 from London, England. Extracting the signal from the noise. It's theCUBE, Covered, Discover 2015. Brought to you by Hewlett Packard Enterprise. Now your hosts, John Furrier and Dave Vellante. Okay, welcome back everyone. We are here live at HP Discover in London, England. This is theCUBE SiliconANGLE's flagship program. We go out to the events and extract the signal from the noise. I'm John Furrier, the founder of SiliconANGLE, Joe McCose, Dave Vellante, founder of Wikibon.com. Our next guest is Tom Bradich, GM and VP of Hyperscale Services and IoT Systems at HP Enterprise. Welcome back to theCUBE, good to see you again. Thank you, it's great to see you. That's a mouthful, we've got a lot of light, a long title there, there's a lot going on in IoT. It's converged, it's hyper-converged, it's now composable, it's energy. And of course, IoT is the killer app now. Lot of hype, lot of reality coming in mainstream. Give us the update, you must feel pretty good. IoT's central in the transformation areas. Let me say, I believe IoT has achieved celebrity status today. So there's a lot of people talking about it. There's a lot of sidling going up because it's hot. It's similar to the new dot com, the new big data, the new cloud and IoT. But it does have tremendous value and we have some experience as a company, as HP Enterprise as well, I'll tell you about. But just think of it like this. Imagine being perpetually connected to things and people. This notion of perpetual connectivity gives great value and companies are discovering that, even our own company as well. And I can detail some of that for you. But my kids were young, I used to read that book to them. Thing one and thing two. And I always say internet of things is not just machines, it's people. People are things too. And so that's a big part of the mobility story. What's your thoughts? No, indeed it is, and not to dehumanize people, but it is the T, the things. So are tractors, so are turbines, so are factories, et cetera. But the people dimension is interesting for health and safety reasons as well. So there's an entertainment dimension, health and safety. And we like to portion the internet of things into two broad categories, the consumer internet of things, and something that would hang on you like a wearable in your home perhaps, and then the industrial internet of things. So having those two dimensions and looking at them, there's a lot of core technologies that apply to both. But again, if we go back to perpetual connectivity, it offers three main values. And I like to call them the three M's, such as you can actually monitor and understand the behavior of things and people. The second is you can maintain, and the famous example there is you can upgrade functionality in a car, a smart car that's connected, or perhaps a cell phone. And then the third M is monetize. You can actually promote an action by a person or by a corporation or a company to sell up, to move to another particular product in the line, et cetera. Dr. Tom, I got to ask you the question. Advice for CXOs, because let's just hypothetically, I'm a CIO, CXO. You know, I have a lot of assets. I got campuses, I got machinery, I got manufacturing, I got retail. I want IOT, and I get these turbines and some stuff outside the perimeter of maybe the network that I might want to put sensors on. Sure. What do I do? What's the action item for me? What would your advice be to me? Yeah, I would say to a first procure a consultant who has experience, who's done the particular deployment, let's just pick the energy sector. I personally had experience there in deploying at power plants a smart monitoring situation to look for issues and avoid brownouts and blackouts and maintenance. So number one is to procure some with some experience and then also start small. Procure some sensors and look for something like to be specific of vibration and what the vibration can tell you about the maintenance of that system, whether it's going to fail or whether it needs to be maintained or have a service or something like that. Start small and then you can grow out with the number of sensors to include perhaps temperature, moisture, particulates, et cetera. Start small, get some expertise and join the IoT expertise with the domain expertise on how energy works and is delivered. So connect what you have? Connect the unconnected, as we say. Well, so can we talk about that? Yeah, sure. Because you've been talking about it all morning as John and I have. So it seems like there's a lot of assets out there that aren't connected. Yes. So can you talk about the dynamic in the business case of, okay, do I go and connect those assets or do I try to define my perimeter and try to get as much leverage out of the perimeter as it exists today? What are you seeing in the market space? Yeah, actually there's a big move to connect the unconnected because the assets are invested and they exist. And we do have technologies to do that. So holding that technology thought, it's to gain, you connect a thing in the internet of things to gain insight from that thing to do some sort of an activity that will help your enterprise. So it's all about insight from the thing. Let's take a turbine that burns fossil fuel and generates electricity. If I connect that thing, I want to gain some insight. There are three types of insight. Business insight, engineering insight, scientific insight. Business insight would be, well, how much inventory do I need to move to a particular location? Where is my sales force going, et cetera? Engineering insight is I'm conditioned monitoring the asset and is it going to fail? Or is it an optimum use of electricity? Scientific insight is, and I consider the medical world in there as well, is the tumor benign? Is the health status of the individual? Or is that a new subatomic particle that we've discovered because that's the thing we're looking at? So you have to determine, to answer the question directly, what insight do I want to derive that will affect the action I take in the business? And that will determine the business case as to whether or not I connect the unconnected. Right, and sometimes they're airtight business case. And it's not necessarily trivial to connect that on. Explain the airtight business case. You mentioned the, it's so obvious that you have to do it. That's what in by an airtight business case. Some are unproven, so there's no history. But to understand, let me put it this way. Think about the business case with converting an unplanned outage, a failure of your car, your dishwasher, your computer, to a planned outage. In other words, wouldn't it be great through the perpetual connectivity of the Internet of Things to know if your automobile will fail and when? Then you can say, oh, it won't fail on my way to an important meeting. Rather, I can take it in and get it fixed. Oh, my washing machine won't fail when I need that shirt for tonight's meeting. It will fail effectively taken out when I can know about it. So that will cause consumer satisfaction to go through the roof and therefore the business case begins to fall out there. I just love Internet of Things because it brings the predictive, prescriptive analytics which is a big data opportunity and also marries the cost reductions or Moore's Law, if you will, for this hyperconversion now composable. So it's kind of geeking out in kind of two areas, right, so yeah. And now you've got a whole nother area of the outside of IT. You have an amazing, you got Telematics, Scott Weller was just on earlier this morning. He had a Telematics background. You have a background in IoT and another outside of IT. Those worlds are colliding. Share your vision on that because you had an interesting perspective on that prior to going live here. Share your thoughts. It is a new convergence and I'm going to pick up on what you said earlier if I can give you a geek term. It's cognitive prognostics. Well, what does that mean? That means predicting the future. Who does not want to predict the future? You want to know, you know, when you're young, who you're going to marry with job. When the Cubs are going to win the World Series. That's easy to break. And this has led to all kinds of things like, you know, palm reading, crystal balls, and astrology, et cetera. But the world is enamored with prediction. So in the world of IT and operations technology that are converging, it's about prognostics and predictable behavior. I want to predict the asset. I want to predict the sale. I want to predict the location. I want to predict the weather on and on, et cetera. So that dimension to do those prognostics, they run algorithms on IT, computer systems. Well, the HP Enterprise Corporation, as you may know, is number one in compute as well. So us taking that compute expertise and imputing it on IoT applications gives us that advantage. Because again, a lot of it is about running software algorithms to do predictions. If I can predict where I can drive the energy, I can predict when I'm going to discover this particle, I can predict the health status through this and that. And I think you know what I'm saying here. That prognostics makes me think. It's real time, you've got flash memory. All this stuff kind of comes together under the hood. All those technologies. Yes, yes. All right, so I got to ask you the next question. So perpetual connectivity, love the term, love the term cognitive prognostication. I'll have to use that at a cocktail party this weekend. There's more to come. I'm going to weave that in somehow this weekend, my kids across the tournament. It only works with Palo Alto. The parent dinner is always good. It sounds good. Sounds impressive. No, but love the cognitive computing, certainly IBM's all over that. But perpetual connectivity is a good concept. What about perpetual power? Because that's another issue in IoT. Power, battery. You're right. The sensors that connect to the things where we derive the insight, if we look at the chain here at the food chain, have to run on power, as you said, and energy as well. So we're actually, I'll tell you an example, get a little geeky here again. We've created a new system called the edge line. And in it has what we call power on ethernet technology, which means if you think of the old phone systems that plugged in the wall, you didn't plug a separate plug in the wall. You had one system that had both the signal, the voice as well as power. So that was power over line. We have the same concept. We can plug a sensor and it gets the power not from a AC socket or a DC source, but rather from the ethernet as well. But that's a big deal. And the fact that low power technology is becoming more and more ubiquitous and affordable is going to mean more things will be connected through sensors. And that's- BLE is a great example on the Aruba stuff. Bluetooth light or low energy, right? As well. So talk about the power thing. That's interesting. I love that power on ethernet. I see this great example. But the issue there is the truck roll. The thing now we're getting out outside of IT, I got a provision, an actual network connection. Yes. And we're getting out on that potentially. Yes. So IT guys have to prepare for this. Understand the fantasy of the costs, the hidden costs that might occur. What are some of the easy tripwires to identify out there that you could share for practitioners out there? I immediately go to security. And an IT professional, one of his, if not the greatest concern, is securing the data center or the cloud. And by the way, I believe a cloud is just a data center that nobody's supposed to know where it is. So when I talk about clouds and data centers, the security dimension is somewhat controllable because it's in four walls, et cetera. I'm not saying it's trivial. I'm just saying it's controllable. Now we move to the internet of things and we're taking compute capability out to the edge. I like to call it the other off premise where the cloud of public or private away is the first off premise. The other off premise is out at the edge. When you take IT assets out at the edge to do edge computing, and I'll explain why that's valuable in a minute, you have security concerns and issues. So the IT professional has to find the way to extend and graduate that out to the edge. And actually we're doing that by imputing HP Enterprise capability at the edge. So the same capability to protect and manage systems in the data center, we will actually impute and extend out to the edge. That's our strategy to do that. But that's the biggest thing is you got to make sure it's secure because you could be collecting data that's extremely vulnerable and extremely important to the enterprise. Or it could be an entry point surface area attack for hacker. That's absolutely true. And making sure that is protected because now you're exposing the thing to the hacker where before when it was unconnected it was not possible or at least feasible. So you have hyperscale in your title. Can you talk about the relationship between hyperscale and... It is a generic term, it's not an exclusive term but it has to do with lots and lots of things. The scale, hyper meaning above and beyond scaling up. And the notion of scaling up and out is an IT term. And it's been round for... Not hyper converges relative to this definition of that. And that's a, you can actually do both. You can have a hyper converged, hyperscale system because when you converge things you're pulling them together in a single unit and the value proposition there is one part never to manage less quality issues, et cetera. But to do the scaling is bringing it out and out. So for example, we're looking at opportunities where we're selling by the thousands because again, the more connectivity you get the more systems you need and they scale out. Now, one quick thing. Why do you want to collect a lot of data? Why is the internet of things a big data problem? It is, by the way. That's because we go back to prediction. If you have large data sets your predictions and conclusions are more accurate. If you have small data sets the chances of you missing the prediction is higher, et cetera. So the more data the better. So can we talk about that data model? Yes. And what's changing in the data model? What's enabling a new data model? Yes. Let me start there. Sure. First of all, the data from the internet of things is unique. It's actually big analog data. National Instruments has actually trademarked that term. It's big data because it's voluminous, it's variety, it's very, very fast, et cetera. But big analog data means it comes from nature and physical sources. Such as motion, particulates, voltage, moisture, acceleration, location, GPS location, for example. These are all analog phenomenon. How do you process an analog phenomenon of how much you're sweating or your heartbeat? Well, you have to take that analog phenomenon and do what's called an analog to digital conversion. That's a technology that a partner will provide for us in the solution. So that's the notion of the data. Now, big analog data is the oldest, the fastest, and the biggest of all other big data combined. You might say, well, you got to justify that point of view, and if you'd like, I will. But that's what's amazing about it, the enormity. I know a big analog data IoT solution that is 40 terabytes a second. That is huge amounts of data coming into a sensor. So I don't question it, but I want to learn more. Can you justify that? Well, let me start with, it's the oldest. Well, if you believe in the, let's go back to the creation of the universe. If you believe that the universe was created by the Big Bang, I would argue that's an analog event. It had motion, it probably had noise, it had light, it had magnetism, it had velocity, it had acceleration. These are all the things we measured today. If you believe in a creation model and you believe God said, let there be light, I would argue light is an analog event. In fact, light is the very thing we're collecting here. So no matter what you believe, it's the oldest. So what is better term? Data lake or data ocean? If motion is a big, data universe, exactly good. Data universe, all right, go with me. So that's oldest, right? You said it was the oldest? Yeah, now fastest, just think about video. You're in the professional video business. Well, that's one of the fastest data, having the number of frames per second in the resolution, right, et cetera. But just think about the electrons moving in your body, your heart beats the blood flow, the vibration on the air conditioning vents here, they're all moving at extremely fast rates and fast frequencies. That's the fastest. The biggest, again, is I have never found an application and I say this every time I speak and do lectures and keynotes. Has anybody seen more than 40 terabytes a second? And if you have, that would be the biggest and that is big analog data. Again, it has to be converted to bits, ones and zeros to be computed, but it comes from an analog source. So think about the enormity. We ain't seen nothing yet, as they say, with respect to the amount of big data that's going to come from the IoT. Okay, but we're not going to move all that data to some kind of box, right, central location. How are we going to deal with it? Okay, I know we don't have a lot of time, but there are seven reasons not to move the data. But to collect, you have to collect it. Collect it at the edge and the phraseology I like is compute at the edge, accelerate insight. In other words, don't compute it all back at the data center or the cloud, but compute it at the point of capture. Now, we don't have time for seven reasons, but if you want to learn more, we'll do that, but let me give you a couple. Number one is if you compute at the edge, you accelerate another phrase called time to insight. How fast can I get insight from the data? Well, I have to move it all the way back to the cloud. That's called latency, it takes time. How fast do you want to know if your asset is going to catch on fire? I would think immediately. Right now. Faster than the time it takes to transfer the data. Yeah, or in a military situation, how fast would you like to know a hostile weapon is aimed at you? Right now. Mealy, so that's the notion of latency. Another one is bandwidth. It costs, and the second one is, even if you can afford the bandwidth, you don't want to tie it up necessarily with all this big analog data coming from the IoT as well. So there's several more, but there's a lot of good reasons. So there are reasons. Edge line fits there. That's the edge line. That's why we've created the edge line family. We're going to compute the edge. It's a data depot. Kind of store it right there. So insights at the edge. Right. And then what? The insights move. Yes, all right. Let's talk about three As. There's the acquisition of the big data at the IoT. There's the analysis. And the other A is the action. So when you acquire it, you're going to do the analysis, the second A at the edge with the edge line product line. And then what action will you take? Will you take a business action? For example, in the Duke Energy example, Duke Energy's the largest energy provider in the United States. When we instrumented their turbines, the action to be taken would be to actually go into the financial systems in the IT dimension. And actually budget money to do a repair or replace action. Now think about the value of knowing where your money's going. If you manage your own home, you can make a type one error, which is I ran out of money. Darn. A lot of people make that error. I have two in the past. But the type two error is equally bad in the business, which is what? You're hoarding it and you don't need it. So you could have used it to send another, a Super Bowl commercial or hire another professional. So to avoid a type one and type two error, the financial case of that data being able to predict and you look at sale, I see you have $6 million allocated to do your replay. How do you know that? Well, my industrial IoT solution told me that by this time it would end. And if you care to know how that works, it's pretty interesting, but it gets kind of geeky here. Love it. Okay, so the bottom line for HP is what? The bottom line is to make it easy to deploy these complex solutions. If you listen to the conversation, it can get complex. There can be up to 30 different vendors creating this cake. But the end user wants a cake. The time to value of an already baked cake is pretty interesting. You open it, you put it in your mouth and there's value. If you want to bake your own cake or make your own, you risk what's called integration risk. And the famous integration risk of baking a cake is barring a cup of sugar from your neighbor. Right, you got to go, oh, I'm making a cake, I don't have sugar, can I bar a cup? That happens in the world of IoT and IT, the missing a cable, I'm missing a driver. So we're going to make it easy to be the baker to bake the cake. Now again, we don't provide every solution ingredient. We have good flour, we have really good sugar, but we don't have frosting, so we will partner with a company like, for example, Intel National Instruments, for example, PTC. These are some of our partners that will help us with the solution ingredients. But we will, as HP, make it easy to add to this transformational journey around a data-driven organization. It's data-driven now with traditional IT data. It will be data-driven as well with IoT data. Talk about Moonshot, one of the things we were very impressed with when they launched it, but it's got one use case, it's mostly big data and videos, it turns out. So now Antonio O'Neary was saying yesterday they're going to start breaking that out because it has low density, a lot of power. What's that, are you involved in that whole Moonshot? How is it going to be decentralized? I am the general manager of the Moonshot business, and I can tell you it's an amazing product. I can tell you also that it relates to the IoT, which is part of our discussion here. How does it relate? Virgin Racing, NASCAR, racing actually uses the Moonshot as an edge computer. So the Moonshot is so dense, so hyperscalable, that it can take a large portion of the data center and put it out on the edge, and you get deep compute at the edge. Another buzzword or phrase. You get deep compute at the edge and the sensors connected to the car send the data in deep computers at the edge. So the edge is the pit in the racing opportunity. So Moonshot in general is doing well. We just had several deployments of, for example, not in the IoT space, but actually on Wall Street traders using Moonshot to do trades and do it remotely. In fact, there are Moonshots sitting in New York City and the traders are in London doing real trading, and they can actually re-reference back to the Moonshot. So you're going to federate that out. You see yourselves going, okay, hey, let's not have the whole chassis because I'm pretty monster chassis, but the back plan or anything else, can you just take those out and just put them as devices? You have a good point. We have an amazing asset that we've invested in. From a financial perspective, we're going to get an ROI, a return on investment by employing the Moonshot technology in the IoT space. Now the Moonshot as it is today is a big data center and a box concept as well. And that's not appropriate for everything. So therefore we announced, as you may know yesterday, the press release at 8 a.m. London Times said HP edge line family. So the edge line is obvious. We're going to have compute at the edge. So we have a scalable family. So there's Moonshot in edge line? Yeah, I'm sorry? And Moonshot is in edge line or is that? Yeah, there's two business units and I happen to be the general manager of both. Okay. And the aforementioned synergies are why. Yeah, yeah, okay. And synergies is not a different group, right? The synergy product is a different blade product that is not mine, another general manager. So you envision Moonshot being tightly coupled with some capacity edge line. That is correct. That edge compute device area. That is correct. And it's very, now why? Because Moonshot is the best in the industry at high performance at low energy. It's the best in the industry at high performance and small footprint. Both important for edge IoT. And it stores a lot of stuff, so good for big data, which is the analytics. Excellent for big data. So all that is applicable at the IoT and we just want to get another return on our investment for our shareholders and our customer value. Dr. Tom Brash, thanks so much for joining us on theCUBE. Thanks for the insight. It's fast moving, high velocity, bringing a lot of signal here at the edge of the network. We're in London all the way across the world from the U.S., bringing the European perspective here at HPE Discover this theCUBE. We'll be right back with more after this short break.