 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 Brick and Stu Miniman. Hey, welcome back everybody. Jeff Brick here with theCUBE. We're live in downtown San Jose at Open Compute Project Summit 2016. It's really the heart of the cloud. It's the heart of the internet. It's where the big iron comes together now around this Open Compute standard in the hardware space. So a lot of excitement going on into the fifth year of the show. We're happy to be here for the first, or excuse me, the third year and really get the smartest people we can find, bring them on the show, share their insight with you. I'm joined by Stu Miniman from Wikibon and our next guest. Happy to have Leo Leung, Vice President, Corporate Marketing of Scality. Welcome. Thank you. Good to be here. So what are your impressions of the show? It's one of our best shows. It's the easiest conversations we ever have because everyone understands distributed systems here. Everyone understands scale-out. So we get right into the meat of what we do on top of hardware. So for the folks that aren't so familiar with Scality, give kind of the quick overview for the people that haven't been familiar with it. Sure, so we were software defined before software defined was the term. Pure software play for storage. Runs on standard Linux, standard X86 and we take all that hardware and those resources and pull them together into a storage pool. Serve up object file all natively and it's great because it supports legacy types of file-based applications as well as newer cloud and open-stack types of applications. Yeah, Leo, I liked your observation about just distributed nature, scale. People understand that here. The other show that I go to that we understand that's usually AWS re-invent. Of course, Amazon's not at this show, at least not this year. What do you see in those audiences compared to the traditional enterprise company that you might find it? You know, some of those other shows that we sometimes go to. Sure, absolutely. I was just having a conversation, the same thing, right? We're talking about the folks that are building the newer systems, right? You know, the Gardner, Mo2 types of people. Yeah, same expression. But it's true, they have a very different mentality about what scale means, what type of infrastructure needs to be deployed and the software and hardware combination. And yes, it's a similar set of people. Obviously, there's a very AWS-oriented slant, but same thing, very easy conversations there. Yeah, so I might've given a little eye roll on some of the Mo2 stuff because we find the conversations we have with users is I can't have two solutions. I have what I had today and I need to get to where I'm going tomorrow, but it's not, you know, slow this down and start that. There needs to be a bridge between those. You know, we see companies like Goldman Sachs and Bloomberg and Netflix. Netflix, of course, is usually one of the lighthouse accounts for some of the really cool forward-looking stuff, but do you find, are there OCP people looking at OCP that are bridging it, or is it mostly, I mean, the big guys just are built this way to start and understand, you build a new data center, you start that. How do I get from where I am today to these kind of environments in the future? Yeah, I mean, I think there's definitely two patterns. I think OCP has been very important to push all the manufacturers down a path, right? I think they've proven out that there's value in these different form factors, there's values in efficiency, and it's forced everybody to change, right? But there is adoption now happening outside of the hyperscale providers, certainly as proven by the financial institutions that are part of OCP. But what's interesting, what's been interesting for me is the appearance also of other more traditional enterprises in life sciences, in manufacturing, and other places. And usually where this kind of stuff happens is during a refresh cycle, right? When there's a big refresh, there's an opportunity to bring in things like OCP. So are you finding some of those users here? Give us your feel as to the attendees of the event. Yeah, I mean, there's absolutely, again, more of your classic enterprises here than ever before, because I think they have to, right? They need to at least take a hard look at it. Whether they end up buying OCP hardware or not, they need to understand what's happening in that industry. So we're definitely seeing that. And in parallel, they need to consider the software implications of moving to that environment. All right, so before we dig on the software side, Open Compute started very much as a compute discussion. Networking has taken a lot of front of mind conversations here with the news, with Microsoft, with Sonic, Web Switch getting into the bottom environments. In the keynote yesterday, I think it was Jay from Facebook said the next thing we're going to have to tackle is storage. Where do you see storage fitting in the whole conversation? And how do we expand Open Compute to more of the full stack? Sure, so IDC just put out their server numbers this morning and the biggest growing piece is hyper dense, right? Dense form factors. So to me, the translation there is storage, right? That's where things are going. Certainly the market has told us as a software defined company that there's a broad range of form factors that are going to work for them. If you can squeeze 80 and 90 drives into 4U, fantastic, we'll do that. In some cases, the ratio is a little bit more even. But the use cases dictate those form factors. So it's nice to hear Facebook talking about that because frankly the storage projects I think have lagged versus the network projects and compute projects. It'd be actually very cool to get involved in that from a user specification perspective to help guide that. Yeah, so Leo, the conversations we've had at storage had tended to be more on the performance side of things. So we have companies like Micron and Sandis. You see all the SSDs, NBME technologies in the Facebook. What about the capacity side of the story? How does that play into the OCP discussion? Yeah, I was having this conversation this morning with some other analysts. The reality is 80% of that data is capacity oriented, right? You're just not going to throw any of these low latency technologies at it, doesn't make sense. You need the pairing of the two for it to work. But if you think about that ratio, that's where all that hardware is going to go, right? So it's interesting that that is not usually discussed. Maybe people think the problem's solved, but it's not entirely. But I think we've seen some great hardware innovation in that space, right? Extremely dense form factors, much more modularity, great. And then bringing the software piece to pair it. I think we've gone down the road a little bit, but maybe it's sexier to talk about MVM and all that stuff. I think that's what it comes down to. So you wrote a good blog post coming into this event, talking about the software piece of it. Right behind you, we've got the big iron racks of devices from Facebook. So how does software fit into this whole discussion? Yeah, it's not a big conversation sometimes because as you know, Facebook wrote all their own infrastructure software, right? Microsoft wrote all their own infrastructure software. But if what if you're not Microsoft or Google, what do you do? Or Facebook, right? So that's where folks like us and other software-defined types of players come in at the networking layer, at the different layers to provide that type of scale-out architecture, right? And because people can't write it themselves. That's what it comes down to. What do you see in the field the impact of Flash beyond the obvious? I wonder if you can share any use cases of how leveraging that performance with the software-defined is going into places that maybe had no thought it would be going there that quickly. Yeah, I mean, I do agree that with Flash, with MVM, there are going to be newer applications that weren't possible, right? If you increase your memory space by that extent, absolutely, there's going to be some cool stuff that comes out of that. But I think in our case, it enables us to be both a eventually consistent system as well as a highly consistent system within one architecture, right? So by having our indexing functions, our lookup functions on Flash, we can provide the kind of performance that people are looking for when it comes to consistent POSIX types of applications. So that wasn't possible before, right? You would have to throw some type of very heavy-duty server at that type of architecture, but it certainly enabled us to do what we do. So one of the big discussions here is, of course, this is open source, there's lots of partnerships going on. What do you see as some of the important pieces of the ecosystem? Who's scality partnering with to help put together all the pieces for users? Sure, sure. So I think we've built up a very strong relationship with HPE and Dell. Recently, we announced an even furthering of the relationship with HPE. HPE just announced another of their Cloudline products. We think that's going to be a great fit for us as well as a hardware platform for scality. So certainly the large brand system vendors still have a play, absolutely. Still driving a ton of revenue and growth. I think OCP's keeping them honest, right? I think having that discussion, having those types of specifications out there and the ODMs creates a nice balance for customers and the customers win at the end of the day, right? They have a broad, broad, broad choice of options. Yeah, I'm wondering Leo, we look at it and say, what motivates someone like an HPE or a Dell to do this? How much of it is it that they've got partnership with Microsoft and Microsoft saying this is what we're doing and they need to support what they're doing both on-prem as well as they want to get a footprint inside that giant Microsoft environment versus kind of the rest of the customer base. How do they differentiate in this market? You know, help us unpack a little bit, you know, what these guys would think. Yeah, I mean, it is a tricky balance, I think, right? I think that, but the reality is, if people are going to move into this model, as I said in my blog, they have to realize that the software is going to have to take a lot more responsibility for things like availability, reliability, et cetera. You're not going to just move out your, you know, higher profile types of server, quote unquote enterprise types of servers that have specific capabilities that let you manage resiliency and recovery to a bare metal or a bare bones type of architecture without bigger considerations. So I think that that is going to slow adoption down a little bit, right? In terms of the mainstream enterprises. They have to take into consideration the entire solution. Okay, so what brings skeletonity to this show? I mean, you look around the show floor, most of the companies that are exhibiting are component pieces. I don't see kind of the big storage companies, you know, the big six or seven of them. They aren't here. Do you expect to see your competition here? Is there something special about what you guys do that make it a fit? I mean, it's sort of a case of who's really, who's really software, right? And who's an appliance vendor? This is sort of telling, right? We can play in this space and we don't have to make any kind of compromises, right? So as I said, at the end of the day, the customers that we're talking to here, the architects, right? The people that are building out their new data centers, the people that are building the updated systems are all here. And it's an extremely easy conversation to go to transition from, yes, we're buying 1,000 or 10,000 servers to, well, what's running on those servers, right? What are the applications? What are the workloads? What's the storage requirements? Super easy. And we've met some very, very nice prospects out of these discussions in the financial sector, in life sciences, et cetera. So another place that Scality has a big presence is the OpenStack Summit. You help, you know, are there connections between, you know, OCP and OpenStack? Do they just sit, you know, different parts of the stack really? But, you know, how do you guys see connections? There's some loose connection. I mean, people, again, it's more of the mindset of we're building a cloud infrastructure, we're building a scale in infrastructure. So oftentimes these conversations go right from OCP to OpenStack, but it's pretty loose. I think when people are thinking about actually deploying and implementing these environments, it's not necessarily one plus one. You know, there's also different levels of maturity and different levels of abilities to actually implement these things, right? Certainly, I'm not the only person to say that there's a lot of projects with OpenStack. So there's some challenges there. Whereas OCP, in some cases, is simpler, right? If you know what your data center looks like and what kind of energy you want to consume, you can get down the menu and be able to implement something, so. So, Leo, we're getting towards the end of our time, unfortunately. What are you working on? What's the next grade hill that you're working on at Scalely that you need to take down over the next six months, 12 months as you look down the road? Sure, so I think that we're continuing to grow. We're over 200 people now, over 100 customers, starting to get real penetration into the mainstream enterprises. So, just like this show, just like OpenStack, when you're starting to deal with the mainstream, the mainstream production workloads, that's an exciting place to be, right? That's where the company can really scale. At the same time, we're seeing some fantastic new use cases. So, recently brought in our first IoT customer. We've been hearing about IoT till we're blue in the face, but it's very interesting now to see how this company is actually building an infrastructure to go support that. And as we've talked about before, what is that going to look like? I think it's being built as we speak. It's very interesting from a layer perspective, as we were talking about, Stu. There's lots of different pieces going into that technology. There's Cloud Foundry, there's us, there's an IaaS underneath the covers. Almost all the trends being pulled into one to serve the IoT workload, which is really a whole gamut of things, right? Everything from healthcare imaging to engine data and industrial data. So, very cool stuff. No shortage of demand for software to find storage in the immediate future. All right, well Leo, thanks for taking a few minutes out of your day to stop by. Appreciate it. Thanks very much. Absolutely. Leo Leung from Scaldi, Vice President of Corporate Marketing. I'm Jeff Rick, he's Stu Miniman. We are live in San Jose at the Opal Compute Project Summit 2016. We'll be back with our next guest over this short break. Thanks for watching.