 Inside the Cube, this is Silicon Valley live in Silicon Valley at San Jose Convention Center. This is Hadoop Summit 2013. This is the Cube, our flagship program. We go out to the events, extract the signal from the noise. I'm John Furrier, the founder of SiliconANG. I'm joined by my co-host. I'm Jeff Kelly from wikibond.org and we're joined by a Cube alum, Bruno Aziza from SciSense, overseas marketing there. Welcome, Bruno. Thank you for coming back to the Cube. Thanks for having me back. So guys, lots of updates here. So first of all, when you guys were a startup, I visited your office. You came down from Microsoft. Big, flamboyant, prolific, social media, maven, you are, met in your offices and you showed us the product and was like, wow, that's pretty compelling. Love how you guys are doing that. All of a sudden, explosion, massive innovation. You guys just secured $10 million in funding from battery ventures. You had blown the socks off the marketplace with a great technical product. Good market fit. So give us the update. What's going on with the company and what's going on here? Well, thank you. So we have quite a few things going on. I think just like you said, we raised $10 million in lead by battery. We also just shipped a new product called Prism 10x which opens basically to the world the innovation in what we call in-chip analytics which is this idea that you can have software that's optimized to run on CPU and essentially give you two X, the performance that you can get out of any analytical technology and allows you to go beyond the typical processing and data size you can do on one node. So when everybody here at Hadoop Summit is talking about distribution multiple nodes, we're actually trying to show that you can optimize on one node. And this has worked very well for us last week. We got the top 10 CIO big data company and so we're getting lots of recognition for the technology. Why is that? So give us, I mean I've seen the tech, I want you to explain to the audience why all the attention, what is the core disruption enabler that you guys have that are driving a lot of the innovation? So there are three components. The first one is this idea that the space when it comes to data processing and data analytics has evolved from going from disk to relying on RAM. And RAM is a great evolution from disk because it provides you with great responsiveness from end users but it's limited in how much data it can hold. There's been some great innovation of the last five years in CPU architecture. Intel's pushing very hard to allow pass-through from the CPU all the way to the chip. There's innovation in the L1 cache of the chip. The components that are closer to the core of the machine. And our technology is actually built by hardware engineers that are building software. And we have figured out a way to leverage CPU and we're riding that wave where a CPU is providing more and more processing power and is better and more appropriate for data analytics. So that's disruption number one is from an end user standpoint, a customer, you deploy this on one node, you get the equivalent of what you would get on 20 Hadoop nodes. The second one is the fact that we're an integrated stack. So in the past when you had to buy database, ETL and visualization, we are one. So all in one, so you get a database, you get ETL and you get visualization. And then the third one, which is quite disruptive, is the customers have been used to perpetual licenses, very expensive, complicated pricing structures and our packaging is very simple. It's one product, it's like the Salesforce.com of business analytics essentially. It's a monthly subscription per end user. And those three components I think put together and the fact that they work on any commodity hardware, we're software purely so we don't require you to buy a particular appliance, it's quite disruptive in this market. So what is the theme that you're seeing here at Hadoop? Obviously we were talking about in theCUBE today. I see you are out and about talking to partners. You got the fresh finance, you got the growth strategy you guys are executing, what's the trends out here that you're seeing at Hadoop Summit this year? So there are a few things and you know that we live in a hybrid world. At least from our business, we focus on the terabyte range problem, so we're not so much looking at the pitabyte range data storage problem. I think what's going on here is that the industry is now starting to finally get a handle on two key trends. The first one is understanding what Hadoop is really good for and a lot of the conversations I'm having here with customers and Sears and Netflix and so forth. You really look at the way they use Hadoop is really optimization around ETL processes, shortening how data is processed in and out to other systems. The second trend that we're seeing is that people are really understanding now that we have a level of maturity just like Merv was saying on the data platform where it gives us now the opportunity to leverage that to really build killer apps. For us of course, we think the killer app of big data is analytics and we think there, there's still some explanation that needs to happen because the data platform Hadoop doesn't have all the tooling and certainly is not the answer for analytics. It's the answer for data processing transformation but not for analytics. So I think that's what's going on here. Well, so put that in context of what we're hearing at the show because of course we're having, we're hearing a lot about bringing SQL to Hadoop and doing interactive analytics and queries against data in HDFS. So it sounds like, so you're saying that's really architecturally probably not the best approach. So expand a little bit on how your approach differs and why you think it's a better way to go about getting doing big data analytics. So what we see is that I think the logic and there was a great panel yesterday on this on SQL and Hadoop, there's a lot of contention here but I think if you step back and you look at what are customers trying to do, you have two paths of looking at it. One is the assumption that everything is going to end up in Hadoop at some point and in which case then you're going to have to have tooling only that work on Hadoop and are dedicated to that. That's not the reality we see today and it's actually not, I think the reality where customers are going to be in the next five years. They're going to need to have an environment that handles Hadoop type of data workloads as well as traditional data sources like the traditional databases that we've been used to. And so I think it's a little bit, I don't want to say myopic, but it's a little bit isolated to think that the entire world's moving to Hadoop for everything. I was telling Merv yesterday that one of the things I think the industry really needs is guidance on what to use, what when. And I think we don't have clarity on that. There's a lot of customers that I talk to that have terabyte range data. We're dealing like 10, 20 terabytes and they're looking at Hadoop for that. It's probably not the best way to do it. When I was at Microsoft we had this thing called the center of design for products. You can take a rocket ship and go buy a baguette with a rocket ship. But it's not the most efficient way of doing it. And I think we still have a little bit of that contention going on and customers need guidance. So all right, so let's dig in on that a little bit. So talk about what is your sweet spot? And what is your workload that you guys are really focused on and when is Sysons the best choice? So what we see is the space is divided in three categories. You've got on the high end the pitabyte range type of workloads that we're talking primarily here, processing and take advantage of the fact that storage is becoming very, very cheap. We don't necessarily play in that space. Although we have customers that are storing pitabytes but want to analyze terabytes. Our sweet spot is the terabyte range workload. And it helps customers that are used to business intelligence metaphors where the legacy business intelligence players just can't stretch themself. You know, I think I talked a little bit earlier about the in-memory restrictions. You know, these in-memory technologies they might have been great for a gigabyte range, maybe less than 10 terabytes. But then when you start moving up into 20, 30, 100 terabytes, then they just choke. And it's not that these tools are bad, it's just that the architecture doesn't support the amount of data. And so the big void that we see is that customers that have BI tools that they can't stretch up to the terabyte range and customers that are looking at a dupe for storage they still want to do analytics. And then move it down to really the workloads that they're analyzing data for is the terabyte range. That's probably where there's research that actually on this on EMA research that showed, I think, that the sweet spot for big data analytics is between 10 to 100 terabytes. And that's the market we go after. So talk a little bit about some of the specific types of analytics you're seeing customers do. So when they're looking to do analysts on that size data, what are we talking about in terms of the actual use cases? Maybe if you could mention some vertical industries and talk through that. I'll give you some examples. We have quite a few customers now. We have over 500 customers in 49 countries. We have the big guys like the Target and the Merck. And then we have the smaller guys like the Uber. So Target, what's their problem? And they're looking at theft prevention. And so how do I bring in data types for multiple data sources? And how do I do analysis to figure out where theft is going to happen? So that's the scenario. Merck on the healthcare side, they're using our technology for vaccination. And actually Merck is an interesting scenario because while Target is on premises, Merck is in the cloud. And so we're also seeing that because of the technology we're deploying, the advantage of in-chip analytics is that it can work on commodity hardware and fewer nodes, which makes it easier for us to partner now with cloud vendors differently than other companies. So last week we announced that we were now available on the Rackspace private cloud, which basically allows you, you could be a small company like Uber with a very small analytics team and compete against a big guy and now have access out of subscription bases on Sysense, as well as Rackspace. You get 24 seven support, you get one hour redundancy. I mean, so I think that's what's going on is you get all types of scenarios from retail to healthcare, business, marketing, sales operations. And the way these solutions are being deployed is becoming less and less heavy than it used to be maybe two years ago. So for the tech geeks out there watching, dig into the tech a little bit, the in-chip technology and how that enables you to really, you mentioned you're focusing on really being more efficient on a single node rather than distributing it across multiple nodes and potentially not using those nodes as efficiently as possible. So dig into the tech a little bit for me and explain how that works. So the way it works is actually pretty straightforward. If you open up any machine, you'll find you'll have three components. You have disk, you have RAM, and then you have CPU. Disk, we use all three components, but basically our technology is optimized for CPU architecture. That means it's cache aware. We use something called the L1 cache inside the CPU. It uses vectorization, which allows you to basically take in data from disk to CPU, storing instructions or what needs to be reused in order to give you the best performance. That's unique and really is only available now because of the innovation that's going on in CPU. And so we kind of get the benefit of in-memory. You know, you get as an end user, you get the same return on performance, but what you will get here, you get to use more of an intelligent design because not only did it pass through the data that's required, but it learns from the queries. So one of the things that CPU based architecture as offers as uniqueness compared to everything else is it's a little bit more intelligent than just moving the data into RAM and just querying that data. It actually allows you to do something called prefetching, which when you do a set of queries, we actually watch the types of queries and when I come back and I do the same query that might be using the same instructions, it will reuse those instructions and as a consequence, give me another boost of performance. And so the weirdness that customers get by using our technology is the first query you might pay a lot by a lot, I mean 10 seconds, and then it goes down. So it's six seconds and three seconds and two seconds. So every new query will reduce the amount of time you need to wait, which is kind of contrary to what BI tools have done in the past, where the more you query, the more it attacks the database and the slower it gets. Right, the performance starts to take a hit. The more queries you perform. So talk to me a little bit about the customers that you're selling to. So the users of your technology are often, it would sound like, business people, not necessarily technical people. So are you selling to them? Do you find yourself selling to IT? Typically, who do you go, who do you sell to and have you seen any kind of evolution or change in that demographic? Yeah, two thirds of our customers are business folks and there's definitely an evolution in how business is buying software. First of all, they have an increasing amount of budget and so when you're talking to people in operations or in marketing, they're just more used to buying this type of software. And the truth is that it's extremely easy for them to procure this type of software because if you think about it, all you need is really one node. You need to be able to deploy that if you don't even want to buy software, sorry, hardware, you can rent that on either Amazon, Microsoft, or Rackspace through us. So it becomes very easy for them to get it deployed. That's really the evolution that we see in our customer base. Like I said, we've been growing 520% last year. Our deal cycle, we're an enterprise software company. Typically within this type of software we sell in six to nine months, our deal cycles about 19 days. And so it's quite disruptive. Bruno, thanks for coming on the queue. I want to just get one last question we got on the time hook here. Obviously we love having startups on, we love having you on. We've had continuity on earlier, platform as well as the big guys in the ecosystem but now that you get the 10 million dollars in funding, I mean you got to have Act One done which you got that plague playing out. Act Two is always the car that hits the table for the B round, the C rounds that come on. What's the plan? What's the growth plan for the company? What's the key action lines for you guys as a team? So the growth plan I think we've shown that there's a huge demand for the type of software that we're doing. The way we're going to use this new investment is really to accelerate our evolution. We're opening now an office in New York. We have our sales team that is going to be deployed here. We have marketing in San Francisco and then we have R&D in Israel. And really the two things that we're going to solve is one deployed more of our sales force into the US market. We got to where we got by primarily selling online. And so I think that's pretty unique. And then secondly, we're building a good, a channel for partners. So if there are partners listening to this and they want to resell our software and partner with us, we're doing those deals right now. Great, this is theCUBE, we're live at Hadoop Summit. This is the coming down to the end of day two, wall-to-wall coverage. If you want to watch these videos, go to YouTube.com slash SiliconANGLE. Of course, go to SiliconANGLE.com for all the reference point in tech innovation. And of course, wikibond.org for free research. Again, we're open source content. We love sharing, we love doing theCUBE. This has been a great venue so far. And again, coming down in day two, stay with us because we have a wrap-up coming at 5.30 or 5.20. Dave and I will wrap this up with Jeff Kelly. But more guests to come. Got the CEO of Datamere coming on shortly. Stay with us. And again, if you want to watch all the on-demand videos, go to YouTube.com slash SiliconANGLE. And of course, we're monitoring the hashtag Hadoop Summit. I'm John Furrier with Jeff Kelly here, inside theCUBE with Dave Vellante as well. We'll be right back after this short break.