 from San Jose, in the heart of Silicon Valley. It's theCUBE, covering Big Data SV 2016. Now your host, John Furrier and George Gilbert. Hey, welcome back. We are live in Silicon Valley for our third day of Big Data Week coverage, which consists of Big Data SV and Strata Hadoop. I'm John Furrier. This is theCUBE, our flagship program. We go out to the events, extract the signal from the noise from my host, George Gilbert, and also wikibon.com and our next guest, Boudard Soutkar, VP of GM at Splunk, entrepreneur, sold two companies, been in the business for Big Data. We've had conversations going back years by doing, we hit very active on CrowdChat on this event and all the other events. You're at the front end of all the innovation, which I love, why would love to talk with you because not only are you, now we're in the big company, but you're an entrepreneur, so you can, you always see things early. And I like to get your take on a few things. The first thing I want to get out on the table right away is the valuations of the unicorns. We had, holding on earlier, she had a shirt that said, I will cut you as a unicorn, look like an angry unicorn. The unicorns are not happy right now. Cloud Air got chopped down in valuation, 38%, and you're seeing a lot of other companies getting chopped down as well, mainly because this overhyped. But the revenue model, so is the Big Data business model working out and then what's your take on that? What's your thoughts on this overvaluation, the liquidity shortage, and softening, if you will, the capital markets? It's definitely a good question. I think from my view is, if you look at the other Big Data players, I think the issue is the markets in general have taken a hit, and that's probably the same shaving people are applying to the unicorns and the Big Data companies, but there is a real Big Data problem to be solved here. I think we just have to step aside from, as John, you've been telling me for the last four, four years, the movie now, the playbook here, is not about technologies and bits and pieces. The next movie here will be finding the Big Data applications, solve a real problem, a real use case, and that I see is a big value that customers are willing to pay and the top dollars, and that's kind of something worked for me also. If I look at what we did at Caspira, we didn't focus on the technology. We solved a cybersecurity problem, and at the end of the day, if you solve a problem and go address their value prop, customers will come and buy and there's a business model to be made. Yeah, it's easy to go to the low hanging fruit and talk about valuations and make a generalization across the board, which I just did, but I want to kind of go down a little bit deeper. You mentioned solving problems. We talk about that, but that's the ecosystem play. Now, the ecosystem is not just a dupe. It's not just Cloudera and Hortonworks anymore. We've shown this 10 years old how so much more meat on the bone has come around the ecosystem and it's not just a dupe. When we had one guest on earlier said, Spark cut the head off a dupe and the clouds cut in the legs off. So all you got is a torso. So the nicknames now Bob can resist that joke. But no, seriously, the torso is the head dupe and stuff wrapped around it. Right. What do you take on that valuation of those companies? Will they make money and what's the business model? No, I think the Hadoop stack, I call it Hadoop ecosystem stack is different. Like if you take, if you look at Spark has become a part of the Hadoop stack. I see the view from the Splunk side is look, Splunk became the machine data fabric for a while. It's been there. I think that what has happened with Splunk is being applied and also in the Hadoop sector today. So today, if you look at the data is being sent to the Splunk. There are apps being, Splunk is a data fabric. We call it a machine data fabric. There are apps being built. I think that same movie will be played out or playbook in the Hadoop ecosystem. Spark is adding a layer on top of that but you still need to add a machine learning libraries that you talk about and app on top of it. But what we see is we think to your point earlier we see Splunk and Hadoop as a coexistence. We are an operation store. We are a batch system and a real-time in Splunk but there are enough workloads that can be done with us and apps built on top of Splunk. They're actually moved also to Hadoop. So we want to coexist with both of this. I see customers doing both. And we've seen terror data on, Informatica's been on, all the big data warehouse and guys want, then they're making their changes. But the answer is really clear. The coexistence really means there's more tools not just a hammer and a nail or manual saw. You have all kinds of tools in the tool chest here and Hadoop is one of them. So the question I ask you is what do you see as the rationalization of these new tools? Because you mentioned Splunk has this already so why should this happen in Hadoop? So you're seeing Hadoop almost getting trimmed down. It seems like the Hadoop and the Hadoop players have overplayed their hand in a lot of areas where services already exist. So the coexistence and integration is the play. Right. No, I think you nailed it. I see the view as I talked about if you take the view of the Splunk, Splunk Query Language Works Splunk also supports through Hunk the Hadoop platform. Right. So from a customer perspective they go to an application where you build a healthcare application and insurance or security that we are an app revolution. Right. App will really change last four, five years. Every customer that I talk to, they want an app. Underneath whether the technology pieces exist or not, it's okay. They need a platform to build an app. That platform could be Splunk, it could be also Hadoop ecosystem, but they want to build an app. People consume applications, people consume users and they want to build an app economy right now. Before George gets his question, I know he wants to get a question and you had a tweet, I want to read your tweet, I want to get your thoughts on it. Let's have a debate discussion on encryption for data in Hadoop. Cassandra, MongoDB, Oracle, question mark. Customers keep asking about encrypting big data which is valuable and sensitive. What are people's views on this Hadoop world? So we've been polling people and the general opinion is kind of deer in the headlights. Look, they're like, what's going on? What's your thoughts? I mean, you're in this here, Splunk, obviously, big on security. What is the security equation in Hadoop, in Mongo, in Cassandra, in Oracle? No, I think that's very important. I think given that, at least the last one month, given the debate between Apple and FBI in encryption, encryption is on the top of everybody's mind. I think that movie will pay out in all industry. That means you need to secure the data that's sitting in MongoDB. You need to secure the data sitting in Hadoop in all of them. Now, encryption alone is not the problem. As you know, it's the key management. Who owns the keys? Do you want to give the keys owned by the vendor? Does the customer own the key? That whole thing has not happened in the big data. And that's what I was talking about. Well, Larry Ellison said on stage, I was there at Oracle Open World, encryption is always on. The default is the encryption and the exception is turning it off. That seems where the Oracle wants to go. No, I completely agree. I think encryption is beyond. The question is who owns the key, though? It's not the on and off the encryption. Do you manage the keys as a customer and user? Does the vendor is going to own your keys for you? In that case, what will happen? So I think that's a big debate that's happening right now. So that goes back to the app. You can encrypt things too, but you still have the insider threat. Yeah, you still have the insider threat. But the question is, good news is, if you have encrypted data, and even if an insider takes the data, if he doesn't have the keys, then your data is still secure, right? I mean, there's angles to play out there. You were talking about applications, and with Splunk, it's been generalizing more to be more an application-centric thing than just a sort of out-of-the-box single app or a dual app sort of solution. And now that Hadoop is so widely deployed, and Splunk as well, how would you tell customers when to think about building in Splunk an application versus when to think about building on Hadoop? That's a good question. I think the way in which I see it is today, Splunk has been the machine data fabric for almost 11,000 customers we have, right? And they have premium applications, we call it whether Caspira is a company, we have enterprise security, we have IT ops, ITC. So there are enough premium applications, and we have ecosystem of partners who are building around Splunk. Now, if customer wants, we give the choice to customer. If customer feels that the data should go into Hadoop for archival reasons, there are tools with Splunk and other people will help you move the data from Splunk into Hadoop. But the app still works. The key, again, goes back to the user experience, the application that you are trying to solve. That's what I think. I think of the Hadoop ecosystem is playing almost like what I call in the storage world, the tiers of the storage. What you do with your data, how do you get insights? How fast can you get your operational insight? What can you do? What kind of actions can you take on? What's the workflow? Those are the areas that Splunk does a very good job out of the box quickly. If the Hadoop ecosystem vendors want to do that, they need to solve very similar problems like Splunk is doing, but also integrate Splunk. All it is doing that in applications. So I think if you do that in applications, whether the data sets in Splunk or in Hadoop, it'll be a complementary solution. Matt, you talk about the startup landscape. Again, you have a unique perspective. You're at the big company now. You see what the Splunk machine's doing in terms of the management and the product roadmap and whatnot over there. Obviously secure. Looking forward to .conf again this year with theCUBE there. At Splunk's conference, a little plug for Splunk. I've got theCUBE coming in this year. But the startup landscape is certainly still dynamic and strong at the series. Seed and series A. Series B's are very difficult. If you're a series A company looking for a series B, it's pretty tough market right now to get that funding because the traction has to be on the revenue side. And so what would your advice be? And what's your take on the overall landscape right now for those entrepreneurs out there? Because this is where the breakout brand's gonna come in. I was saying, but Peter Bear showed a slide, no one really has 10% share of anything. So the games wide open, the next Uber-like companies might come out of this crop of innovation. Right, I think this is very good. So here is my mantra that I've applied at least for my startups. In this day and age I call the Moore's Law. In 12 to 18 months, if you're a series A company and you don't have a million revenue, you're in trouble. Because technology changed so far. 18 months, a million revenue a month or a year? A million revenue in a year. Okay. Right? If you're a startup and if you have not reached a million revenue run rate in 18 months and you're applying for a series B, I mean you're going to get downroads or maybe you'll be not getting a chance to even finance it. So all the series A companies, the best thing they can do is find customers. And again, revenue of million dollars don't get them for $10,000. I would even say that in this market it might be tight. It might be a million dollars a month or a $10 million run rate. Because you're seeing so strict revenue focus right now. It seems to be. It is. If you have technology, they'll give you a pass so they can see that. Right. And also the type of customers you get it. Again, there are two ways I call it. Get the big game hunters. So getting a million dollar deal or half a million dollar deal customer is more important than getting 10 customers of $20,000. Right. So it's important if I'm a startup company, I want to get the big account. See, can I repeat them? If you get four or five deals of a half a million dollars, a million dollar account, then you know there's a business model to be there. It's got to be repeatable, which is cool. So the final question on the entrepreneur thing is, the barrier to entry. It used to be when I started my first startup over 17 years ago, getting into the enterprise was very, very difficult because you had established companies. That was the application boom during the people soft days. And so the bar to get in was seriously high. No one buys from startups. That was kind of a thing. Huge sales costs. Then the cloud era came during the era of Amazon, startups had easy spinning up of solutions and we're getting penetration in the enterprise. Now we seem to be going to an era where it's like now hard to stay in the enterprise. It's not hard to get in. It's hard to stay, which brings up the integration conversation. What's your take on that? Do you agree with that statement? And could you share your thoughts on how a startup can win, get into the enterprise and what it takes to stay there? No, very good question. I think my advice to startup there is don't be an island. I mean, this is a great example of what you bring in is if you walk into an enterprise, become, you tie yourself with all the other products. You tie yourself with Splunk, you try with Palo Alto, tie yourself with Oracle. I mean, you cannot be an alone. If you solve a piece of problem and you're an island and the user is using that product and he's not, has a workload with other products, you'll be taken out immediately. So I think a good startup is something you have figured out all the ecosystem of the partners they want to work with. And they have a tight ISV integration both at a technology level and a business level. That is a value that customers are looking for. That is one plus, one plus one is not a three, it's 33. And that's what Peter Bartz was saying. Yes, they're about freeing the data and that's making the data frictionless. Again, back to the value of the digital capital, the data. That's right, right comment. Great. What's up with you? Anything you're excited about these days? Final question. It's like, what are you kind of looking at right there in the marketplace? What do you got your telescope aimed at? I think that my big thing is big data apps. I'm a big on apps. That's what my message today was here is I think people are too much focused on technology. Even with the machine learning libraries that we talked about in the crowd chat also, I want people to build app. Underneath is the libraries and technologies, et cetera. People need to focus on a problem and solution first and the apps. That's great vision. I would agree with 100%. Great to see you. Madhu, entrepreneur, very successful entrepreneur. Repeat entrepreneur, now working at Splunk as a senior executive. And we'll be at Splunk on September 26.conference is there, annual conference. It's always going to be a blast there. Last year, we were doing so many Cube interviews. I thought I was going to pass out. We might have to have two Cubes there this year at Splunk conference, so it's a great company. We'll be back more with live coverage here in Silicon Valley for the Cube, extracting the signal from the noise. Go to Twitter and search Cube Gems. Hashtag Cube Gems and search hashtag cube cards. You're going to see photos with quotes and video highlights from these interviews. Of course, go to youtube.com slash silicon angle for all the videos and silicon angle.tv to check out where the Cube's going to come next. We'll be right back with more after this short break.