(techy music) >> Hi, everybody, this is Dave Vellante from theCUBE's Palo Alto studios, and welcome to this CUBE conversation. You know, theCUBE and SiliconANGLE/Wikibon have been documenting the evolution of data and storage over the last decade or so, and what we've seen is the simplification of storage. Going from hardware consolidation with conversion infrastructure and we saw hyper conversion infrastructure and sort of software-defined come on the stage, but now we're, you know, in the heart of the cloud era, and what we're seeing emerging is true cloud-like models for data services. So, we've asked Brian Biles and Sazzala Reddy from Datrium to come back into our CUBE studios and talk about this a little bit. Brian and Sazzala are both co-founders of Datrium. Brian is the chief product officer and Sazzala's the CTO. Gents, let's get into it, thanks for coming back on, and let's talk about that a little bit. So, your model, as we've talked about in the past, is a pure SAS model. You're accessing data services in a SAS-like, cloud-like experience, and people might say, "Well, isn't everything SAS today?" But in the storage world that's not the norm. Typically you would either install a box, you know, and that box might have a very rich set of software-defined services on top of it, but it's not really a cloud experience. We're starting to see certain models pop up. You're seeing some companies actually delivering that. You guys started there, that's your DNA, so let's talk about what you're doing and how that's different in the marketplace, Brian. >> Sure, you know, the way to maybe start the conversation is imagine that you're already, you know, embracing a multi-cloud, you know, plan in your IT organization, so you know, you might have a little Amazon, you might have some SAS. You know, software company, stuff going on, and you have some on-prem experience, and you want to make that as simple as possible-- >> You just described everybody. >> We want to make it as unified... (laughs) Yeah, as unified as you can, as simple as you can. You know, at that point you want to think about, you know, what is the highest leverage, simple thing to do on-prem that connects to that, you know, world of services in the cloud. How can you align that as closely as possible, so what Datrium is doing is trying to do that. We have, you know, our on-prem software is very Amazon-like, it has two layers, it operates in a very similar way, supporting many types of frameworks from VMware to Docker to Red Hat. What we've done with how to store, manipulate, mobilize data and orchestrate, you know, transitions between clouds is, it ends up feeling fundamentally different from other types of ways that you can deal with on-prem infrastructure. It's just much simpler, much more coordinated, and it allows more flexibility over time. So, Sazzal can maybe tell you about it. >> Okay, so but Nutanix, Sazzala, would say, "Okay, well we're cloud, we're creating "a cloud-like experience for on-prem," how are you guys different? >> I guess the fundamental difference is how we think of the problem. We want to say our goal is to run, protect applications in any cloud, because we cannot be in the business of building the infrastructure because that's an investment. There already are three players. What most customers want is to commoditize the cloud. They could care less if they're running on Amazon or Azure. In fact, they care that they're tied to one cloud vendor, so our goal is to make that cloud, commoditize the cloud, make it all seamless so they can move from one place to the other, whatever agreements you have. Tomorrow Google may give you credits, say, "You know what, I'll give you one year free, come on over." What you want is a one-click and move everything over to their stuff over the weekend. That's kind of where we are, that we want to provide that level of simplicity, run, protect your workloads in any cloud you want. So, Nutanix is, I think it looks like from at least from what we read from their press releases, is their cloud. They say one OS, one cloud, we are seeing any cloud. So, that's, I think we want to give that flexibility for people to not be locked in by any cloud vendor, that you can take advantage of it. You know, tomorrow Amazon may not be doing that well or tomorrow Amazon may be enemy to your business, so you want to click it and move it away to some other cloud-- >> Separating the data services from the underlying infrastructure. >> Yeah, that's right, so I think you have to separate the data services and the data management to the best, abstract it so it's so high level that then you don't care where it runs. It runs on-prem, it runs on Amazon, so it looks the same experience for you, that's what we're aiming for, that level of simplicity, but remember, to do these things you must run and protect. You can't just do run only, you must also protect because it's part of your data, you know, your IT philosophy that you must protect your data, you must have copies of it to guard against ransomware and other things, but compliance reasons, right? You want to manage your data, so it has to be a holistic view of the entire end-to-end lifecycle of your data. It cannot just be, "Run my apps here and there." >> How about Pure, how would you differentiate from Pure? Let's say pure wants to, say, OEM its stack to a cloud service provider, how is it different? >> So, the common denominator in cloud services is the workload, the instance, the VM. >> Mm-hmm. >> All of the coordination between clouds is going to be on that granularity. That's what we focus on, so, you know, we have a catalog to show relationships between VMs so that, you know, when we DR you can restart in a certain order or you can validate, you know, workload granularities, have policies at a workload granularity. That's how clouds', you know, behavior is sort of itemized today. If you buy separate parts, like you know, a SAN array, you have to buy something else to do that work. So, it's fundamentally limiting. You know, if you just take VMware because it's so well understood, you know, VMs are going to be put into a LUN as, you know, a file system of VMs. So, to transfer a LUN to a cloud, and then what do you do with it? You know, are... There's no instance to restart. So, you know, it doesn't, it just doesn't operate on the same granularity. >> Speaking different languages, essentially. >> Yeah, so you're either, you know, an ingredient to somebody who's building a cloud who's assembling lots of things to get to the level where Datrium is offering it today, or you could just be simple and... >> Ultimately are you a software company, are you a hardware company, right? That's the thing and the difference is how we are a software company. You have to think about it as a software scaling. You can scale and make it all scale quite well. >> So, let's talk about some of those services-- >> Yeah. >> Which are all software, so let's list some, and we've talked in the earlier segments about data reduction and... >> Right, so I think the company has built that background that we're going to enable the services one day, so the first service we enabled, and this, the beginning on the year, was backup and data archiving. So, it's a SAS platform, it's a multi-cloud services, this first one. Second one we are building right now, we're going to ship it pretty soon, it's something called Cloud Shift. It's a DR orchestration, app mobility orchestration kind of framework. You can just click, move your workloads anywhere you want, any cloud you want. It's a big piece of our next offering. The third offering we're going to be doing is how do you manage all these different data sets you have across multiple places you have, so we're going to offer that next. So, we also have something called Providence built into the system, like every object knows where it came from, where did it, like all these apps, they kind of know we have all the data, we kind of know where they came from, so that's the next one, we call it a Global Ledger, how do you keep track of all this stuff. And the fourth one is we have all this data now, we have all this metadata, how do we provide governance for the end user, because ultimately they do care about compliance, they do make sure, they want to make sure that they're not moving data to the wrong place, that they have made the SLAs, so that's the ultimately kind of like where are we going to, kind of that's a two-year road map-- >> Mm-hmm. >> Idea. >> Okay, so I've got mobility, discovery, there's analytics in here-- >> Analytics, yeah. >> You've got governance and compliance, obviously backup is something we talked about. >> Yep. >> Now, these are discreet services that I can acquire separately-- >> Yeah. >> Is it all included-- >> Well, in a SAN, yeah, you'd have to buy them separately. >> Yeah, okay, right, right, you do. >> In a cloud approach like ours, they're just automatic and always on, so you don't have to think about them. Global dedupe is an example, if we always have that on you can't turn it off, that helps it locally for cloning so you don't have to move data from server to server-- >> Mm-hmm. >> In a developer shop, for example. It's just, you just boot, you know, start it up and it all has access and it's very fast, or across clouds we don't send all the data when somebody says to move it. We look for the deltas between site A and site B and only send those in a compressed, encrypted way. So, having that stuff just be fundamental and always on means cloud mobility gets a lot easier and a lot faster. >> And I, backup's another good example. I don't need to go buy backup software from a backup software-- >> Or hardware. >> Vendor, or... (chuckles) >> Yeah, that's right. >> Or hardware, right, it's there. >> It's just standard. >> Yeah. >> It's self-protecting, so you know, when you think about cloud mobility it changes the way you think about the problems. For example, if you want to, you know, enable a context for automated DR from prem to cloud, there are a lot of risks in many of the current systems. I don't want to go through the whole, you know, problem set because it's bad and we're solving it in our own way, but just take the conversion problem. If you have to move from point A to point B, you know, 90% of the time if you convert a VMware VM to an Amazon instance it'll kind of work. Well, for DR that's not sufficient, so we're taking a much more sort of thoughtful and open approach to how we deal with, you know, stack providers. So, you know, we'll be able to... In the VMware case, for example, move things straight onto their cloud from our S3 data so that you don't have to convert, so it just always works. >> And I'm interfacing with your SAS, it looks the same where it's on-prem, whether it's in the cloud-- >> Yep. >> It's the same experience. >> We're hoping you have to do less work and less interaction because it's all built in, it all just works. >> Okay, so that, the vision is sets of discreet services separated from the underlying infrastructure-- >> Infrastructure. >> Able to call those services as needed, run on any cloud, on-prem-- >> Run, protect, any cloud. >> Full set of services. >> Right. >> Right. >> Integrated-- >> Right, and as time goes on all our sort of operating software and analytic software and governance, and so on, will actually be, you know, literally SAS in a cloud. That makes it much easier to control a multi-cloud deployment, to control stuff in the cloud, but it also means you don't have to update software, we do it for you. It's just way simpler, so as time goes on, you know, on-prem infrastructure, in our belief, will become more and more the, you know, the thing operated by the cloud and the sort of puppet master will be outside. >> And performance, can you address performance? >> Yeah, so we ran, so basically our system scales quite well because of the way we built it, and we ran, you know, benchmark to take some of the vendors because we wanted to prove it that we're really good at this stuff, and we are the fastest probably on the planet. Our performance is really, really, very good, and it's not because everybody wants it, it's because you don't have to think about it anymore. You don't have, you don't, like it's one of those things again, don't think about it, just works for you, the performance is super high. >> Hm... >> We have customer validations, via the way, we have gotten our reviews from customers who are really, really five-star. We have, like, raving fans for our product. >> Excellent, well guys, thanks very much for helping us parse through that and appreciate you coming back on. >> Okay, thank you very much. >> Thanks for having us. >> All right, thanks for watching, everybody, this special CUBE conversation from out Palo Alto studios. This is Dave Vellante, we'll see you next time. (techy music)