 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, Stu Miniman. Welcome back to Nutanix.com here in Miami. I'm Stu Miniman with wikibon.com. This is SiliconANGLE TVs, premier live coverage, all the best industry tech events. We go out to all the events, help extract the signal from the noise. We're really excited to be able to geek out in this session. We've got Sunil Poti and Vinnie V. Vinnie with Nutanix, Vinnie Gill. You guys did the bikino this morning, unleashing a torrent of new discussions. So Vinnie, welcome back to theCUBE. And Sunil, thank you for joining us for the first time. Got it, happy to be here, Stu. All right, so just so people know, Sunil, you're the SVP of engineering and product management, and Vinnie is the chief architect. So Sunil, I mean, you went through a lot this morning. I'm wondering if you can just, for our audience, we've teased out a bunch of it from ourselves and with Dheeraj and some of the others, but high level, what's the big news that you guys unveiled today? So the big news was that we, Nutanix raised the bar in terms of announcing a whole new platform called XCP, Extreme Computing Platform. And essentially elevating infrastructure from compute storage complexities to actually the entire data center construct. And that's why we're calling it as the next generation of computing. And so XCP is the product family. It's made up of two primary products, Acropolis and Prism. And inside Acropolis is the core distributed storage fabric, which is essentially extending our core distributed file system to the next level. In addition to that, we've come up with a net new application mobility fabric, which frees up application workloads, leveraging the distributed storage fabric to actually have cross hypervisor mobility, but also taking it to the next level to get into containers and hybrid cloud. And then of course, the last, you know, least important piece is the fact that we also enable the application mobility fabric to land on our own native hypervisor called the Acropolis hypervisor as a starting point, which we will extend to other hypervisors such as Hyper-We, but also Amazon and Azure. Yeah, it's interesting. You guys, you know, fit into what many people have called the hyperconverged infrastructure space, but, you know, at Wikibon, we've talked about, you know, first of all, we're very early in kind of this transition infrastructure, but it's a lot more than just kind of an appliance. Because, you know, those of us who've been in the industry while, you know, appliances are kind of usually a bridge to something else or they're short-lived. But in many ways, when I saw Dheeraj in your presentations, you know, Nutanix is a software company. Yeah, it's definitely a software company. In fact, I think today, as you know, we ship mostly an all-software product in the form factor of an appliance just because we want to retain the high bar of customer experience and support. I think just to speak a little bit about why we've announced what we've announced to, right? You know, you know, before Binny gets into, the Binny gets into some of the technical details. I think the single biggest thing that people who come to us and say is, look, you've done this great job on hyperconverged web scale. But really, aren't you selling enterprise infrastructure in a world when people are moving to the public cloud? Because the users are going outside because of consumerization and after going outside because of public cloud, why is infrastructure still inside? And if you really think through that and people think about public cloud as headwinds to enterprise infrastructure companies like Nutanix, you know, the big companies and so forth, but it turns out that we truly believe that public cloud is actually a huge tailwind. Here's why. People obviously are tasting the wine of public cloud, is how I call it, using elastic workloads, bursting workloads, DevOps and so forth. But the enterprise industry's actually broken up into predictable workloads and elastic workloads. Right now, it's sort of lost in the noise of what is exactly elastic versus predictable. Everybody thinks public cloud is good for everything, which is, by the way, they're really good at, if you break down all the marketing noise, it's good at the fact that you can actually start quickly, start small, and skill forever, right? The thing is that the economics, though, of a public cloud, it's kind of like rent was released versus buy. If I'm buying a place in Hawaii for five years versus renting a place for five years, clearly the economics of buy are better than the plan. The problem, though, is today that in the enterprise market, the legacy stack, the legacy virtualization, virtualization management, the whole stack, the entire stack, makes that economic delta not as clear. So that's really the value proposition of the Nutanix's XCP, frankly, is with acropolis and prism, we essentially take the design point of what powers Amazon and Azure, essentially to create an A block for the enterprise. So the design point is around the next generation building block for enabling a cloud native architecture inside the enterprise to get you the best economics, but at the same time, enabling you to leverage Azure and Amazon as you need for the last quarter. Yeah, so, Vinny, I want to pull you into this conversation here. What we've looked at is the center of the universe has been shifting over time. I think back, what server virtualization helped us do was it was no longer about a chunk of server. We kind of moved up the stack a little bit, and virtualization for a time became kind of center of big ecosystem, we grew out VMware, of course, drove that, and when I first got introduced to Nutanix, I thought of it as part of that ecosystem, but really the center of the universe should be the application, it's the business. So, can you maybe talk about, from an architecture standpoint, I've heard a lot of talk about what your field does, but from a development standpoint and solutions you have, how was it, talk about that journey some and what you guys are doing. So if you look at how technology is evolving, like 10 years ago when VMware made virtualization quite popular, applications used to fit in a virtual machine, and that gave you great consolidation benefits. Now if you look at the next generation applications, they are actually spanning multiple hosts, even Netflix is a single app. So now suddenly the construct that used to be a virtual machine is not really the construct on which you want to take actions, you want to take actions traditionally on applications. So if you power on a VM or power off a VM or take snapshots and backups, that's not really doing it for your application. So that's the reason why we want to focus on applications with our app mobility fabric as well. All the actions that you do need to make sense at the application level, right? So when you understand that, what comes out of it is that what the hypervisor does now needs to be actually made simpler and lean. The hypervisor needs to focus on being secure and being fast and reliable, right? So, and there needs to be a higher layer which we are calling the app mobility fabric that takes care of the higher level constructs like applications and services. Yeah, so it's interesting when I think back to what sits on top of the hypervisor. VMWare, the vast majority of environments, it's Microsoft that sits on top of that. When we looked I think at the analyst day when you guys unveiled it, it's kind of the options that I have on top of that, you got your Linux, your Windows, containers and even other type of clouds look like they start to fit in. Can you talk a little bit about that ecosystem? How important that breadth of ecosystem of applications is for what you're doing? Right, so the ecosystem is really important. We are working with a lot of partners and they've shown great interest on supporting their applications on Acropolis Hypervisor. And we continue to support VMWare and HyperVecosystems and also allow you to migrate and move these applications from your private cloud to the public cloud and back. So our philosophy is that you should be able to pick whatever hypervisor you are comfortable with based on your applications that you're running and give you a seamless experience and mobility so that you're not stuck to any silos of the past. All right, and I think from an ecosystem side just to add to what Vinny's talking about. So we break it down into like three categories, right? There's the actual hardware and run times whether it be Dell, Supermicro, eventually other hardware vendors, but also the new run times such as Docker or CoreOS. And then building on top of that to the core operating system layer, whether it be Ubuntu, Canonical, as well as Windows Server technology as the core. And then obviously layering on top of that to a widespread series of applications technology, such as Citrix, such as even VMWareView, such as F5, such as CommWall, Avaya, UC. It's essentially what we are now doing is because vendors of these application stacks are recognizing the value of this new web scale building block, I think there's a good marriage of synergies there so that we can work together so that those application workloads could be certified and validated on the A block for the enterprise. Yeah, so Sunil, I wonder if you could help clarify for our audience, you know, I remember a couple of months back it was like, oh, Nutanix is getting into the hypervisor business. I mean, I think VMWare didn't create the VM market, but they made a hypervisor that was easy, low-hanging through consolidation, utilization, and they got a ton of OEMs. I mean, you mentioned a couple of pieces, but are you trying to become the new hypervisor for the entire market? So obviously not. I think right now, I think the way to think about our strategy is, as I said, about building towards the Amazon-like or the Azure-like building block in the enterprise. If you think about the Amazon block in about 2009 to 2010, it had S3 and EC2. S3, did it have a proprietary SAN in it? No. Did it have proprietary compute or some hardware in it? No. Did it have a hypervisor that was restricted from an ELA perspective? No. Did it have management software that was built for a traditional stack? No. So from our perspective, the new stack, therefore, in the enterprise, needs to mimic that same architecture. So really, from our perspective, we believe ESX is a great hypervisor. The problem lies in the software above the hypervisor. It's in the management layer. It's in the hierarchy of operations versus management versus migration capabilities that were architected as piecemeal technologies on top of a design point that was a decade ago. That's really what we want to do with Acropolis and the application mobility fabric. The hypervisor happens to be the first instantiation of that app-loanability fabric. As we'll talk more, we fully intend to support other hypervisors like Hyper-V, as well as new cloud runtimes like AWS and Azure as first-class citizens of Acropolis. Yeah. All right, so, Vinnie, when we heard a little bit of the details about the Acropolis hypervisor, the two things that were kind of missing from just open-source KVM, I think we said were management, storage. Can you maybe give us a little insight as to what kind of effort was involved, how much is, you know, you're taking something to open-source, you guys giving back to the community? Is it, you know, public as to what's going on? What can you share? Yeah, of course, anything we are changing in the open-source, GPL, code-based, we are giving back. But what we are focusing on is making the hypervisor lean and secure. A lot of the intelligence is actually sitting on top in our controller virtual machine. So all of the storage constructs are abstracted out and made available through our storage virtual machine. And the hypervisor essentially is stateless now because the state we keep in our own metadata Cassandra layer. So it's a different model at looking at hypervisors. These are like consumable and replaceable units that you can use. On the management side, we have innovated, again, I think for the first time in the enterprise, a management plane for virtual machines that is truly distributed. So it is never down. Even when you're upgrading a management plane for virtual machines, you have complete uptime all the time. And with management, we have analytics that combines on the same Cassandra-based layer. So when you're doing VM provisioning, as you saw in the keynote, if there's, you know, you're running out of space, you can quickly drop down into analytics and see where your VM sprawl is, what you need to do all from a single fabric. And that's what's exciting about how we've done the hypervisor. All right. The other thing that is interesting, but I want to understand more is some of that mobility. You know, some things we're talking about, like, is it kind of like vMotion, but, you know, migrating from one hypervisor to another might be one click, but there was a number of steps that went through. You've got to reboot a server. It's not nondisruptive to what you're doing. Can you walk us through, you know, how it works and what use cases you really see customers using this for? Right, so I mean, when you go from one hypervisor to the other, you have to go through a reboot because, you know, the hardware is represented in memories is different. But we make it seamless and also fast. And the reason it's fast is because we've invested five years on the distributed storage fabric. And that makes any piece of data available anywhere you want, whether it's in the private cloud side or the public cloud side. Now, once you have the data, what you need to do is first convert the vM metadata, which is actually very simple. We do it instantly. And then you have to convert the data format, you know, vMDK versus vHDX and so on. Being a storage layer, we actually can do it on the fly. We can inject metadata and do it. And finally, it's about drivers' compatibility and all of that. So we do all of that seamlessly in the back end. There's a lot of engineering involved. But the reason we do it is we want to, again, not lock in our customers into any one particular hypervisor, including our own hypervisor. You need to have complete mobility. Yeah, it's interesting. When I think back to, you know, when we first got vMotion, everybody was super excited. But the problem you had is that to do virtualization while you needed network storage, and storage was broken. So we had vMotion and then took years later before we had storage vMotion and coordinating that stuff. Do you guys, with an architecture that kind of the hyperconverged, if you will, architecture in Nintana, specifically, does that give you an opportunity to just do something different with the storage fabric? I think just to add to what Vinny said, in fact, in the keynote, somebody was telling me that when they had heard about the Atmobriddy fabric, there was a tweet, I guess, which said the last time that they were so excited about a new technology was when they first heard about vMotion at one of the VM worlds about 10 years ago. I think the promise of Atmobriddy fabric, while it says Atmobriddy, the point of it centers around the fact that we actually want to make a lot of functions around infrastructure management, simple. Remember, that was the core point that we wanted to simplify. It so happens that to make them simple, we wanted a runtime that could be simpler to manage and simpler to deploy. And that's why we started with the Acropolis Hypervisor. And we will absolutely invest in similar capabilities of extending the AMF fabric into something like Hyper-V. But the functionality is just not about migration. It's about simple things such as VM placement, right? The ability to actually spin on VMs, power on VMs, to clone. So these are all functions that we believe are part of application mobility fabric with the goal that it frees up IT, elevates them to focus on business app development and deployment versus the actual nuts and rolls of infrastructure deployment. Yeah, it's a true, one of the architectural arguments we've been having for the last year or so is kind of the stack owner versus the stack user. So VMware with vSAN, some of the other guys that were using KVM before said, when I own the whole stack and I can control it, I can do some more there. Nutanix has been arguing the other side because it's flexible, I can use it. Performance is always a discussion. But, you know, so now would you agree that the stack owner can do that? I think stack users, it's another good way. I mean, I don't know how they'll go down politically, but I think from a state of being of what we think people want and desire is to ensure that the integrity of their job is protected from an SLA perspective. But they actually want their weekend back. If anything, that's our new tagline, right? As you've heard, is I want my weekend back. And I think Bob loves the one that coined that term. We're going to have him on later today. So sounds good. And so I think, I think from our perspective, you know, invisible infrastructure should elevate IT. That's our, you know, as you've heard that ad nauseam here. And so therefore the concept of using versus owning and owning as in having to deal with the upgrades, having to take time away from the family, having to work on triaging, finger pointing. Those are all the things that we believe we want to abstract by elevating IT. It's all right. So Sunil yesterday said, you know, simplicity is really being genius. And you know, it's, you know, complex. You know, changing from that complex to simplicity is kind of hard to do. What's your quote on that? That was Benny's quote, I wouldn't say. That was a Benny's quote? Yeah, go ahead. What's the quote on that one? Yeah, I mean, I've looked at- Complex is complicated, but simple is genius. Right? It's, yeah. Complex is competent and simple is genius. I mean, when I looked at a lot of technologies that are out there, including software that engineers write, right? Over time, you know, even with software, you say, I've written it, you know what? Let me refactor it. Let me do it better. And in the end, it becomes simple and yet does the same thing, right? So it takes a lot of effort and iteration. And that's also what geniuses do. I mean, this 99% perspiration you know, one person is inspiration. I think that's what we want to do at Nutanix. You know, strive every day. That's a true North. We want to create a simple interface that is as powerful as a complex one would have been and yet very easy for people to use. And if you look at all throughout our stack, you know, the Prism user experience, our support user experience, and our product, you know, serviceability experience, it's all driven by that. Yeah, one of the things I like is, you guys aren't saying that Nutanix is the only solution or everything should funnel in. You go listen to Amazon, of course, Public Cloud is the only solution, the cheapest solution. It's a one-way road. Please give us all your data now, now, now. So you guys understand it's a multi-cloud, it's a multi-hypervisor world. And even more at this show, I'm kind of surprised how fast your ecosystem is building. Can you talk from an engineering perspective, how are you partnering, open APIs for everything? I mean, what's, how do people work with you? We had foreseen this quite long back. I think we have been focused on open REST APIs, you know, where the documentation is automatically generated. We have created commandlets, you know, RESTful commandlets for Microsoft ecosystems. And we've invested heavily in automating all of that. And the reason is that we don't want our partners and our customers to lag behind when it comes to keeping up with our technologies. So all these SDKs and APIs are readily available on the day when we release a new product. All right, so I'd be remiss if I didn't bring up the container discussion. Talk about, you know, you gave a demo of it. You know, where are we along that? I mean, it's still early days. What are you seeing? What are your customers asking for? And what, you know, give us a little bit about where the products are. I've been talking to a lot of our customers. I would say, you know, a fourth of them are actually experimenting with containers. There's a lot of excitement. I myself am excited about it. And the reason is it separates the, you know, cleanly separates the, you know, the boundary between developers and operations. Operations can focus on the infrastructure and the operating system. Developers can focus on application. And that's really nice. But what still needs to be done is there's a lot of maturity that needs to go in into containers in terms of networking, in terms of security, and also in terms of, you know, controlling storage and managing storage. I mean, containers don't do stateful, you know, containers right now. Absolutely. Today, you take stateless applications. There's lots of usage for there. But, you know, just like, you know, virtualization before it cloud even, you know, that we expect to see some maturity along those applications. Exactly. And we plan to lead that, you know, especially from the container, stateful container perspective. Yeah. I guess, Sunil, you know, I look along that lines. We're in a radical change of not only what's happening at the infrastructure, but the role of the hypervisor, even the role of the operating system, you know, all those are changing to deliver some of those new applications. Yeah, I mean, I think, look, I think value is going to, the true north, as Dilij would say, right, is always about caring for the apps. And to the extent that you can make things simpler and simpler to not just deploy but to consume, I think the faster you can get there. So I think hypervisor, software around the hypervisor, and beyond that, you know, if you just think back to what did Amazon do since releasing the first EC2 and S3? And if you just reverse engineer the roadmap, you'll see that there's a whole bunch of other services that are needed to be re-architected on this new fabric, right? Some will come from partners, some will come from us. And that's really what we mean by this three-phase journey of making storage invisible, making virtualization invisible, and then making cloud invisible. But we're not going to do them seriously. We have to do them in a semi-parallel fashion. Yeah, it's a good vision. So, you know, we spent a lot of time talking about Acropolis. Vinnie, you know, I know you've been involved, your team's been involved in a ton there. When people leave this show, give me a nugget or two that they should go dig into a little bit more that might get missed and, you know, some of the big discussion that you guys are really proud of. I said, besides the Acropolis stuff. Besides, I mean, we've done a lot of innovation just on the storage plane as well. I mean, you know, Sunil just announced this morning support for a NAS filer kind of concept. I mean, if you had bare metal Oracle running, we can expose now our storage into that ecosystem as well. What we're trying to do is cover all the use cases of storage that you have in your data set. Yeah, so Sunil, you know, last question I've got for you is, it's a challenge, you know, when they say Nutanix started, your competition would say, you know, it was an appliance that ran for VDI. The good thing about that is if you could really focus. So now, you know, as I said, filer, you're going after a whole scale-up file market, you know, it's multi-cloud, multi-hypervisor, expanding your ecosystem. Yeah, yeah, yeah. How do you balance the focus and the question? I think the key is to be, you know, pretty grounded and real in terms of product execution. The aspects of vision is to make sure that not just from a marketing perspective, but from an architecture perspective, that the incremental steps that we are taking are towards that long-term vision, right? So I think, you know, to Vinny's point, you know, we did invest a lot in turning our dispute file system into a storage fabric, you know, like the scale-up file server and all, but at the same time, we're also investing in making sure that the current hyper-conversion environment is scaling out. For example, as I mentioned, you know, we've gone from fixed platforms to configured order, we've given you next, you know, capabilities like this, the size of that simplifies configuration. But also, you know, the other point that we've talked about is, look, we've not stopped at Ddupe and compression. We believe that the hardware-driven storage market is going to be commoditized increasingly in the enterprise. Just like, if you look at Google's going to keep dropping storage prices, Amazon should follow, so the private cloud should follow suit, especially if we aspire to be that. So that's why we came up with this next generation erasure coding technology, which is, I think, a leap forward, frankly, from a storage app. And these are things that we've not taken lightly there. We've taken a couple of years to build. So the key thing that keeps us, you know, sort of focused is the fact that we are continuing to hire extremely talented people, and having this practical, you know, unpeeling of the onion towards the larger end gate. Great. So, unfortunately, so many things to go into, but we're going to have to leave it there. Really appreciate it. I mean, you know, congratulations, guys. You know, I wrote a recent research report, and I said, you know, first of all, it's early days of hyperconvergence, but we need to elevate that discussion. It's about systems. It's about those new applications, because most customers today are running the same virtualized workloads. And I think you guys have set forth the vision here. So, Sunil, I'll give you just kind of the final word here. You know, people walk away, you know, think about the technology. What should they think about from here? No, I think, look, I think invisible infrastructure truly is about ensuring that the IT, you know, person, and you know, in charge of whether it is the VP, whether it is the architect, whether it is the infrastructure administered, and at the end of the day, you know, they buy equipment once, they use the features maybe a couple of times, but most of their, you know, sort of like career-spent operating it. And that's really where the real pain and the real gain is from invisible infrastructure. Yeah, so, you know, Dave Vellante has said many times over the last five years, it used to be the question, does IT matter? We know that IT matters, and Nutanix is trying to help elevate ITs. So, Benny and Sunil, really appreciate you joining with us. Look forward to having more conversations with you. Thanks, dude, thank you. Thanks for joining us here from .NEXT. Be back with much more coverage over the next day and a half. I'm Stu Miniman with Wikibon. We'll be right back.