 From the Walt Disney World Swan and Dolphin Resort in Orlando, Florida, it's theCUBE. Covering Splunk.com 2016, brought to you by Splunk. Now, here are your hosts, John Furrier and John Walls. Hey, welcome back everyone live in Orlando. This is SiliconANGLES, the CUBE, our flagship program. We go out to the events to extract the single noise. We are here for DotConf 2016, as part of Splunk's seventh annual event. I'm John Furrier, my co-host, John Walls, and we are going to have a Shazam of a second. You're a music guy, aren't you? I mean, you're a big music guy. I love music. You've got Shazam on your phone. I've got Shazam on my phone, yep. All right, we're going to mess with this a little bit. To help us do that is Chris Cameronman, who's a senior infrastructure engineer at Shazam. Chris, thanks for joining us. Good to see you. Good to be here. First time that I'm here. First CUBER. We need to come up with a song for that. We need a CUBE alumni. You've written it already, or done it already. So we were talking before we came on. We're going to have a Shazam app integrated to the CUBE experience to bring that community together. It'd be pretty cool. Obviously the CUBE has a great community and thanks for coming and joining us as a CUBE alumni. Great to have you. It's a pleasure. So Shazam's got some cool tech. I honestly started as a real kind of like total geek project and music wasn't even on the agenda. With identifying music was a unique algorithm. Matches of certain frequencies, ties in to the database. Hello, it's a Google search for audio, whatever you want to call it. Now expanded, a lot of infrastructure involved. First give us the overview of what's going on with Shazam. How big, what's data, what goes on under the hood. So we have over 100 million monthly active users and we need cloud infrastructure and on-prem infrastructure to support that. So we have a mixture of Amazon. We have a mixture of a third party managed data center. We have our own in-house data center. But we have, without giving too many numbers away, obviously hundreds of servers. We have some EC2, some DynamoDBs and Amazon Redshift. We have about 30 spunk indexes, sort of. So a lot of cloud. A lot of cloud, a lot of cloud, some on-prem. Actually, because of our requirements to do the recognition infrastructure using GPUs, GPUs are a little bit hard to put in a cloud infrastructure because they generate a lot of heat and also they're slightly slower because you share them. Facility issues is classic car valve for on-prem, have that on-prem. Yeah, exactly. So our GPU recognition cluster actually makes up the size of the state in a managed data center. So a lot of us are basically web-scale and the modern era mobile users. I got Shazam here, you want to show a demo? You want to tease this up real quick? So I'm here, I'm at theCUBE and you want to know who this guy that's speaking on-screen is. He's a bit crazy, he's Australian, he has an accent. So let's just pretend we forgot who you are. Yeah, I'm somebody. So if you press the Shazam button right now, press it, magically through the interweb and the magic that is Shazam, it will actually recognize. That's you. Who I am. Yeah, it's me. It's me in virtual form. And it's an app that I created for son.com. And it tells me where I'm speaking, what I'm doing and it's a way for you to interact with me. If you want to link in with me, there's a button to link in. You can get a bio on there too. We want to know who you are, what you're talking about, where you are, what time of day, all of the time. Exactly. So this brings up a good point. Obviously the directory, music, people, it's audio. Yes. And obviously with Siri, AI, audio cues are now huge. Yes. Right, so you're saying you guys can do this. Yes. For people and anything. Absolutely, anything. We do it for almost anything and everything. So our audio gets used at live events, gets used in stores and retail outlets so that we can help people and help brands engage with their users. We do it visual as well. So you can take pictures of Coke cans and magazines and other merchandise. So I think even a Formula One car might even be shazamable because we have a relationship with one another. I mean, Barco's skating's been around for a while. There's been all these devices, the Q-Cat going way back and you really know your web-dage history. But now with QR codes and Snapchat, people are actually getting used to using the native device for capture and then discovery. Exactly. Obviously Siri with the iPhone and other technologies, cars are going to have it. So the interface of the voice is there. What are you guys doing now to create more headroom? Obviously, what do you currently have in market and what's in the pipeline? So we do audio recognition right now. We do visual recognition and we've got Bluetooth recognition. So we can do hyperlocal, we can do anything. This is our ethos to be able to recognize anything. And we're trying to do some, without giving away too much of our pipeline, but the next visual implementation is going to be awesome. Everybody's going to love it. Is there anything because of the mobile environment? I mean, does that change at all? I mean, either how you harvest data or capture, whatever, because it's a unique animal. Yeah. Mobile poses its own challenge. Wi-Fi drops in and out and we're saturated here at Splunk.com. Sometimes I can't get a signal. Sometimes I can't get a response as to what I've just shazammed. Mobile, it's a mature environment right now. It's not like the 2007 years. So it's really, in some ways, no different to a normal weather. But it's our native application format. So no difference for us. So what's the coolest thing that you've seen with Shazam? Obviously, my daughter uses it for music all the time. I see her with other Shazams. She puts it on the radio. Yeah. I was playing and she gets the Shazam. Great for music. Everyone gets that use case. I see that on TV. Well, what other cool things can you share in having access to all the data? So cool use cases involve being able to sing your friends a song by karaoke. It involves, we did some Oculus stuff with a soft drink company. We have some in-app experiences like for one of the movies. You shazam the movie poster and your phone becomes part of the set. And every time you move around in three dimensions, like it feels as though you're in part of the set. And one time we're in the Jaguar car and it felt as though you're in the car and you press the button and the engine just goes boom. And you felt as though you're in the car. It was a really good interactive experience. This brings the augmented reality gesture-based discovery. I mean, I don't know what to call it, but we'll just say a gesture, voice, picture. I'm initiating a progression that you guys are facilitating. I mean, in a way, that's an opportunity, right? It is an opportunity. And I'd say watch this space for the play on realities. Hopefully that it'll be coming and be announcing that some stage in the future. All right, so there's a big opportunity for you guys. You can see a lot of future, you can see the vision. No doubt. How do people work with you guys? Do you guys have an ecosystem as you guys doing it by yourself? How do people make money with Shazam? How do people get involved? Obviously you're here at Splunk. You Splunk some things for the demo. Can people play with Shazam? Do you guys have any open APIs? Any kind of developer environment? So we have partners that we work with and we also have our own sales force. We're a data-fied business, so we're trying to sell data to other people as well. So you can, if you wanted to buy a Shazam-enabled advert or a Shazam-enabled experience, you can speak to our sales people. They're quite easy to find. Shazam.com slash brands, check it out. And we have in-country representatives in various parts of the world. But yeah, also if you want access to our data, you can look on some sites. I think data at Shazam.com. You can buy some of that. I'm sure we'd be happy to talk with you. We have relationships with media, like the record labels, you know, individual artists, radio stations. So there's plenty of ways to get access to that. So it's just as if you're a cubit if you get to Shazam and find out who the hosts are. Yeah, I'd love to be able to do that for you. Give me 15 minutes, I'll make a call. So it's all about me, you know, all about user behavior, you know, about engagement. These are points that you're refining with the data that you're gathering, you're aggregating out of your collections. What are you finding out, like in general, that you find actionable, that you're finding useful to fine tune your services? So one of the things that help, so we have a set of people in such as them who's able to help somebody if they buy Shazam-enabled experiences, to how and best you can get users to interact with Shazam and perhaps lead to purchases. So whether that be choosing the coolest music or, you know, placement of call to action signs, we have some insights as to what people do in shopping centers and what they're likely to be interested in or buy. We have a whole lot of data that we can actually mine and help other companies monetize. But the people who do that are the research people, not me, and I forget some of the useful insights. Yes, question. So here's just kind of thinking about what we do in theCUBE. We kind of connect the dots, we riff on stuff. So let's just say that I have a CUBE interview we're doing this here and it's on our video, so it's watching live, hello everybody. But it goes on demand for YouTube. We do CUBE gyms, we also have our podcasts on Friday, 8 a.m., Pacific Time, listen to my podcast live, also go to soundcloud.com, CUBEcasts, the weekly podcast of fourth week. So I can see a use case where I would like to have, see if I can load a Shazam and say, anytime someone's Shazam's a CUBE video, that they can pull up derivative data around it, like related videos, which I have meditated on, I might have top stories that might be trending in the genre, same thing we're talking about, big data at Splunk, someone's Shazam's this interview and says, wow, cool interview related to Shazam and Splunk with IT, I mean, I'm making it up. But is that possible? I do that. Absolutely, because we can pinpoint anywhere in the audio stream as to how far, two minutes and 21 seconds through the audio and based on that, we can generate your custom result and you're right. So I give you a metadata playlist or something or a storyboard or? We need to work on the solution, but yeah, essentially. It's possible. Yeah, absolutely. And we're kind of doing things like Shazam for conferences and I've been doing a bit of Shazam for Splunk live London where basically throughout the day, the results change and you can Shazam during the day, during the conference, during the talk and you can get different results. So it's quite helpful. So what if I wanted to build into a mobile app? Because we're coming out the Cube app soon, our native mobile app and it's going to be, I'll playlist with like top videos. Can I put a Shazam button in my app? Because I'm on my phone, I'm mobile with the consumption. Yeah. An analog makes sense, radio or TV. Is there a Shazam button I can natively put on my mobile app? Without giving too much away, the answer is yes, you should talk with us. That's coming. No, that's it. Because you can't listen, you can't say, hey, you want to say, hey, Shazam this iPhone. Yeah, Shazam this interview. Yeah, so Siri will work, you know, hey Siri, what's that song? But, I'm sorry, I probably just turned on all our phones by saying that. Well, if I hit the button Siri, I turn off when I'm playing or if it's playing in the background and Safari browser sucks for that. And you don't want to develop your own app too and your own tech to do the audio recognition. You want to make use of it. But you guys nailed the use case and the user expectation of Shazamming. Absolutely. Which is like typing in a key word or Google or, you know, so that is a good brand value. Absolutely. So I don't want to have to reinvent an image bar or any kind of brand. I just love to write on Shazam's coattails on that. So, and his audio. Absolutely, so I'll give you a use. So, is it playing in your store and you want people to engage with your brand? You don't want to develop your own in-store app because I'm going to walk into the store and I don't want to download your little app which I'm going to use once. So, what you should, and app real estate is a prime commodity. People have 100 apps on their phone but they use 10. And one of those 10 is usually Shazam. So don't invent your own. All right, so we're going to do a Shazam button on our mobile app. This is a great deal. And like again, first of all, Shazam like Splunk has a unique brand. It's a verb. Yes. C-Verbs has brands, Kleenex, Splunked it. Yeah. You know, Googled it. Shazam did. Yes. You guys are there. So you crossed over. Yes. Now where's the growth? How do you get to one billion? Good question. And we're always working on that. You're the interesting guy. Is there like a certain... You can just say, hey, my job is to make sure that everything runs for a billion users. That's, you know, you're good. Well, I'm really excited to work for Shazam's. Oh, you know, this is all very exciting to me. Is there a security player though to me with recognition, identification, those kinds of things? So maybe there is an opportunity there for a whole different sector for you to play in? Yeah. The sectors that we can play in is virtually limitless. It's just up to us to choose where we're going to go. But we do like, we use Splunk and we kind of use Splunk for security and we have security issues with our app as anybody does. And we've got to care about that. We have celebrities on our hosting. Twitter-like messages would be bad but that got hacked. But we can go anywhere with this app and if the future is our oyster, the load is our oyster. It looks like I go into the queue. All right, so we will work on that. We'll continue. Great, congratulations. I got to ask you the kind of the big question, kind of looking at the landscape. This is not really a Shazam question, but just in general. The trends in tech that you're most excited about, knowing what you're dealing with at Shazam. A lot of pressure to be a hyperscaler, web scaler now. You got cloud, obviously a lot of Amazon. We have a lot of Amazon as well. It's a lot of goodness in there but scale is really, really important. Cloud gives you that. You got some on-prem. What do you worry about? What are you focused on or are you like? Is it more in memory flash? Do you like Linux, kernel? Actually, what's kind of exciting to me and some of the stuff we're doing at Shazam is machine learning and the making real data science use of the data. So I've been playing with Splunk Machine Learning and what's great about Splunk Machine Learning is it's democratizing big data and it's democratizing data science. So somebody like me who's not a data scientist can actually go and play with it. You've solutionized it. You've got some work. You've made a solution out of it. Simplified it and I can actually go and play with Shazam data, machine learn it and I can predict the next number one song and I've got a prototype of that happening right now and it's awesome and I don't need a PhD to be able to do this. Well Chris, great to see you in the cube. Love the Mojo. Love Shazam. Big fan. You've got Avery Wang who I know. Shout out to him. Great company. Great success. You're a verb. When you're a verb, it's a good thing. Shazam it. Hopefully coming soon to the cube videos on our app. We're going to work a deal. We've got a media company at Discount Park. Speaking of verbs, I've been cubed. You've been cubed. Exactly. Exactly. She's been cubed. Been cubed here. It's Splunking. We're Shazamming. We're cubing here in our land. We've got more live coverage after this short break.