 Live from Las Vegas, it's theCUBE, covering HPE Discover 2017, brought to you by Hewlett Packard Enterprise. Okay, welcome back everyone. We're here live in Las Vegas for our exclusive three-day coverage from theCUBE, SiliconANGLE Media's flagship program. We go out to events and we extract the civil noise, talk to the smartest people we can find, CEOs, entrepreneurs, R&D lab managers, and of course, we're here at HPE Discover 2017. Our next two guests, Andrew Wheeler Fellow, VP Deputy Director, Hewlett Packard Labs, and President Nicker Fellow and VP, Chief Architect of HPE Labs, was on yesterday. Welcome back, welcome to theCUBE. Hewlett Packard Labs, well known, you guys doing great research. Meg Whitman, really staying with the focused message and one of the comments she mentioned at our press analyst meeting yesterday was, focusing on the labs. So I wanted to ask you, where's the range in the labs and in terms of what you guys, when does something go outside the lines, if you will? Yeah, good question. So if you think about Hewlett Packard Labs and really our charter or role within the company, we're really kind of tasked for looking at things that will disrupt our current business or look for kind of those new opportunities. So for us, we have something we call an innovation horizon. And it's like any other portfolio that you have where you've got maybe things that are more kind of near term, maybe one to three years out, things that are easily kind of transferred or the timing is right. And then we have kind of another bucket that says, well, maybe it's more of a three to five year kind of in that advanced development category where it needs a little more incubation, but it needs a little more time. And then we reserve probably a smaller pocket that's for more kind of pure research, things that are further out, higher risk. It's a bigger bet, but we do want to have kind of a complete portfolio of those. And over time throughout our history, we've got really success stories in all of those. So it's always finding kind of that right blend, but there's clearly a focus around the advanced development piece now that we've had a lot of things come from that research point. And really one of the kind of- You're looking for breakthroughs. I mean, that's what you're, some incremental improvement, simplify IT, electric stuff. But you guys still have your eyes on some breakthroughs. That's right. How do we differentiate what we're doing? So, but yeah, clearly, clearly looking for those breakthrough opportunities. And one of the things that's come up really big in this show is the security and chip thing was pretty hot, very hot. And the actually Wikibon's public, true public account report that they put out sizing up on-prem, the cloud market, which is- True private cloud. True private cloud, I'm sorry. And that's not including hybrids, a $265 billion dam. But the notable thing I want to get your thoughts on is the point Dave pushed was over 10 years, $150 billion is going to shift out of IT on-premise into other differentiated services. Out of labor. Out of IT labor. Out of labor. So this, and I asked him what that means, as the analyst said, that means it's going to shift to vendor R&D, meaning the suppliers have to do more work so that the customers don't have to do the R&D, which we see a lot in cloud where there's a lot of R&D going on. That's your job. So you guys are HP Labs. What's happening in that R&D area that's going to offload that labor so they can move to some other high-yield tasks? Sure. Take first? Go ahead, take a step at it. You know, when we've been looking at some of the concepts we had in the memory-driven computing research and advanced development program, the machine program, you know, one of the things that was the kickoff for me back in 2003, we looked at what we had in the Unix market. We had advanced virtualization technologies. We had great management of resource technologies. We had memory fabric technologies, but they're all kind of proprietary. But still, it got us thinking, and back then we were saying, how does RISC Unix compete with industry standard servers? This new methodology, new wave, exciting, changing cost structures. And for us, it was a chance to explore those ideas and understand how they would affect our maintaining the kind of rich set of customer experiences, mission-criticality, security, all of these elements. And it's kind of funny that we're sort of coming back to the future again. And we're saying, okay, we have this move. We want to see these things happen on the cloud. And we're seeing those same technologies, the composable infrastructure we have in Synergy, and looking forward to see the research we've done on the machine advanced development program, and how will that intersect hardware composability, the converged infrastructure, so that you can actually have that shift, those technologies coming in, taking on more of that burden to allow you freedom of choice. So you can make sure that you end up with that right mix, the right part on the full public cloud, the right mix on a full private cloud, the right mix on that intelligent edge, but still having the ability to have all of those great software development methodologies, that agile methodology. The only thing the kids know how to do out of school is open source and agile now. So you want to make sure that you can embrace that, and make sure regardless of where the right spot is for a particular application in your entire enterprise portfolio, that you have this common set of experiences and tools, and some of the research and development we're doing will enable us to drive that into that existing conventional enterprise market, as well as this intelligent edge, making a continuum, a continuum from the core to the intelligent edge, and something that modern computer science graduates will find completely comfortable. And the one attracting them is going to be the key. I think the edge is kind of intoxicating if you think about all the possibilities that are out there, in terms of what, just from a business model disruption, also technology. I mean, wearables are edge, Brainiam plans in the future will be edge, the singularities here as Rakers, while we say, I mean, but this is the truth, this is what's happened, this is real right now. Oh, absolutely, we think of all that data, and right now we're just scratching the surface. I remember it was 1994, the first time I fired up a web server inside of my development team, so I could begin sending out design information on prototype products inside of HP, and it was a novelty. It was like, what is that thing you just sent me an email, w, w, whatever, and suddenly we went from, like almost overnight, from a novelty to a business necessity, to then it transformed the way that we created applications for the enterprise. But since you brought it up, it's a historical trivia. HP Labs, you look like Labs had scientists who actually invented the web with Tim Bernsley. I think HTML founder was an HP Labs scientist. It's a pretty notable trivia, a lot of people don't know that, so congratulations. And so I look at just what you're saying there, and we see this new edge thing, it's going to be similarly transformative. Now today, it's a little genomic-y, perhaps it's sort of scratching the surface, it's taking security, and it can be problematic at times, but that will transform because there is so much possibility for economic transformation. Right now, almost all of that data on the edge is thrown away. If you, the first person who understands, okay, I'm going to get 1% more of that data and turn it into real-time intelligence, real-time action, that will unmake industries, and it will remake new industries. This is the applied research vision, and you got to apply R&D to the problem. Correct. That's what he's getting at, but you also got to think differently, you got to bring in talent, the young guns. How are you guys bringing in the young guns? What's the honeypot? Well, I think for us, the sell for us, obviously, is just the tradition of Hewlett-Packer to begin with, right? So we have recognition of that level, and it's not just Hewlett-Packard Labs as well, it's just R&D in general, right? Kind of the DNA being an engineering company. So, but it's, I think it is creating kind of these opportunities, and whether it's internship programs, just the various things that we're doing, whether it's enterprise-related, high-performance computing, I think this edge opportunity is a really interesting one, is a bridge because if you think about all the things that we hear about in the enterprise in terms of, oh, I need this deep analytics capability, or even a lot of the in-memory things that we're talking about, real-time response, driving information, right? All of that needs to happen at the edge as well for various opportunities. So it's got a lot of the young graduates excited. We host hundreds of interns every year, and it's really exciting to see kind of the ideas they come in with, and they're all excited to work in this space. So Kirk, you have your machine button. You got it. You got the, of course, you got the logo, and then you got the machine. I got the, yeah, the Labs logo. The Labs. I got the machine logo. So, when I first entered, you talked about in the early 1980s, when I first got into the business, I remember Gene Amdahl, the best I.O. is no I.O. Yeah, that's right. We're here again with this sort of memory semantics centric computing. So, in terms of the three that Andrew laid out, the three types of sort of projects you guys pursue, where does the machine fit? Is it sort of in all three? Or maybe you could talk about that a little bit. I think it is. So we see those technologies that over the last three years we have brought so much new, and it was the critical thing about this is, I think it's also sort of the prototyping of the overall approach, our lean-in approach here, is that it wasn't just researchers, right? Those 500 people who have made that 160 terabyte monster machine possible weren't just from labs. It was engineering teams from across Shield Packard and Enterprise. It was our supply chain team. It was our services team telling us how these things fit together for real. Now, we've had incredible technology experiences, incredible technologist experiences, and what we're seeing is that we have intercepts on conventional platforms, whether it's the photonics, the persistent memories. Those will make our existing DCIG and SDCG products better almost immediately. But then we also have these whole cloth applications, and as we take all of our learnings, drive them into open source software, drive them into the Gen Z consortium, and we'll see probably 18, 24 months from now some of those first optimized silicon designs pop out of that ecosystem, then we'll be right there to assemble those again into conventional systems as well as more expansive exascale computing, intelligent edge with large persistent memories and application-specific processing as that next generation of gateways. I think we can see these intercept points at every category Andrew talked about. Well, and another good point there that kind of magnifies the model we were talking about had, you know, if we were sitting here five years ago, we would be talking about things like photonics and non-volta memory as being those big R projects, right, those higher risk, longer term things, but right, as those mature, we make more progress, the innovation happens, right, it gets pulled into that shorter timeframe that becomes advanced. And Meg has talked about that, wanting to get more productivity out of the labs, and she's also pointed out you guys have spent more on R&D in the last several years, but even as we talked about the other day, you want to see a little bit more D and keep the R going. So my question is, when you get to that point of being able to support DCIG, is it a handoff, are you guys intimately involved when you're making decisions about, okay, so Memrester for example, okay, this is great, that's still in the R phase, then you bring it in, but now we've got to commercialize this, and you've got 3D NAND coming out, and okay, let's use that, that fits into our framework. So how much do you guys get involved in that handoff, you know, the commercialization of this stuff? It's very, we get very involved, so it's at the point where when we think we have something that, hey, we think, you know, maybe this could get into a product, or let's see if there's a good intercept here, we work jointly at that point, right, it's lab engineers, it's the product managers out of the group, engineers out of the business group, they essentially work collectively then on getting it to that next step, so it's kind of just one big R&D effort at that point. And so specifically as it relates to the machine, where do you see in the next, you know, the near term, let's call near term next three years, or five years even, what do you see that looking like, is it this combination of, you know, memory with capacitors or, you know, flash extensions, what does that look like in terms of commercial terms that we can expect? So I really think the palette is pretty broad here, that I can see these going into existing, existing rack and tower products to allow them to have memory that's composable down and compute that's composable down to the individual module level, to be able to take that facility to have just the right resources applied at just the right time with that API that we have in one view, extend down to composing the hardware itself. I think we look at those edge line systems and want to have just the right kind of analytic capability, large, persistent memories at that edge so we can handle those zettabytes and zettabytes of data in full fidelity, analyzed at the edge, sending back that intelligence to the core, but also taking action at the edge in its timeframe that matters. I also see it coming out and being the basis of our exascale high performance computing. You know, when you want to have an exascale system that has all of the combined capacity of the top 500 systems today, but one 20th of their power, that is going to take rather novel technologies and everything we've been working on is exactly what's feeding that research and soon to be advanced development and then soon to be production and supply chain. Great. So the question I have is, obviously we saw some really awesome Gen 10 stuff here at this show. You guys are just seeing that, obviously you're on stage talking about a lot of the cool R&D, but really the reality is that's multiple years in the works with some of this root of trust, silicon technology that's pretty getting the show buzzed up. Everyone's psyched about it. DreamWorks Animations talking about how some pivoting and inorganic opportunities and they got the security with the root of trust, NIST certified and compliant. Pretty impressive. What's next? What else are you working on? Because this is where the R&D is on your shoulders for that next level of innovation. Where do you guys see that? Because security is a huge deal. I mean, that's that great example of how you guys innovated and bring, because that'll stop the vector of attacks in the surface area of IoT. If you can get the servers to lock down and you have firmware that's secure, it makes a lot of sense. That's probably the tip of the iceberg. What else is happening with security? So when we think about security and our efforts on advanced development research around the machine, what you're seeing here with the Proliance is making the machines more secure, the inherent platform more secure. But the other thing I would point to is the application we're running on the prototype, large scale graph inference. And this is security because you have a platform like the machine, able to digest hundreds and hundreds of terabytes worth of log data to look for that fingerprint, that subtle clue that you have a system that has been compromised. And these are not blatant, let's just blast everything out to some dot, dot, x, x, x subdomain. This is an advanced persistent thread by a very capable adversary who is very subtle in their reach out from a system that has been compromised to that command and control server. The signs are there if you can look at the data holistically. If you can look at that DNS log graph, a billion entries every day, constantly changing. If you can look at that as a graph, in totality, in a timeframe that matters, then that's a very powerful thing for our cyber defense team. And I think that's one of the interesting things that we're adding to this discussion. Not only protect, detect, and recover, but giving offensive weapons to our cyber defense team so they can hunt. They can hunt for those advanced persistent threats. One of the things, Andrew, I'll get your thoughts on the reaction of this because I'll make an observation, you guys can comment and tell me I'm all wet, fell off the deep end or whatnot. Last year, HP had great marketing around the machine. I love that Star Trek ad. It was beautiful and it was just, the machine is a great marketing technique. When you use the machine, so a lot of people set expectations on the machine. So you saw articles being written, maybe people didn't understand it. Little bit pulled back, almost damper down a little bit in terms of the marketing of the machine of the other than that been. Is that because you don't yet know what it's going to look like or there's so many broader possibilities or you're trying to set expectations? Because the machine certainly has a lot of range. And it's almost as if, if I could read your mind, you don't want to post the position too early on what it could do, that's my observation. Why the pullback? I mean, certainly as a marketer, I'd be all over that. Yeah, and I think part of it has been intentional just on how the ecosystem, we need the ecosystem to develop kind of around this at the same time, meaning there are a lot of kind of moving parts to it, whether it's around the open source community and kind of getting their head wrapped around what does this new architecture look like? We've got things like the Gen Z Consortium where we're pouring a lot of our understanding and knowledge into that. And so we need a lot of partners. We know we're in the day and an age, we're look, there's no single one company that's going to do every piece and part themselves. So part of it is kind of enough to get out there to get the buzz, get the excitement to get other people then on board. And now we have been heads down, especially this last six months, of getting it all together. You think about what we showed, essentially first booted the thing in November. And now we've got it running at this scale. That's really been the focus, but we needed a lot of that early engagement interaction to get a lot of the other members of the ecosystem kind of on board and starting to contribute. And really that's where we're at today. And it's almost you want to let it take its own course organically because you mentioned just on the cyber surveillance opportunity around the crunching. Kind of don't know yet what the killer app is, right? And that's the great thing of where we're at today. Now that we have kind of the prototype running at scale like this, it is allowing us to move beyond, look, we've had the simulators to work with, we've had kind of emulation vehicles. Now you've got the real thing to run actual workloads on. We had the announcement around DZNE as kind of an early, early example, but it really now will allow us to do some refinement that allows us to get to those product concepts. I wonder if we can just ask the closing question. So I've had this screen here. It's like the theater. And I've been seeing these great things coming up. And one was Moore's Law is Dead. Come on, Lurba. Oh, that was my session this morning. Another one was Blockchain. Unfortunately, I couldn't hear it, but I could see the T's. So when you guys come to work in the morning, what's kind of the driving set of assumptions for you? Is it just that technology is limitless and we're going to go figure it out? Or are there things that sort of frame your raison d'etre that drive your activities and thinking? And what are the fundamental assumptions that you guys use to drive your actions? So what's been driving me for the last couple of years is this exponential growth of information that we create as a species. That seems to have no upper bounding function that damps it down. At the same time, the timeframe we want to get from information, from raw information to insight that we can take action on seems to be shrinking from days, weeks, minutes. Now it's down to microseconds. If I want to have an intelligent power grid, intelligent 3G communication, I have to have microseconds. So you look at those two things and at the same time, we just have to be the lucky few who are sitting in these seats right when Moore's Law is slowing down and will eventually flatten out. And so all the skills that we've had over the last 28 years of my career, you look at those technologies and you say, those aren't the ones that are going to take us forward. This is an opportunity for us to really look and examine every piece of this because if it was something we could have just, can't we just dot, dot, dot do one thing? We would do it, right? We can't just do one thing. We have to be more holistic if we're going to create the next 20, 30, 40 years of innovation. And that's really what I'm looking at. How do we get back exponential scaling on supply to meet this unending exponential demand? So technically, I would imagine, that's a very hard thing to balance because the former says that we're going to have more data than we've ever seen. The latter says we've got to act on it fast, which is a great trend for memory, but the economics are going to be such a challenge to meet and balance that. We have to be able to afford the energy and we have to be able to afford the material cost and we have to afford the business processes that do all these things. So yeah, you need breakthroughs and that's really what we've been doing. And I think that's where we're so fortunate at Hewlett Packard Enterprise to have the labs team, but also that world-class engineering and that world-class supply chain and a services team that can get us introduced to every interesting customer around the world who has those challenging problems and can give us that partnership and that insight to get those kind of breakthroughs. And I wonder if there'll be a tipping point if the tipping point will be, and I'm sure you've thought about this, a change in the application development model that drives so much value and so much productivity that it offsets some of the potential cost issues of changing the development paradigm. And I think you're seeing hints of that. Now we saw this when we went from systems of record, OLTP systems to systems of engagement, mobile systems, and suddenly new ways to develop it. I think now the interesting thing is we move over to systems of action and we're moving from programmatic to training. And this is this interesting thing. If you have those zettabytes of data, you can't have a pair of human eyeballs in front of that. You have to have a machine learning algorithm. That's the only thing that's voracious enough to consume this data in a time-enough fashion to get us answers, but you can't program it. We saw those old approaches in old-school AI and old-school autonomous vehicle programs. They go about 10 feet, boom, and they flip over, right? Now, you know, they're on our streets and they are functioning. They're a little bit raw right now, but that improvement cycle is fantastic because they're training. They're not programming. Right. It's a great opportunity for the breakthrough to your point about Moore's Law, but also all this new functionality that has yet to be defined is right on the doorstep. Andrew Kirk, thanks so much for sharing that great insight. Love, you look back at labs. Love the R&D conversation. You know, it gets us a chance to kind of co-play in the wild and dream about the future. You guys are out creating congratulations and thanks for spending the time on theCUBE. Appreciate it. Thank you. theCUBE coverage will continue here live at Las Vegas for HPE Discover 2017. You look back at our enterprises annual event. We'll be right back with more. Stay with us.