 Live from Miami Beach, Florida, extracting the signal from the noise, it's theCUBE. Covering .NEXT Conference, brought to you by Nutanix. Now your host, Dave Vellante and Stu Miniman. Welcome back to Steamy Miami. We're here at the Nutanix NextConf, Nutanix inaugural user conference, almost a thousand people here, a lot of energy, a lot of disruption going on. A little tension between VMware and Nutanix, which is, Nutanix is embracing. Matt Eastwood is here. He's Senior Vice President at IDC, focusing on all things infrastructure. Matt, it's great to see you again. Thanks for coming back into theCUBE. Thanks, Dave, it's a pleasure. It's fun to be here. So yeah, it's good, good conference, you know? Stu and I were saying we do a lot of these and usually the inaugural ones are just like kind of big hackathons and you know, this is very well orchestrated, nice hotel, great food, good customers. What's your take on, on Next? I thought the keynotes were very well done today and so it was, you know, as an analyst, it was really nice to see them lay out a vision but also product behind that vision that really kind of extends their story. Right up front, they were very clear that they were uncomfortable kind of being boxed in with a definite, you know, kind of a defined market space around hyperconverged and clearly they see themselves as players, you know, really across the data center, across, you know, the layers of the data center but also extending out into the cloud, across that virtualization layer into the cloud. So that was really fun to see and I think the audience really, really received it well. The company's like, you're trying to kind of challenge your business because you've got to, you know, you got to put things in that have very tight definitions. If you don't have tight definitions, then what's the point of your numbers? So how do you, I mean, I don't know who coined hyperconverged this. I heard this morning it was a room in Asia. I thought it was Steve Chambers, it could have been you, I don't know. But anyway, regardless, it doesn't really matter. You guys use the term. I think you had a market scape up there. Nutanix are showing that today with the term hyperconverged. So do you size that space per se? We've, yeah, so we're setting up the tracking of that market but at this point we've done some forecasting. So we forecasted this market to grow pretty dramatically from I think 800 plus million in 2015 to a billion and a half next year and then it kind of builds on out beyond that to north of three billion. But the broader space, the overall, what market that people would refer to as converged or IDCs use the term integrated, it's really incredibly fluid. So you have these big plays, what VC did with VBlock and what NetApp did with FlexPod and then of course Oracle with their Exa. But now you have this kind of really, all this momentum in the market in the middle that's really around these infrastructures that are increasingly software defined that are going to put an awful lot of architectural pressure on what had been a market that was really more about integration than anything else. But so these markets are changing really fast. So the hyperconversion is subset of the overall integrated, what you guys call integrated, right? So obviously integrated is going to be much, much larger. We've talked about this before and TAM for integrated is essentially everything. All storage servers and networking. What you're thinking on, what percent, think out even, I don't care, give me a number of 10 years, you know, spitball number, what percent of that business do you think will be integrated in some way, shape or form, even from a reference architecture all the way to a, you know, close to a skew or hyperconverged, what percent do you think are going to go in that direction? Pretty significant. So I would start first by also kind of picking up on a few themes from this morning where if you think about what X86 has meant to the server market, massive amount of transformation in the market, which led to lots of standardization, but also lots of merchant software running on those systems. And then of course we had the, you know, the Linux and open source movement on top of that. And then more recently the virtualization movement on top. So if you kind of build off of that, it's a very, very significant part of the market that's addressable in this space. I think what I heard from Nutanix yesterday is if you take the kind of the hyper, I guess hyper scale piece of this off the table and you know, assume that the Amazons and Googles and the Microsofts of the world are going to do things in a somewhat different way. And you look at the more traditional enterprise market, the mix here becomes pretty significant I would say. You're looking at upwards of 50% of the market that will move in this direction in the next five years. And then the guys who don't move, why are they not moving? It's just they like to have sort of a roll your own. Yeah. Good seats to the socks game. I mean, a combination of things. I mean, I think, so part of it is, in any end user conversation, I think we've talked about this in the past too, you really want to kind of ground the conversation in a couple of different planes. One is what's happening kind of on-prem, what's happening off. And then what's happening kind of in a traditional vein and what's happening in a shared set of pooled resources. And what you see of course is all the growth is in the upper half of that kind of quadrant and the shared resource. And more and more of it's actually been happening off-prem but what you really need to do is give the end user a lot of room to run in terms of how they create efficiencies around their traditional applications and move them into these pooled infrastructures, make them easier to manage. A lot of the attack cost here is actually people. It's going after the CIS admin expense of managing these very discreet infrastructures. Yeah, yeah, it's interesting. When we've talked about hyperscale and baiting the enterprise, one of the numbers you and I've talked about is how many servers can I manage per person? Facebook at the Open Compute Summit said that they have, it's quarter magnitude, 10, 20,000 servers per admin but that's not a guy that really configures and does more. I mean they bring it in by the rack, they have a bunch of guys that do it, when it's done they turn it off, they throw it out. So the role of what compute is and how I manage it's different. Our forecast actually that I think within the next five years, we think two thirds of stuff that customers own will be in some sort of converge because people want simple. And if you take, you've tracked a lot of the how much is going into the hyperscale guys, that means how much is left is just a standalone, I'm going to buy a server where a couple of servers has changed. I mean what do you see kind of macro level that hyperscale and service provider and enterprise? Well I think you touched on an important point Stu which is you still see massive differences between what's going on in enterprise and what's going on in the hyperscalers. The enterprises will tell you that they're managing 7,500 servers per admin and the hyperscalers are well in the thousands. They also tend to, I think the other thing we should probably see up here is what might happen next in the infrastructure itself in terms of you have a bunch of different kind of competitive threats and forces on the market, one of which is disaggregation. You have to move towards a more of a disaggregated set of resources where you can start to think discreetly about how compute, memory, and disk is applied to a workload. So HP last week even talked about composable, that kind of a story you hear Intel talking about rack scale. So there's a lot of that happening in the marketplace. But I think in terms of really what the end user wants to do and I think the Nutanix team has said this quite well, the focus on differentiation really moves into data and into apps. So how do you differentiate what the data you have on your customers and on your market and how do you differentiate using new types of applications say mobile engagement. And the differentiation's not in the infrastructure. So you want to make that easy and you want to make that something that is, you know, can really be done super simply to allow the speed that needs to happen in the business on top to go faster. Yeah, I want to get your opinion. One of the criticisms of hyperconverged as a category is that, you know, it's your compute and storage together. Nutanix talked about how it's not a homogeneous single block that they have flexibility. They have compute heavy, they have storage heavy. But at the end of the day, there is some linkage between my compute and my storage. And, you know, what part of the market does that hit? Are there, you know, certain environments where I want to just scale compute and scale storage separately? What's your take on that next? I mean, you're right. They laid out that kind of two by two today where they looked at compute heavy, storage heavy. They looked at, you know, kind of the IO, the IO officer, the IO intensity on the workload itself. I do think that is very much how end users think. But of course, when you're using these standard building blocks, you can't hit all those range of workloads with a single, with a single physical piece of hardware. So today, I think what they're laying out in terms of, I don't know, was it five or six different physical configs to cover the landscape is completely reasonable. Even in Facebook, which you mentioned earlier, they have five physical deployment types that they use for their range of workloads. And you could argue that their range of workloads is a bit more homogeneous than what you might find in an enterprise. So I think that's a very reasonable starting point. I do think that over time, as some of these next-gen architectures build out, on the physical architectures build out, you'll probably see the ability to mix and match on the fly, the resources to the workload in real time a bit more. But that's still out there three, five years. Yeah, so right. In Facebook, it has five configurations at any one point because they do refresh that at a pretty regular base. What's the impact on all this on the chip makers themselves? I mean, Intel, there's been a run on consolidation in the chip world, Intel, the Avago Broadcom stuff. What's the ripple to those component pieces for this change in the marketplace? Well, the biggest, I mean, if you kind of look over the last 10 years and what's happened with the shift to virtualizing Consolidate, what that tended to drive was higher end Binsk use on the CPU side. So Intel, even though they weren't selling maybe the unit volumes that they thought they would back 10 or 15 years ago, they actually saw a pretty nice margin opportunity in that consolidation shift that was happening. I think Intel will see, they'll see some potential pressure on their data center business in places where they don't have a strong footprint today. So networking would be a good example where you could start to see, as the networking fabric begins to standardize, you could start to see influence more on the arm side there. You will start to see even some of the edged here computing layers also start to see some pressure from alternative technologies. But the acquisitions that are happening in the space are very much of the mindset that you're going to have bigger physical footprints that are consolidated in these mega hyperscale data centers, which can be increasingly optimized, not just through software, but also through how you attach acceleration hardware with the CPU GPU model, if you will. And so there's a real opportunity to differentiate there as well for Intel. So we heard the Gartner guy, Ray Puckett, I think his name is, talking about mode one and mode two. It sounded, I tweeted out, it sounded a lot like IDC's platform two, platform three. I don't know who beat who in the punch, but you've been talking platform two and three for a while now. They're similar in concept. D. Raj, when we asked him about it, said, we're in your parlance, platform 2.5, kind of a David Goulden statement. They're using mode 1.5, which I found interesting. You would have thought that they said, oh no, we're platform three, but his point was open stack is platform three and look at what happened to Nebula and others who sort of over rotated in there. So what are you seeing in terms from a standpoint of the practitioner? How are they dealing with, again, Puckett talked about that this morning, the sort of transition from one to the other, this bimodal organization. Are you guys seeing that and how are organizations putting a veneer, are they putting a veneer on the old stuff to make it look like the new stuff? Well, I think that's kind of where the struggle is in the market today. So if you think about that two by two construct I made around on and off-prem and traditional and shared, three of those boxes are really addressable kind of extensions of the 2.0 to 2.5 where 2.5 really sort of suggests that you're trying to get more efficient, you're trying to automate a lot more of the management of the overall shared pool, if you will, without using the cloud word. But when you start to think about really where we're heading and where the investment is going to happen more and more is in these cloud native or next gen applications which are increasingly built for frameworks and really aren't really kind of coded down into the hardware, into the systems. And so I think that we're still very early days in the enterprise and developing some of these really what we would call third platform type applications. And IDC also likes to use the third platform story as a way to kind of indicate where the innovation in the market could move on top. And so we talk more and more about innovation accelerators on top of that third platform. So things like robotics and cognitive computing and other kind of advanced uses of technology, really that will extend deeper into the business. And so I think we're very, very early days in kind of what technology will mean to the business as you start to drive instrumentation deeper and deeper into the business and change the business model to one that's really more around lifecycle services that are really optimized through some sort of digital transformation. Yeah, so that's a great point. I mean, it's the combinations of technologies on top of that platform that are ultimately going to drive innovation. And so then it begs the conversation, okay, what does that infrastructure look like? I would think just in listening to you speak that, okay, think about who's well positioned. Obviously the guys like Amazon who were doing virtually, they started in platform three and improving a business model that are well positioned. But you look what IBM's doing with BlueMix, on top of Cloud Foundry, HP to a certain extent, although they don't have the developers, Microsoft for sure is in a very strong position, one would think, to bridge from past the future. John Furrier asked Pat Gelsinger, are you a halfway house? I thought Pat was going to close line, John, when he asked him that, but it's interesting to watch VMware sort of evolve from what you call platform two to platform three and be that 2.5. And it's also interesting to see Nutanix trying to accelerate some of the things that maybe VMware is putting the brakes on. VMware talks about OpenStack, they talk about containers, they talk about all these sort of platform three innovations and they join, but ultimately one looks back and says, maybe they're trying to freeze the market, slow things down so that they can catch up. What are your thoughts on what's going on between Nutanix and VMware? Well, I think you touched on a couple of, I've even heard Joe Tucci kind of refer to VMware as a bridge from second to third, so maybe not a halfway house but still kind of a bridge. But I think if you're an end user or you're a vendor that's trying to attack this market, and this is very much what Nutanix is trying to do, you want to play to a number of core advocates that you have in the data center or in the infrastructure that can help you with that transition. And so if you think about what VMware did, VMware built a really nice ecosystem of CIS admins that were empowered by the technology and they really pulled that technology in. And then what's happened more recently with some of the shifts towards this conversion or integrated building block has resulted in the need for very dynamic shifts in how the organizations are set up and managed so they compute the storage, the networking teams needed to be brought together and you needed also to bring in a high level architect or VP of infrastructure to really kind of bridge some of the gaps that were happening. I think what Nutanix is playing on or could play on is the ability for that CIS admin down at a grassroots level to continue a transformation in the data center that happens in a way that's used their term a bit more invisible. It just sort of happens because it happens at a price point or at a friction point that's a lot lower in the organization than what maybe it takes to drive a really high level million dollar plus kind of integrated strategy into the data center. So that will be an interesting thing to watch is just kind of how fluid this plays out in the data center. Now the other thing too is that there's the little rift I call the urinary Olympics between VMware and Nutanix, small Chuck Hollis blog type of stuff. But with Evo Rail, we had D Raj last year said, hey, what do you think? He's like, okay, game on. Justifies what we're doing. But Nutanix chose not to be an OEM for VMware so they don't resell VMware licenses he was on theCUBE today. He said, because we don't want to compete with our partners. Okay, what he didn't say, which is also, I think the fact that I'd love to get your thoughts on it maybe it's not a fact, my opinion is that they don't want to be a reseller of VMware because it's a crappy business. And you look at HP now, I mean, HP used to claim Iowa with the largest reseller of VMware, IBM used to claim the same thing. And I think they've got to be looking at their businesses now and going, so what? You know, we want to be a platform company. We don't want to be a reseller of technology. So Nutanix has made that decision. And I think that's pretty bold. I wonder what you would think about that. I think it's wise. I mean, there's certainly, one of the things I also heard loud and clear this week is that one of the ways that Nutanix intends to grow their business is by allowing their partners to differentiate more on top of the platform. So, you know, if there's a ding from Nutanix towards VMware with Evo, it's that Evo doesn't allow as much differentiation in the platform. It also doesn't allow, you could argue, as much opportunity to share in the profit pool with VMware in that market as well. So I think that's a big part of what Nutanix is trying to operationalize around, is give a smaller group of strategic partners. So what I kind of heard from Nutanix this week is, yes, they've got Dell in place. They clearly see an opportunity for another global OEM partner. And they also see opportunities for regional OEMs to help them in China or Japan or Europe, say. So they're very focused on building partnerships that allow those partners to really differentiate in how they bring the technology to market. But yeah, they don't want to, clearly, you don't want to just get in the business of reselling somebody else's technology. In fact, really what we're talking about here is control. And so, you know, the control point continues to move up the stack. And you want to be at front and center at that point of control as that moves up. It gets control on its margins. Stu, what the VMware perspective on this, though. So I can imagine VMware saying, well, we know that homogeneity is an advantage for the hyperscale guys, and we're trying to drive homogeneity. And we can be the iPhone of the enterprise. What do you think they're taking? But the problem is, is if you look at the reality, if we, you know, we talk to users today, most customers have multiple hypervisors. And most customers have multiple clouds, multi-hypervisor, multi-cloud world. And there's a huge opportunity in the management space of all this, which is where Nutanix is going all in with XCP, Prism, and Acropolis. So, I tell you, Acropolis, I think is going to get a lot of conversations. It's, you know, KVM-based hypervisor. Maybe I can move around and everything. Most customers I talk to that have VMware are pretty happy with it. They're not necessarily fleeing. Sure, there's costs there that maybe I can save, but, you know, OpenStack tried to pull that away for years to KVM. So, you know, yes, there's some opportunity there. I thought, you know, when the full vision's unveiled and I have my virtualized workloads, my containerized workloads, and my cloud-based workloads, if I can manage all those together, you know, that's pretty special. I'm curious, what's your whole take on Acropolis? What's, yeah. Yeah, I mean, I think, I mean, ultimately, if you want to play that customer choice card, that's the right way to go at the market. And, you know, I think, if you kind of look at how virtualization's played out over the last 10 or 12 years, what you frequently heard, even from VMware customers, you know, even eight or 10 years ago, that it was great technology, it was helping them save an awful lot of money, but it was expensive. And where they actually start to feel that expense is when you get to a point where you're, you know, 85, 90, 95% virtualized, and all of a sudden, virtualization's no longer a tool to help you save money. It's something that's baked into your baseline cost. And I think that's an area where Nutanix has the opportunity to kind of pivot and help shift that conversation a bit. And so what you typically tend to see when people get, you know, kind of, maybe start to feel that way, is they shift into a more of a converged infrastructure as a way to get even more efficient with the infrastructure itself, or they start to look at multiple hypervisors and kind of broadening out how they think about usage and for various workloads. And I think Nutanix is trying to play both ways. They're giving customers choice on the hypervisor, but they're also giving them an opportunity to chase a lower cost infrastructure that maybe has higher rates utilization. Yeah, I just, I wonder that management play is a pretty lofty goal, because if you look, you know, VMware's, you know, has their management, they've bought a number of companies. Cisco's bought a number of companies. Microsoft wants to be the center of your universe. Yes. And, you know, nobody has done it. I, you know, I joke always as an analyst, I can always say, you know, that the two things that are always holding us back are management security. And, you know, does Nutanix have the right and the talent to be able to be an important part of that discussion? Yeah, I mean, I think that's, I think that's a really well asked question. And probably not something we can answer in the short segment on theCUBE, but it's a, I think they've done a nice job creating a lot of buzz around their business, around their technology, and now they're extending the conversation beyond just the, you know, what they talked about yesterday was making storage invisible was kind of their first phase. Second phase is now extending into making that whole virtualization layer invisible, introducing more choice and then broadening that out to the cloud. I mean, I think there's a nice story there, but obviously as you broaden that story out, the number of points of intersection competitively in the market grows significantly. And, you know, these are complex data centers that we're talking about that like it or not are actually on some evolutionary path. They're not generally revolutionary, rip and replace type of place. Well, I think the other big challenge that Nutanix will have, maybe IPO will help a little bit, but it's the big guys will just steal their messaging. Are you going to hear the same, you already are hearing it from VMware and their customers, okay, well, and I think Vinod Kosla said it and said everybody's sort of risk averse and IT are risk averse, that's not going to Nutanix, not going to change that alone. What could change it is, you know, Amazon, the cloud, Docker, you know, DevOps, I mean, those things, but this is still slow to change. And so EMC, IBM, HP, VMware, Microsoft, they're all going to be saying the same thing and they've got huge footprints, right? Trust me, you know, we can do this too. And they don't really have any incentive to drive that transformation any faster than it needs to happen, right? So they're going to play to that conservatism where they can. So Ravi Matre, I asked him, well, will we see a multi-billion dollar company emerge from the core enterprise space, not Amazon eating into it or Google or Facebook, but from the core, he said, absolutely, the conditions are there. I'm not so sure, I mean, I see this oligopoly, you know, they get big, Cisco buys them, HP buys them, IBM buys them, EMC buys them, or whomever, IBM, VMware, Microsoft, what do you think? Do you think we'll see? I mean, we all want to see it. But will we see another NetApp emerge? Yeah, that's a good question. And you go through these waves of the last round of, you know, storage startups all were acquired and now you've got a couple of interesting plays out there in the market now, whether it's Nutanix or Pure or Nimbla, you know, where are they going to end up in the end of the day? I think really what Nutanix is signaling, I mean, I think a lot of people kind of misread it initially and saw it as kind of a client's software and hardware together, they're clearly a software company. And I think you could, you know, a lot of people rotated to it and said, well, who's going to, who would acquire Nutanix? You know, there've been a lot of rumors about Cisco, there've been a lot of rumors that, you know, Delgun and early, but you could argue that Microsoft would be, you know, that could gain as much as anybody from acquiring an asset like Nutanix over time. Microsoft, IBM, you know, but they're not going to over, IBM tends not to overpay for an asset like that. I was just, Matt, I love your point there because when Deera's compared himself against the competition, it's not a storage discussion. And really the opportunity in this space is to be that software, be that, you know, systems level. Deera says they're not quite ready to be called a platform company. There's a lot of work they need to do. We like that kind of humble nature, but yeah, absolutely, if they're, you know, a big software player, because I mean, you know, multi-billion dollar company, VMware, you know, came out of, you know, where they started and is now, you know, a major force in the environment. Yeah, absolutely. Five-billion dollar run rate, I guess. Yeah, I asked, will there be another VMware? It would have been a better question. And in the number of software companies that get to run rates north of a couple of billions, pretty limited. I mean, for VMware to do what it did and get to five billion is impressive. Now, what VMware needs to do is, you know, continue to drive the transformation, you know, well above the hypervisor, you know, NSX like you mentioned, Stu, and certainly what VSAN, and even with VCloud Air, I mean, they're talking about the same types of transformations in the data center, but the story gets a lot more challenging because the competition changes when you get into any one of those veins. It's complicated, it's challenging, you know, users are not complaining that, okay, every time I turn around, VMware is, you know, taking attacks or they're putting a new product in my face, but that's natural. VMware has to grow. If you're Pat Gelsinger, you got to keep expanding your team. Exactly, yeah. So as long as they add value, I think VMware is the gold standard, I mean, there's no question about it. I mean, they're hands down, you know, the best at what they do. And there's a lot of good customer loyalty there too, so. Yeah, so. All right, Matt, we have to leave it there. Awesome segment as usual. Thanks very much for coming on theCUBE. Thanks guys, appreciate it. Appreciate all the insight that you're providing and the framing of the market sizing, so really great, great work that you guys do, so appreciate it. All right, keep right there, everybody. Stu and I will be right back from Miami. This is the Nutanix NextConf. We'll be right back. This is theCUBE.