 Live from the MGM Grand Hotel in Las Vegas. Extracting the signal from the noise. It's theCUBE, covering splunk.com 2015. Brought to you by Splunk. Now, here are your hosts, John Furrier and Jeff Rick. Okay, welcome back everyone. We are live here in Las Vegas at splunk.com 2015. This is Silicon Angles theCUBE, our flagship program. We go out to the events and extract the signal from the noise. I'm John Furrier, the founder of Silicon Angles. Joe, my co-host Jeff Frick, the GM of our CUBE operations. Our next guest is Snehal Antani, CTO of Splunk. New CTO on the job. Welcome to theCUBE and welcome to Splunk. Thanks, appreciate it. Like, I'm welcoming Splunk. I don't work there, but we've been here for four years. We're like incumbent, like, you know, a tech in the ecosystem. We can't get rid of us. We love Splunk. We've covered it for four years. You're new. Great. What's it like in the job now? You know, four months. What attracted you to Splunk? Why did you want to work here and tell us what happened when you got here? Yeah, so it's an absolutely awesome job. I feel like a kidney candy store. So first week on the job, I've got customers coming in talking to me about high-end lasers doing chip manufacturing and streaming data for preventative maintenance. First day on the job, learning about a completely new industry. Second day on the job, I've got a big bank coming in talking about fraud detection and money laundering. Third day on the job, I've got government agencies inside a threat. Fourth day on the job, it's connected cars from the largest manufacturer. Same exact product, completely different use cases, and it wasn't just PowerPoint. They had already done these problems or solved these problems. They wanted to know how to take it to the next level. So I mean, just imagine completely diverse problems, completely versatile technology, and, as I said, kidney candy stores. It's like the coolest ride so far. So Splunk is a great product because it was a four different use cases, core platform, but now technology is enabling other creative things. What other things technically are inside the platform that are going to, I mean, let me rephrase it. The platform is enabling technology. What is coming on top of Splunk? We see analytics, that's pretty obvious. The killer app, what are the technologies that are out there that are on your radar and Splunk's radar in terms of the key enablers to connect into the core platform? Yeah, so one thing is, let's level sit a little bit on, what do we mean by core platform? So it's kind of the story I've been telling around the history of Splunk, if you will, is I think about, imagine a mainframe environment. In a mainframe, it's a converged infrastructure. You have all these VMs running, and you could route all the logs from those VMs into a single Z log stream, and everything was right there. It was all fit, you know, there was no moving parts, there were very few moving parts. It worked. Well, then 10 years ago, we had SOA and distributed services. And then the Yahoo's and the Google's of the world, you had very large-scale distributed searches. And suddenly, not all that data was in one place anymore. And when I was the CIO before this, 34% of my operational tickets were requests for logs. So the origins of Splunk were bringing together all of that data, making searchable, indexable, being able to do an analysis, and really facilitating that investigative loop. So if you think of an outage, you get this error, and you have no idea what caused it. So showing all the logs errors from five minutes before and three minutes after, what just happened? Well, that's interesting. Let's go search that instead. So it's this investigative cycle. And you can start doing really cool, advanced correlation as a result. So that's the core. And then what we did was the investigation for an outage was very similar to the investigation for hunting a breach if you're a security person. So we suddenly found ourselves solving really interesting security problems, and we didn't even have a security product, right? We won Best Fraud Detection Award a few months ago. We don't even have a fraud product, which is cool. So the core is very versatile, and we've got these lenses into the core. So we've got a security lens with capabilities in Splunk Enterprise Security that it's a user experience, a set of questions oriented around that security analyst, that hunter. We've got an IT operational lens into that same data that's oriented around that ops person and the kinds of things that they deal with on a day-to-day basis. So what we have is this platform, very versatile, and different lenses into it. So that's kind of the Splunk company and the product today. What's cool is IoT is an example. It's just another data source for us. We can already ingest data at scale. We can already speak or ingest data for many languages and protocols and dialects. We can do very high performance sense and respond. We can do geographically distributed search. Things that are critical to IoT. So that's just another data source and all the magic and power Splunk just plays right to our question. You mentioned data is not everywhere. You mentioned some hyperscalers, whatever they're called these days, web scalers, hyperscalers, the large web company, Yahoo, Google, we know who they are. But now that's not the enterprise. But DevOps now is coming on strong. You've got this agile rapid development environment. You've continuous integration and then you got the IT ops converging in on the dev. That's causing a great migration to the cloud. Cloud and analytics kind of go together. So do you see that as an opportunity for Splunk and if so how? So last year I was on stage as a customer and I talked about the transformation I drove at GE Capital. And I'll tell you about DevOps, for example. One of the mantras I pushed out was, in IT we're not in the business of selling religion. We're in the business of selling candles because every religion needs candles, right? So DevOps transformations fail or agile transformations fail because the agile team will argue against the waterfall team and it's a very subjective, very religious debate. Or the DevOps team will argue with the ITIL team and a very subjective, very religious debate. So one of the key things that we did and we were successful in our transformation was we sold candles to the app developer platform. If you want to go live, you've got to have secure code and you've got to make sure that you've got zero vulnerabilities. Well, you can either wait till the end of the release and get a whole bunch of findings or I can give you a report every day through this automated service. And then path of least resistance, developers use this automated service. So you artificially create bureaucracy to create a path of least resistance to drive the right behavior. Well, one of the cool things we did was I had a bunch of automated steps to validate code was of the right quality and a completely automated deployment pipeline and I splunked all that data. So I can tell you in real time, my best developers, my worst developers, which developers struggled to write secure code, which developers struggled to write performance code, my best and worst contractor vendors. I'd picked them up against one another and say, you know, contractor A, you had the worst code this quarter, you better do better, you better improve or you're not going to get the contract renewed. So the key thing in here is transparency. Because you can't pull that Jedi mind trick of there's nothing to see here, all the boxes are green because I've got the data at my fingertips as a decision maker. So I think that really changes the game and how we develop software, how we run an organization and that level of transparency allows the business to move at market speed, where they can quickly make decisions and pivot and iterate based on the actual data at hand, not some subjective piece of information. That's awesome, that's awesome. So I got to ask the next question, this is great, great conversation. We could go all day actually, this is so much fun. Data silos, big, big issue that's kind of not being talked about much. I mean, silos in general love to be talking about, people love talking about breaking down silos, but data in particular, people have recognized that data is a competitive advantage. Data driven, data driven sales, data driven organization, blah, blah, blah, goes to a nice punchline, but people are hoarding their data. So Twitter doesn't share their data with Facebook, Facebook doesn't share their LinkedIn, organization, that's my department. How do you get an open data environment? Because data needs to interact with each other and that's what Splunk has proven, that works really well. Where are we with this whole data open data issue? Is it early stages? Is it a crisis? Do we worry? What needs to change? So let's put that into two pieces. The first is, within an enterprise, you've got these data cartels, which is absolutely idiotic, right? I mean, your job is to serve the business and drive revenue and better serve your customers, yet you've got these kings and queens of the organizations hoarding their data and keeping everyone else off of it. I kind of joke that, you've got three groups of people in an organization. You've got those eager to collaborate. They believe in what you're trying to do, they want to do what's best for the customer. You've got the Eeyours or the cynics that don't think anything's ever going to work, it's too hard, you know. You've got those eager to be offended, which are the kings and queens of the org, looking out for number one. And it's those eager to be offended that tend to hoard the data and create these data cartels. And so what's important is, and this is more of a culture at the top, is if you've got a culture of transparency, if you always put the customer first, and you make everyone measured on customer experience and customer success, you'll naturally start to break these walls down. And you'll naturally start to share because it's these advanced correlations. It's connecting a nugget of information from this data stream, and another piece of information somewhere else that seems completely unrelated, when combined gives you incredibly interesting insight. And that is how you drive much greater value to your customers. I'd say you drive much better profit margins, much greater capabilities, and so on. And so within an organization, we absolutely have to break these cartels down, and those that have cartels put themselves first, not the business first. And I think that's a dying culture. I think that people starting to recognize that that's just not going to respond. Now, separately, I think we can think of them as walled gardens, right? Twitter's a walled garden. If you're within the Twitter ecosystem, you've got access to their data, you do interesting things. Facebook's a walled garden, you've got, if you're inside their ecosystem, you've got access to their data, and you've got high value. I don't know how long it'll take for these walled gardens to better work together. We see that, right? We see open-gov movements and open-data movements within government as a starting point. We see that amongst the web companies. But I think as time goes on, customers are going to demand that level of integration, and I think the industry will catch up to true. There'll always be someone who will have an innovation or an invention that will unwind the data cartels. Right. At some level. It's just a matter of time. You really have to, it's brute force. To get rid of those. But I also think that, I think the value proposition has changed. When data was sparse, data cartels were a source of tremendous value, right? I remember in your sales guide in your early days, you know, your Rolodex, literally the stack of cards was your goal. But now the data, there's so much more data, there's much more value in sharing in unique and innovative ways. So you actually have more power being a data share now than you used to be by being a data hoarder. And I just don't think those data hoarders and the data cartels have kind of figured that out. They can actually have more power by sharing that data in unique and innovative ways. Yeah. You know, let's take an extreme example for a moment. So I had a mentor once who was driving to work one day, had a heart attack on the road, veers off and hits a telephone pole. The impact from the pole, followed by the impact from the airbag restarted his heart and he lived. If you think, it's an amazing story, right? If you think about that problem for a moment, technically now, I've got a smartwatch. I don't right now, but you know, you can imagine a smartwatch. And on that smartwatch, you've got time series data of your heartbeats. You can start to look at anomalies. Your heart rate has suddenly changed or something, doing something different. Well, that on its own, isn't all that valuable. Separately, I've got my skin temperature. Well, suddenly changes in skin temperature on its own don't mean anything. But maybe I'm instrumenting data from the car. And on its own, I've suddenly accelerated or slammed on the brakes or swerved with lateral Gs on its own that doesn't mean anything. But if I correlate those three sources of data together, I've detected an adverse health condition, I can take action, slow the car down, turn on the hazards, dial 911. So you've got seemingly independent sources of data, independent streams of data. You can do advanced correlation and actually solve really interesting, very high value, very high impact problems. And that's all over the place. In every aspect of the organization, in every business and in every problem, you've got these innocuous or completely disparate sources of data that we can integrate, correlate and actually take action. I want to get your thoughts on a trend that's happening now. And I want to compare vis-a-vis the web. And I noticed you worked at IBM in your past life. So the web was a big part of the growth of e-commerce, e-business back then. And I use IBM as an example because I liked their strategy there. E-business was e-commerce, basically, the web. And they had a good read on that. But now IBM's talking about social business and as an indicator. So I'm bringing that up as a discussion point. This new fabric that's developing, the sharing economy as Jeff was kind of hinting out, you got Airbnb, Uber's out there, always being on every keynote, Uber of this and Uber of that. But what's happening is a new social web is being created, a new fabric, a data fabric, a new kind of mobile-first dev ops, all that stuff's kind of coming together. What is this new phenomenon? Is it something as big as the web was as a platform? Because no one goes to websites anymore. I mean, a website is like, okay, it's there. But I also got a mobile app, Native, iOS, Android. I have Twitter, Facebook, LinkedIn, other tools and sharing devices. So a whole new fabric of infrastructure is developing. How do you look at that as a CTO, as someone in the industry? How do you make sense of that? What do you do to view this? So you've described a bunch of technology trends, but for a moment, think about this through the lens of the business. If you're a bank, you only talk to your customers when they owe you money, like if you're a credit card company, you only talk to your customers when they owe you money or they've missed a payment, right? It's a very negative interaction with your customers or if you do leases and loans and so on. And how do you transform that to be a much more positive and ongoing conversation? And so if I can somehow understand that you've leased a piece of equipment, you've leased a fleet of cars for me as an example. And if I understand how you use those cars, I can come back to you in an ongoing basis and say, you know, the fleet of cars that you have aren't the right mix for your business. You know, you do a lot of city driving, you do a lot of SUVs. I think we should change your fleet up to do something else. You become much more proactive in your conversation with the business. And you become relevant. And you become relevant. And you're not just a financier, you are a business partner. You're driving higher value outcomes with your customer base. I think that's what's really important is that the reach of you, of the business is far beyond anything we've had before. The line between digital businesses and physical businesses are blurred. And that reach gives you a chance to transform your conversation to not just be some negative, you owe me money, but really reach out and drive these higher value interactions on an ongoing basis which allows you in business terms to either protect your premium and charge higher prices, increase the spend of your existing customers, capture net new customers. So I think that's what's really important. When you look at an Uber or you look at Airbnb or you look at any of these other new and emerging digital properties, what are they doing? They have this ongoing continuous conversation with the customer. And they're always trying to find a way to enhance that customer experience and drive interesting higher value outcomes as a result. My next final question, because we're running out of time, but it's more on the development side now. So I totally love that answer, great answer. In 2008, I wrote a post on my blog called Data's the New Development Kit. Back when there were development kits. Back now it's just open source, right? So we laugh because we know what that means. So, but now let's flash forward today. If you're a developer, you're dealing with data as a development resource versus I'm going to create a database to store data that's part of an app. So share with us your vision and how you see data as a development tool per se, because data's living, right? So data, just because it'd be active or passive or cold or warm, whatever. But data is now a part of the development. How are developers using data? Yeah, that's a great question. So let's take kind of, I talked with this business operations center and running the business in real time. So I've got a bunch of streaming data coming into my environment. And they're really interesting real time questions to ask that data. But at the same time, as the data gets older, I'm going to age it out and start to ask historical questions of that data and start doing historical analysis. And they start entering into this data science development loop. I've got a ton of data. I'm going to be doing models, building apps in R, checking out my R squared values to make sure the models are correct. And what happens is you've got a ton of data that's historical. You want to start building and developing those models through real time information. So you start sampling that real time data. So what happens is your data science development platform isn't just a ton of data over the last 10 years. Think of like stock markets. You can get 10 years of stock data and do a bunch of back testing. But eventually your development lifecycle is, well, let's start getting real time feeds and let's start testing this model off of sample data. And as you harden that model, you want to push that model to the edge to as close to real time as you can so you can get that fast response. So injecting real time in with historical data gives you a better prism to look at the results. And it's exactly, you get a much better prism, a more complete view. You ask real time questions. You ask historical questions. And sample data has always been a struggle. So you start to really iterate and you have a much more robust data lifecycle strategy, which I think is really lacking in organization. And you'll see things that by mashing the data up like that, it'll make, it makes the overall data more valuable. Right, absolutely. They start enriching that data with your existing BI environment or other pieces of like open Gov data. You start decorating this information and now you've got your core data that's incredibly valuable to you. You're adding or decorating or enhancing it and suddenly you've got a much more complete view of what's going on. You have much more interesting insights as a result. That's awesome. So data will be part of the development process. Another really important point though, and I had this experience when I was at Capitol, is when you've got a spectrum of users from very tech savvy to not very tech savvy at all, which is very typical in the world. I mean, we've got people that run their business on three by five cards. We've got people that run their business in the most advanced technologies. How do you start telling stories with your data? I think that to me is a really interesting problem. So one of the things we did in my inventory finance business, which is a incredibly cutting edge technology platform that we had at GE, we ended up having this data storytelling funnel. So I wanted 90% of my insights to be consumed or my insights to be consumed by 90% of my customer base. That's spectrum of tech savvy to tech illiterate. So the what, what was the insight had to be really easy to use, easy to consume. Think of the email you get after your fantasy football weekend. You should bench Tom Brady this Saturday because he's not good with the strong insight rushing the rain, right? And even though there's very complicated data science behind that. I keep radio out there personally. I know. I'm a past fan. He's on fire. It's absolutely. But he's been tearing it up. But what's interesting there is you've got complex insights and analyses that were done to draw that conclusion, but it was delivered to the customer base in a very consumable way. Now you've got the what. So the what might be the demand for speed boats is going to go up next year. And the dealer's going to adjust the kinds of inventory they're going to order. Well, half of those people are going to say, why? Why do I think the demand for speed boats is going to go up? Now you need traceability in that insight. Traceability in the outcome. Half of them are going to ask, what if? Well, what if gas prices take? What effect does that have on the demand for speed boats versus sailboats? And then half of them are going to ask, well, well, let's go have a strategic conversation. But I think that within the data science world, within big data and within BI and so on, it's the real-time historical analysis combined with data storytelling that really starts to become a differentiator. Yeah, we're getting over the hook here. We're getting the hook. We're way over, but excellent. We can go for another hour. Of course. Love to continue this conversation. Thanks for coming on theCUBE and congratulations on your new role. They're psyched to have you, I'm sure. Appreciate it. I'm glad to be here. It's a great time. Thanks for your time here, too. Thank you. I guess theCUBE, it's our fourth year and we love talking about Splunk. So much diversity and so much excitement. It really is a, not great for a kid to be in a candy store like this. So we'd love it, too. We'll be right back more with you this short break.