 from London, England. Extracting the signal from the noise. It's theCUBE Cover Discover 2015. Brought to you by Hewlett Packard Enterprise. Now your hosts, John Furrier and Dave Vellante. Hello everyone, welcome to day three of coverage with theCUBE here at HPE Discover Live in London, England. This is SiliconANGLE Media's theCUBE, it's our flagship program. We go out to the events of SiliconANGLE, the signal from the noise. I'm John Furrier, the founder of SiliconANGLE. I'm going to go to Dave Vellante, founder of wikibon.com, wikibon research, free research, and also now subscription research. Go to wikibon.com to get all the latest in research from, in big data cloud, and also the infrastructure. And go to siliconangle.com for all the news around tech, from gaming to enterprise. And of course, go to siliconangle.tv for CUBE coverage. And if you want to join the conversation, go to crowdchat.net slash HPE Discover. That's the conversation. Engagement container technology that we built and also go to crowdpages.co slash HPE Discover. That's our new crowd pages, aggregated page for all the social content and videos there. Okay, Dave, to day three, we're kicking off wall-to-wall coverage here for HPE Discover. And it's really all about the enterprise. And it's interesting to see HPE really step out and establish themselves in the new company. This is their coming out party of the new entity. A lot of spring in the steps, smiles. People are pretty confident here at HPE, certainly the management team. Great presentations yesterday on the keynotes. Love the interactive presentations. A lot of flair and glam on stage with Meg Whitman, Antonio Neary, crushed it on the keynotes. Just overall, the old-timer HPers and on the new blood really seemed pumped up and they seemed competitive. And I think that is the big takeaway for me is that there's a competitive swagger at HPE. They feel like relieved. This relief, it's over. This cloud that's been around them, around the debates going back to the board debacles, going back, you know, Carly Farina, Mark Hurd, Leo Apatecker, and now Meg Whitman coming into Clean House, turn around, multiple years to takes. It's been a slog for HP. But the enterprise group has been doing a good job over those years. We've been covering it. And I got to say, you know, they've been on task. They may have to tack a few times, use a sailing analogy, but for the most part, everything that they've been talking about for the past five years has been right on the mission. They had Moonshot. They had GNA. They had all the stuff in converged infrastructure. Now with the big data, obviously, the acquisition of Autonomy was a little bit of a side show. The Vertica integration was a fairly solid one. Aruba Wireless Acquisition, they're geared up. I think I see a strong HP. I feel good about these guys. You? Yeah, well, I think you're right. I mean, Meg has had to deal with a legacy of, you know, partial integrations. Even going back to the HP Compact merger, she's had to deal with, you know, years of cost-cutting and R&D squeeze. And she had to deal with some misplaced acquisitions with Autonomy and clean that up. And she said it was going to take five years. And I really do think now finally we're seeing the light at the end of the tunnel. I've talked all week about the balance sheet issue, which I'm very excited about. And I think you're right. Ostensibly the reason for this split was focus. And it's interesting to note this year, we're not listening to, you know, a bunch of discussions on personal systems and printers and the ink business, right? So there really is a focus on the enterprise. And so that's all goodness. I think now the issue is how do they grow top-line revenue from a constant currency standpoint, which I think is fair because the headwinds of currency, how do they drive operating margins? How do they continue on innovation? How do they clean up the services area? And then really importantly, how do they get into new markets? And John, one of the things that all these vendors are talking about is internet of things. And internet of things is a huge opportunity because it's a trillion dollar opportunity, GE's talking about it, IBM has their app, certainly HP has theirs. Everybody's salivating, especially around the data. And so they're all looking toward the future saying, hey, we can capture that data, we can analyze that data, we can provide infrastructure. But I want to get your take on this. You really have, I take a good beat on things like wireless, all over the Aruba acquisition. And it seems to me, John, that there's some steps that the industry has to take before all this revenue that all the vendors are talking about can be generated. I wonder if you could sort of take us through your view of what has to be done from an infrastructure standpoint in order to mine that gold. Yeah, I saw George Gilbert, Wikibon's analyst, put out a post on wikibon.com about IoT, the IoT 3.0, big data 3.0, I think he called it, really run IoT, kind of a directional post, not so much, not a lot of meat on the bone, but kind of the first wave of a directional framework. Yeah, quick hit, quick hit. And he lays it all out there, but I think one of the things that's not clear in that post is what the action item is. What do people do? IoT certainly has been hyped up and we see that. But there's real tangible value in IoT, but there's a lot of fantasy as well. So we've been tracking that. You know I've been tracking the IoT pretty heavily. We have a big data practice here at SiliconANGLE Media. You know, we do a lot of DevOps and Amazon and BlueMix and SoftLayer and other, among other platforms. We're in the media because we're very DevOps for media, we're DevOps for data. So we kind of, I got a good feel on this, but here's the thing about IoT. There is a fatal flaw right now in IoT execution and the action item is for practitioners on IoT is don't buy into the hype of the fantasy. And the fantasy really is, oh everything's going to be instrumented. That is accurate, it will be in the future, but doing it right away, you could get ahead of yourselves, you could get over your skis and have a yard sale in terms of execution. And here's the fatal flaw in IoT right now. If you try to push your network beyond its true edge, and we'll get to that in a second, in my opinion, to devices that are on far away, remote places like sensors, if there's a network connection and a battery required, that's the fatal flaw. Until those two issues are solved, IoT at the far edge of the network is an absolute fantasy. And what I mean by that is, in the old days of doing truck rolls in the internet and pulling the last mile broadband, that's the cost. So the benefit of wiring up all these sensors at the edge of the network really comes down to the cost and also the ability to actually get the data. So if it requires an internet connection, you need an air interface, wireless, or you need a network connection. Two, you need power. And power is either real power or a battery. Battery's going to run out, network connection's going to be spotty. That is a real fatal flaw in that execution piece. Now, action item for practitioners in my opinion is to establish the edge of the network. And when I say the edge, the true edge of the network, and that is where you have power and connectivity that is either wireless or wired and power should be connected power. So that is really the first step. Understand your true edge of the network. And once you understand your true edge of the network, then IoT everything. Because then you get the data. That is then going to come into some of the things that HP talked about, this edge device, their edge servers, they're going to bring in, they get the moonshot. So all that's kind of known, compute and storage and networking stuff. That's really the goodness of IoT. So what you'll see in my opinion is phase one of IoT. So if you define that edge and that edge, what you're saying, implying anyway, that that edge might not include the windmills and all the infrastructure and the industrial internet to use a GE term, is that correct? Yes, that's correct. Okay, so let me ask you some questions about that. So the windmill, when I fly out of San Francisco, I fly towards Sacramento, I see all the windmills. There's no internet out there is what you're saying. Well, if this wireless connection, obviously LTE, you're going to need to have that or some sort of broadband wireless or some actual physical connection, that requires a truck roll. Well, there's, but there, hold on, but there's power out there, right? I mean, there's power, there's self-interpreting power. You assume there's power, yeah. The theory, okay, so they get power and there may or may not be internet is what you're saying. Yeah, yeah. And so, when you say it requires a truck roll, what do you mean by that? So the whole purpose of IoT is to minimize truck rolls and labor. So you want to get device information so you actually can be more efficient in your management of that cost of windmill maintenance, other things. That's the benefits of the sensor data. So if the windmill has accurate connectivity, so if they actually have internet to the windmill, you got to wire it up and that's going to require, where's the connection? Is it a physical connection or is it wireless? And two, the power is going to be connected with AC and or battery. So those things are an issue and it's not that clear and it's more complicated than cost to deploy than the benefit that you'll get out of it right away. So if I define my edge and my edge doesn't include all those assets that I want to instrument, the action then may be to provide internet connection to those assets, but that's not trivial is what you're employing. Not trivial, if you don't define your true edge. True edge is where in your network you actually have connectivity, opportunity, and power to your devices. So Antonio Neary kind of hinted at it yesterday when he said that they're going to go after a lot of the campus networking on the Rubik side and also you've seen a lot of the IoT action in plant facilities, campuses, so manufacturing, where you actually have assets that are connected and have power. That is phase one of IoT. Absolutely that's real. There's no fatal flaw in that and that's what I call the true edge. That could be your manufacturing, it could be remote. Wherever your edge of the network is, that is the first thing you got to do and then you IoT everything. So that's where the phase one action will be and that is absolutely not hype. That is truly real opportunity today. Okay and then presumably it's relatively straightforward enough and there probably already exists to put sensors and microprocessors in there. The issue is the last mile, that connection. It's the truck roll, it's the cost of the labor to actually put a network connection in. Have the power, is it plugged in? Is there a battery? Is the battery dead? Is it a long life in the battery? What's the form factor of the device? All these little nuances are a cost issue so if you're going to IoT everything, there's a scale point and cost benefit breakthroughs. So phase one again is have the vision of IoTing everything but understand your true edge, then IoT everything which means deploy devices, get the data and then bring it back to the central, get the data back and that's what George was hitting on in his piece. Yeah or, okay now I want to ask you about that too, that's what George was hitting on or maybe not bring it back to the center, maybe bring the compute to the edge. This is a variety of techniques, you can trickle it in, you can store it, you can cache it, you can send it when you updates, all kinds of capabilities and once you're in the network, that edge, that true edge, you can do a lot of different things. Depending on. There's different approaches, cost, topology, methodology, whatever the operational footprint looks like from a workflow, a work stream standpoint, that's what the network guys will decide. There's no one playbook on that piece but once you get your true edge, then you deploy your methodology for how you collect the data. And to your point, maybe it's one dot five is sort of anticipating when there's going to be failures. So if I understand you today, the way it works today is you roll the truck out to get the data and figure out what's going on and then you got to go back and get whatever, parts to go back out and fix what's broken. Okay so you want to minimize that lack of efficiency. Okay and then beyond that, there are other things that you can do to sort of optimize the workflow around a variety of things, whether it's logistics or traffic flow or optimizing energy efficiency, et cetera. And that's really what George is, I think, getting to in his piece. I mean, it all comes down to cost benefit. If you've got windmill maintenance is through the roof, then it might be cost effective to send someone to provision devices on all the windmills. That's simply how they do it today is they look around, they have device, they have instrumentation, they have to collect the instrumentation data, they got to understand the life cycle. Sometimes it's just random schedule based based on average, mean times to failure, things like that nature. But when you actually have real time data, that is the value, getting the real time data. So if you have no battery or internet, you're not getting real time data. You might be good enough. So there's a good enough question. But again, the edge of the network truly defines it, Dave, because then that's really a critical asset. Plant, equipment, manufacturing, you're seeing that stuff happen now. Airplane's another good example. I'm on flights all the time. It says, hey, welcome. We have wireless and the wireless doesn't work. No, it must be your device. It's funny, nobody in the plane can get on the network. So still some basic infrastructure issues is what you say. What can we learn from the cable industry? Anything? I mean, is it, you know? Well, we learned from the breaking up of the telcos where people want to go the last mile. It's very expensive to provision to a central office to a home and cause the truck rolls. That is the same paradigm you're going to see in IoT. At the end of the day, you do have to connect to some network point inside the core network or edge core. If you will, or towards the edge. And then also actually provision the device. So at scale, it's a little bit more difficult. Again, it's a little bit bigger task, but the benefits definitely can be there, Dave. So, you know, that's the, the action item is, you know, understand the hype as a vision statement, but the tactical thing is the edge of the network and then power and battery, power, battery issue and network connectivity. So define your edge. Understand the assets that are in the edge that you want to bring into your IoT umbrella and make the cost benefit analysis on whether or not it makes sense to actually bring internet and power to those assets. Or can you get enough leverage out of what's already at your edge? That's really, that's where the analysis should start. That's the analysis. And that's phase one real market today. That's what Antonia was talking about basically yesterday, which is there's tons of campuses out there that are doing refresh to get wireless. That's Aruba. There's a ton of opportunities around actually manufacturing and physical facilities for businesses that could be retail, that could be manufacturing. And there's all the subsystems already in place, the network, the CRM, the ERP system. So you have infrastructure, critical infrastructure in place for those low hanging fruit items and those are real opportunities. I think the other thing about IoT that's not talked about a lot, and we learned this when we did industrial internet study, David Floyer and our team at Wikibon, we went out and talked to numerous engineers, like engineers who run the windmills, who install them and maintain them. And there's a lot of concern. I mean, it's a new buyer, let me put it that way. You know, it's not the head of IT, it's not the CIO, it's not the CMO, it's the engineer who basically owns that piece of infrastructure and it's not going to let you mess with it. So there's a new sort of whole class of buyer which most traditional IT vendors aren't used to selling to. I mean GE clearly is. Well this brings up the other issue we mentioned services, right? So one thing that's happening here at HP Discover that's notable and it's kind of hypers, but it's really announced it is that the composable infrastructure is the real deal. It's definitely the direction of the market just like converged infrastructure was in the early days. HP nailed converged infrastructure, the jury's out there, time is the ultimate arbiter of who was right, who was wrong. HP was 100% accurate on their vision of converged infrastructure and they deserve props for that. I think the similar thread here is the composable. It's a little bit half-baked, but directionally it's absolutely relevant. DevOps is about composing applications using infrastructure as Lego blocks, infrastructure as a code, this is where we've been using since 2010. We've seen this movie before Dave. That is the next converged infrastructure opportunity that you're going to see HP have the right answer on. I will predict that their vision of composed infrastructure is going to be very accurate. Yeah, I want to pick up on that because I mentioned it yesterday during the discussion with Antonio Neary when I was at the IBM Pure launch and Steve Mills was there and he actually said the words. I said this yesterday. This is evolutionary. Essentially, look, this is a natural step and when you talk to Antonio Neary and the rest of the HP crowd, they see this as a big technological leapfrog. That's what they believe. I don't believe that. Really? Yeah, I just agree. So you see it as more evolutionary of the blade server market space? Yeah, I didn't want to debate Antonio Neary because he's in a good mood, he's on the cube. I'm still agrees with you, by the way. I don't think it's revolutionary. Well, evolutionary, they're building on top of Converge Infrastructure. It was revolutionary. The Converge Instructure project will be out of business tomorrow, right? Or three years. So do you think that Converge Infrastructure product line for HP will be out of business in five years? No. Then it's not revolutionary. Yeah, so. So to me, evolutionary in a sense. Or VCE or others. Or VCE or others. So it will be a game changer. So that's good. I wanted to get that out there. And then the other thing I wanted to address is in our discussion yesterday, I don't necessarily see this as a zero-sum game. I agree with that. And so, when you talk to Oracle, their wheelhouse is database. You talk to HP, their wheelhouse is infrastructure. You talk to Dell and EMC, it's going to be low cost and it's going to be storage. You talk to all these different, IBM is now moving up the stack into analytics. So I think there's opportunities for all of these guys to really thrive. It's a big market. Well, we're going to ask people that today. And I think to your point, I do think that it's not a zero-sum game because Dave, the mature market's still producing revenue. And this is a growth opportunity. And HP's not talking about it, but that's their innovation strategy. Composables of growth market, even on storage that what Minish Goyal is saying, same thing. Okay, this is theCUBE. We're kicking off day three. That's John and Dave's editorial analysis of what's happening in the industry here around HPE Discover. We'll be right back with more after this short break.