 Okay, we're back here live in Las Vegas for Splunk Conference, dot-conference 2013. This is theCUBE, our flagship program. We go out to the events, extract the symptoms from the noise. I'm John Furrier, Joe with Dave Vellante. What a run, we have Oracle Open World Day, VMworld before that, now it's a Splunk Conference. Just a great run here. Our next guest is Robert Osborn, Director of Application and Development Systems for Homes.com. Robert, welcome to theCUBE. Thank you, glad to be here. We love having folks who are in the trenches. We were talking before we came on about your site, Homes.com. Sounds like you're really excited about some of the site activity, site thing technology. Great name. Great name. Got the best name in the business. I was just playing with the map resizing. Really good, nice job. So congratulations. Thank you. Application, web-based application, obviously a lot of JavaScript, a lot of new frameworks coming out there, cloud computing, a lot of development frameworks. So the application market's exploding, but a lot of the stuff on the back end is needs to be managed, right? So Splunk is one of those. So first, talk about your relationship with Splunk and let's get into the how you're using Splunk because you have a lot of data. You have millions and millions of pages a day. If users don't like your site, they're going to go somewhere else. So you got to keep a good value proposition. So talk about Homes.com and as an app guy working with Splunk, what are you guys trying to do? So last year we experienced 40% year-over-year growth. We had a bit of a tough time scaling to support that level of traffic. We were saturating everything in our data center from NetApp IOPS to switches on the A6. You name it, or A6 on the switches. If you name it, we were probably saturating it. So we really needed to bring in a technology that would allow us to aggregate all of our logs and better monitor what we have. And not only were we growing then, but the plans were to continue to grow. We recently spent $4 million in TV ad campaigns on an HDTV network and DIY network, product basement spots. So our key focus was let's get our Homes.com architecture and infrastructure to be able to scale. And Splunk was going to be a huge part of that. So talk a little bit more about your environment, paint a picture for us so we can sort of connect to what's there. Sure, we use for our backend storage NetApps. We have five load balance everything. Everything's virtualized using VMware running on Cisco UCS blades. We also are working on bursting to the cloud where if we were to run into the problem of saturating some points in our data center, we can burst new 10, 20, 30% of that access traffic out to the cloud using Amazon's AWS. We positioned our data center and the same data center that Amazon uses and one of their availability zones. So we can run fiber directly into the cloud for low latency lengths. Okay, and then talk about your application, part of your job is application development systems. I mean, that's pretty fundamental. Talk about that, the application development organization that you're involved in, that you manage and how that all operates. Sure, we're a very small team. So what's cool about homes.com is we're a smaller business that's owned by a bigger business. Homes.com is owned by Dominion Enterprises and Dominion Enterprises is an umbrella company. Under there we have about 45, 46 different business units, homes.com being one of them. So what's so cool about our environment is that we run very lean. We only have a team of five programmers working on the homes.com portal. We have three systems administrators and we have about four or five different products people developing ideas for the platform. So we run very lean and efficient. At the same time, it's also very exciting. So your business is changing quite dramatically. I mean, you got, you know, sites like yours really disrupting the real estate market. You got, you know, realtors participating but not that, you know, reluctantly but it's really changing the nature of that business. What if we could talk about that a little bit and how that is putting pressure on your technology infrastructure? The whole landscape is changing within the real estate portal. The amount of inventory that's on homes.com now is far different than what it was five years ago. We just have a lot more listings on the internet and that means our data footprint is growing. That means we have more data analyzed and as we start growing more and more and we get more and more page views and visitors, we have to really analyze, well, what's engaging to them? What are they most interested in? How can they be more engaging? It's all about determining the engagement factor of the customers and how to improve that. Now, is that where Splunk comes in? I mean, what role does Splunk play in terms of helping you guys understand the engagement metrics? Well, if we step back just one small step, when we were looking at purchasing Splunk for purely our operational data, our CIO challenged us to think outside the box. You know, sure you can use Splunk for operational data but what can it do for product? How can Splunk drive revenue to homes.com? So we took that challenge on and... Literally, your CEO asked you, how can we use Splunk to drive revenue? He was that close to the situation and the technology. Yes, and Joe Fuller really encouraged us to innovate and he saw something in Splunk that maybe not a lot of other people saw and that this product can drive revenue and he really wanted us to step up and figure this part out and so he did and he kind of made it not a full condition of purchasing Splunk but a pretty big partial condition. Part of the business case. So how did that all come about? You guys said, okay, we'll take on that challenge. You had no idea how you were going to do it and then just how did that manifest itself? So big data, the whole concept is new to homes.com. So our first challenge was explaining to our product people what big data is. Here's the types of data that we have. Here's what we're not doing with it. What are some things that we can do with it? So if we talk, if we start thinking about what's the most valuable data that we collect in the operation side into Splunk? Well, that would probably be our access logs, the footprint of our users, the tall telling sign of how a user's engaged with our website and then how can we drive product value by using that data and I think probably the coolest idea that we have that we came up with and which we implemented mid Q4 is that we're going to develop a recommendation engine. So the idea being when you look at a detail page on homes.com, you're looking at a property, you're looking at the price, how many beds and baths. Well, every event that you do on that page leaves some type of engagement footprint. For example, if you look at, if there's 10 images and you look at all 10 images, well that means you're very engaged with that website. If you only look at four images, we might not be as much engaged with it as a person who looked at all of them. If you print a listing, well that's a very engaging factor because you're actually printing on physical medium and to use it for later. So every click that you do on homes.com leaves an engagement footprint and what we do is we calculate that and determine an engagement value on that property and then recommend listings based off of that. So one follow up, I just want to flip it to John after that. So in terms of how you get paid, you have some advertising in the site but what you're talking about is other back end monetization strategies, right? Sure, so one of the ways that we really monetize the website is we match real estate agents with serious home buyers. So if you go to homes.com and you're in the market for a house, you look at a property, you like the property, you want more information about it, we have forms that you can use to contact the real estate agent. So then what we do is we collect that data and we sell subscription plans to real estate agents and we'll send, we call them leads, 10 leads to you per month or 20 leads, whatever package you sign up for and that's our big value driver, just marrying real estate agents to serious home buyers. So I got to ask you about the mobile conversation. We had the guys from New Zealand on this that said 40% of the traffic's mobile. You're probably obviously seeing a mobile object depending on the demographics, they'll be shifting and growing depending upon which segment of the market you have on the site at any given time, but mobile will be a part of it. That brings the whole data log question, what kind of data exhaust are you seeing? Are there new types of data that's entering into your purview that's now part of this blunt equation? Absolutely, mobile's huge and it's just getting bigger and bigger and it's important that we really think outside of the box and figure out ways to innovate in that space and how can we make our product better and we just have a bunch of great ideas. The amount of data exhausts that we're getting I think is most interesting with the geospatial data because with a device, with the mobile device, we can pinpoint where they're located geographically and then we can marry that data to our listing data and maybe you can have a real estate agent send an alert to someone when they're near their property saying hey, you're near this property, you're in the market for a house because you have our app and you did some recent searches and according to your recent searches, this is a property that you're interested in. Why don't I buy you a cup of coffee and let's talk about this property? This is where the application framework intersects with web design so can you share with folks out there that are looking at the convergence of mobile and web apps? You know, obviously web apps has been evolving very quick. Obviously you're seeing Mozilla, Chrome, native browser integration. Obviously that's always been a hassle, the different browser types but now you've got mobile apps, right? So you've got native apps versus frameworks. How are they colliding? Because many in the industry are saying web and mobile are one, will be one in the future. So how do you get your arms around that? Well I think the first important problem that's tackle in this space is you're dealing with a bunch of different technologies. Developing native apps on a mobile device is a lot different than developing web applications. So our first strategic position is to unify our programming language. So we're going to a complete node.js architecture where JavaScript is going to be the core competency that we require of all of our developers. And from there we're going to develop dynamic feature rich backend and front ends using JavaScript and then we're also using titanium to develop mobile applications written in JavaScript as well. That's awesome. So what are the future things you're seeing with Splug outside of the challenge from the CEO which by the way is a great testament to the value proposition that Splug brings to the table because hey you know go solve some problems but now it's driving revenue. That's the holy grail, that's what people are really tasked with and that gets IT folks excited. But what's next, what else are you looking at? What else are you thinking about with Splunk? Specifically around, obviously with Noble the node is going to be some great stuff. How are you thinking about the next leg of the journey? So we really remain focused on developing products features around big data. What I would like to see from Splunk is more direct connect features into our databases primarily the ones that do MapReduce such as Mongo. I think, I'm really looking forward, I know Splunk's tackling this problem and I think when they do it's going to bring a tremendous amount of value to homes.com and I'm sure other business units as well. Give an example of what you mean by that because a lot of folks don't know the new ones with Mongo. Sure, so Mongo is a non-relational database much like Hadoop. Well the challenge is and the challenge that we have right now is when we're building a recommendation engine, we store all of this data in Splunk and Splunk's very good at aggregating the data and sorting and shifting through it very, very quickly. But what it's not so good is handling 10, 15, 20,000 concurrent connections because for every query it uses essentially one core and a CPU. So you have to move that data aggregated into some type of transactional database which would be Mongo. Right now we need to have jobs run once every minute to do that for us. I think going forward from Splunk we're going to see more of a seamless integration that they're developing now with things like Conquere. The data just goes directly, goes right into Splunk and then directly into our Mongo database. You know we've heard that from the Splunk executives. Integration is a big part of it. You got the ecosystem partners here. Final question for you for me is I want you to just take a step back and share with the folks at home. What's it like here? I mean obviously it's our second year at theCUBE here at dot-conference. We've been following the company for many years in Silicon Valley and watching it grow. Explain to folks a vibe here. To our right we've got the partner of Pavilion. You've got the keynotes going crazy. What's the vibe here? What's it like here? It's a lot of fun and it's centered really around learning. I've gone to a lot of the different sessions and I always learn a lot. It's just amazing. It's a great place to be. I really enjoy it. I look forward to it every year and I hope to keep coming back here after a year. Robert Osborn inside theCUBE. This is what it's all about, learning. That's why theCUBE exists. So we want to go out and get the extractive ceiling from the noise and share that with you. We're open source content at Silicon Angle and Mookie Bond. We love that last line, epic sound bite. That's a tweet. I shouldn't put that on a tweet because it was so good. All the best conferences these days is about open source, learning and sharing. So congratulations for taking questions for that comment. This is theCUBE. John Furrier with Dave Vellante. We'll be right back. We have this short guest to wrap up day one of the Splunk Conference. This is theCUBE. We'll be right back.