 Live from San Francisco, celebrating 10 years of high-tech coverage, it's theCUBE. Covering VMworld 2019, brought to you by VMware and its ecosystem partners. Well good morning, welcome to day three of our coverage here on theCUBE of VMworld 2019, we're in Moscone Center North, here in San Francisco, kind of a, well not kind of it, it's a really cloudy day, but I kind of expect that we've been talking about clouds all week, right? Multi-hybrid public-private, you name it. We've been talking about it. John Walls of Dave Vellante, good to see you this morning. Good to see you John. Yep, we're joined now by a couple of executives from Kevin Arrio, Josh Epstein, who's a CMO and I'll DeVeed, who is the CTO of Kevin Arrio. Good morning gentlemen, good morning. Good to see you guys. Great to be here, great to be here. Yeah, first off, let's just talk about the show. I know you've got a presence down on the floor, just you're feeling about the traffic, the kind of traffic you're seeing at your booth, what the questions are coming from customers and maybe what those answers are. I don't want you to jump on that. So first of all, it's great to be back in San Francisco for this conference this year. You're here. Agreed, definitely. And I think it's very clear that definitely cloud is the name of the game and especially how do you implement a hybrid cloud? Customers are all on their cloud journey and the big question is how do I do that? How do I take these new technologies, the cloud containers and how do I take my applications and my data services to the next step? And it's kind of all over the place, all the sessions, all the customers are asking about this is where the focus is, where their interest is and it's great to be in the center of all of that. You made a big decision or big announcement about a month ago. You said, okay, public cloud, that's where we're going. Josh, the driver behind that and kind of what the early fall out. Sure, sure, no. I mean, we started our journey, really from the beginning of coming out. Coming out was about 10 years old, entered the data storage market as a traditional all flash storage array. The past 24 months, you've really pivoted that the business model towards first 100% software or got out of the appliance business, started really focusing our business on doing these large software-based implementations, moving into more subscription-based revenue, kind of delivering that cloud-based economics experience. And then over the last several months, we've been focusing on taking our core architecture, which fundamentally decouples the data services from underlying infrastructure. Think about how that might actually look on public cloud. Doing the same thing, kind of creating this sort of shared storage experience, delivering all the traditional enterprise-class data services, but sitting on public cloud infrastructure. It's been a really interesting journey. So let's double-click on that, because it's clear that this space is not about the media. It's about the business model, it's about the additional value that you can add for customers. So maybe you could add a little bit of color as to sort of how's that going, where you guys are differentiating in the marketplace, where you're winning. Sure, I mean, I think you're jumping in. Yeah, so I think it's, as you said, it's not about the media. It's all about how do you help customers have a uniform experience around any deployment model? So they want to deploy on-prem. They want to deploy in the cloud. They are actually seeking for a uniform way to do that without too much heavy lift. There's some challenges in going to the cloud. If you are not born in the cloud, you need to re-architect your applications. You need to kind of learn some new skills. And there's a big challenge, especially if you have big data-intensive applications. And that's where we focus, delivering that uniform experience around orchestration of resources and data services across your on-prem, off-prem, and public cloud implementation. So you guys decided not to ship a box anymore, you know, the Silicon Valley show, right? Where's the box? And so I'm interested in the technical challenges of doing that, but also the customer feedback. Because sometimes people want an appliance. So how are you able to transition through that and what's the feedback been? Yeah, I think, you know, for us, I mean, our core business, our core customer has really been cloud-scale applications for the last five years, right? So this is large SaaS providers, e-commerce platforms, FinTech, HealthTech, kind of these large, mature software companies, right? That their core business is delivering a cloud-scale application. And for them, you know, many of them were born kind of before the age of the public cloud. So they've actually heavily invested in application architectures that rely on enterprise-class shared storage. That said, they see the draw towards the cloud. They see the benefit of the cloud-like economics, subscription-based, consumption-based economics, and then the overall capability to scale up and scale out like the cloud does. But that said, they need that bridge from where they are today with traditional data-centric architectures to this cloud world. You mentioned FinTech, and there's an interesting case because when the cloud first really started to gain momentum, a lot of the financial services companies, the big guys, especially said, you know, we can build our own clouds. And then they realized, well, we can't build them as fast as Amazon can build them. And so they sort of pulled back in that, but then they sort of put their foot in the cloud and then went all, and then they said, wait a minute. So what are you seeing in terms of like the, call it the private cloud. You know, we've kind of swung back to that. Is that gap closing? Are they able to get close enough? The key part of that is obviously the pricing models and the pay-by-the-drink. I wonder if you could add some color to the sort of on-prem cloud business if we can call it that. Some people might object, but that's... Yeah, definitely. So the way we approach it is that we want to bring the simplicity, the agility, and the flexibility of the cloud model to this on-prem data center to deliver the same performance control of a dedicated resource, which is exactly what these type of FinTech customers are looking for. So in our basic architecture, which was already decoupled from hardware and already decoupled performance from capacity, we're able to do that in extremely flexibility. You can get the same flexibility of the cloud in an on-prem solution with all the benefits. And you can also decide on your own pace in your own terms what you actually need and make sense to run on a public cloud infrastructure. So scale is obviously a big deal for your customers. That's kind of been your focus since day one. What's the bell curve look like? Are we talking about scale in terms of just the ability to scale quickly, or is it also just the sheer size? And what does it look like? I think it's about performance at scale. It's about control or performance at scale. It's about control over availability at scale. And it's obviously about cost at scale, right? I mean, there's so many different ways to look at that the economics of public cloud versus on-prem. If you look at the pure dollars, clearly building on your own dedicated on-prem infrastructure is cheaper than paying Amazon or Google or whoever to do it. But there's clear benefits to kind of going in that direction in terms of agility, in terms of hands-off management, in terms of really just staffing expertise. But I think it does come down to control, right? And when you talk about scale, when you talk about petabyte scale, it's easy to lose control. And this is the benefit of shared storage models and this is where we think there's a real opportunity. Can I follow up on that? Because you said there's a clear benefit of, if I understood it correctly, of building out your on-prem infrastructure at some critical mass. There's obviously people that, like Andy Jassy would disagree with that. So what's your data showing? I presume it's weighted toward large customers. Absolutely. Maybe you could add some color to that. Turn it back up. Yeah, we've done some good research into good analysis on this. I think if you're talking about, we're talking about certainly over 500 terabytes to a petabyte to multi-petabyte scale, data-driven applications. We're talking about business critical applications, big block storage, heavy analytics. If you compare just raw economics, right? And the thing is there's a lot more than just the raw economics, but the raw economics of an infrastructure built on common RIO versus the equivalent infrastructure built on one of the block storage resources from one of the public clouds. It's about literally about one third the cost to build out that your own dedicated infrastructure, leveraging a good high-quality colo, good high-quality hardware underneath it. So raw economics, it's clear where that sits. Okay, so we're comparing the acquisition cost versus the sort of some N number of years, right? That's correct. Yeah, exactly. And not really going into the labor cost at that? Not going into the direct labor cost of managing the storage, yes. There's clearly interesting benefits to going to 100% cloud model. What that does to an organization when you kind of hands off, you don't have the same kind of in-house IT resources, you're outsourcing a lot of that. Well, except what Al was saying before is you're trying to bring that cloud model to the data. So to the extent that you can close that gap, then you can substantially mimic. We saw the opportunity to extend those capabilities into the public cloud. Delivering a high-performance storage solution in the cloud today is expensive. We focus over the years of taking these commodity components and comprising them into a high-performance shared storage solution. We can do the same in the cloud. We take it. But I think the key is multi-cloud, right? The key is that there's not one-size-fits-all and it really is about creating this mobility between your on-prem data and public cloud number one and then public cloud number two. One of the key concerns about moving a business-critical application to a public cloud is lock-in, right? And if you can create this infrastructure, you're decoupling that data-services stack that the application relies on from the underlying infrastructure. You get this mobility between clouds that becomes really attractive. So you're kind of answering that question. It was on my mind. I said, how are you selling that to customers? I mean, the fact that we're having this very robust discussion about this fundamental shift and you get it because you're kind of providing this service to your whole client base. But if I'm a client, am I head-starting to spin a little bit, right? And I've got big decisions to make. So how do you sell that that this is not a little shift. This is a fundamental way in the way you're going to do your business. So in the simplest form, we tell the customers that we significantly lower the barrier of entry into the cloud. You don't need to re-architect everything. You don't need to be worried about performance management or control or orchestrating resources. We do all that for you and we do that in the same way that we did it for you on your own on-prem data center. We can do it on any of the public cloud. So the barrier of entry, the risk of actually doing that transition. It's lowered significantly. And you can do that on your own base in your own terms and make some smart decisions later on about what needs to reside over time. So when we think about multi-cloud, we think about, okay, I'm going to have data on-prem. I might choose Azure for my collaborative workloads. I might put my dev stuff in AWS. I might put some analytics in Google, you know, whatever. My business is going to decide what to do. I'm not going to have this grand multi-cloud strategy. It's just kind of going to happen. And then I'm going to be, then IT is going to be called in to clean up the crime scene. But so what we're envisioning this architecture that's shipping metadata and maybe compute to the data versus moving data, do you agree with that or do you see it differently? We see, I think, two types of customers. Some behave just as you described but some have a very specific decision not to be locked into a single vendor. So they'll say, I'll put one business unit on Google Cloud and I'll put the other business unit on Azure. I'll put this certain type of applications on one cloud and the other type on the other cloud because I want to make sure that I am cloud agnostic. I'm actually mandating within an organization that I can run anywhere. As a hedge. As a hedge, as a definite hedge because they're concerned about locking to either of the vendors. And in that sense, they later on make the decision, okay, where is the kind of, where is the core of the data? Where is my mission critical data which always has some gravity and how do I make sure that it's in the right place at the right time? Doesn't that add complexity for the client? I mean, if they've got a workload here and here and here, it'd be a lot easier if it were all here or most of it were here. But that adds, I mean, I just said, I'm wondering if that- You're absolutely right. But what we see is this rapid shift towards this embracing the multi-cloud model. So let's just take an example. You know, you have a classic Cloud Scout application that might have an active act of data centers in two parts of the United States. Sort of serving up the production application. You have DevTest requirements, right? So they want the ability to rapidly spin up an environment to mimic a problem or reduce the development. Public Cloud's a great example for that. You have DR requirements. You have backup requirements. They want to be backing up. They want the ability to rapidly spin up an instance in a public cloud instance. And no matter what, in every organization somewhere, even in the most sophisticated IT organizations where they have tremendous control over the data centers, there are some C-level execs somewhere that says like, in five years, I'm 100% in public cloud. I want no, I want nothing, right? So you have to sort of service that as well. And what we're doing is it's saying, listen, you can continue to focus on building out a world-class next-generation data center, but at least on the NVMe, all NVMe fabric, and still have the ability to do certain things in the cloud and still have this path, if it makes sense for your organization, to migrate the entire thing to public cloud and not get locked in, to be able to sort of surf the clouds as it actually- So technically that means you have to speak Azure, API, S3, whatever language of the cloud. And so I'm trying to understand sort of, technically what you have to do and then where you add value, where you pick up from, whatever, VMware or whomever else is trying to be the control plane. So that is exactly the point, and to address the question about what the complexity of this multi-cloud world, this is exactly where we see the rise of this next-generation orchestration framework, either from VMware or from others, that strive to give you this uniform experience. So we deliver that at the data services layer. We connect that to the orchestration layer that allows you to do seamless workload mobility, seamless data mobility, to wherever it makes sense for that, those applications or business workloads to run. And basically the customers expect today that we encapsulate all that complexity for them. They want to be able to put their Google Amazon or Azure credentials and then forget exactly where it runs. And this is a lot of what's going on in the floor, right, this week. And that's exactly where we connect to the rest of that orchestration scene within the data center or the public cloud. So in that context, you primarily, I know you sell to a lot of different people, but is it the cloud architect or the architect that's actually determining that throughout the organization or is it, again, cleaning up the crime scene kind of thing? It's usually a customization with that CIO who's kind of on that cloud journey, building his cloud strategy. And even if he made it, the decision was made to in five years beyond the cloud. And as the question is, okay, what's happening in the meantime? How do I actually do that? One of the core things that's happening in the meantime is most of our customers is in this perpetual state of data center consolidation, right? Most of these large task companies, they're growing through acquisition, they've got nine DAF data centers, they all have a plan within two or three years to be consolidated on three next generation data centers and then have cloud mobility. And so what we're able to do, this is leveraging our software model as well. I say, listen, let's do an enterprise-wide kind of unified licensing scheme where you're paying on consumption based on actual data stored. And then you can build the underlying infrastructure wherever you want, right? You can base it on your traditional infrastructure you might already own. It might be on next generation NVMe, NVMe over fabric connected data centers. And then a piece of it now might be in the public cloud. So you're talking CIO, Dave, you're talking CSI. I'm just a little confused. Gentlemen, thanks for the time. We appreciate it. Great discussion and continued success downstairs and on down the road. Great to be here guys. Thank you. We are VMworld 2019 here on theCUBE.