 From San Jose, in the heart of Silicon Valley, extracting the signal from the noise. It's theCUBE, covering OCP US Summit 2016, brought to you by OCP. Now your host, Jeff Frick and Stu Miniman. Hey, welcome back everybody. Jeff Frick here with theCUBE. We're heading into the final stretches of day two wall-to-wall coverage here in San Jose at the Open Compute Project Summit. It's really where all the hardware manufacturers that power the cloud and deliver the infrastructure that powers the cloud and enterprise data centers are coming together. So we're really excited. You know, we like to talk to the big companies, but it's always fun to talk to startups. It's where a lot of the innovation happens. And in fact, now we're going to get in before the startup actually, I guess you started, but you haven't come out yet. So we're a little bit stealth mode. We're excited to be joined by Mark Fleischman, CEO of Deterra. Mark, welcome to theCUBE. Thank you very much for having me. Absolutely. So not quite out yet, you're kind of drip feeding a little bit of information to the market. Everybody excited? Yes. So what can you tell us about Deterra? What can you not? So you're going, you got some people. Can you tell us VC funded, I assume? Can you tell us who your VCs are? Yep. So we are a company that we've been funded and founded about two and a half years ago by folks like Vinod Koshla and Andy Bechtelzheim, a couple of other investors as well. We are getting very close to coming out of stealth mode. Product is shipping. We're in production with a number of customers. Hopefully we'll have a lot to tell to the end, very shortly, and I guess that's why we're sitting here. Okay, so I won't nail you to a date, but are we talking about months, quarters? Weeks? Very few weeks. Okay, very exciting. So two and a half years, you guys have been at it for a while. What was the business problem that you were trying to solve? What's the opportunity that kind of started this journey? The basic business problem is that if you look at infrastructure as it moves forward, you really want to flip it upside down, so to speak. The old enterprise, the traditional enterprise has been very infrastructure centric, had obviously a lot of people running around with scoot drivers trying to fix hardware that would act up all the time. Google and Amazon and a couple of other cloud providers have very impressively shown us that that's not the way to build systems. They literally flip infrastructure, I would say, on its head. They make it DevOps centric. They make it, you know, some people call it invisible. They just change the operational model and the development model, the usage model of infrastructure, and we take a critical part out of that, a crucial part, and we do that, the same thing for that as well. So Mark, if maybe we can pull the lens back a little bit, and look at your background. You've been involved in a lot of open source activity. That is true, yes. And I think about infrastructure usually is not very open source. You know, we're at the Open Compute Summit, so lots of discussion about networking, opening that up, talking about storage is one of the next big challenges, so how do you see that interplay between what's been going on in infrastructure and all of these open source initiatives? Well, I mean, I would actually say the open source content in infrastructure is increasing pretty rapidly. If you look at OpenStack or Ceph or other pieces in the infrastructure stack going forward, there's an increasing amount of open source in there. Now, obviously, open source might be great because the price model is very attractive and the service model typically is pretty attractive. It has other challenges, I would say, maybe more from a vision perspective, from a particular roadmap, from an execution perspective. So it's not going to be obviously all open source based, but it's an important factor. And frankly, where we are coming from, yes, we do have an extensive open source background. I can at the very minimum share that we contributed the entire SCSI target, which is Block Storage Subsystem to Linux. That's called LIO. That's widely used by now a number of billion dollar companies. I would just mention one, for instance, Pure Storage, a couple of other companies as well. We looked at a lot of those companies about two and a half years ago seeing what they did with it. Frankly, taking proprietary hardware structures, functionality and putting it in software open-source software partially on less proprietary hardware, but not really changing the consumption model, okay? And that's ultimately why we set out to build that data, really fundamentally changing how you consume infrastructure. Interesting. So I worked for one of the big storage companies for quite a few years and when I joined Wikibon about six years ago, I had to go through some retraining. So the hang is upside down for a couple of months and let some of the Kool-Aid drip out. But I think about the enterprise mindset. Let's harden it, duly redundant, give it its own proprietary network, build this whole stack because it needs to be reliable and it needs to work. When you look at kind of the hyperscale, when you say distributed architectures, let the software handle all of that resiliency and can actually get to more reliable, maybe more secure environments. As we said, hardware will always, it will eventually fail, software will eventually work is one of the tropes that you see out there in the marketplace. So what do you see out there? Yeah, no, it's very true. I mean, I think, again, it's inspired by the likes of Google, Amazon, Microsoft Azure and so on, Super7, right? That's clearly the future of infrastructure. It's self-optimizing, self-healing, all the kind of autonomic attributes you look for there. And what that ultimately means for even the traditional enterprise, although you're right, they kind of are typically a little slower in adopting that. It really changes the consumption model, the purchasing model, the economic model. Ultimately, these trends are now getting so powerful. And since these guys have shown that you can build economically very attractive systems with those technologies, that it's clearly the name of the game right now is, take, I'll call it hyperscale, by that I mean more an architecture or a philosophy rather than scale itself and drive it into the enterprise segment by segment. And obviously we're taking one of those and we're doing the same thing. All right, so where do you see is kind of the white space out there because there's a lot of companies that have been saying this for a while. I think, I cover converged infrastructure and the hyper-converged guys, that's their message. It's this new architecture, they put the whole stack, they make it really simple because I would say simple IT used to be an oxymoron. Absolutely, absolutely. So hyper-converged is a brilliant example how to take those concepts and package them for I would call the mid-market enterprise and make them extremely easily consumable. The problem with hyperscale is that at a certain level of scale it is not efficient anymore because resources are tied to each other and you have stranded capacity if you scale them linearly with each other. So at a certain level in the enterprise at a certain scale level you're going to have to disaggregate. So to speak and start scaling compute and start separately again. However, following very similar principles, hyperscale principles, okay? I would say the gap in the market, for us at least, how we see it is more on the higher end market. I would generally say the compute orchestration problem is very well solved. There's a lot of people who are competing for that spot, Mesos, Kubernetes going forward, Docker, of course, OpenStack, CloudStack. Perhaps the underlying reliable infrastructure making easily composable and making it very easily consumable for these compute orchestration frameworks, that seems to be a little bit a less solved problem. And that's what we were talking about. All right, so it's interesting, you threw out some of those, the containerization, some of those more emerging cloud native-type environment. Yes. Storage tends to be one of the last kind of problems that are solved along those lines. Absolutely, yes. And I think back to the VMware days. I mean, VMware helped greatly with efficiency, but boy, did it break my storage environment. Yes, yes. So how does storage kind of catch up and not be the anchor in the data center? Yeah, so clearly storage needs to be, we thought, and I don't want to share too much quite yet, but first of all, I think it needs to unify existing classic infrastructure and we still have a lot of virtual machines, we still have a lot of bare metal, we still have a lot of traditional databases, but at the same time it needs to be, to provide a platform to go forward and actually host these cloud native applications as well. And while it is scale out, it still needs to be very high performance or at least I would say low latency, right? Because if you make compromises on its efficiency, ultimately you impact the cost, you impact performance, you impact density. So lots of challenges to be solved there. All right, so I know we can't ask you too much, I know Amber's out there, she'll get on us if we go too deep, so Amber, we know you're there, thanks for watching. So just we'll wrap it here, the real thing's coming out, they're gonna launch to the world a couple of weeks, where do people go, where do they watch, just go to datera.com, what should they do? Watch our Twitter handle, watch our website, datera.io, it's very stealthy right now, admittedly, but we'll keep everyone posted, as I said, we're getting very, very close. All right, awesome, well Mark, thanks for sharing a little bit of insight with us. Thank you very much. We'll keep an eye out and look forward to the announcement in weeks, not months. Absolutely, thank you and thanks for having me. Absolutely, so I'm Jeff Frick with Stu Miniman, Mark Fleischman's stopping by, giving us a little inside scoop on what's yet to come, a little glimpse into the future, what you'd like to do. I'm Jeff Frick, we're coming to the end of day two, here, wall-to-wall coverage, Open Compute Project Summit, San Jose, California, stay close, we'll be back with our next guest after this short break, thanks for watching.