 Live from London, England, it's theCUBE. Covering Discover 2016 London, brought to you by Hewlett Packard Enterprise. Now, here's your host, Dave Vellante and Paul Gillis. Come back to the London docs, Excel London. This is HPE Discover 2016 and this is theCUBE, the worldwide leader in live tech coverage. We've been extracting signal from noise and we have another session with bloggers, analysts, independent folks, we'd love to have these guys on. Paul Teich is here, he's the head of TIRIS Research and founder of that organization, correct? No, Jim was founder. I thought you were a co-founder. I'm a principal analyst. Principal analyst, great. And Dana Gardner, who I think is the founder, am I right, of Interarbor Research. Gentlemen, welcome to theCUBE. You guys are here as influencers this week as part of that sort of influencer program and let's unpack it. Paul, let's start with you. Well, first of all, I know it's not this cold in Austin. You know, it certainly is this cold in New Hampshire and Massachusetts, but talking about the crisp, beautiful weather here all week. But Paul, let me start with you. What are your sort of initial impressions after three days of Kool-Aid injection here at HPE Discover? HPE has been doing, I think, a credible job at trying to simplify their message. I think they're creating a set of overarching themes for ostensibly what we'll call a bag of parts, right? They're a huge company. They have a very extensive portfolio. What we haven't seen in previous years is kind of what ties that all together. Trying to be everything to everyone has its limits. And so, you know, the CSC spin merge, that's what they're calling it, the micro-focus spin merge. It's important to say what you're not doing as well as to say what you are doing. So I think there's some focus here. Didn't, that was, I didn't actually expect to see that focus. And some of the hardware and development stuff they're doing with around the machine architecture because it's not a complete product. It won't ever be product-dised as one thing. But the machine architecture, it's also very interesting. It's much more advanced than I thought it was. Yeah, we want to come back and get some of your thoughts on that. And Dana, a lot of the areas that you cover are getting spin merged. You know, the big data piece, you know, the cloud piece, spin merged, but pieces of it actually did, the open stack distro. So what's your take? Well, I think it's interesting to consider that they're not selling them. The spin merge is different. They're retaining ownership and ultimately control of the companies that they're putting these assets into. And I think it speaks to the idea that it's an ecosystem that can support a hybrid cloud environment, not one provider. So that's true for the services, not true for the software, correct? For micro focus it is actually. They're going to have 51% of micro focus will be a part of HP, I believe. Yeah, I think, I thought it was some weird nuance. It'll be owned by the shareholders. Like the HP shareholders. The HP shareholders are going to own it. The HP is not the biggest shareholder of HP. Anyway, I think micro focus has the controlling interest, but not, yeah. Well, it's fair to say that they're going to be cooperative and competitive. Absolutely, yeah. No question. And they have a stake. It's a mutual stake. They've got skin in the game. Their existence is interdependent. Yes, right, absolutely. Like a marriage, maybe they're just common law marriage, right? Yeah. Nonetheless, the fact is that it's going to take an ecosystem approach with many partners and attracting other partners. It was interesting to me that Meg Whitman had on stage startups. I hadn't seen that before. That HP has become an incubator, and they called it a curator of Silicon Valley talent. So to think of HP as a monolithic entity is gone. The new idea is that it's an association of companies that have affiliation and that they're incubating and thinking about where the next technology innovations come from doesn't necessarily come from HP Labs. It comes from the companies that they're investing in. It's an interesting point, Dane, is making. Paul, I hadn't really thought about it. What do you think about, remember the EMC Federation? It was a noble attempt. Ultimately, I guess it failed to live up to the objective. It threw it out. Right, but this is sort of a new business model. Your premises, HP will be able to turn some knobs and gain leverage. It's the hub of a spoke entity that is an ecosystem that has interdependence rather than one monolithic decision coming from one board. And they haven't positioned it that way, really, have they? Not, I think Dana has articulated it better than HP has, frankly. Okay, some free advice on the cube. I want to ask you about this partnering thing, though, because it's so critical to where they're going. I mean, you know partnerships are a dime a dozen, right? We've got nothing else to do. You announce a partnership. They are basing their company on the ability to carry off this partnering strategy. Do you feel that's a solid strategy going forward? Paul? It has to be. Where I live, close to the silicon in system design, even. The idea that you'll have one processor architecture, one accelerator architecture, even one class of memory and storage is gone out the window. Okay, so when we talk about heterogeneous architecture, even within one vendor's product line, there are, you know, there are going to be relationships with graphic processor vendors, relationships with FPGA vendors, relationships with processor vendors, and it's not the same kind of monolithic infrastructure that we've seen for the last 15 or 20 years. You, when you came in here, we were talking about what does this all mean, all this automation to IT? Computer Weekly has this article, and I'll quote from it, get your reaction. Hewlett Packard has become the latest vendor of many to sound the death knell on the IT department. An imminent and presumably universal move to cloud computing and proclaims happily will inevitably lead to the termination of rafts of technical IT jobs as the automators become the automated. Did they say that? Computer Weekly said that. And we were talking about it earlier. Will IT change? Will it need to adapt and become something different, perhaps more, even higher in their elevation and in their value in an enterprise? Yes. Does that mean they go away? No, not at all. Fact are going to be freed up from having to check on the heat status of racks and start checking on the viability of services that support major applications. So we have to think about the complexity curve. It's steep and it's getting worse. More microservices, right? Containerization, APIs that are bringing services in from someone else's data center that you can't control that you don't have a red light or a green light bead on. All of a sudden you have more responsibility but less authority. That's a tough place to be in. But somebody's got to do it. So I don't think that IT is going away. I think they're becoming more indispensable. They're elevating in their value to the organization. And if HP knows what it's doing, it'll become the best friend of that new IT organization. So, Paul, provisioning LUNs is not a long-term career growth path. Let's, this is part of an actual larger discussion on automation of jobs, right? And so, what's happening is the skill set has to move up the stack. And so, your IT admins aren't going to be checking to make sure, in a software-defined network, they're not going to be unplugging ethernet connections on a daily basis, right? But, you have to educate your staff. Continuous education, up-leveling your skill sets in IT is going to be critical as we do automate all, automate the low-level stuff, and as an IT admin, you're going to have to stay ahead of that wave. Because this wraps into machine intelligence, deep learning, and applying these new learning techniques toward more effectively, efficiently, and optimally operating your infrastructure. So, where you run jobs, right? It's not just which processor is more efficient, which processor is faster, which service costs less, which network am I going to be on, and how much does that cost? And so, the overall cost and efficiency and performance of running a workload in a hybrid cloud model, that's good, in real time, is going to take a lot of automation skills. And putting all that together and doing the supervising functions requires a different skill set from IT. You might think of IT as becoming the master of the algorithm, and the algorithm is which decides what workloads run where across a continuum of public to private and multi-cloud public, rather. Then they have to also have an algorithm that decides the cost issues, right? And the T's and C's, and whether they're adhered to. It's a contract, it's a procurement, it's a supply chain function. It can't be done at a microservices level by human decisions. It has to be an algorithm run by machine, AI and machine learning brought to bear. So, IT is becoming sort of the, I don't know, overseers of the algorithm, some sort of a chapel on the hill. I also think they need to become renaissance people. So, I had to wait for my room when I got here to London, and it was near the British Museum, so I had a great time spending a few hours there. One of the things that's there is a renaissance room or the enlightenment room, but the founders of the museum's collections are. And these are people that in the mid-1700s literally knew everything about everything. And that's where IT's got to go. They got to know everything about marketing, R&D, back office applications, what the cloud providers are doing. This is going to be a very important job, and I think people should gear up for an excellent career if they can be a renaissance IT person. It's a great point. Vinny Mershandani's book, I don't know if you've read it, the new polymath has made that point exactly. Coming out of this conference, how confident are you that HP has an edge up or has an advantage in being able to achieve this kind of automation and simplicity? They talk a lot about simplicity here. Simplifying the IT operation. We look at where R&D takes place in the industry, and over the last decade or two, R&D is consolidated. There are only a few companies really investing in the core research and development of new concepts. HPE inherited HP labs. And so as we, on the show floor here, they're showing the progress they've made on the components of the machine architecture. And so if you look at the stuff they're doing with photonics, with VIXOLs, with silicon photonics, if you look at the architectural improvements they're making, it's, that gives them a leg up. Especially when the ODMs who are following Intel's R&D machine, HP has to differentiate from the true commodity suppliers as they move into the cloud, trying to display some of that infrastructure. And as they move traditional IT into private cloud, bursting into public cloud, that whole hybrid model. So I think that, yes, HP labs does give them a leg up. The work they're doing there is very forward-looking, both on hardware and software front. I would just add that it's not just HP labs, it's got to be an ecosystem of IP, intellectual property and capital. And another interesting thing on main stage was, Andy Bechvilscheim, if there's any God in Silicon Valley, he's one of them. And the fact that Arista is making networks new from scratch, basically. But his prophecy was that we're nowhere near plateauing when it comes to technical innovation at the very core, which is speeds and feeds. So if you think that we're done with improving processing and memory capacity, think again, and I believe him. And if I had money, I'd put it on him because he's been incredibly successful. So if that's the case, then you can't just sit back and expect that whatever the status quo of the leading public cloud provider is the end-all-be-all. What it means, though, is you're only as good as your competitor or your competitor can only be as good as what that public cloud provides. It's sort of a leveler. It makes all enterprises operate at the same pace that that cloud provider services operate on. Some companies, a lot of companies, are going to want to have a competitive advantage. And that's only possible when you deploy your own systems using those cutting-edge capabilities that Andy was talking about. Thoughts on the business model for HPE? I'm essentially becoming a partner-oriented company and a reseller of many companies' technologies. I mean, they've always been comfortable reselling, not always, but recent history, Intel and Microsoft, and then making margin on top of that. It's a lower margin business now, which maybe is a good thing because they can compete effectively, more effectively with Amazon. You certainly see the Dell EMC merger being a lower margin business. But having said that, reselling risk, the relying on partners, the spin merge relationships, it's not as lucrative a model as an up-the-stack Oracle or SAP model. How would you explain your stock being at 50% this year? Well, because it's still trading at 76 cents on the revenue dollar. Nowhere to go but up with any kind of improvement. And that's a good point because it's cash flow is what's driving that. So these are good, strong cash flow businesses, not a huge profitability from a software margin comparison. So that was a deliberate choice that they made. And interesting to juxtapose that against what Dell is doing with EMC, the mega merger, and obviously what St. Oracle is doing, giving away hardware and trying to just get people on the software side, thoughts? First, I agree with the ecosystem approach. I think what's often overlooked is that there's a tremendous amount of traditional IT still in the field. It dominates the current enterprise landscape. When we talk about cloud, we're talking about adding new functionality on top of some IT shops today. As much as Amazon and Google and Microsoft have made money on their cloud services, they're just scratching the surface. The real opportunity for the next decade or two, at least, is private cloud. Okay, it's upgrading traditional IT infrastructure to your own differentiated infrastructure. The only reason you would own your own infrastructure is because you have some overarching security or performance or operational concern. Okay, if you don't have that, then you might as well go to the cloud, but most companies do have something they consider to be a core advantage. It's typically not operating their own IT, okay? So the idea of easy to operate and deploy private, somewhat custom delivery of those services, I think is going to be huge. So on-prem stays as a thing, okay? We don't all move off-prem. It's not all managed services, but having your own private infrastructure I think is where a lot of the innovation I've seen here really takes hold, okay? Building the structures to deliver that effectively. So it sets up this interesting sort of discussion where this week is reinvent as well, as you guys well know. You talked to Andy Jassy and the scenario is very few companies will own their own data setters in the future. You talked to the Meg Whitman's, Michael Dell's, Joe Tucci, the world, and their perspective is there'll be way more private clouds than will be public clouds. And you're like, somebody's got to be wrong there, but maybe not, I don't know. So help us square that circle. Yeah. We just know where the dividing line is. Nobody knows where that dividing line is. And I would posit that that dividing line is going to be constantly moving. So we've already seen some companies that went all cloud and that did the economics and saw the bill come in every month and said, maybe we can do better. And the fact that performance and price on private cloud infrastructure is going down that Moore's law is still at work, if you will, in the data center level, at that macro level, it means that some of them made the decision to go back out to their own data centers. There's never been a better time to build a data center. It's often the size of a refrigerator, not the size of a football field warehouse. The cloud providers are building football-sized warehouse data centers, but you can do just about anything you want in a data center, the size of a refrigerator or freezer, and then you can put multiple of those in geographies where you need to have data sovereignty issues ameliorated. Or you want to have backup and recovery. If you're in a hurricane zone, you'll put one in a place where there isn't. If you're in a tornado zone, you'll put one in a place where there isn't, you'll mirror them. So the choices are there. We just don't know where those choices are. And I think it's going to be an algorithm that decides. I guess. Just quickly because we're almost out of time. They introduced this storage service on-prem storage at a rentable cost, implied that they're going to be doing, broadening these kinds of programs to other products. Do you think they're onto something here? Yes. But again, with partners, okay, they can't all be in HP infrastructure. So I think the, we saw it with Microsoft at Ignite, the idea of the Azure Stack, taking a bit of the cloud, bringing it on-prem to where, in the case of storage, where your data resides is almost irrelevant to you as long as the right policies are set. Okay, but you've got the right redundancy, geographical protection. It's, I think the secret ingredient is collaboration within their partner network to make it all economical. And frankly, they can't see all of the technical innovation as much as they have great foresight. A partner network lets you take advantage of everybody's thought process. You get to pick the best innovation from an ecosystem. I think another point to consider too is that the way that you purchase technology is going to be as a service on an ops basis. You want to buy things per unit, per time, whether that is physical in your organization or not. And then to make apples and oranges comparisons work better. If you're going to really decide on a cost-benefit analysis, whether you want to be in a private or public cloud, they're going to need to be modeled on a business level in a similar way so that you can price them in a similar way. So if I'm selling metal, I might sell it in a fashion like GE used to sell aircraft engines, you'll lease it, you don't buy it. That also gets you into an interesting upgrade path that you can have some control over as the supplier. Just like Apple is going to obsolete certain MacBooks, I'm going to be forced to buy a new MacBook whether I want to or not. There's benefits to having an operational economic model and I think that's what we're starting to see across the private cloud continuum as well. Excellent. All right gentlemen, we have to leave, we'll go on for another half hour with you guys. Really excellent segment. Thank you so much for coming on theCUBE. Really appreciate it. Thank you. My pleasure. All right, keep it right there, everybody. We'll be back with our next guest. This is theCUBE, we're live from HPE Discover in London. Right back.