-to-wall three days of live coverage. Our next guest is John Gilmartin, GM and VP of the Software-Defined Data Center Business Unit. Welcome to the Cube. Thank you. Glad to be here. This is an area that maybe mainstream is not on top of, but we love to geek out on the Software-Defined Data Center. That's right. A few years ago, maybe three years, feels like ten years ago, this year acquisition, and Martin's been on multiple times. Software virtualization really has set the agenda for what's going on in the data center. Remember, it was very much a buzzword, SDDC, Software-Defined Data Center, but now it's becoming a reality. So one first question, get your perspective, is where is the meat on the bone right now this year with Software-Defined Data Center? What is materializing right now in market that's available and happening? Yeah, it's been fantastic because if you think about our customers, they're all trying to move to this notion of self-service cloud, help their developers be more agile, be more productive. And Software-Defined clearly the right architecture to go do that. And the last year has really brought us the last couple of pieces to go make that a reality. Obviously, network virtualization is a huge component. Delivery of NSX has really brought us a kind of a leaps and bounds for it around that, and the adoption of that has been great. And then now with virtual sand, as well as bringing Software-Defined into the storage space, you're seeing a tremendous amount of interest. You take all that, you fully virtualize your infrastructure, and then you bring management on top of that and automate on top of that. And really now we have the ability to build these self-service clouds inside the enterprise, start to meet the developers and have this kind of very self-service, agile IT infrastructure. It's almost as if you're changing the airplane engine out at 30,000 feet with the Software-Defined Data Center, as some people said, on the cube. And it's very difficult, but I want to get your perspective of where the pressure points of innovation are coming from. Is it coming from the apps? Certainly the containers show that apps are setting the agenda, workloads now of diversity, another variable. It used to be the infrastructure would enable, on top of it now, we seem to be pressing down from the top. And this dynamic provisioning environment seems to be this DevOps requirement. All that's in place. So how do we talk about the innovation of the pressure point? What are those pressure points? Well, as you point out, it's really about the applications and the requirements of applications and pushing down on the infrastructure. And in particular, as you look at kind of new style cloud-native applications, which tend to be a bit different than traditional apps, they're asking different things in the infrastructure. And that's, you know, developers are asking to do different things than necessarily what kind of traditional apps are required. Developers are looking for portability, they're looking for agility, they're looking for a different set of tooling, really. And, you know, they want that experience where they can go to a website, they can go to an API and programmatically spin up infrastructure. And so that's really what enterprise IT organizations now are challenged to go do, is to go provide that type of infrastructure that can support that for developers. Technology today is, you know, fundamental to the business model of every company out there. It used to just be about back office operations. Now it's really about the marketing organization, the sales organization, product development organizations. Every part of the business depends on technologies, changing business models. And therefore this is really what's asking IT organizations to be much more responsive and to do a lot more than they necessarily ever have in the past to support the business to move quickly. So in terms of network virtualization, so how much is that a part of this new model? Yeah, network virtualization clearly a critical component to this, right? And, you know, it's super interesting when we first brought out NSX last year, a lot of the value proposition was around the speed and agility. And if you look at the big cloud providers, if you look at the big financial firms, that was the kind of primary motivation initially. And we still see that a lot. It's been interesting over the last year to start to see the value proposition for network virtualization really shift. And if you look at more of the mainstream now, it's really a lot about this notion of micro segmentation and this notion of how do I bring security from that, you know, used to just be on the perimeter and start to bring that inside the data center. And that's been driving a lot of the interest and being able to get security controls all the way down to the VM and the application itself. So on Friday's pregame crowd chat we had Steve Perich chimed in on the security question. I asked, what are the big opportunities for startups? And he said security. And it's really not about perimeter security anymore. It's about something else. Could you describe what he means by that? I see perimeter security was the old way. Secure the perimeter, but people are getting in, the APIs and all kinds of things. Protect the queen with a moat. Exactly. What does he mean by that? And why does he say there's opportunities there? Yeah, I mean that's the traditional model of security in the data center. You put up this big, as you said, moat around the data center. And you hope that no one would get over that. The problem was if someone did, then it's all exposed on the inside, right? And so the notion now is how do we bring security inside the data center? Protect those applications. But in order to do that, you know, that traditional models are doing that are just too operationally complex or too expensive. It just can't do it with physical systems. So the beauty of network virtualization is you can start to bring that in inside the data center, bring those security controls to the VM and do so with enough automation and policy-based automation that it's operationally feasible to manage. Well, what about the flip side of that when the queen wants to leave her castle? How do you secure that use case, if I'm making sense, right? Yeah, I'm not sure I fully understand. So, okay, so you get the queen being the data, let's say. And the data by its very nature is distributed. So, okay, protect the perimeter. That's not enough. Now I can go deeper inside the data center and provide tools to make it simpler to deploy or if I can, you know, find a problem faster to solve that problem. But as the data starts to become dispersed, how do I create a security model? And does the software-defined data center help me do that? That can accommodate that dispersed data, that distributed data model? Yeah, because I mean, the great thing is as you bring security controls into software instead of the hardware, then it can travel and be part of that application. And actually, as the application moves or the data of that application moves, it can tie those security policies to the application itself. So it's an application-centric, data-centric security model. And it's a platform also that, you know, an ecosystem is building on top of to go bring even deeper set of security capabilities on top of. To talk about the startups you were talking about a second ago, you know, it's this whole platform now for innovation on top of that you can bring really interesting ways of thinking about new models of security. So two years ago when Pat Gelsinger took over as the CEO of EMC had a financial analyst meeting and Pat was part of that. And your new CFO stood up and talked about the TAM and gave a really good crisp presentation around that. I'm sure you're seeing these slides a lot. We see them as analysts. It's a big, big opportunity for VMware. And a huge part of that opportunity is the software-defined data center. I wanted to dig into that a little bit. Specifically, when I look at things like TAM, I say, okay, what's the business case? Because the business case is going to ultimately determine the degree of, you know, the rapidity of the adoption. So I wonder if you could talk about the business case for the software-defined data center. Maybe compare it to sort of phase one, which is, you know, virtualizing compute. Business case was enormous. It was a 10x, you know, value proposition. Absolutely. Is this bigger, similar, smaller, twice as big? I wonder if you could talk about that a little bit. And when you say business case, obviously, you're thinking about it from the customer perspective. Well, it's either I'm going to cut costs or I'm going to create some other kind of incremental business value. Either I'm going to drive revenues, I'm going to reduce cycle times, or I'm going to reduce the lap times, time to value, et cetera. The interesting thing is software-defined data center is really kind of hitting on all of those things. One of the key motivators is really moving faster. It's to be able to reduce cycle times. Instead of four weeks to deploy an application, let's get it down to a couple of minutes. Let's be able to meet the needs of developers to do DevOps-style software development. So it's all about speed and kind of driving revenue. But on the back end, if you start to think about the operating expense and capital expense associated with the IT infrastructure, you can start to address those pretty aggressively. If you think about VirtualSAN, for example, it's all about a different operating model for deploying storage for virtual machines. It's application-centric and VM-centric. And so you can reduce the amount of time that administrators are spending managing infrastructure and get them focused on managing kind of applications. So one way to address OPEX. Or if you think about the capital expense, what we see now, you've done quite a bit of analysis, is by virtualizing network, virtualizing storage, you can actually get down to anywhere between, I think it's 35 to 49% reduction in the total capital expense of building your data center. So really significant opportunities to reduce costs, both on the operating expense side through automation, but also on the capital expense side by moving more intelligence into the hardware itself. So just like with VirtualSAN, if you go back five, 10 years, VirtualSAN was a very simple capital expense story. We're here now where we have a story that's well much broader than that, but still inclusive of all those kind of capital expense benefits. I got to ask you about competition. Just checking out what's going on around the conversations. Obviously VMware's taking their claim out. Amazon on one front. But Microsoft's a player in the enterprise. So what do you guys do? These are the Microsoft partner, they're a software player, they got cloud. So how do you guys look at those guys? Are you guys too far down the stack? Yeah, so with Microsoft, at this point, we still are, let's see ourselves, it's really kind of leading the way around virtualization, and that's really been the kind of core and the foundation which we started from. And we still have a tremendous set of capabilities there. And so that's kind of a starting point. And then you build off on everything we're doing around network virtualization, everything you're doing to find storage. Really a very differentiated set of capabilities. And a really unique set of capabilities for being able to build that whole virtualized infrastructure. Then your app is set of management capabilities on that that are increasingly heterogeneous in nature. And we have this ability to kind of extend the data center in unique ways. Manage and automate it here, but extend it out to the cloud as well. So pretty powerful set of kind of technology. So Carl Etchenbach said that VMware is a data center automation company. Should he add an orchestration to that too? Or talk about that. What does data center automation company mean? Because obviously he's referring to the Self-Defined Data Center. And cloud certainly is automation and orchestration in the cloud. But in your world, what does that mean? Automation is clearly about taking a lot of the manual activities that an IT administrator or anybody else who's spending time with the infrastructure does. And let's run that in software. And let's not tie ourselves to operations that are specific to proprietary pieces of hardware. Let's get to a model where everything can be automated through software. We can get the scalable models of deployments and operations. That's really what automation means. Automation then allows you to start to move at the speed of business rather than being tied to kind of infrastructure and the hardware and everything else underneath. So the other quote from Carl was awesome, by the way. Great interview. He said a lot of customers are still on-prem. Okay, I'd buy that. These have a zillion customers. Yep. A lot of them on-prem. Why not private cloud? Or private cloud is today? Or is private cloud the halfway house or way station to the hybrid cloud? So talk about that dynamic. Software-defined data center, at the end of the day, could be software-driven. But at the end of the day, it's still a data center. You still have a data center somewhere, whether you're on cloud or on-prem. Talk about that on-premise dynamic. Yeah, so ultimately, if you think about the hybrid cloud, hybrid cloud is really the combination of assets that you own inside the data center along with assets that are sitting someplace else. And the motivations for that are, I want to be able to think about how do I optimize my costs? I want to think about how do I optimize my choices of placement for projects that are either short-lived, et cetera. And so there's a set of applications or projects where it makes sense to go rent capacity. But if you actually look at the total cost of ownership inside the data center, you can actually get to much better economics by owning the assets yourself and building on top. So there's definitely a ongoing and continued role for the private cloud. But there's a very clear set of use cases for extending periodically into the hybrid cloud. So really, let's combine both of those except the best of both. So let's do that in a way that's seamless. So we really treat the management and the operations of everything as the same, regardless of whether it's inside or outside, right? So I mean, the buzzword bingo is all getting reset because the new names are coming out, certainly the renaming convention. But I've got to ask you about kind of specifically around the suite that Pat talks about. Pat talks about the suite. So I just don't understand how that parses out relative to hyperconverge. And describe to the folks, what is hyperconverge? I mean, that's the new buzzword. I know hyper scale. Is it hyper scale with convergence? Is it web scale? So what do you guys define hyperconverged as? Hyperconverged is, in our mind, really kind of the coming together of a prescriptive hardware definition with software that's pre-installed and tightly integrated so that it's really easy to get to time to value, right? So you can get up running virtual machines in less than 15 minutes. And do that all with, you know, kind of a prescriptive design guidance, prescriptive kind of price understanding, and a single support organization that you can call and get support if you need help. And that's really built also- That's a key definition right there, up and running within- With 15 minutes, right? And one of the key enablers of that is, you know, the key enabler is virtual sand, and really building all that and types inside one of these kind of off the shelf. Why not call it package converged? Pre-package converged. Purpose-built converged. But basically- That's fine. That's essentially- But that's where it's going, right? That's what they mean. That's where it's headed, right? So it's really about making it easy for an organization to get up and running, get virtual machines deployed super quickly, and then be able to expand that in a building block way that expands very quickly and easily. John Gilmartin, who's the VP and general manager, software-defined business unit for VMware. Tell the folks out there, the last word here to end the segment. What's the biggest misconception of software-defined data center in the context of VMware? I think the biggest misconception is that it's something that's far into the future. The reality is, this is something that people are doing today. Technology exists. We can build this. And, you know, this is the way and the architecture that everyone's headed down towards. And what's the one thing that you could share that they might not know about you guys that's a very positive thing? Well, you know, hopefully people saw all the announcements and work we're doing around OpenStack, for example, really looking to bring these types of open APIs, middle-neutral APIs on top of this software-defined platform. And, yeah, that's a big news item for us. I want to make sure that everybody saw that, and it's a big part of where we're headed. Yeah, I mean, OpenStack and Docker are two big relevant news pieces. Exactly. It gives the app developers essentially access to infrastructure without being infrastructure guys, right? Correct. That's fundamentally something to find out. Yeah, we're, again, helping the enterprise guys set up an infrastructure that developers can access programmatically. That's the objective. John Gill-Martin inside the Cube. We're here live in San Francisco for VMworld 2014. I'm John Furrier with Dave Vellante. We'll be right back after this short break.