 Live from Las Vegas, it's theCUBE. Covering VMworld 2017, brought to you by VMware and it's ecosystem partner. Hey, welcome back everyone. Live here in Las Vegas, this is VMworld 2017, exclusive CUBE coverage. I'm John Furrier with theCUBE. My co-host Dave Vellott, their next guest, Brian Biles, who's the CEO and founder of Datrium, former entrepreneur, founder of Data Domain, entrepreneur, great to have you. Great to be here, great to see you guys. Big launch, Datrium, you guys are out there as the crest of the wave is hitting. The crest of which wave? The wave of cloud, data, cell waves. So many waves. What's the, okay, question, what's the biggest wave that you're riding right now? Within, so there are two big ones. One is, enterprises are shifting some of their deployment to public clouds. So, you know, we're helping navigate some of that. The other is within private clouds, they're moving to converged infrastructure. So our business starts with that but is starting to have dimensions on public cloud. And what does that mean for the customers? People out there right now coming to VMworld this year, they go, okay, I know what cloud is. It's a true private cloud on premise operating model with bursting to, with Amazon now, I got clear picture. Everything else I'll fill in the blank from there. Yeah, there's, let me start sort of the other order. Within convergence, there's a, I don't know, some like 10%, 20% of infrastructure has kind of moved to that model and it's growing. So traditional sort of specialized appliances are declining in favor of that. Most of our business is in that world. And as we're able to combine compute primary and secondary storage in one scalable infrastructure. And that's never been done our way before. So we have a lot of great things to talk about there. Within migration to cloud, the sort of first thing that happened, if you're not moving like wholesale, a bunch of applications there, if you, if you're doing a hybrid thing, the first thing that moves is like backup because it's a simple thing to move and you know, it's a low cost store. So we are showing at this show, migration of snapshots in a pre-compressed, pre-duplicated encrypted way to S3 for restore to your on-prem site. So as an entrepreneur, we're sitting around kind of like on the beach playing golf. And I want to just do a start. What was the motivation for you to do Datrium? For Datrium. So you know, you guys have seen me for a while. I was a founder of Data Domain and Data Domain got bought by EMC. Frank Slutman who was running ServiceNow used to say as our CEO that after that acquisition, it's sort of like being eaten by a Python every day. You look a little less like yourself. So, you know, we all sort of decided to move to a different thing. And so the CTO group and I left about the same time, decided to do something different. We met two guys who were principal engineers at VMware, had been there for, you know, 10 years and knew everything about the hypervisor world. We got together and Datrium came out of that set of discussions. And it turned out that was in 2012, hyperconvergence was just starting to emerge. So we were looking for the next thing after that. And then cloud was becoming big. You know, how do we branch onto that? So all of these discussions were relevant to kind of early thinking even then. And now it's just become more clear. But Brian, this notion of bringing compute, primary storage and secondary storage together. Where did that come from? Nobody really does that. No, and it's partly because of our background as people. Data Domain was a backup centric product. When I went to EMC, I ended up, you know, doing product strategy for the network and Avamar products as well. And we'd been in the backup world for 10 years. So we knew that space pretty well. Doing a storage system that was good for that on either disk or flash was something we knew how to do. But the people that were purely doing that couldn't figure out how to get it to work for primary storage. We took our understanding of that and with the help of our founders from VMware who knew how VMs expected primary IO to work put together a different kind of framework. So we use hosts with flash as a primary store and a secondary set of chassis with spinning rust for persistence of data. So it's like a peer on every host writing to kind of a rubric for persistence. So if you actually had hyperconverged and a scale out backup system, you'd have too much of everything. You'd have like too many motherboards, too many persistence drives. You'd like have duplicates of all kinds of hardware. By doing it together, we stripped it down. So it's way more cost efficient and it turns out way more scalable. So you have to share those resources and then still solve those unique problems. Right, efficiently. That's right. So we can be as competitive as any data protection architecture as well as being faster than the fastest all flash array. So there's a lot of debates going on, mostly on the all flash data center side. People say all flash, it's here to stay. Flash is cheaper than spinning disk. There are a few people who say, that's not true and I think, you're one of them. Well if you look at an actual implementation of an all flash array, it's never alone. You're always doing backup to something else. And the something else is probably disk. We all the time see some all flash array partnering with some disk based backup thing and that's the composite solution. Well, you're not getting rid of spinning disk in that case. You're only doing it for the primary IO, but the secondary IO is going to disk. Or it might be going to, there are risk management cases where you might be going to Amazon with that backup data, you might be going to tape. But you're not getting rid of something low cost. And if you look at just the sort of list prices of an all flash array versus the list prices of a scale out backup array, it can be a quarter of the price. Because even if both have D-dupe and compression, you're paying for insurance in the backup side. You don't want it to be super expensive. Yeah, so okay. And so you see that gap staying indefinitely. You want to protect your data and you want it to be low cost. We've just engineered it together. So if we up level it beyond the media, Datrium data is obviously fundamental to your name. Everybody talks about digital transformation. We see digital businesses as defined by the way in which you use your data. So how does Datrium fit into helping companies better leverage their data? Well, you can sort of go through the life cycle of the data. At the end of the day, we're kind of a data management company that also lets you go fast. So in data management, you mostly want to work at a small granularity, like VM or container level, at an application level, not at a sort of storage artifact level, like a LUN. You want to deal with apps. You want to deal with, and so the convenient approximation of that is VMs or containers. That lets you deal with it easily. So when we have policies in our product for data protection, we don't do volume level snapshots. We let you do container level snapshots, persistent volumes or VM level snapshots. Replicate them across any number of sites, including AWS. Keep more than a million of these granular snapshots in a single system of shared data. So if you snapshot a VM on one host, it's immediately cloneable on another one in the rack, or the same with a container. And otherwise, you can't really do that with a normal storage system. And this just makes it much more flexible, as it turns out, a lot faster, because the read IO never leaves the host. It's always on local flash. And we first had you on theCUBE at VMworld, I think two years ago. But you're moving beyond your VMware roots. Quite a bit. Into Linux and other areas. Maybe talk about that a little bit. Yeah, thanks for bringing that up. This year we've had a lot of announcements. We changed the product considerably, the company. So not only can we do all this data management stuff, we can simultaneously support VMware hosts, Red Hat virtualization on just Linux hosts, as well as Docker on bare metal. So that multi-hypervisor, multi-container ability in the same shared storage space is very unique. Even hyper-converged vendors don't do that. So yeah, that's opened up a lot of doors for us. What's the impact of customers? That's a unique opportunity for you guys. Why go that way and what's the impact to customers? Ultimately, people want to consolidate to converged infrastructure. But different application environments are optimized differently. If you want to consolidate, you have to support all the ones that matter. And that's why we're doing this. So here's the update on where you guys are right now as a company. We are an employee headcount. What's the vision? What's the next couple of milestones for your success? We're about 140 people. We have sales in the U.S. predominantly, a little bit in some other locations, but mostly U.S. We'll be expanding that. We've just, over the course of this year, done quite a lot. So we've gone from, you know, we've expanded our capacity and performance by more than an order of magnitude. We've expanded our just number of snapshots and so on and at the same rate. M by N replication across sites, including Amazon. So that's a lot. A lot of product work. A lot of product tech getting done. Yes, digestion of that and selling of that. We've moved to everything from sort of commercial mid-range, hyper-converged or array sales to sort of telco service provider class, rack scale infrastructure opportunities. There's a lot. A lot of where we're investing is data management across multiple clouds. So we're showing Amazon replication today. We'll get to other clouds over time and as well do more things than just store backup data. We'll be doing, you know, more migration and disaster recovery and a lot of things in the future with clouds. Well, it's the most exciting thing happening that gets you excited right now in the industry. In the industry, as you guys look at your opportunity, because you're the founder, CEO, you get the 20 miles there. You kind of do a lot of product work. Product CEO, you get to see the vision. As you guys have got that trajectory going as a starter, you're forging new ground. What's getting you excited out there? In a sort of physical infrastructure sense, it's really interesting to see where NVME over fabric is going and how to leverage that because it's, first of all, going to be a different way to compose host instances. It has some problems. It's not a sand. It doesn't let you do shared data. You can't do vMotion using it because it's not a shared facility and it requires hosts to do the erasure coding for the off-host storage. And that's not a normal model. It turns out it's what we do, so it's interesting. And that could be an opportunity or a challenge. Yeah. It's also really interesting to see where customers are going with their interface of data management across multiple clouds. It's very much an emerging territory. Similarly, with containers. Containers have always been this sort of lightweight, stateless, it comes up, it goes away kind of thing. It's within a host. You don't really think about, you know, orchestrating container management across hosts in a data management sense. But because we can do it, you can snapshot a container persistent volume on a host with us and then clone it immediately in the same namespace on another host. For rapid development, no one's ever thought that through and now we can offer that. So we're in a bunch of emerging dialogues on that stuff that no one's really had to offer. It's interesting, bringing us a little adult supervision to stateless apps. Yeah, because containers are super fun, you know, useful as a development paradigm. It's lightweight, it's small, it allows migration in these really interesting ways. But persistent, so people are trying to apply it to persistent data as well as stateless data. And that's giving a bunch of interesting new energy to the whole policy approach. Brian, great to have you on the queue. Great to see you, congratulations on the venture. It's going great. And you had to launch recently on big announcements. Keep on innovating your pioneer and looking forward to our next chat. Great to have you on the queue, appreciate it. Thank you very much. Thanks for the time. More live coverage, three days wall to wall, covers day one, coming to an end. And here at VMworld 2017's theCUBE, we'll be right back with more coverage after this short break.