 Okay, we're back here inside the Cube. This is EMC World's exclusive coverage from SiliconANGLE and Wikibon. I'm John Furrier, the founder. This is the Cube, our flagship program. We go out to events, extract the ceiling from the noise. I'm joining my co-host for this segment is Jeff Kelly from Wikibon. Wikibon analysts and big data, and this topic in this segment is going to be around big data, but also storage. There's a lot of stuff going on around images, et cetera. And we're Ken Malero, senior director at Pixia Corporation. Welcome to the Cube. Thank you, thank you. We are here in the ground floor in Las Vegas to hear the noise all around us. There's people yelling all over the place and the blogger lounge behind us. But you guys do something very, very interesting. You guys, talk about your application first. Introduce your application and we'll dig into some of the questions. Sure, thanks. So, appreciate speaking about it, but Pixia Corp is a software company. We specialize in data access for large data. So unstructured data, mainly raster data. So, heavy pixels, satellite imagery, also motion imagery, and also video data. So, obviously, we heard from the folks here at EMC, Joe Tucci, David Goulden, Pat Gelsinger today. Internet of Things is hot and Paul Moritz highlighted in his keynote today. GE invested $100 million in the Pivotal company. So, obviously, Internet of Things is going to throw up more data. Surveillance cameras, satellite images, these are stuff that's been around. That's kind of legacy, but now you're going to have a lot more data. So, talk about the dynamics of what that does for infrastructure and applications. And what are some of the state-of-the-art bleeding-edge things that you guys are seeing? Well, we work a lot with the government, defense and intel industry, so they have a lot of unique sensors. A lot of those unique sensors are starting to be used domestically. We've heard a lot about that stuff in the news, but what we provide is a lot of the data access and as those data sets grow in size, the problem for them is that they're really becoming just unwieldy. So, you have terabytes and petabytes worth of this imagery data that you need fast access to small random block reads to, and you need solutions for providing data. So, are you storing the data or are you just providing software to access data, or both? A little bit of both. So, we provide the software that sits on top of, specifically EMC Isilon, and it allows you to basically access the data as a logical container. So, you have a lot of that data that's all containerized into a single layer that's homogeneous, and you can access that very quickly. So, can you talk a little bit about to the extent that you can? Some of the types of applications that folks are building on top of this data, is it mostly just to access the images to view them, or are there also some deeper analytics taking place? Yeah, there's a lot of analytics. On the government side, there's all types of algorithms that they're running on the data. Very simple things from building tracks, from moving objects, building information of what's changed in the environment. It's not only defense and dental, but there's also a lot of natural resources types of applications for it. So, changing of the environment from a city as urban, cities start to grow. So, we're talking about analysis of the images themselves, not necessarily the metadata that's associated with those images, but actually the images themselves. So, we're actually providing up the raster data, specifically the heavy pixels. Talk a little bit about what the challenge involved there. I mean, it could be a goldmine for not just for the Intel community and government, but I could certainly see commercial applications as well. So, what are some of the challenges, though, that you had to overcome to make this data available for analysis? Right, right. So, the large volumes of this data, be it if it's satellite imagery or if it's video data, is the size. I mean, you're talking about terabytes or petabytes of imagery that you're really looking at a very small area of interest in. So, to be able to pull that out, that small area of interest out of such a large dataset is a huge problem. And so, that's really where we focus on. The other thing is all of these algorithms and all of these applications are, need a standard way, an open standards way of accessing that data. So, we provide web services, restful APIs on top of that dataset so that people can access that with whatever tool they're using. What are some of the storage challenges that you guys see? Because, you know, you're on the bleeding edge and this is going to be a use case that's going to be pretty defined for other people. I mean, image is not so much, but big data, well, big heavy could be images, files, whatever. Unstructured data. Unstructured data, loosely structured data, not the rigid, well, you got to ask a lot of questions you don't know yet. So, we like to break things down to known queries and unknown queries. For that time, you need to ask an unknown query. You don't want to have to reset tables and do all kinds of loading the data. So, what are some of the storage challenges that you face and how did you overcome them? So, one of the things that we partner with Icelon on is the scale out capability of their NAS. So, as this dataset grows in size and as the number of users grow and the number of requests, right, queries start to grow, we need to be able to access that data very, very quickly. And so, when you're talking about terabytes to petabytes, the ability to scale and maintain that consistent performance is an issue that we've tried to address. Where are you guys taking the solution, your application and solution, and what are the storage needs here that are around the corner for you guys? Video. I mean, so much more video. I mean, you see not only commercial applications like YouTube and Facebook and others, but on the government side as well, we have unique sensors again that are collecting more and more video. The video's a huge problem. And the video's a huge problem. You indexing that? Yeah, yeah, we do. We're indexing that and we're providing quick access to that to all types of analytical algorithms. The Icelon reps must love you guys. Do more video. They do, they do. Back up the truck. Yeah, exactly. Sign the P.O.'s, man. The video's a hard problem, I mean, how do you tackle that video? I mean, honestly, we're doing, we're live streaming right now to hundreds and hundreds of people. How do you handle that ingestion and how do you do all the indexing? Well, we have a unique container. It's a logical container, like I said. You know, I equate it to kind of a zip file without compression. So we're kind of like this database within a file. So we're able to take all those streams, concatenate them together, piece together all the metadata, and provide kind of a single view to the dataset within that logical container. So you're big data and big storage. Big data means big storage. Exactly, exactly. So lots of times, we'll say we're big data and we're not necessarily in the sense of analytics, but we're in the sense of very large files, right? You've got to come in like a laser focus and pick a block piece of data out and bring it in very quickly and being to do that on an ad hoc basis. Exactly. You're talking about a query, say Google Maps, right? And a query, and you're doing a query of a little area and you're talking about 100 kilobyte view that you're getting back, but you're pulling it from a three or four or five petabyte set of data. Yeah, and it's got to happen like that. And it's got to be a millisecond response, yeah. Wow, that's huge. So I'm curious, so you mentioned analytics. We talked a little bit about some of the analytic capabilities on top. What's your, can you help us understand how kind of what you do, the storing and making images available to the enterprise and to government clients. Where does that fit alongside something, things like what Pivotal's doing with Hadoop and bringing in some of more of the text-based analytics. And I imagine there's got to be some correlation between the images and the text. Definitely. How do the two fit together? So that's a great point. So Pivotal is one of those relationships we're starting to grow. We're trying to provide our customer solutions to bring in together the metadata that's being collected, so which is text, in the Pivotal system and marry that up with the data that we're deriving or that we're storing in PIXIA containers. So that when you build an algorithm, a data scientist goes out there and builds an algorithm, it can take in those various types of data sets. Not only the metadata for the imagery, but the metadata, the text metadata that may be collected for other things that Pivotal's bringing to the table. So we're really trying to provide that kind of integrated solution to some of our customers with Pivotal, yeah. Ken, my final question, do we're getting tight on time here? We've got the end of the day here on day two. Thanks for coming on theCUBE, by the way. We really, really appreciate it. Great application. Go to PIXIA Corp. Is there a website? Yeah, PIXIA.com. PIXIA.com. That's right. Share with the folks out there what you've learned. I mean, you're touching a lot of different things. They're kind of spread a little horizontally. It's not so much vertical per se, but you've got big data problems, you've got software, you've got all kinds of coding, you've also got storage. Share with the folks out there what you've learned and advice that you give to your peers out there about if they had to face the same road that you went down. Great question. Well, really open standards. For us, open standards... Get a big, fat checkbook. Yeah, exactly. Be ready for the storage bill. Exactly. No, open standards is a big deal to us. Being able to have access to all of that data from a variety of algorithms, GUIs, tools, visualization, platforms. But the data comes in so many different ways in so many different formats. We want to normalize all of that, but we want open standards. So really, restful web services that are open standards, when you build an as-a-service web service, build it with open standards. Okay, well, we really appreciate it, Kent. Thanks for coming inside the CUBE, pick, see a corporation, check them out. Doing some really interesting things with big data and images. I mean, think Google Maps, doing things like that is what these guys do. And it's a real world problem and this is part of the new era. This is the CUBE. We'll be right back with our next guest after this short break.