 Extracting the signal from the noise, it's theCUBE, covering VMworld 2015. Brought to you by VMware and its ecosystem sponsors. And now your host, Dave Vellante. Welcome back to Moscone Center, everybody. This is theCUBE. We've been going three days here live at VMworld 2015. This year we're in Moscone North for the first five years. This is our sixth year here. We were in Moscone South in a nice, so spot, but VMware has really opened up the space here in Moscone North. We've got about at least a quarter, maybe half of the lobby over here. It's been fantastic. We've been talking a lot today about storage. Really excited. Randy Arsenault and Brian Carmendi are here from Infinidat. Randy's the CMO. Brian's the CTO. Gentlemen, really excited about Infinidat. You guys have come out and have some tremendous traction. So thanks very much for coming on theCUBE. Thanks, Dave. Have a nice trip. So of course, by now anybody who knows storage knows the story. Moshe and I's company, they say Moshe. It's like, oh, instant credibility, but it ain't that easy, is it? So we want to talk about what you guys are doing and the innovations. First of all, let me say, let me ask you, so what's going on across the street? You guys got a presence. It's kind of your coming out party. At first you had a sort of a soft coming out party, but now this is really a big deal. You're ready to rock. What's happening? It's been great. I mean, so this is our second year at VMworld. So we were here last year as well, obviously in a much smaller footprint with a smaller presence. But since that time, obviously a lot has happened in terms of the company growing, the install base growing. We had a big funding announcement in April, so this was a great opportunity for us to really get out and tell the story on a much broader scale, which has been phenomenal. So we've had a lot of activity in the booth, a lot of interest, a lot of people coming to kind of, we have a system in the booth, so people are coming and checking it out and asking the architectural kind of questions. We've also had a really full briefing schedule, so there's a lot of interest. So it's been a great opportunity for us to not only establish more of a presence for the brand, but also to kind of tell the story to the analysts and to the end users in a more detailed, kind of more comprehensive way. So it's been great. All right, so let's get into it. Brian, can I ask you, actually, let me stick with Randy for a second, then Brian, we're going to get into it. Just give it the high level, sort of for the company and the sort of architecture, and then we'll go deep with Brian. So a company was founded in 2011, as you mentioned by Michae and I, kind of the godfather of enterprise storage. So this is really our third project together as a team. So a lot of the leadership team, a lot of the development infrastructure has been sort of a long for that journey. So it's a really interesting opportunity to build on a lot of foundational technology and capabilities, a lot of IP, and deliver it in a way that's really broadly applicable to a wide, wide range of workloads. So it's a enterprise class, very high aerial density, high performance storage array. Brian could talk a lot more about the architectural components, but it's come into market at an interesting time and there's a lot of kind of noise and confusion around whether it's convergence, hyperconvergence. There's a million players here, new companies. Large enterprises are looking for cost effective, high performance tier to put their mission critical workloads on. There's still a real demand for that. So what this represents for them is not only a great consolidation platform, so it's a great way for them to consolidate multiple frames worth of workload into a single floor tile, but delivering very, very robust, very highly differentiated storage services on top of it. So it's a really good enterprise story. So it's coming to market at a time when enterprises are looking for a way to kind of get them from where they are today to the point where they can begin talking about real hyper scale enterprise transformation. So they need to buy some time essentially to get from here to there and then this is a perfect solution to do that. So Brian, from our private conversation that we had last week and thank you for that. It was really helpful, especially to prep for this event in a lot of fronts, not just talking to you guys, but just seeing what the developments are in the industry. So you've got this architecture, it has flash and it's got disk in it, that's not the interesting part. The interesting part that I learned from our conversation is the way in which your team has figured out how to use the appropriate media the way it was designed to be used. So you guys know a lot about storage. So talk about the architecture, talk about some of the things that you're doing that are unique. Yeah, so I think one of the central concepts that makes our vision and the direction we're going in, a little bit different from a lot of the other storage companies that you would see here pitching to the enterprise space is we're taking a more modern approach. We're doing what Facebook and Google and a lot of hyper scale storage engineers are doing is we're using flash as a caching layer for storing hot data. Nan flash is exceptionally good at servicing random, small, traditionally challenging IOs, the stuff that disk drives aren't good at. But we don't use it for primary data storage. We don't use it as a durable store. For that we use large capacity, near line SAS drives. These are the slow, boring 7.2K RPM drives. And when you combine those two things together and you get the algorithms right and you figure out how to break the data into small pieces and how to disperse it properly, you get the best of both worlds. You get performance that's like an all flash array, but you get to be a hundred times more reliable than anything that's conceivably possible. And you can scale a hundred times larger than anything that's available today. And Randy kind of touched on the pedigree of the company. A lot of our development team as the original team at EMC that brought Symmetrics to market. And at the time people thought that they were crazy and they said, why on earth would you ever need a terabyte scale database? I met those guys by the way, they were crazy. Yes, I can personally attest to the insanity of our development team. We're at that same point now, X number of years later, where the interesting problems in IT, big data is this thing that's thrown around, big data is, or hyperscale data is, petabyte class computing, that's the new terabyte. And what we see happening is we are trying to enable petabyte scale computation the same way that it was done with Symmetrics in the early 90s, and that changed the world. I mean that ushered in the end of the mainframe era and the beginning of the internet era and the dot com boom and everything like that. Today it's about the industrial internet of things, it's about genomics, it's about cybersecurity, and opening up these new possibilities. And that's what we're super laser focused on. So you talked about, you referenced some of the algorithms, that's really, part of anyway the secret sauce is the algorithms. I wonder if you could just sort of briefly touch on those, sort of what makes them unique, where the IP exists, I call them the bunch of grapes algorithm. But take us through that. Yeah, sure, so one of the things that's different this time, as opposed to the earlier days, is we are not, and nor should anybody be in the hardware business. Back then you got great performance by designing chips and interconnects and everything that were fast. That technology all exists and it's awesome. So Infinibox is built on Melanox Kinect X3 cards, the fastest Infiniband that you can get. These are Intel Xeon CPUs, these are Dell servers that we favor. So step one is get out of the hardware business and let people who are super innovative kind of run with that. So like most of the storage companies, our intellectual property is focused on the software layer that sits on top of it. Now, the way that our technology works is as data comes into our system, we break it up into small little 64 kilobyte objects that we call sections. We tag those with metadata, which tracks locality of reference, multi-dimensionally and timestamps and all this stuff. We disperse it widely across large numbers of media devices on the flash layer and also on the HDD layer. And when you make a problem hyper parallel like that through wide dispersion, you can achieve amazing levels of performance, amazing levels of redundancy if you get the math and the algorithms and the distribution schemes right. And I think that was the core computer science breakthroughs that allowed systems like Infinibox to start coming to market. And I think you're going to be seeing a lot more technology like this as the other vendors catch up. And so Randy, as you described this to customers, analysts, bloggers, what's the reaction? I mean, are they getting it, are they excited about it? Yeah, well, I mean, it runs the gamut, as you can probably imagine. I mean, we were now going through kind of the second and third iteration in some cases of briefing the analysts. So those who have already heard the story are now kind of just interested to hear what's coming next. So they're kind of bought in and they understand the architecture, they understand the value proposition. So for those that we haven't briefed yet and we've done a number here at the show, I would say they're initially kind of ambivalent. They sort of, they see the numbers, they see the two petabytes usable in a floor tile at 750,000 sustained IOPS. So they see the kind of spec sheet Olympics and they think, well, okay, that sounds interesting, but what's different about it? And then we start to get into the core IP and talk about the data placement algorithms. We have the mechanisms that Brian's talking about. We talk about our one of a kind SNAP architecture. So as we start to then drill down into some of the features and services that we provide with the system, that's when the light bulb starts to go off and we get analysts who really light up on those specific capabilities because they are fundamentally different. It's funny, we were talking during one of the briefings, there's capabilities included in the system that we take for granted. We consider them sort of foundational table stakes like our SNAP function, for instance, which are highly differentiated and really unlock legitimately valuable use cases and workloads that would not be possible otherwise. And sometimes it takes an outside observer to kind of hear the story and say, you know, that feature would enable you to do these five things that people are dying to do and they can't do them right now. You know, unfortunately for customers, I think the technologists in the industry are kind of used to a lot of BS and so they take everything with a grain of salt. So one of the most common things that I hear from CIOs and CTOs in the industry are if this thing does what you say it does, then this is a game changer for my business. And then, you know, we're setting ourselves up for success. The thing demos very well, POCs very well and I think that looking beyond the tech itself, the reason why we're seeing the astonishing growth of the business, we've doubled our global footprint since April. We're up over 100 customers. We're up over 300 employees worldwide. I think part of the reason why we're seeing this phenomenal growth is not only that we got the technology right, but stepping back, we listened to customers. And the number one thing that I hear from every CIO in every industry and companies of every size is not take my latency down from microseconds to nanoseconds. There are use cases for that in high-frequency trading and stuff like that, but what everybody is asking for is make my storage reliable and cheap enough that I can afford to store everything forever, to be googly, to be like these companies, like Apex big data companies. And we give them that. We give you petabytes of cheap, reliable storage. You have to figure out yourself how to compute on it and how to extract competitive advantage on it, but we can give you the foundation. That's your job, right? Yeah. Okay, and Randy, you mentioned some use cases that wouldn't be possible, otherwise, can you give me some examples? Yeah, so for instance, our SNAP capability is very unique and Brian can go into the details, but just kind of my 10,000-foot view of it is we have a very, very unique SNAP mechanism that does not stall I.O. on the box, which is very unusual. So we're able to take multiple SNAPs, and we demo this all the time. We can fire up a loop that does SNAPs on a volume as fast as it can execute, and from zero SNAPs to 100,000 to a half a million SNAPs on the volume, while we're simultaneously reading and writing to it the performance curve is absolutely flat. Dead flat, it does not change at all. So when you can SNAP that way, that enables you to do things like manage multiple environments very, very easily. So you're a DevOps environment, or you're rolling out unit test environments. The example that we were speaking with a customer about is they're a service provider that does a big ETL process and generates these huge analytic data sets which they then make available to their clients. They're only able to do it once every 24 hours because of all the computational overhead and just the data movement required to make the data available. He said, with this kind of a capability, they could offer it on a five minute basis at no additional cost to them and enable a whole premium service around just that capability. So there's a lot of examples where, again, things that we somewhat take for granted, I guess, as foundational features can really open up some capabilities that might not otherwise be awesome. And Brian, you mentioned some inspiration, essentially from the hyperscale guys, Facebook and Google. Can you add some color to that? I mean, where's that come from? Is it you bring in engineers from those places or just sort of reading papers? So Google is actually our number one competitor for software engineers. And we win half the time, they win half the time. They have more snacks. They have better snacks. We have better stock options, more outside. So, you know, at a high level, I think the challenges with recruiting and talent is that they're these modern problems that we're trying to solve and the technologies you use to do it, these are new ideas. And that's why a lot of the real talent for this hyperscale stuff, these are guys and gals who are pretty much all under age 30. And so I'm not gonna go Zuckerberg on you and start insulting. I am on the other side of that, so whatever. But this is a young person's game. A lot of these programming paradigms didn't even exist 10 years ago. And I think that's why the competition skews so heavily toward younger folks and why it's companies like Google and Facebook and Infinitab that are competing so heavily for that. What's kind of interesting from our perspective, though, I think is that we're able to bring, you know, really innovative and proven design patterns and kind of turn these youthful, crazy developers loose on those, and it's a magic one. I was just going to say, I mean, you got the best of both worlds because you have systems expertise. Here's the way that you did it 10 years ago. You hired 100 storage developers. You allocated a few, you know, $100 million in capital to build your lab. And then you went and built your storage product. That's what all the hyperscalers did. They all stopped buying storage systems a decade ago from the industry, which by the way is a testament to how terrible we are that our biggest customers all stopped buying systems from us. So, as Randy said, we captured these design principles. We built it into a box that you can load off of a truck, load into your data center, hook it up to your power and your network and have petabytes of storage that never stops serving IO. There's no conceivable maintenance operation that ever requires unavailability. This is what people are used to. If you look at any product that's delivered in the cloud, downtime is something that is the exclusive province of old school enterprise IT companies. And it doesn't fly anymore. And so we like to think that we productized a lot of these concepts that previously required hiring development teams and DevOps people. And we put it in the box. We give it a nice GUI. And that's the way it works. And you talked about it. I think 100 customers, is that right? Over 100 customers. Any you can share or types of customers, industries? So the early affinity has been kind of the usual suspects, financial services, banking, insurance, healthcare and life sciences, cloud providers. The nature of these companies is that of course none of them are willing to be public references. We do have a handful of them that are public references, verisign, tricor. So if you go to the website, there are some case studies and some videos that we're doing. We're bringing more into the fold and convincing them to be public references. But again, the large players in financial services are pretty wary of ever being public references. So. Of course. Last question. Moshe, those Moshe and the four or five crazies that change the storage business forever and then many others thereafter, people wonder how involved is Moshe? And is he just a figurehead? Is it just a way to raise a bunch of money? How involved is Moshe? So Moshe is in the office every day if he's not meeting with customers. He has a lot more helicopters than he did back in the IBM and EMC days. So he takes a lot of customers up. We do an awful lot of business and helicopters. 20,000 feet, yeah. You'd be surprised how many deals get done up in the chopper. And maybe it's the threat of being pushed out. I don't know. But, you know, Moshe, it's funny. He has the vision in his head. He is without a doubt. He has the ability to see long-term IT trends as clearly as I see the hand in front of my face. So he has the vision and he has the capital to make it happen. And the passion. It really is the leadership team and the developers, the group that he's been able to get to follow him and say, okay, this is the direction we're going and go make it happen. And when we have epic battles where there's two architectures and two ways of doing it, he'll make the final call. But I would say that it's the developers. It's the guys and gals behind the screens typing that really make what we do possible. But he also, in addition to that, he's just, he's very passionate. I mean, people presume I think sometimes that he's been very successful. He's established in the industry. You know, do I want to kind of step back and he just, he's not ready to do that yet. He is the most tenacious person I have ever met in my entire life. He can argue either side of an argument and beat you both ways and that's part of his charm. Yeah. When I get a ride in a helicopter, call Infinity. There you go. We're handy and Brian, thanks very much for coming to theCUBE. Great day. Congratulations on the success and really excited about watching the ascendancy of Infinity. We have very ambitious plans for the future. Awesome. So hope we get to more and watch this space. Good deal. All right, keep right there. We'll be back with our next guest right after this. This is theCUBE, we're live from Moscone, 2015, VMworld.