 from our studios in the heart of Silicon Valley, Palo Alto, California, this is a CUBE Conversation. Hello everyone, welcome to this CUBE Conversation here in Palo Alto, California. At the CUBE Studios, I'm John Furrier, host of theCUBE. We're here for a company profile, company called Diamante. We're here at Tom Barton, the CEO. As VMworld approaches, a lot of stuff is going to be talked about. Kubernetes, applications, microservices will be the top conversation, certainly in the underlying infrastructure to power that. Tom Barton is the CEO of Diamante, which is in that business. Tom, we've known each other for a few years. You've done a lot of great, successful ventures. Diamante's new one you got on your plate here right now. Yes sir, and I'm happy to be here. So I've been with Diamante just for about a year or so. I found out about the company through a headhunter, and I have to admit, I had not heard of the company before, but I was a huge believer in containers and Kubernetes. So I was already sold on that, and so I had a friend of mine. His name is Brian Walden. He had done some massive Kubernetes cloud-based deployments for us at Planet Labs, a company that I was at for a little over three years. So I had him do technical due diligence. Brian was also the number three guy at CoreOS, and so deeply steeped in all of the core technologies around Kubernetes, including things like EtsyD and other elements of the technology. So he looked at it, came back, and gave me two thumbs up. He liked it so much that I then hired him. So he is now our VP of Product Management. And the cool thing about Diamante is essentially we're a purpose-built solution for running container-based workloads in Kubernetes on-premises and then hooking that in with the cloud. So we believe it's very much going to be a hybrid cloud world where for the major corporations that we serve, Fortune 500 companies like banks, like Energy and Utilities and so forth, a lot of their workload will maintain and be maintained on-premises. They still want to be cloud compatible, so they need a purpose-built platform to sort of manage both environments. Yeah, we certainly, you guys have come up on our radar, but I was really curious to see when you came in and took over at the helm of the CEO because your entrepreneurial career really has been unique. You're a unique executive, both entrepreneurial and as an operator. You have an open source and software background, and also you had very successful companies and exits there. As well as on the hardware side with Rackable, you took that company and went public. So you got, I mean it's unique, and then open source, right? You got software open source and large hardware, large data center deployments at scale, which is essentially the hybrid cloud market right now. So you kind of got a unique perspective. You've seen the view from all the different sides. And I think now more than ever with public cloud certainly being validated, everyone knows Amazon, if you're green field, you start in the cloud, but the reality is hybrid cloud is the operating model of this next generation of companies, probably for the next 20 to 30 years. And this is the biggest conversation, the most important story in tech. You're in the middle of it with a hot startup with a name that probably no one's ever heard of. Right, right, we hope to change that, obviously. Why did you join this company? What got your attention? What was the key thing? Once you dug in there, what was the secret sauce? What got your attention? Yeah, so to me again, the market environment, I'm a huge believer that if you look at the history of the last 15 years, we went from an environment that was 0% virtualized to 95% virtualized with VM based technologies from VMware and others. I think that fundamentally containers and Kubernetes are equally as important. They're going to be equally as transformative going forward and how people manage their workloads both on-premises and in the clouds, right? And the fact that all three public cloud providers have anointed Kubernetes as the way of the future in the Docker image format and runtime as the wave of the future means good things are going to happen there. What I thought was unique about the company was for the first time, surprisingly, none of the X86 vendors in companies like Nutanix that have hyper-converged solutions, they really didn't have anything that was purpose-built for native container support. And so the founders all came from Cisco UCS. They had a lot of familiarity with the underpinnings of hyper-converged architectures in the X86 server landscape and networking subsystems and storage subsystems, but they wanted to build it using the latest technologies, things like NVMeBase Flash, and they wanted to do it with a software stack that was native containers in Kubernetes. And today we support two flavors of that one that's fully open source around upstream Kubernetes and another that supports our partner Red Hat with OpenShift. I think you're really onto something pretty big here because one of the things that Dave Vellante and my and Stu Miniman and our team have been looking at is, we're calling it Cloud 2.0 for the lack of a better word, kind of riff on the web 2.0 concept. But Cloud 1.0 was Amazon. Okay, DevOps, Agile, great, check the box. They move on with life. It's a great resource, it's never going to stop. But Cloud 2.0 is about networking, it's about security, it's about data. And if you look at all the innovation startups, they all have one characteristic. They're all playing in this hyper-converged hardware-meet software stack with data and agility, kind of to make the original DevOps model go better. The one data, which was storage and compute, which were virtualization playing. So you're seeing that pattern and it's wide-ranging. It's security, it's data, everything else. So that's kind of what we call the Cloud 2.0 game. So if you look at VMworld, you look at what's going on, the conversations around microservices, it's an application-centric conversation in an infrastructure show. So do you see that same vision? And if so, how do you guys see you enabling the customer up there saying, hey, you know what, I have all this legacy, I got full-scale data centers, I need to go full-scale Cloud, and I need zero disruption to my developer community? Yeah, so this is the beauty of containers in Kubernetes, which is they know it'll run on the premises, they know it'll run in the Cloud, right? And it is all about microservices. So whether they're trying to adopt the modern database, something like MongoDB or MariaDB or Crunchy Postgres, whether it's on the operational side to enable sort of more frequent and incremental change, or whether it's on the developer side to take advantage of new ways of developing and delivering apps with CI CD tools and so forth, it's pretty much what people want to do, because it's future-proofing your software development effort, right? So there's sort of two streams of demand. One is refactoring legacy applications that are insufficiently kind of granularized and behave and fail in a monolithic way, as well as trying to adopt modern, cloud-based native solutions for things like databases, right? And so the good news is that customers don't have to refactor everything. There are logical breakpoints in their application stack where they can say, okay, maybe I don't have the time and energy and resources to totally refactor a legacy consumer banking application, but at least I can refactor the database here and serve up container in Kubernetes-based services, has microservices, database as a service to be consumed by the application. They don't need to kill the old to bring in the new. They can use containers in an orchestration layer like Kubernetes and still be positioned for whether it's service meshes or other things coming down for that piece of the infrastructure and everything else could run as is. Right, and there are multiple deployment scenarios for containers. You can run containers bare metal. Most of our customers choose to do that. You can also run containers on top of virtual machines and you can actually run virtual machines on top of containers. So one of our major media customers actually runs Splunk on top of KBM on top of containers. So there's a lot of different deployment scenarios and really a lot of the genius of our architecture was to make it easy for people that are coming from traditional virtualized environments to kind of remap system resources from a VM to a container at a native level or through a VM. You mentioned the history less than there around virtualization. How 15 years ago there was no virtualization now virtually everything's virtualized. We agree with you that containers are going to change that game for the next 15 years. But what's interesting about VM where what made them successful was they can add virtualization without requiring code modification. And they did it kind of under the covers. And that's a concern customers have. I have developers out there, they're building stacks, they're building code. I got a preexisting legacy. They don't really want to change their code. Do you guys fit into that narrative? We do, right? So every customer makes their own choice about something like that at the end of the day. I mentioned Splunk. So at the time that we supported this media customer on Splunk, Splunk had not yet provided a container-based version for their application. Now they do have that, but at the time they supported KVM but not native containers. And so unmodified Splunk, unmodified application, we took them from a batch job that ran for 23 hours down to one hour based on accelerating and on our hyper-converged appliance and running unmodified code on unmodified KVM on our gear, right? So some customers will choose to do that. But there are also other customers, particularly at scale for transactionally intensive applications like databases and messaging and analytics where they say, you know, we could preserve our legacy virtualized infrastructure, but let's try it as a bare-metal container approach and they discovered that there's actually some savings from both a business standpoint and a technology tax standpoint or an overhead standpoint. And so as I mentioned, most of our customers actually choose to do bare-metal. So there's real efficiency there. It's in the batch as a great example. All right, so let's dig into the product and technology differentiators. What's the big secret sauce? Describe the product. Why are you winning in accounts? What's the lift in your business right now? You guys are getting some traction from what I'm hearing. Yeah, sure. So like at the highest level of the value proposition is simplicity. There is no other purpose built, you know, complete hardware software stack that delivers Kubernetes, you know, production Kubernetes environment up and running in 15 minutes, right? The X86 server guys don't really have it. Nutanix doesn't really have it. The software companies that are active in the space don't really have it. So everything that you need, the hardware platform, the storage infrastructure, the actual distribution of the operating system, CentOS, for example, we distribute. We actually distribute a Kubernetes distribution upstream and unmodified. And then very importantly, in the Kubernetes landscape, you have to have a storage subsystem and a networking subsystem using something called CSI container storage interface and CNI container networking interface. So we've got that full stack solution. No one else has that. Second thing is the performance. So we do a certain amount of hardware offload. And I would cite Amazon's purchase of Anteperna. So Amazon bought a company called Anteperna that's basis of their nitro technology. And it's little known, but the reality is more than 50% of all new instances at EC2 are hardware assisted with the technology that they bought. So that's what you mean by hardware offloaded? Yeah, exactly. So we actually offload storage and network processing via two PCIe cars that can go into any industry server, right? So today we ship on an Intel white box. They're hyper-converged containers. We're hyper-converged containers. Yeah, sort of, exactly. So you're selling a box? We sell a box. With software, that's the... With software, but increasingly our customers are asking us to unbundle it. So not dissimilar from the sort of journey that Nutanix went through. If a customer wants to buy on Dell, we'll support Dell. If a customer wants to buy on Lenovo, we'll support Lenovo and we'll just sell on top of it. So have you unbundled yet? Are you unbundling? We are actively taking orders for unbundling at the present time. And this quarter we have validated Dell on Lenovo as alternate platforms to the Intel white box. And you have subscription revenue on that? We do not yet, but that's the goal that's quarter right. That's Nutanix struggled with. So, and they had to take their medicine. They did, but they had to do that as a public company. We're still a private company, so we can do that outside the limelight of the public markets. So I'm expecting that you guys are going to get pretty much, I won't say picked off, but certainly I think your doors are going to be knocked on by the big guys. Certainly Dell at Dell EMC, for instance, I think he's starting, you said you're doing business with Dell EMC? We are doing as a channel partner and as an OEM partner with them at the present time. There, I wouldn't call them a customer. How do you look at VMware? Obviously they're in the VMware business and Pat Gelsinger's on the record. He'll be on theCUBE. He said, you know, Kubernetes is the dial tone of the internet. They're investing, they're doubling down on it. They bought Heptio for half a billion dollars. They're big and cloud native. We expect to see at VMworld tons of cloud native conversations. Good bad for you. What's the tailwind? I think it's good. It legitimizes what we're doing, right? And so obviously VMware is a large and successful company. That kind of legacy and presence in the data center isn't going to go anywhere overnight. There's a huge set of tooling and infrastructure that VMware has developed and offers to their customers. But that said, I think they've recognized in their acquisition of Heptio is indicative of the fact that they know that the world's moving this way. I think that at the end of the day, it's going to be up to the customer, right? The customer is going to say, do I want to run containers inside of VM? Do I want to run a bare metal? But importantly, I think because of, you know, the impact of the cloud providers in particular, if you think of the lingua franca of cloud native, it's going to be around Docker image format. It's going to be around Kubernetes. It's not necessarily going to be around VMDK and VMX and ESX, right? So these are all very good technologies, but I think increasingly, you know, the open standard and open source community is going to bring those. I mean, we see people putting Kubernetes on switches directly is no need to have anything else there. So I got to ask you on the customer equation. You mentioned you got some of your taking orders. How are you guys doing business today? Where are you guys winning? Give an example of why you're winning. And then for anyone watching, how would they know if they should be a customer of yours? What's, is there like, is there any smoke signs and signals inside the enterprise? You mentioned batch to one hour. I mean, that's just music. There's a lot of financial services years, for instance. You know, they have timetables and whether they're pulling backups back or doing all the other things, timing's critical. What's the profile customer? Why would someone call you? What's the situation? The profile is heavy duty production requirements to run in both the developer context and an operating context container and Kubernetes based workloads on premises that are compatible with the cloud, right? So increasingly our control plane makes it easy to manage workloads, not just on premises, but also back and forth to the public cloud. So I would argue that essentially all Fortune 500 companies, global 1000 companies are all wrestling with what's the right way to implement industry standard x86 based hardware on site that supports containers and Kubernetes and is cloud compatible, right? So that is the number one question that essentially. So I can buy a box and or software, put it on my data center, and then have that operate with Amazon, Azure or Google. Which is the beauty of the Kubernetes standards, right? As long as you are Kubernetes certified, which we are, you can develop and run any workload on our gear, on the cloud, on anyone else that's Kubernetes certified, et cetera. So you know that there's no. Give an example of a workload that would be indicative of the nine. Well, I'll cite one customer, right? So the reason that I feel confident actually saying the name is that they actually sort of went public with us at the race and Gartner conference a week or so ago. And the customer is Duke Energy. So very typical trajectory of journey for a customer like this, which is a couple of years ago, they decided that they wanted to refactor some legacy applications to make them more resilient to things like hurricanes and weather events and spikes in demand that are associated with that. And so they said, what's the right thing to do? And immediately they picked containers in Kubernetes and then they went out and they looked at five different vendors. And we were the only vendor that got their POC up and running in the required timeframe and hit all five use case scenarios that they wanted to do, right? So they ended up refactoring core applications for how they manage power outages using containers in Kubernetes. So a real production workload. Real production. We had the develop and stand up. Absolutely. In a sandbox, push it into production and be working. Absolutely. So you sounds like you guys are positioned to handle any workload. We can handle any workload, but I would say that where we shine is things that are transactionally intensive because we have the hardware assist and the IO offload for the storage and the networking. You know, the most demanding applications, things like databases, things like analytics, things like messaging Kafka and so forth are where we're really going to show you. So large flow of data. Yeah, absolutely. Transactional data. Yeah. We have customers that are doing simpler things like CICD, which at the end of the day involves compiling things, right, and managing code bases. So we certainly have customers in less performance intensive applications, but where nobody can really touch us, and what I mean is literally sort of 10 to 30 times faster than something that Nutanix could do, for example, is in those. So you're saying you're 30 times faster than Nutanix. Absolutely. In transactionally intensive applications. Just when you sell a subscription, not to dig into the small little bit, but does the customer get the hardware assist on that as well? To date, we've always bundled everything together so the customers have automatically gotten the hardware assist. On the hardware box. Yes. And that'll involve where I got to load it on a machine. That's right. But does that machine give me the hardware assist? It will not unless you have our two PCI cards, right? And so this is how, you know, we're just in the very early stages of negotiating with companies like Dell to make it easy for them to integrate our two PCI cards into their server platform. So the preferred flagship is the device. It's the device. If they want the hardware assist, if they still need the software, it's maybe not that intensive. That's right. If they don't need to have 30 times faster than Nutanix, they can just get the software. Right, right. And that'll involve our CSI plug-in, our CSI plug-in, our OS distribution, our Kubernetes distribution and the control plane that manages Kubernetes clusters. Tom, it's been great to get the feature on New Company. Give a quick plug for the company. What's your objectives? What are you trying to do? Obviously probably hiring, get some financing, any news? Any kind of things you want to share? Yeah, we will be announcing some news about financing. I'm not prepared to announce that today, but we're in very good shape with respect to being funded for our growth. And consequently, so we're now in growth mode. So today, we're 55 people. I want to double that over the course of the next four quarters and increasingly just sort of build out our sales force, right? We didn't have a big enough sales force in North America. We've got to establish a beachhead in Amia. We do have one large commercial banking customer in Europe right now. We also have a large automotive manufacturer in APAC, but the total sales and marketing reach has been too low. And so a huge focus of what I'm doing now is building out our go-to-market model and sort of 10xing the amount of marketing. So standing up a lot of field, going to market. How about on the BizDev side? I'm going to imagine that, you mentioned Dell, I can imagine that there's a large appetite for the hardware offload option. Absolutely, so BizDev boils down to striking partnerships with the cloud providers really on two fronts, both with respect to hardware offload and assist, but also supporting their on-premises strategies. So Google, for example, has announced Anthos. This is their approach to supporting on-premises Kubernetes workloads and how they interact with Google Cloud, right? As you can imagine, Microsoft and Amazon also have on-premises aspirations and strategies and we want to support those as well. This goes well beyond something like Amazon Outpost, which is really a narrow use case and point solution for certain markets. So cloud provider partnerships are very important. X86 server vendor partnerships are very important and then major ISVs. So we've announced some things with Red Hat. We were at the Red Hat Open Summit in Boston a few months ago and announced our OpenShift project and product that is now GA. Also working with ISVs like MariaDB, MongoDB, Splunk and others too. So solid tech team, product team. Great. You guys are solid. You feel good on the product. I feel very good about the product. What about the skeptics that are out there just to put the hard question to you is, man, it's crowded field. How are you going to compete? What are your chances? How do you like your chances known that it's a very crowded field? You're going to rely on your fast balls, what do you say on the speed? What's your thinking on? Well, it's unique. And so part of the way or a proof point that I would cite there is the channel, right? So when you go to the channel and channel is afraid that you're going to piss off Dell or EMC or NetApp or Nutanix or somebody, you know, then they're not going to promote you. But our channel partners are promoting us and talking about companies like Lifeboat at the distribution level, talking about companies like CDW, SHI, you know, WWT, these major North American distributors and resellers have basically said, look, we have to put you in our line car because you're unique. There is no other purpose built. And why that life? They get more services around that. They wrap services around it. They want to sell the hardware. They want to wrap services around it. Absolutely. And they want to do migrations from legacy environments towards microservices, et cetera. Great to have you on, share the company update. Just want to get personal, if you don't mind, personal perspective. You know, you've been on the hardware side. You've seen the large scale data centers from Rackable and that experience. You've also been on the software side, open source. What's your take on the industry right now? Because you're seeing, I talked to a lot of CISOs around the security space and, you know, they all say, oh, multi-clouds, a bunch of BS because I'm not going to split my development team between four clouds. I need to have my people building software stacks for my APIs. And then I go to the vendors and say, support my APIs or you can't be a supplier. Now that's on the CISO side, but the mega trend is, there's software stacks being built inside the premises of the enterprises. That not, I mean, they had developers before building, you know, cobalt apps in the old days, mainframes to client server apps. But now you're seeing a renaissance of developers building a stack for the domain specific applications that they need. And that requires that they have to run on an on-premise, hyperscale like environment. What's your take on this? My take is it's absolutely right. There is more software based innovation going on. So customers are deciding to write their own software in areas where they can differentiate, right? They're not going to do it in areas that they can get commodity solutions from a SaaS standpoint or from other kinds of on-prem standpoint, but increasingly they are doing software development, but they are all 99% of the time now, they are choosing Docker and containers and Kubernetes as the way in which they're going to do that because it will run either on-prem or in the cloud. I do think that multi-cloud management or multi-cloud is not a reality. Our primary modality that we see our customers choose is tons of on-premises resources that's going to continue for the foreseeable future. Then they pick one preferred cloud provider because it's simply too difficult to do more than one, but at the same time they want an environment that will not allow themselves to be locked into that cloud vendor, right? So they want to potentially experiment with a second public cloud provider or just make sure that they adhere to standards like Kubernetes that are universally shared so that they can't be held hostage, but in practice people don't need multiple clouds. Or if they do have multi-cloud, it might be applications like if you're running Office 365, that's Microsoft. It could be, yes, exactly, on one particular point. So domain-specific cloud, but not our core cloud. Have a backup, use Kubernetes as the bridge. Do you see that, do you see that? I mean, I would agree with, by the way, we agree with you on that. But the question that we always ask is, we think Kubernetes is going to be that interoperability layer the way TCP-IP was within IP networks where you had this interoperability model. We think that there will be a future state at some point where I could connect to Google and use that Microsoft and use Amazon together, but not to do it. And so nobody's really doing that today, but I believe and we believe that there's a future world where a vendor neutral, vendor neutral with respect to public cloud providers can offer a hybrid cloud control plane that manages and brokers workloads for both production as well as data protection and disaster recovery across any arbitrary cloud vendor that you want to use. And so it's got to be an independent third party. So you're never going to trust Amazon to broker a workload to Google. You're never going to trust Google to broker a workload to Microsoft. So it's not going to be one of the big three. And if you look at who could it be, it could be VMware, Pivotal, now Cisco's got an interesting opportunity. Cisco's got an interesting opportunity, Red Hat's got an interesting opportunity, but there is actually, it's less than, the number of companies can be counted on one hand that have the technical capability to develop a hybrid cloud abstraction that spans both on-premises and all three And it's super early. If you had to peg the inning on this one. First inning, obviously. First inning. Yes, really early. Yeah, we like our odds though, because the disruption, the fundamental disruption here is containers and Kubernetes and the interest that they're generating and the desire on the part of customers to go to microservices. So a ton of application refactoring and a ton of cloud native application development is going on. And so, you know, with that kind of disruption you can go to the company. So you're talking application refactoring that needs to run on a cloud operating model. On-premise and public. That's correct. In a sense, DMATI really brings the cloud to the on-premises environment, right? So for example, we are the only company that has the concept of on-premises availability zones. We have synchronous replication where you can have multiple clusters that are synchronously replicated. So if one fails, the other one, you have no service disruption or loss of data even for a state full application, right? So it's cloud-like services that we're bringing on-prem and then providing the links, you know, for both DR and DP and production workloads to the public cloud. A lot to unpack with you guys at DMATI. We want to keep track of you made a stateful date. It's a whole nother topic as, you know, stateless data is easy to manage with APIs and services when it gets state. That's when it gets interesting. Right. Tom Barton, the CEO, the new chief executive officer at DMATI. How long has you guys been around before you took over? About five years. Four years before me. I've been on board about a year. Great job. Looking forward to tracking your progress. We'll see you next week in San Francisco at VMworld. I'm Barton, CEO of DMATI here inside theCUBE. Hot startup. I'm John Furrier. Thanks for watching.