 Live from the Wynn Resort in Las Vegas, it's theCUBE, covering .next conference 2016. Brought to you by Nutanix. Now here are your hosts, Dave Vellante and Stu Miniman. Welcome back everybody, Jeffrey Snowver is here. He's the technical fellow at Microsoft and Sunil Padi's chief product officer amongst other things at Nutanix. Welcome back. Good to see you again, Jeffrey. Thanks for coming on. Yeah, it's great to be here. So great keynote this morning. I got to lead off with my favorite quote of the morning was the cloud is not a place, it's a model. And we talk all the time about it's not a place to put apps and data, it's about how it affects your business, how it affects the operating model. So what did you mean by that? Yeah, I mean that's literally it. A lot of people think the cloud is some public cloud thing and it's not. It's a way of doing business. So I contrasted the cloud model from the traditional model. The traditional model is one where we would have dedicated infrastructure and hardwired things and the cloud is much more of a model where we use the magic of software to get increased reliability, availability, and scalability using industry standard parts and common operating systems, et cetera. And then you can do that in the public cloud like you could not run the public cloud unless you did that. But we can take exactly the same techniques and in many cases exactly the same software and run that in service provider clouds or in customer's own data centers on their hardware. And Sunil, you have an ambitious goal of trying to essentially replicate that type of simplicity and functionality for your customers, but it's not trivial, is it? Yeah, yeah, definitely not. I mean, that's what makes our jobs all the more interesting, frankly, because we're still in the, I would say, the second innings, if any, of our journey here, right? So, but I think, as Jeff mentioned, I think completely agree with the analogy that you had there, and the way we have internalized it is that I think cloud could become a lifestyle for IT, not a thing or a place. It's a lifestyle thing for IT. And Keith, I think, said it yesterday well, which is people talk about cloud as one of the aspects of consumerization of IT, but it's actually the humanization of IT. And if we can help bring the power of Azure and Amazon to mainstream enterprises, either on their terms, on their own infrastructure, or on their service provider partners, then hopefully, you know, their lifestyles can evolve and dramatically change to focus on value about the stack. So, Jeffrey, I mean, Microsoft's always been a company that's had some examples of trying to simplify infrastructure. I mean, for a software company, you want infrastructure to be simple. Yes. Inexpensive, make my apps run faster. Right, so what makes so-called hyperconvergence a good platform for Microsoft workloads? Well, again, one of the benefits of the hyperconverged approach is in this lack of drama, right? So when you are customizing your hardware for your application, well, you might get it right, but you might get it wrong. If you get it wrong, well, you just spent all your money. You spent your budget. And even if you get it right, well, what happens is things change. Volume goes down, volume goes up. Like if it goes down, you're underutilized. If it goes up, now what? And so you have a lot of drama. The hyperconverged allows you this nice, uniform substrate, again, using the Magica software where you can scale up and scale down without any drama. You talked about Windows 2008 as the catalyst that sort of led the journey into the cloud. Now, Windows 2016 is essentially the platform for the future. Can you take us back? What was it about Windows 2008? How was it the jump off point for cloud? Yeah, so basically that's what we did, was at that time we, in fact, we were getting into the public cloud. And the reality was that 2008 wasn't up to the requirements for the cloud. So we took a snap of that code base, we split it, and then we did our public cloud on a version of Windows called Red Dog. And of course we didn't want it, that was a temporary thing. We took those learnings and started bringing them back into the operating system. And then with 2012, now Azure runs on it. Again, we're taking the learnings from their design points and their operational needs, and we're incorporating those into our next version of the operating system, Windows Server 2016. So it's really, we're called eating our own dog food, right? We're doing it ourselves. It's one thing to say, well, like I produce the CD-ROM, I give it to you, you tell me you have a problem. It's like, oh, okay, you had a problem. This guy had a problem too. What do I do with that, right? I'll fix some of them. It's another thing to say, well, I just released some software and now I'm getting my pager going off at two o'clock in the morning. Now, I'll tell you what, I'm motivated to make it really great because I don't want the pager going off at two o'clock in the morning. So that's what I mean. We've been doing that for many, many years and evolving the software to eliminate the drama. And now we're taking it, making it available to everyone. Jeff, one of the things we look at in this whole space is so many people can't really grasp the difference between just virtualization at the server level and what we're doing in cloud. The term I heard that really resonated from Microsoft a few years ago, it was cloud first. It was, as you said, you build that experience up in the public and then you're driving that down. How's that, can you talk about that difference between the virtualization and what you drive from the experience of the cloud? So virtualization really is just a variation of that traditional IT model, right? It did not change the model. It made it more efficient and more agile. The cloud really is this different model. You have different software architectures. You have different hardware architectures and different operational views. Now the point of that is that while it provides great benefit, it has a strong point of view, right? It doesn't allow you to do everything. It allows you to do the cloud things. And when you do it that way, you get fantastic benefits. But not everybody's going to want to do that. A lot of people are going to say, well, I need control over this or that or fine tune things. That's why I believe that most people that do deploy a private cloud solution will also continue to deploy a virtualization solution because of virtualization, they'll have all the control that they want or need. But they're also operating at a pretty low value stack, layer of the stack. One of the benefits of the cloud, as Sunil mentioned, is the opportunity to improve their lives by raising up the level of which they operate. You know, you remember before plug and play, right? Remember, you're like, I was one of the first guys at digital to get a CD-ROM and boy was I an excited guy. Oh, and then came the bent paper clips and the dip switches and the tears and the IRQs. Two days later, I got it working, but I felt like the stupidest guy in the world. Well, you know what? A lot of IT is like that. Just operating at such low levels, are they really adding any value? So I talked to a customer today and I said, well hey, does your company sell more product against your competitors because you're doing a great job at virtualization? And the answer was, no, is it right? No, but with the cloud, we take care of all that for you and you're able to focus in on enabling your developers to deliver better customer value through better servicing, et cetera. So Sunil, people talk about, well, where do I, you know, what stays, what goes? Where do I put what? Is your shared vision to make that transparent or make it a function of where the data lives? Yeah, I mean, it's, well, I think there were a couple of questions there, but let me start with the first one. It's a shared vision, right? I think absolutely. I think, in fact, that's one of the reasons why we have worked together so closely on the Azure pack integration as the foundational step before we get into deeper integrations that are on Azure stack and the evolution beyond. But I think, look, we all know that based on the type of workloads, independent of security, compliance, and so forth, from a pure economical perspective, at least from our perspective, it makes sense if you're elastic workloads, an OPEX model, just-in-time provisioning model, you know, start small, start quickly, stop as I can go, right? All work, which is the classic example of why you would use public cloud or various reasons for it. If you have a predictable model, whether it's the app, inertia, politics, inertia, but inevitably, in our opinion, if the workload is predictable, if you can foundationally provide the same architectural foundation as what powers the public cloud, then economically, things should work to your advantage for predictable workloads. So the customers, in our opinion, and I think this is the shared vision, I think are quite okay with that trade-off, right? And I think they're starting to internalize to use the right cloud for the right workload, but they still covet the same tooling because they want the same interface. They don't want to go back to like three, four different interfaces like they had between servers, networks, and storages. They don't want to do this across five different clouds or, you know, between private and public and so forth. So I think that's where, I think that, as I mentioned before, the true north of our internalization of what is needed, and we're going to execute on that in a joint approach as we have demonstrated with our first release of Azure Pack is the fact that, look, we should take a real step into providing that single pane of glass, the common tooling across your private deployment as well as your public department. Basically, whether I choose a CAPEX model or choose an OPEX model, should be transparent from a tooling perspective. I want to follow that up a little bit because there's two ends of the spectrum. Microsoft's always been a partner-friendly organization from day one. The other end of spectrum is Oracle. They have this thing called Same-Same, and they point to Microsoft often as a big competitor, and they say, well, they can't do it because they don't own everything. They have to do it with partners, et cetera. What do you say to that? I mean, is that Same-Same? I mean, it's their messaging, but basically that's what Cineal just described, a consistent experience, whether you're on-prem or in the cloud is what your shared vision is. So what do you say to those who say, well, Microsoft can't succeed at that because it's multiple companies? Yeah, at the end of the day, the proof of the pudding is in the eating, right? The reality is that there is some challenges here, and that's why it is a joint-engineered solution, right? We work closely with our hardware partners to make sure that all the seams work out. So that's the answer. Jeff, maybe can you help unpack for us? There's CPS with the Azure Pack, and then there's Azure Stack. You can just explain, because that gets even closer to having the exact same stack in the public Azure versus on-prem. Sure. Basically, the cloud is a journey, right? It begins with a set of steps. The first step is cloud platform solution, running on great hardware. Originally, we had cloud platform solution premium, which was a fixed set of hardware that you could choose one, two, three, or four racks. That was your choice. And now we have our second step is cloud platform standard, where you have a set of hardware choices and they start at a much smaller. How small do we start? One node, one U, in a three node cluster, it's going to be a one U box, basically, or a two U box, yeah. So you can start off three nodes and then scale up from there. That's what we have today. And then we're working on the next iteration, and we gave it a new name, Azure Stack. Marketing guys got earned their money. But it's basically the next step in that evolution. And with each step, we have more and more Azure consistency and we'll have more and more services that we take from Azure and we make available on-premises. Look, last year, Azure implemented over 500 new features. These guys are cranking stuff out. So it's going to take us a while to be able to find out which of those components, which of those services customers need the most, and then we'll be making them available. So Azure Stack is really just the next iteration. As you scale, what are you learning about the economics of your public cloud versus what you can do on-prem? Is it similar? Is it based on renting? Is it more expensive than owning? But then again, you see the impacts on business. What do the economics tell you? What have you learned? Yeah, I don't think I have the answer to that. I think the, you know, we're still finding it. Don't know yet. Too early to tell. Yeah. Well, actually, I don't know the answer. I think we know the answer. The answer exists. They might, but I don't know it. Do you have a thought on that? No, I think, look, what we've found anecdotally, but it's not mainstream is the fact that, look, you know, if there is true elasticity, if there's an obvious elastic nature to the fact that it's an elastic map reduce or a Hadoop-like workload where you have to consume a total amount of capacity for a small period of time during the day, then the economics are clearly in the favor of using Azure and Amazon or public cloud delivery. Where things become gray is the fact that, look, okay, yeah, it's, as I said, predictable workload. But it's not as easy to say that, oh, therefore they should just stay in the house. Because there's also, I'll give you a simple reason why people are using it. When we go and canvas customers and partners and so forth, there's an interesting reason for the advent of more workloads going to the public cloud is the fact that it takes a while to get approvals, provision space in a data center, get all the rack and stack going, even mundane stuff like that takes months, right? Sometimes, especially if you're out of space and stuff like that. So in many cases, especially in large enterprises, you'd be surprised many of the times the initial reason to use a public cloud is simply that, oh, yeah, it might make sense to come back here. But look, you know, it's just not, not only is it just easy, it's just the only way to get to time to value to hit a particular business user's requirements. And then, look, six months later, 12 months later, once they have the right assessment done, they may consider moving it back. Sometimes in some cases they fall in love with how it is. They don't mind, quote unquote, economics looking worse, sir, but the value that they're getting in terms of the lifestyle drives that economically different. So I think the economics of when to put a public cloud to use is a lot more straightforward. I think it's a little bit more grayer, at least from what I can see, and so what stage it has. I wouldn't underscore that and say that, you know, as you think about public cloud, private cloud, hosted cloud, you know, economics is one thing, or economics always matters, but reality is it's far more complex. There are a lot of other aspects to it, right? Data sovereignty, there you are regulated environment, et cetera, and even just pure economics. If it's just a pure lift and shift, I know exactly what this workload does and it's not going to grow and it's not going to scale, you can work out the economics. But a lot of the cloud is about, well, hey, I'm going to run this and then it's going to surge and then it's going to go down. And in the cloud, you stop paying for things. On premise, you still have that capacity, but then you have the ability to use the capacity for something else. So it gets pretty complex pretty quickly. And, Sunil, your point about predictability, I think, is a good indicator. I was watching some CUBE interviews recently and John Furrier and Peter Burris who heads our research organization, were talking about, Peter made the observation that historically, I think ERP, we had known processes, but unknown tech. But today, we have pretty well understood tech, but the processes are unknown. You're talking about data sovereignty, you're talking about this digital transformation. There's so much we don't know about the business process and so maybe that's where agility and elasticity and flexibility becomes more important. Some of these known processes are maybe better fit for on-prem. I don't know, any thoughts on that? Yeah, I mean, I think, look, I think, probably in the right direction. I guess the way to probably maybe summarize some of this discussion is, you need the optionality to be wrong or right. Which is the fact that even if you choose to stay on-premise, if it doesn't make sense in 12 months, you should be able to use the public cloud for certain services, or for the whole service. If you choose to use the public cloud for the first 12 months or six months or two years, you reserve the right to actually not be stuck there in a folk-left way. This is where the hybrid tooling becomes so much more important, is frankly that it's not as important maybe where you land on the first three months, six months. It is more important to make sure that you can sustain the workload with the right economics, the right agility in the next three years, four years, because that's what you're buying when you're buying infrastructure today. You're buying something for the next five, six, seven years. So why are you not making a similar sound decision when you're actually choosing a cloud? Where you're not stuck on something. If the economics change, then you should reserve the right to actually have the flexibility to move back. Yeah, you know, that's the thing I love about what we're doing together so much. You know, I've got friends that are in the public cloud only business and competitors and friends who are in the private cloud only business and they will talk to their customers about the benefits of their offerings and those are true. And then they'll start to throw rocks at one or the other. And they know in their heart they're not telling the truth, right? And with us, we, like I don't have a dog in that fight, right? We offer public cloud solution, you know, hosted cloud and private cloud. So what that enables me to do is to go to a customer and say, let's talk about your particulars. What makes sense for you? And then say, okay, great. Therefore we think the public cloud or therefore we think the private cloud or a hosted cloud. And as Sunil mentioned, the wonderful thing about our approach is that it's okay to be wrong. Like if you go and you did something and it turns out, oh, you know, that wasn't the right thing to do. I can move it from one place to another with very little drama. Our competitors, not so much. You got to get it right. And if you get it wrong, that's going to be a bad day. You know, I feel like we've had this discussion now for the last five years, but we're starting to get clarity on it. And I think one thing's for sure, we can, I think most of us agree with the exception of Andy Jassy that it's going to be a hybrid world. Okay, but so the guys we're going to win are the ones who can actually emulate that functionality, you know, on-prem as well as in the cloud. That's what you guys are doing. What gives you confidence, Sunil, that you'll be able to stay ahead of the competition that whatever lead you have today you'll be able to maintain? Yeah, I think speaking for Nutanix and then, you know, partnering with folks like Jeff and team there gives us a little bit more confidence. But look, I think our clarity is that, you know, what we talked about two things. One, from an architectural perspective, because this is a little bit of a paradigm shift in terms of how things are consumed, not just delivered. We believe that we have the luxury of time behind us over the last five, six years of building the right product without having to jam things over layers and layers, right? So I think some of the competitors who are trying to patch things on top of their existing solution set, you know, we'll run into some issues over the long term. The other thing is, as I mentioned, it's also the luxury of not having, you know, an incumbency to worry about. That actually complicates matters. So we don't have a large sand footprint, we don't have a large server footprint, we don't have a large storage switching footprint, you know, and so I think some of this is also correlated to what Jeff was talking about, which is as part of Azure Stack and those kinds of paradigms where they're actually taking some of the core tech that has proven that they've spent years trying to build it, tune it, delivering the right level of consumption model, and then using that technology to justifiably power the enterprise workloads. I think from the infrastructure app, that's really part of the journey that we've been on is to take a pure approach while also not being burdened with the sort of legacy of incumbency, right? And I think between those two factors, I think that creates sustainable differentiation for us long term. And of course, the fact that, you know, we, because we don't have much to lose, we also have a lot to gain with partnering, right? And the partnership that we have right now is a case in point of about how while we have AHV, while we support ESX, we don't mind doubling down on hyper-rebase solutions if that's the right answer for many customers, especially if they're bridging this uniformly across private and public labs. All right, Jeff, we'll give you the last word. Pick your poison, you know. Microsoft, Cloud First, the new company, the partnership.next, final talk. Yeah, I'd say that, you know, Microsoft really is a different company under the leadership of Sacha Nadella. And every day, you know, he just enjoins us to go listen to our customers, figure out what our customers need to do, break out of the old way of thinking, create new partnerships, support new platforms, do whatever it takes to make customers successful. And I'll tell you, it's a great from a business perspective, and it's great from a personal engineering perspective. We all want to work on stuff that change people's lives, and that's what we're doing under the new Microsoft, so it's a good day. Jeffery, Sunil, thanks very much for coming on theCUBE, appreciate it. Thanks, guys. All right, keep right there, we'll be back with our next guest right after this is theCUBE. We're live from Las Vegas, right back.