 Okay, welcome back everyone. We are live in New York City. This is Silicon Angles theCUBE, our flagship program, where we go out to the events and extract the signal from the noise. We are in New York City as part of the Big Data NYC in conjunction with Strata Hadoop. We are a hundred yards from the Javits Center, where it's the center of the universe on the Big Data conversation where this big story is beyond Hadoop, Hadoop next. It's no longer Hadoop. Hadoop World has been sunsetted out of the vocabulary. It's now Big Data, it's called Strata plus Hadoop. We're going to start to see Hadoop World phase away. I put the question out there. Dave should be called Spark World, our Big Data World. Anyway, we've got all the action covered. I'm John Furrier, my co-host, Dave Vellante. Our next guest is Cube, I'm Jim McHugh from VP Marketing at Cisco. Great to see you again. Good to have you. Good to be here, guys. So great to have you on. You're like a coach that comes in from the, you know, just one of the sewer bulls being guest analysts with us. So I want to pick your brain first. It's like an industry analyst type. Okay. The Hadoop word is kind of becoming regulated into just storage layer. And we're hearing that theme where even Cloudera announced that, you know, hey, we've got this new fabric, access. You know, customers want to do whatever they want to do for access. It's kind of here. But then the conversation is shifting to the analytics piece. Yeah. What's your take on that? Well, I mean, the reason why it's shifting to the analytics piece is we do all this to deliver business outcomes. And data management and Hadoop is extremely important. And it was so difficult for years that that was the focus. But now people need to actually get insights from the data to pay for those deployments or you're not going to be able to do it. And it's interesting because even in our world, a lot of the revenue is going to come from the data management and the Hadoop components of it. Just a few nodes of analytics actually deliver so much insight and so valuable that really a lot of our analytics partners' business is taking off. And Cisco, obviously, your number, your market share at UCS, congratulations. You guys got some good performance and doing well. But I want to get your take on as someone who's an incumbent in the market, certainly on the networking side, moving up the stack to the edge of the network with Internet of Things is going on huge for you. But in the old Hadoop world, when the ecosystem started six, seven years ago, it was the picks and shovels game. But not all picks and shovels are working. So I got to ask you, you're in that business ultimately under the hood, there's some stuff going on. What's your take on that? You're doing well performance-wise, so customers are resonating with your solution. Which picks and shovels are working? What's not working? Well, I mean, so I think what you're seeing is people are looking at it for different scenarios. I mean, I used to tell all the Hadoop vendors, especially when it was so difficult and early on, people would vendor hop, like, oh, it didn't work with one. I'm going to go to the other, go to the other. But I think you're starting to see so many deployments that are working now and that it's becoming reliable. We've all made the switch, everyone said, you know, it was MapReduce, now it's Spark, right? Spark is the new thing coming and it's actually, I'm a big fan, I've been a big fan of Databricks guys for a while and what they're doing. So I think that will take off and that will continue to change. Companies like Platform of Embrace Spark and it's starting to pay off for them. We're seeing it in our customers when they go in. Because it's simplifying the process. Nobody really wants to go through, oh my God, I got to do ETL, then I got to bring in integration, then I got to go the whole component. So that's why Spark's paying off. But I think there are other vendors who are doing it as well. I think people are starting to get the data integration, data prep, you're getting a lot of insight coming from the trifactas, the Skytrees, the Puxatas, all these guys are doing a really good job of saying it's not just about managing the data, you got to clean it up and get better preparation out of it and then the analytics guys come in and are getting a lot of value too. So I wonder if we could come back to, John sort of threw out IoT there. It seems like a lot of the internet of things, the internet of everything you guys call it, is going to happen in the cloud, the data is going to live in the cloud, in mega data centers. I wonder if you could give us your take on that, particularly as it relates to the cloud and mega data centers and where Cisco plays. I think the way we look at it is, first off, 40% of the world's data is being created at the edge. It is going to be internet of things, you know, internet of everything is just adding the people process to that. But the internet of things is going to create an incredible amount of data, incredible amount of insight, that's where the machine learning comes in. A lot of mega data centers. I think it varies based on what people are trying to accomplish, right? So honestly, I think it's not going to be mega data centers when it comes to IoT, it's going to be small-fault computing that actually is going to let you actually get it. Let's talk about mega data centers. I came up multiple times now in the past six months on theCUBE, which is the death of the data center, the mega data center is dead. That's what the quote, headlines on the blogs are. That's actually not the case. The mega data centers are getting beefier. If anything, there's more of a distributed data center model going. Do you see that trend? I mean, debunk that myth like, okay, the mega data center is not going away. You obviously would agree with that. You're in the data centers. At the end of the day, it's the applications that are mattering really still. And the data is great, and it's really important, but it's the importance of the application. So when you have high application utilization, you have a big data center to support that. When you don't know what that application utilization would be, you put it in the cloud. That's what customers do. And once it starts getting enough concentration, enough action, and that application, you're going to put it somewhere where it has the cost benefits. So it's swinging because applications are coming and going where you're trying out new applications, new ways of doing it. Internet of things, yeah, it spins up a whole new way of writing applications. Their lifetime is less. Yeah, yeah. Well, you guys nailed it. I got to give, again, I'll let you do some props on the numbers of your performance. Cisco had unified computing, not innate stairs early on. So we're kind of throwing you into the buzz. Ah, there's no way adjacent markets, the chamber's vision of, and going into the service. That's not your core competency, all that has been kicked around and kind of like proven wrong, so that's cool. But the word unifying is coming up. Unified cloud is VMware. You got unification message being taught here. You have that platform, you've done it and been successful. Now you're in this big data world where there's a lot of partners you work with. So integration's critical. So what is the key success to make integration work and to truly be unified in big data? In big data, we have, I mean, a lot of our solutions, even at Cisco Live, our show, where we did vertical. We did big data and analytics solutions inside each of our verticals, healthcare, financial services, security, transportation, and each of those solutions include at least two of our partners. We did healthcare with Cloudera and Tableau. We did security with MapR and Platforma. We did transportation with Splunk, I think maybe that was just Splunk, but each of these components are a different partner coming in because they're different expertise. And I actually think there is a little bit, that integration becomes important. Again, data management without analytics is data management. Analytics without something to take it from is not going to get you there either. So what we're seeing is these partners are working so much better together to drive the solution and it's just part of what is being expected by customers. I was at a partner event yesterday and it was a lot of our joint customers there. So I went up for the reception afterwards and talked to them, some big financial services companies, but also just about every industry you could think of was here and it was part of that partner. So is Cisco providing sort of the leadership to get those guys to work together? Cause they don't sort of just organically start working together. So the honest thing is we've actually been proactive. We've been really proactive is much more than you would think from an infrastructure provider bringing together the big data and analytics community and talking about it. It's like going to be their crowd chat later. A bunch of cool people coming together to talk about this stuff. But I think that has brought it together but it's really customers are demanding it. It is, give me the infrastructure that's reliable, give me the data management that's going to work and then give me whatever I need to bring in the multiple data, so data integration and then I need the analytics. I need the business outcomes and we're obsessed with business outcomes cause our customers are obsessed with business outcomes. Now you're speaking tomorrow, keynote, let's get some kind of cool examples. I wanted to tee that up for us a little bit. Yeah, so keynote of five minutes, by the way. Yeah, that's the way it works. Speed keynote. But yeah, so a little bit we were talking about the edge, what's going on and I'm kind of obsessed with the data that's being created at the edge and so I'm going to be talking about two examples. Our friends at Splunk invited me up to a race day in Sonoma, so I had a Porsche Cayman took it up there for the day and went around the track and we put these little carvoyant dongles into the dashboard and was taking all the diagnostics coming out of it and passing it up through their cloud to the Splunk system and in real time we could see our maximum speed in going from there. So I'll be showing a geospaced map with everything laying that out and talking about how that can work. And you were driving? I was driving, man. So just give you a report card, like how do you get the right stuff? I will say that Godfrey and the rest of the Splunkers drive much better than I do, so I'll give them that. They've been through a couple of times. Did you get a chance to go back through and try that out? I had a point Godfrey buy with his Ferrari twice, by the way, so he's actually a really good driver. And then the other one he's talking about is a partner of ours, Dimension Data, long-term partner, one of our oldest channel partners and Solutions partners at Cisco, is actually putting GPS on each of the 198 riders of the Tour de France and actually giving so much better feedback. I mean, so it's really removing the traditional, I got a motorbike and I'm actually taking video and now that's where we know that poor guy is, so we'll be talking about that as well. So talk about the Spark, non-Spark Solutions, but some talk out that you mentioned, Platform, they also started without Spark, so they're doing actually in memory, but now they're doing Spark. Is there, can you have a balance between Spark and non-Spark Solutions? In your mind? I'm glad, so I think today you can. Obviously Spark will continue to grow in points, but there's other solutions coming. I mean, so the Apache Spark team, the Databricks team, they need to keep innovating or they're going to be challenged by something else. Nothing is here forever unless it keeps innovating. I mean, this purpose builds stuff that's working, engineered software can be in use case. Everyone of them says we support Matt Reduce for BigBatch and then we support whoever. What's your biggest takeaway over the past five years looking at the ecosystem? What's changed the most in your mind? In big data analytics? In big data analytics, we're here. Strata, Hadoop, the word Hadoop is still in there. Strata is, oh, rather they're not really a vendor, they're just an event company, but Hadoop is the ecosystem, that's Cloudera, leader and their number one app on all clusters with Cloudera Manager. You got Hortonworks, they went public. I mean, those are two big firms hoping to be the next Microsoft, not looking good. I mean, or is it? And by the way, the MapR guys are doing pretty well as well. I mean, but they're all doing well, but I think there's going to be more change and I think it's going to be more to the analytics. We'll see how we're, you know, Mike Olson and team and the rest of the Cloudera guys go, but they are strong. I mean, they're doing really great stuff out there. I just think customers are saying it's time to step it up. I'm not just here to manage my data. We've got tons of great examples with them. And it's exciting because we are changing people's lives. We are changing the world. Atlanta Health Care with Cloudera, collecting all the diagnostics off the neonatal for vital signs. And then actually by studying those, we were able to help them understand that they need to change their care practices. So things like that, when you're part of those kind of solutions, you feel, you know, hey, this is actually really compelling and you are doing great stuff. Well, somebody was saying on theCUBE recently that this whole ecosystem reminds them of the old Unix wars, you know, days of the Unix wars. But there's a lot of things that are different. Being an old Sun guy, I don't know, I don't see any command line gray hair guys. So I wanted to ask you, so, you know, what are the differences? What are the similarities? I mean, yes, there are some similarities, but a lot of companies, you know, particularly you guys, IBM, certainly talking about business outcomes, much more so. And the customers, I don't think they're going to let the industry take 10 years to try to figure it out and then realize, oh, well, we just need Linux. I mean, what do you see as the similarities and the differences? Oh, well, I think, well, the main difference is that it's just a whole new world of how we're looking at applications and data. I mean, back in that day, it was big applications, monolithic, you know, we ran it on our machines, like, you know, our big sun machines back then. The box, shove all the data into that box. It was box-centric. And today, I think you alluded to it, it's distributed, right? Applications are distributed across from the edge, data center, you're, you know, out to the cloud. And, you know, let's take a simple example of the parking app. A parking app that allows you to find a parking spot and then pay for it. It starts by collecting the data of available parking in the city, right? So you're collecting all the information. That data's then being communicated to the digital signs around the parking garage and around the city. And it needs to be communicated back to the data center so that it can be communicated to the app, right? All that happens before the user even gets involved. But also what's going on, it's doing real-time pricing in the back. We basically have the equivalent of surge pricing capabilities of making that for parking to try to encourage people to park in different locations. So you have data being collected at the edge, being communicated to digital signs, et cetera. Then it's being also communicated to the data center. But when I go to purchase that, I got a whole transaction processing engine, all these different components, and I may be doing sort of like predictive components in a Hadoop area could be on a cloud-scale basis. All that is one simple app that everyone thinks is like brain dead, but there's a lot going on behind the scenes. So let's sit back to this integration thing again because you mentioned the customer thing and I want to get this Hadoop next concept. At Hadoop, some of that came up a lot, which is it's beyond Hadoop. You guys were really talking about that last time in your team about it's just bigger than Hadoop. What specifically are you seeing, obviously Spark is one great example you have, streaming, real-time, and you can talk about Internet of Things is trickling it down of data, there's real-time data. Well the funny thing is there's a lot of data that will never make it into a database. You don't need, if I want to know how many customers are coming into my store between 11 and one, I don't need to put that into a database to collect all that information. I can do an average and do the analytics at the edge. So if I want to know what's going on with a drill bit and whether I'm actually going to shut it down or keep it going, I'm not going to have time to bring it back into a Hadoop database. Make the call. Yeah, it's like it cut it off now. But if I'm actually doing seismic analysis across, yeah, you definitely want to put that in there as well. So there's so much that's going on in data that you're just going to make decisions on the data before you have a chance to store it. Think about how mind-blowing that concept is. Some data will never actually make it into a database. That was never in the case before. But it's cool. That's what the edge is about. You're going to have to make decisions. And the thing is, the analytics will, a lot of times have to be there, either the machine or the human. And it's going to be really interesting to see where those two play off. It's interesting. Cisco has a core company. I'm a big fan of Cisco. Obviously watching you guys build up and you said you were at Sun. So all the early computer industry pioneers of all of them right now, Cisco and even Oracle's out there at least on the database side. So they're going to be, oh, database, of course. I disagree with that statement, all database. But you guys have always been wire speed. Packets had to fly around the network. And then the edge of the network was just a branch office to Cisco. This is years ago. That now extends out to the wearables, the watch, the computers, wearable computers. The car, the bike. The car, everything, right? So, does that- Somebody said last year whether the car is the biggest wearable there is. Yeah. Yeah. And then you got networking inside the car. You have a subnet inside the car. So that changes the game. What is going on at Cisco? Share some color, anecdotally, it'll be specific and reveal any secrets. But does Cisco look at that edge? They talk internally like they've moved up the stack. Has Cisco thought, hey, we're moving up the stack? Yeah. Well, Cisco believes in connection at the edge. Cisco actually leads the IoT, we have the IoT World Forum. We're a big part of the steering committee and driving from that. Moving up the stack. I mean, that's kind of, that means so many different things with so many different people. I mean, it's like- It means a lot to networking guys. Like, what? Moving up the stack? Yeah. So, yeah. So that- Everything is application-centric. Everything is application-centric. You have to make your network in support of the business requirements of the application instead of the technical requirements of the network. And that's a big difference because that means you're going to policy. Once you go to policy, you're being able to change these on the fly. So my question then is, you talked earlier about how the world used to be just sort of box-centric and now it's sort of focused on the applications at the edge in particular. How are customers sort of moving? I mean, clearly they got to move in that direction, but where are they today? It seems like many customers are still sort of, they can't get out of that old model. Where should they be investing and how do they get some, where they are today to, where this vision meet? You can't get out of the old model if a large part of your business is dependent and running on it. But, I mean, look, everyone's talking about there's the two modes of IT, more and more. Are we by moto or not? That's right, it creates two more stovepipes. I mean, that's what's going on. How do you do that without creating two stovepipes? How do you do that without with having an infrastructure that supports both? Because, I tell you, I mean, there's a lot of talk about, you know, I'm going to spin up this appliance and that appliance is going to solve it. It reminds me of the early days of the app server when you can actually, you could take the reference architecture of the app server, or you could actually take this one that had these little hooks that made your, you know, you could get home faster on the weekend because I could actually call upon them and boom, everything was done. Fast forward three years later when you wanted to try to migrate off that, you couldn't. So I think really the biggest challenge is with people is when they're trying to solve these two worlds of the newer types of applications, the distributive types of applications and maintain their ERP that they don't end up with two completely different systems. There's a false summit there. They get enamored by the short-term gains but foreclose their future scale. That's essentially what you're saying. Exactly. And the pain of doing that is significant. The pain is you can't innovate. What are some of the consequences of that? Just give it to me, Rhianna. You're locking all your budget down for years to come. You won't be able to innovate when something new comes. And look, we've just talked about in the last couple of years we went from MapReduce to Spark, changing the big data space, the analytics has become more important. If my budget is locked down because I made decisions that won't let me do that, if I have to actually keep people around on my payroll because they're the guy that knows how to manage that, whether it's an old Unix machine, or that new appliance model that I went with that doesn't really work, those are the kind of things. I think there is more to open standards and doing that base, which is kind of interesting. People say open standard, Cisco, UCS, is it a, it's not commodity hardware. It's the most commodity hardware there is. It uses Intel, it uses the standard memory. The thing we do different- I think that argument is pretty much over now. I think you guys have put the naysayers. Wait, wait, wait, finish that thought. The thing you do different is- We do is software. Yeah. It's a software defined server. It always has been. We create service profiles, which allow you to actually put the personality of the server in software. I always say it's the SIM card, like the SIM card is to the cell phone. That's what our service profiles are. Jim, thanks for coming on theCUBE, really appreciate it. Great to see you again. I want to give you the final word as a quote, guest analyst on this segment here. What's going to happen at the end of this week? Okay, what are we going to, customers are going to look at us and say, right now, forecast out what's going to be the outcome of this week. Well, hopefully everybody's saying there's that amazing keynote that took place at about 8.45, Wednesday morning. It was far too short. They should have given the guy more time. Besides the crowd chat we're going to do at 11. At the crowd chat we're going to do at one o'clock this afternoon. I think people are going to, it's going to be just that. There's going to be much more demand for the integration of the different providers, the different solutions. I think it's going to go to much more rapid succession. I think we're going to see people talking about how they're going to bring that data from the edge, get that insight. I'm looking forward to the world where we have, I don't know if the right word is a dashboard, but it is a dashboard that lets me know that data that doesn't belong in a database because we're getting the decision so quick. I can see that at the same time. I can see my deep analysis and insights I'm getting from a Hadoop plus analytics provider. The world is changing. Some data will never hit the database as Jim McHugh. That's a true statement that we see that as fact happening. You're watching theCUBE. Henry, if you want to see more of what's going on with Cisco, Cisco has organized a crowd chat with just everybody. It's not a Cisco related crowd chat. It's a bunch of cool people talking about Strata and Hadoop. Go to crowdchat.net slash strata Hadoop or just go to the hashtag strata Hadoop and join the conversation, be a thought leader, be part of the conversation. More after this short break here, live in New York City with theCUBE. I'm John Furrier with Dave Vellante. We'll be right back.