(orchestral music playing) >> Hi. I'm Peter Burris. Welcome to another Cube Conversation. Brought to you by theCUBE from our beautiful studios in Palo Alto, California. Great conversation today. We're going to be speaking with Datera about some of the new trends and how we're going to utilize data within the business, with greater success, generating more value to superior customer objectives. To do that, we've got Marc Fleischmann, who's the CEO and Founder of Datera. Marc, welcome to theCUBE. >> Thank you. >> And Guy Churchward, who's the Executive Chairman of Datera. >> Yeah, thank you Peter. >> So guys, this is a great topic, great conversation, very very timely industry. One of the reasons is we've heard a lot about the Cloud-native stack. Now the Cloud-native stack is increasingly going to reach into the enterprise and not just demand that everything come back to the cloud, but bring the cloud more to the enterprise. Well one of the things that's still something of a challenge is and how do we bring data given it's native attributes into that model more successfully. Marc, what are the issues? So look, ultimately we believe it's all about data freedom, the capability to extract the value of data across the enterprise. However, as long as we continue to think about proprietary systems silos, where data is trapped, where it can't move freely across the enterprise, we're not going to be able to get there. So ultimately what it requires is changing our thinking of infrastructure from a hard for centric prospective to a service centric prospective. Ready applications drive the needs from the data, where it's an application centric perspective that automatically drives how data is actually consumed across the enterprise. >> But the, we've been thinking about that through software defined, ECI, and other, you know, hyperconversion infrastructure in other things. But at the end of the day, we really have to make sure that we're doing so in a way that marries to the realities of data. >> Absolutely. >> Talk to us a little bit about how Datera is providing that substrate that is native to data, but also native to the cloud. >> Absolutely. So I would describe Datera as Datera is to data what Kubernetes is to compute. What do I mean by that? First of all, it's all about data orchestration. We orchestrate the data just like Kubernetes would orchestrate compute. That's the foundation of our platform. Now if we don't deliver enterprise performance, so that we can actually, you know, replace existing storage, we wouldn't be able to actually broadly deploy. So we have enterprise performance as well. And lastly, to get away from a hard for centric model, we offer wide variety, wide choice, future ready choice of Harver. Those are the three key tenants that we actually see as getting to that vision. >> So Guy, you've been in this business a long time. You've looked at a lot of changes in technology, for rays where we were mainly focused on persisting data to now some of the new technologies, we were more focusing on delivering data to new classes of applications. From your perspective, how does this message Marc's bringing line up with customer needs? >> Yeah, I know, appreciate it. I mean that was one of the reasons that when I had the opportunity to work closely with Datera, I kind of jumped into it. You know, because part of this is, as Marc said, data freedom. Unlocking, in other words, unlocking from the boundaries of basically a physical location. I think, you know, we always aspire and believe that we want to move towards a cloud, a pure cloud model. But we're going to be in this transition for five, six, seven years where we have on premise a bit of hybrid and a bit of distributed and things like Intelligent Edge. So in other words, the whole concept is to say how do I utilize data no matter where it is into a fabric or a mesh. And I think that the industry that we all live in sort of, by accident, tries to own the data, you know. It doesn't matter whether you own it in a physical construct of a data center or we own it in a physical construct of a piece of hardware or a proprietary format. But in essence you have these data silos absolutely everywhere. And so for me to move to a cloud, you've got the simplicity you need. You've got the orchestration that you actually need. But you need this freedom outside of the bounds of a physical location or a piece of tent. >> I want to return back to the issues of performance >> Yeah. >> and the need for performance because the world that you just laid out guys, makes an enormous amount of sense to me and the Wikibon community. But it does mean that this data generated by that application in this location may have value to some other applications somewhere else that may have completely different performance action. >> Absolutely. >> So let's talk about that need for ensuring, that again, this notion of a native data approach to incorporating data into the cloud. How does the performance angle really work? >> I would argue where traditional self defined storage, SDS, fell short was exactly on the promise of performance. We saw that we contributed a significant part of the Linux data path itself. The way we architected the system, we delivered true, primary application performance. So that in combination with the ability to orchestrate data across the data center, across multiple data centers, and ultimately across the data center and the cloud gives you the best of both worlds. It gives you primary workloads, the ability to actually serve primary workloads across multiple protocols, but to serve them location dependent, wherever you like, because we orchestrate the data through those places. >> And- >> So- >> Oops! Go ahead. >> Sorry. It's the coffee. It's going to kick in. (Peter laughs) So I mean part of it is not just that, but it's also the life cycle. >> Ah, very true. Right, I mean and, you know, this is the thing that kind of attracts me is, and you mentioned, you know, what you learn with the amount of hair I don't have now and the gray beard I've got is, you know, there's one thing about this sort of data boundaries and things getting locked in. The other one is the speed of which people want to build an application. They need it to be have the enterprisilities, and then they'll take the application down. You know, if you kind of think when we started in the industry and it would last 20 years. And then 10 years. And then five years. And now you look at it saying somebody wants an enterprisility application up and running within two or three months, which is preposterous, but needs to be done. And then it might be down within a month. Because- >> Oh 15 years ago it took us two or three months to create the test data required for the application to follow up. >> Right, and how many people would ever used to tell you never use an application if it's a window zero. But we're talking about, in a window zero period, they're actually going to serve their communities, the most critical thing. Data is it for a company. If you're analytics don't run as fast as your company's competitive space, you're behind. So if you're going to analyze something that application that you bring up to analyze has to be critical to your business. And that's going to go up and it's going to go down. So in other words, it's going to go from test and dev, up into production, tier zero, then tier one, tier two, tier three, and then out into an archive in a period of time that normally a window zero would gestate. And so you need a platform that has that ultimate agility and again it can't be bound by anything. And this is something that, you know, Datera has as unique. This was why I like software defined and why I believe that this market's place is now for this space. Everything prior to SDS is basically what I call new legacy. You know, it doesn't matter whether it's a ray or it's hyperconversion, and they're great and they've got their place. But each one of them has this fixed boundary that allows you to flex but inside of its own control. Businesses aren't like that. They can't be done like that and applications can't be done like that now. So it's all multi-cloud, it's all going to be versed. >> Well let's build on that. So the Kubernetes describes, as you said, a cluster of compute. When you pull away the- It's really a network of compute. >> That's right. >> It's a network of compute resources that Kubernetes has visibility into so we can move resources >> That's right. >> Or move elements where they need to be to be optimally utilized. Let's build on that. So what where is Datera in this relationship between resources as it starts to build a an orchestrator, a manager, a network of data elements, and pull that into something that makes it easier for developers to do what they need to do, operators to do what they need to do, and the business to do what it needs to do? >> Yeah, so you can call Kubernetes the network of compute or a swarm of compute, right? So the power of Kubernetes is that it abstracts the infrastructure to a level where it gets delivered continuously to the application on demand. We do exactly the same thing for data, for the ability to store, manage, and ultimately life cycle data. So simply label based, like Kubernetes is, you specify the service level objectives for every individual application, and Kubernetes pretty much does all the rest of the job, completely independent of the hardware underneath. Again, we do that for data. You have certain access requirements, protocols, authentications, security. You have certain performance requirements. You have certain reliability requirements. You articulate them simply in similar SLO, service level objectives. Datera does all the actual implementation automatically across the data center. So now you get to a point where in the modern data center and the soft defined data center, I would argue we are the data foundation in those kinds of scenarios, we can co-orchestrate data along, since you said Kubernetes specifically with Kubernetes, with its compute. Obviously we work in other environments as well. We work equally well for Enver. We work for some other, a number of other cloud orchestration frameworks. But Kubernetes is a really good example here. >> So who's going to buy it? I mean cause going back to this issue of the orchestrator, the developers clearly need this because they want access to real data, but they typically don't think in terms of underlying data structures. If it's available that's all they care about. Data administrators, business people. Who do you find your customers today are really making that, not the initial contact, but actually driving the adoption of this new data fabric? >> So Marc, I mean I know you'll answer it more accurately than I will. But just from a higher level to step down, there seems to be two types of people inside of large companies. One is a project owner. So for instance, you know, I've been blessed with a job inside of BMW that I have to do, autonomic cars. And I'm tying together a very complicated pipeline that has to be extremely agile. So that's the type of person that would basically look to buy and move us forward. And the other one is an internal service provider to the enterprise. So in other words, instead of being a group that has a physical job, what I'm actually doing is I'm saying I'm now going to be a service provider, or a cloud provider, or a resource provider to an organization that now has complexity that's moving into and embracing the digital economy or digital transformation. So if those are the two types of person inside of an organization, I think if you get a tie kicker, you know, there are places that we struggle with, I think it would be fair to say, is there's always going to be a geek somewhere that wants to kick the latest, cool technology, so we get involved with that. And then by the time you go all the way through it, there's no project there. They just really enjoyed themselves and so have we. But in essence there's enough people now who recognize my business is going through this transformation, I need to get out of my technical debt, I'm throwing business into, you know, this economy. It's normally around machine learning applications, Kubernetes, things that are fast moving, you know. And they need that level of ility that they're used to getting through fixed bounded technology, you know. And so we're actually seeing that as a service provider, both external and internal. But internal, inside the enterprises, is something which we're very key on. >> And let me give you perhaps a few examples. We're looking at Fortune 2000 companies. A good example, for instance, would be one of the top airlines in the world that is replatforming from a more rigid siloed IT to really deliver all their applications to internal and external customers as a service. It would also be digital businesses where there currency really is speed, agility, and obviously data is their currency. So if you're looking here at one of the top travel fare aggregators, that's one of the customers, actually interestingly we are in their tier zero at Storch. That's quite an endorsement of the performance aspect. We are also in one of, I would say, the leading service providers outside of the typical crowd you think, those are one of the up and coming guys. So those are typical markets and customers we're looking at. Really Fortune 2000 companies that are replatforming to cloud, hybrid cloud, and digital service businesses. Digital businesses. >> But it is most people who are basically going from, they're transforming their data center into a metadata center. They're embracing the distribution and then cloud. But they're not going wholesale and just saying (claps hands) we're over. They have this practicality of first thing I need to do is to free up my data, make my data center agile, and then decide how I want to distribute it across. >> Marc Fleischmann. Guy Churchward. Datera. Thank you very much for being on theCUBE. >> Thank you very much Peter. >> A pleasure. Thank you. >> And once again, this is Peter Burris from our CUBE studios in Palo Alto, California. Thanks very much for participating in this CUBE conversation with Datera. (orchestral music plays)

(inspirational music) >> Hello, I'm Peter Burris of Wikibon. Welcome to another Cube Conversation. We're going to have a great time talking over the next few minutes about the role that performance in the data plain is going to play at making possible both the options provided by the Cloud, but at the same time in a way that actually allows us to actually run the applications at the speed and the scale the business requires. To do that, we've got Marc Fleishman, who's the CEO and founder of Datera, and Marc Churchward who's the Executive Chairman at Datera. Welcome back to theCube, guys. >> Thank you for having us, Peter. >> So, Marc, I want to start with what I started with. That at the end of the day we've got this enormous agility that we're provided with in the CloudStack, but you still have to run on real computers that have real constraints and everybody knows that there is no greater constraint than maintaining the state of data and moving data. So how does Datera address those issues? >> Part of data freedom obviously is not only about automation, the the promise of software-defined storage is seamless automation. But unfortunately in many cases with unimpressive performance. In our case, we've engineered the whole data path down to the physical devices, ourselves. The levels of performance we can deliver is millions of IOPS across the data center at less than two hundred microseconds latency. Most importantly, on standard servers, over standard protocols, so nothing fancy in terms of hardware required. That's the true promise of software-defined storage. >> Now you mentioned automation. That kind of performance has got to open up new classes of automation potential, so that the storage or the data resources are that much easier to envision, that much easier to apply, that much easier to exploit by the development community. Tell us a little bit about how automation plays into this. >> Absolutely, once you've made data delivery frictionless, and you've made data orchestration and data automation frictionless, you do unlock new classes of applications. What we're specifically seeing is folks who traditionally run an array of databases on very dedicated proprietary hardware, and then again they get the data trapped in those silos, and they have a real hard time to extract the value off that data, we see a lot of database farms coming on our unified platform across the data center, basically being able to really extract the value of the data across a range of applications. >> Now we've been in the last few years investing pretty heavily in storage area networks and arrays and those types of resources. Flash is changing that, but it sounds as though you guys are actually making it easier to bring servers into the mix of this. What's the real direction you see? Where's this resource going to be managed by and what's the opportunity? >> Ultimately the resource should be managed by the applications, it should be driven by the applications and managed by machine learning. So as we understand the requirements of the applications, every individual application, it should be managed by machine learning in terms of the physical resources on the servers. The server capabilities you put underneath it, and then obviously start rolling the server hardware, as technology improves as well over time. >> So it's really being driven by the server, that's where the market opportunity's coming from. >> That's right, yes. >> The last question I have here is, when we think about new technology, new classes of automation, new trends in the industry, people always immediately go, "Yeah, but new companies?" Where does Datera fit in its lifecycle as it works with customers and as it delivers value out? >> If you look at the market today, server-based storage is already larger than traditional array-based storage. It's growing at five X, year by year. Since we've been on theCUBE the last time, about two years ago, we are now looking at a 240% kegger every year, so the market has clearly come our way. This is the time for this kind of product. >> So the market's good, company's good, trends are good. As we think ultimately about where this ends up in a few years, what role will Datera play within the evolved computing industry? What do you see for it? >> Given that we have the broad data orchestration, enterprise performance and choice on hardware, we really do see ourselves as the data foundation for the software-defined data center. What I mean by that, again, just in an operational model, we are to data what Cooper Ladies is to compute, across a number of operating environments. So it's a really broad data foundation for everyone who wants to deliver ITS service. >> Guy, I have a very simple question for you, very complex answer. One of the places where this seems to be especially important, where the need is especially great, is in that world of analytics. Especially as we try to close the loop between the analytical systems and the operational systems. How does Datera and analytics come together, not just in the use of analytics to make Datera better, but Datera in making analytics applications run better. >> Yeah, and as you said, an easy question, complicated answer. In reality, what companies are trying to do is to run the analytics at the speed of which they're competing in their market space. Which means that it has to get a lot faster. Today's classic environment is an ETL with a data leak, so parking stale data and analyzing it, post-event and tomorrow, in an environment where people are using AI and ML is now in stream and it's in real time. So part of that is you actually have very very fast applications, both from a performance perspective, but also how long their lifecycle is. Because people are doing AB testing on the web, they're doing analytics on the fly, and it really is a kind of a different world. It's a different pace. When I started this business, or when I was in business early and I had hair, we used to look at organizations that had applications that were lasting 10 or 20 years. Now we're looking at enterprise applications that are up and down within a period of months, if not weeks. So managing that lifecycle and not having to invest in infrastructure to support something, that age-old adage of you don't buy an application if it's in 1.0 is gone. Because by the time you're into 1.1, that opportunity's disappeared as well. So, part of what I saw in the attraction with Datera is because it's absolutely software-defined, and all the resilience handles in the software not the hardware. There's not the infrastructure burden. It has much more agility to get up. It can provide tier zero, tier one. Again, you land and expand, so in test endeavor you have the same environment. By a matter of flipping a few switches, you can have tier-one-ilities and then you can drop down in that lifecycle. It doesn't matter whether it's on premise, whether it's a distributed environment or on Cloud. It's the same infrastructure, same architecture, so, back to what Marc said, you have data freedom. >> So we're trying to tie the physical realities of data, to the virtual realities of machine resources in IT, to the Cloud realities of the new wave of applications. >> That's exactly right. >> Marc Fleishman, CEO and founder of Datera, Guy Churchward, Executive Chairman of Datera, thanks very much for being on theCUBE. >> Thanks for having us Peter. >> Thank you Peter. >> And once again, this is Peter Burris, Wikibon, thanks for watching theCUBE. (inspirational music)