 Live from Las Vegas, Nevada, it's theCUBE. Covered, AWS re-invent 2016. Brought to you by AWS and its ecosystem partners. Now here's your host, Stu Miniman. Welcome back to AWS re-invent 2016 here in Las Vegas. This is SiliconANGLE Media's theCUBE, we're the worldwide leader in live enterprise tech coverage. Happy to welcome program, a user in the Amazon ecosystem here, Walter Scott, who's the CTO and founder of Digital Globe. Thanks so much for joining us. Great, thanks, good to be here. All right, Walter, give us a little bit of background for those that haven't heard of Digital Globe, what the company is, what your founder, what was the big, why that you were looking to fill? So who we are, actually most people probably have used Digital Globe imagery without realizing it. If you've ever looked at a satellite image on a mobile device or a web mapping portal or seen any of the awesome map box stuff that you saw earlier on stage, it's Digital Globe imagery behind that. And so we power a lot of location businesses around the world, whether it's governments, oil and gas companies, location-based services, global development organizations. And a lot of what we provide is global transparency. So with Digital Globe, you can see what's going on around the world, literally anywhere in the world, within a few minutes of it actually taking place. So pretty cool stuff. And it was really that idea of global transparency that motivated me to start the company. Really interesting, I love that idea. Can you just give us a little bit, kind of speeds and feeds, how big is your company, how long you've been around, what's the scope, is it truly global and who are kind of your biggest part? Sure, so I started the company back in 1992 when there sort of wasn't really an internet. I remember actually one of the things that our investors were kind of wonky about. One was that we were using satellites launched on rockets, so there was this push button risk. You push the button and it either goes up or blows up. And the other was that they thought that the internet was kind of a fat. So it turns out, obviously the internet was not a fat. Today we're about 1,200, 1,300 people. We are in multiple locations across the world, Singapore, London, a number of locations in the US, Tampa, Herndon, Colorado, where our headquarters are. We're sitting at about 100 petabytes worth of imagery that goes back to 1999 when we had our first satellite up and operating. And that was, I think of that as like a time machine, a time lapse library of the entire planet. Putting it into somewhat different terms, we've covered the Earth's surface many times over and there are some places where we're covering an area once a day, sometimes several times a day. The most interesting parts of the planet, obviously, we're looking at a lot more frequently. Okay, great, and are people like Google, are they now competitors of yours or, you know, what is that kind of thing? Google's a customer. Google's a customer. Google's a customer, Microsoft, Apple, Mapbox, who I mentioned. Great, so your company existed long before we were talking about this thing called Cloud. How does that change your business? I have to think, you know, the speed of what you're doing, the size of all the data you're dealing with, you know, talk us through kind of that journey with Cloud. Well, you think about having 100 petabytes of data and you saw in the keynote that you can now fit 100 petabytes of data in the back of a shipping container, but there aren't very many people on the planet who can actually accommodate 100 petabytes of data. So, up until the advent of Cloud, our data was basically in jail. We had access to it, and we could dribble it out in dribs and drabs, but there were very few people who actually had the ability to do anything with it at anything remotely approaching the scale that we collected it. So, Cloud was a big enabler in getting our data out of jail. It meant that it was possible for us to do things like for a company in Australia, PSMA, that serves the insurance industry, they want to map and attribute somewhere between 13 and 15 million structures in Australia, and we use machine learning and crowdsourcing, leveraging the data in the Amazon Cloud to do that at a price point and with a speed that is orders of magnitude better than what you could do with clipboards and people running around. You know, you mentioned 100 petabytes of data a number of times. The AWS snowmobile was, you know, brought out at the end of the keynote yesterday for Andy Jassy's piece. As you said, it's the extension of their snowball family, which was, you know, talking 100 terabytes, you know, it's 80 terabytes I think of the original, the new ones, about 100 terabytes, this new one, you know, 100 petabytes, but I mean, that's not an instantaneous movement. There's, you know, even with the terabit of, you know, connectivity, if I need to do that, most customers, I mean, it would take a couple of months to load that up and then you got to move it and then you got to unload it. So, can you walk through, I mean, when you say 100 petabytes, how often do you get that much data? Where does it come to and, you know, you're shipping it into the cloud, it sounds like? So, we collect about 10 petabytes a year. And the purpose of the snowball, it's really what is enabling us to go all in to the cloud. So, it lets us catch up. It allows us to take that time lapse library that goes back to 1999 and port that into the cloud. We then have a continuous flow of, on the order of about a dozen snowballs that are always cycling in, as well as a network connection, a direct connect that allows us to take the most current and the most sort of high interest data and pipe that into the cloud immediately. So, it's a tiered approach that's very cost effective. Well, that's fascinating and, you know, how do you determine that that tiering? Is it pricing, is it, as you said, kind of critical time sensitive data? How do you walk through that understanding of the value of data? So, it's hard just to begin with because you never know where in the world something interesting is going to be happening. And it could be an Ebola outbreak, it could be a natural disaster, it could be an Olympics, well, generally you know about those in advance. And the fact that you don't know means you really do need to be able to get access to all the data. So, the snowmobile goes into Glacier and that allows us to have a deep store of all of our historical data. And then we have algorithms to manage the staging of data to various layers. We may use social media to indicate, hey, you know, something's likely to become interesting. So, let's start migrating it up because it's probably going to be of use. And Walter, for the snowmobile itself, were you talking to Amazon about this challenge when they created it? And have you actually had the truck do this? And how often, if so, if you're using it? So, we have it sitting at our place right now. It's sitting in Colorado. You can actually see it from space, by the way, with our satellites. And if you check on our website, you can actually, I think, see a little video that shows you that. So, we had it for the last few months and we've been loading it sort of as a background task. It's not, we're not using 100% of our bandwidth to load it at all. So, it's going to take a few months to load the snowmobile fully. And then we'll ship the data into Glacier. Okay, so, is this the first time you've got the truck at your location then? Yes. I think we may be, we may be either the first or one of the first to be using a snowmobile. You know, when the snowball was announced at last year's keynote, we immediately thought, they're talking to us. But we need a bigger snowball. Going to need a bigger truck. Absolutely. And the cost effectiveness of that, you know, you said it's going into Glacier. How does that going to unlock the value of the data more for you, that kind of repository? It definitely does. You know, what you had, you were able to like get rid of, you know, a data center or what? So, eventually our plan is to migrate out of data centers for our commercial business. But I think there are really two key business benefits of being in the Amazon Cloud. One of them is that it allows us much more rapid access to data that was previously stored on tape. And not just rapid access, but the ability to access it at a scale that was not possible before. So if we had tried to do certain jobs in our own data center, they would have taken four to five months. Operating in the cloud, they take a few days. And that's the sort of time scale that wasn't previously possible. And then there's also the matter of just simple economics, that it's reached the point where it's more economical to operate on these large data sets in the cloud than it is to maintain your own data center. Yeah, I'm wondering if you can kind of unpack that a little bit. People that you say, oh, you know, well, I know the cloud is less expensive for test dev, or if I've got really bursty applications, that makes sense. You've got a big amount of data. Seems like you have a very good understanding of how much data you're going to have. So are you using reserve instances? Is there special pricing? How do those economics work for you? So I'm not going to get into a huge amount of detail about how we manage the cost, but we leverage reserve instances. We do a fair bit of machine learning. So that's on GPUs. We are mixing various compute instances. Our workload does tend to be fairly bursty. And in particular, something that we launched a couple of years ago, and it's been seeing very rapid growth, is a geospatial big data platform, GVDX. And it's what enables geospatial producers and consumers to come together to create location intelligence. It allows for unlocking the kind of finding, counting and measuring that previously you'd have to have people marking individual shipping containers or cars or measuring the height of buildings. And now it's possible to do that with machine learning. So that's been a big unlock for us. And it's intrinsically bursty because you ask a question. We have an application called answer factory. And you go ask answer factory a question, like show me the variation in standing water after this flood. And you've drawn an area of interest and it'll tell you, here's how much compute you're going to need to use in order to do that. Well, you may decide that you want to darken the Eastern seaboard and do that over the entire planet. Or you may decide that you want to focus your area of interest on something that you maybe are more interested in. All right, Walter, I want to give you the last word. For people that didn't make it to the show, tell them kind of the value you get coming to an event like this, your experiences and key takeaways for you. So I think a big thing about this is it unlocks a realization of a potential by looking at what other people are doing. And the ability to see, hey, somebody else has figured out how to do X means we don't have to go through the hard work of figuring out how to do it ourselves. And it's particularly exciting to see many of the services that Amazon is announcing that in a lot of cases it's clear that they're listening to their customers. So you guys, what's next? You guys going to be mapping out Mars or some of the other things beyond the Earth? What's next for the company? We're sticking to Earth, but as we continue to build out our business, we build out applications that are more and more time relevant. And we just launched a satellite a couple of weeks ago. More than doubles our high resolution capacity. So more to come. Excellent. We really appreciate you coming on the program. Walter Scott, CTO and founder of Digital Globe. We'll be back with lots more coverage here of AWS re-invent 2016. You're watching theCUBE.