 And we're back in Las Vegas. We're live, the Cube, SiliconANGLE.TV's the Cube, our flagship telecast. We go out to the events where there's news and events and we'll get the action and extract the signal from the noise. And we're here in Las Vegas for the ULIPAC or ProLiant Gen 8 server announcement. You just saw the press conference live. You can always come back to SiliconANGLE.TV. We'll have highlights and on-demand video there to watch it, but we're gonna now extract the signal from the noise and break down the event. And I'm here with Dave Vellante, co-founder of Wikibon.org, with his co-founder, lead analyst, David Floyd. Guys, welcome to the session here. Obviously, the roundup here, we wanna just go in and drill on the roundup of HP's announcements. We're in Vegas, a lot of action going on. VMWare's here with their partner alliance down the street. You got the Data Warehouse Institute here. Your colleague, Jeff Kelly, is out there scouring the landscape in the Data Warehouse world. And not to kind of pat you guys in the back, but I will. We've been talking about this move to Systems Architecture, David and Dave. IO Centric Infrastructure, David Floyd recently put out that's been very big hit. Jeff Kelly put out the Big Data Manifesto. And SiliconANGLE obviously covers cloud, mobile, social, and big data. So we love this. I mean, this is like for us, red meat. So let's break down the announcement from HP. They announced the Gen 8 storage. So Dave and David, let's break that down. Dave, what's your quick reaction to the highlights here? Obviously the digital world, we're exploding with data. Server architectures are changing. Cloud people, cloud service providers going crazy. So what's your take? Well, I mean, I think the trend was quite obvious to us that data was really the big factor in the industry. And we're living in a data-centric environment. Big data is born out of that. And it's interesting, David Floyd, to hear now a server vendor, HP, the leading server vendor talking about the data tsunami. You usually used to hearing storage. People talk about that. And there's a chart they show up into the right. It was the first point, the first pillar, they called it, HP, that they really talked about. And you've been talking about IO-centric architectures now for quite some time. We've been covering big data at Wikibon and SiliconANGLE. And are we seeing a change in the business model, in the problems that server companies are trying to solve? Yes. The server structures are going towards, away from structured data to unstructured data. The amount of unstructured data that's in the servers themselves is increasing dramatically. And there's a move, in general, from the storage arrays to a lot more data in the servers. There's no doubt about that. And that move means that they have to solve the problems of storage in those server racks. That's a big problem. Let's just break this down. So let's just review the key products. I took some notes here. I'll read from the paper, from the keynote, I mean, the live press conference. Obviously, their number one, they call a pillar, is storage IO. And we're talking about solid state, smart data protect, and data services. That's a storage thing in a server announcement. So they're talking about convergence. So that really addresses some of the dynamic workload. Pillar two that they announced in this server innovation is the whole server management. That's the classic, I got a zillion servers and they're stacked up and there's management of that and all kinds of provisioning issues. And third is the space issue. Data centers aren't getting any bigger. They cost more power and cooling. So break that down. Are those three areas, did they do a good job talking about that? Well, the server area, if you started storage area, that job is being addressed. I don't think this announcement says actually anything very new in that particular area. They've had solid state drives, they've got PCI drives. Yes, it highlights that they've done some work in terms of automating and getting information about that. But the actual products themselves, that's not new. Really, that's an interesting nuance that you're bringing up because essentially the messaging would say otherwise. I mean, getting data in and out. The messaging says we're getting data in and out faster. I infer that they're actually re-architecting systems to do that. In your opinion, that's not the case? There's a little bit of architecture. They've got some nice solutions for data protection within the servers, that's nice. But in general, the industry has got a hold of that and serve HP. I think far more important is they're addressing the complexity of these large numbers of servers in Iraq. OPEX is the real issue that they're trying to address. So would you say that's the core problem? I mean, that seems to be the core issue. Because so much human effort, human time required to provision, to manage, to service, all of these servers, loss of mistakes are made. The biggest cause of downtime is people pulling things out. Human error. I just tweeted that. Yes, human error all the time. And it's absolutely true. So it's automating that. It's trying to simplify, take away things that are not necessary, collect all that data, and make this closed loop feedback. And I think that's an incredibly important. I would agree with you. I would agree with you. I want to just drill down. That's the server thing. Because that's pillar number two. But let's go back to the pillar number one. Storage IO, obviously, I just tweeted it's all about storage and converges, baby. And that was kind of my soundbite. But really, you know. Is that a pound, baby? You know, pound Gen 8. Actually, we've been CloudConnect and Santa Clara where Alice Williams is at. No, but seriously, look, this notion of dynamic workload acceleration, it's science vague to me. I mean, so we've talked about this in the past with Fusion IO as a context at the Node Summit, where Node.js was addressing a lot of those IO latency issues. That's a different app set, or is it? What's your view on that application marketplace? Oh, is it important to increase the IOPS through the servers? Absolutely. I mean, that trend is very, very strong indeed. And it's been happening for some time. My only issue is that there's nothing particularly new in this announcement about that trend. So it's not very meaty. So when they say dynamic workload acceleration, what is that? Buzzwordy? I mean, it's not cloudwashing. It's dynamic workload acceleration washing. I mean, what do you call that? I mean, you know. It's a great set of words. Sounds great. Dynamic workload acceleration. OK. But that's solid state. I mean, we're, you know. It's solid state. Did they miss that? There's some nice things they're doing in terms of monitoring what the workload is and being able to, if you need to, put greater caches in and be able to use caches to monitor that. And that's particularly valuable in the VMware environment. So there's some nice things in there. But that's not where the money is in mind. Dave, don't jump in yet. I want to ask one more thing. Architecturally, are they in a good direction? Oh, absolutely. The architecture of capturing all of this data, being able to take it, give it to their partners, give it to themselves, simple things, but like knowing what server is where, what the warranty is, all of that sort of simple, seemingly simple stuff is incredibly useful in being able to actually service it, whether it be a partner servicing it or whether it be themselves. Who are someone remotely in a call center? Yes, absolutely. The right data. That to me is incredibly useful and the right architecture. And that's going to allow them, if you like, to avoid that together with the consolidation of servers and storage and being able to combine those. It's going to keep them away from other vendors, for example, trying to put in very large mainframe type devices. It'll protect them to some extent from that market. So David, you were talking about, they said earlier, that's not where the money is. Where is the money? You talked to a lot of clients. What are they telling you in terms of where the money is? Specifically, I'm talking about CIOs and admin types. The money is in the operation and services costs. The cost of maintenance, the cost of operations, that's the money. And in the work that I've done, the potential savings, that's where roughly about 50% of the savings are coming. So Mark Potter, we just listened to his keynote. He said that $24 million is spent on manual operations over three years. $29 million in energy costs. This is for the typical data center over three years. Interestingly, the focus on unplanned downtime costing $10 million an hour. I want to come back to that. They're talking about Gen X delivering three X admin productivity, giving 30 days annually back to the admins, et cetera, et cetera. What are your numbers showing? You've done a sort of a back of the napkin analysis of this. What are you saying? Are those numbers valid? Are they spot on? Are they inflated? Some of them are, and some of them are. You know, they're mixing apples and oranges. Which ones are spot on? Which ones are conservative? Which ones are double counting? Cut through that for us. So one of the parts of their announcement was that they could improve the power by 70%. And that was power per compute. Well, I think we have to take it for granted that the power of the servers are getting greater every year. So that's a Moore's Law. That's a Moore's Law over the rest of it. OK, well, that sounds good. But I see, but you're right. Anybody who's on Moore's Law, which is everybody, can claim some of that. Could claim exactly that. So that's a large thing you're saying. The preponderance of that 70% is Moore's Law. No, about half of it is about, but maybe just a little bit over half is Moore's Law. And the other half is Gen 8 in addition. It's genuine, yeah, absolutely. So it's a significant number, but it's not 70%. Yeah, yeah, OK. It's good marketing to say 70%. It's a nice number. It's a nice number, yes. It's 50% is a damn good deal. It's a fact, though. It's a fact. It's a fact. Fact-based marketing. It is a fact, it's true. But that's not what a customer is going to see, one of their customers is going to see. I mean, customers want reliability. The channel partners in their sales force want to maximize the service revenue at the lowest cost point, which is automation. OK, so that helps us squint through that piece. What else? So the three areas where they're going to save money are in facilities. And of course, well over 50% of IT shops don't pay for facilities. So that's a little bit difficult sometimes for them to get to focus on. But it's certainly very important that they should focus on that. Because the acquisition cost, the cost of power in facilities is roughly the same as the acquisition cost over the five years. Did you get a sense from their design innovations? They were throwing some pretty good numbers up to outside. We tweeted $300 million to your project. But we saw the video ad that they're running on SiliconANGLE TV as touting 150 design innovations. Do you have any insight into what that is? There are. There are a huge number of small innovations, which makes it very, very difficult to summarize. But lots of little. Like what? Monitoring big data stuff you mentioned earlier? Monitoring it, for example, being able to put in a new process chip without bending the pins. It's all of this. Human stuff that breaks. Human stuff. There is lots and lots of little tiny ones. I could make a list about a thousand lasts. We live in an air-conditioning. That could be 24-7 on this. So, John, I want to come back to the economic model, because HP is making some bold claims, and I want the audience to be able to hear David's analysis. So what's the bottom line? How much is this actually going to save a customer in terms of server costs? We have a standard model. And it's a billion-dollar company, 4,000 users, roughly spending 3% of their revenue on IT. And a mid-sized company of that sort. We have a standard model, which we use. And if we look at that from the point of view of the overall IT budget, and that's development, everything, PCs and everything, which are not affected by that, my estimate is that this generation will save them around 4%, about a million dollars per year. So over three years, that's $3.3 million. So that's everything, infrastructure, facilities, maybe a little bit of development. Exactly. How about the server piece? What if we drill down into the server piece? So if we take servers, servers are around 20%, 25%, 26%, if you include the facilities cost and some of the software cost and things like that. If you take them out as a group, around 26%. Now, of that 26% of that $30 million, then there's a saving of 14%, 15% of that. That's directly attributable to Gen 8. In my view, they'll be able to make savings of around 15%. Now, sometimes that'll be indirect, because it'll be the services that they buy in. Well, like you said, facilities, for example, or services they buy in for maintenance, or services they buy into, they may outsource the servicing. So they will see that indirectly through more competitive quotes. But if they're doing everything themselves around 14, 15%. How much of a competitive differentiator is this, in your opinion? I actually think it's pretty good. They've put in a lot of effort into addressing this, and it's been a two-year project. Now, obviously, the other vendors, there's IBM, of course, and they've got a pretty good track record of innovation in this area. But HP, in my view, have got by far the biggest push in this whole area of reducing operational costs that I've seen, other than things like VCE, where they've really taken a limited choice. Yeah, it's sort of a green field. It's a green field environment, and they've limited choice, basically, by saying, okay, you can have those two cells. The Henry Ford, you can have any color you want, as long as it's black. Taking another approach, and a valid approach. Given that HP have so many different servers, all of these commodity servers, this is, in my view, the best version that I've seen by a long way. How about the broader context? We're making a big deal out of three major announcements in a hundred days. We saw Project Moonshot, Project Odyssey, and now Project Voyager. Do you, as a person who's observed servers, used to be at IBM, you know that market really well, do you see those as the mark of a new change in this space? I see them as HP being investing strongly in this area. I see the consolidation of servers and storage, being able to put them out as a single skew that move towards that, again, as a very strong way of reducing operational cost and maintaining market share. I think it's a strong move on their part. And IBM, Cisco, and Oracle, I suppose, are the three biggest competitors in this area. So they're going to have to invest as well if they're going to stay up. Each of them has their own sweet spots, obviously. Mainframes are doing very well with IBM, and Oracle is doing very well in the database area. So I want to go back to John's point about this IO-centric architecture, which I think is a really major theme and one that we've been hitting hard. So you've got, we saw last week EMC announced VF Cash, we've certainly covered Fusion IO, HP's a big partner of Fusion IO, we see HP Today talking about Flash, certainly Oracle and others have announced Flash. Where, in your opinion, so we're seeing this flash hierarchy emerge, change the storage hierarchy, where do you see as the logical point of control? First of all, do you need a single point of control and where should that be? Should it be in the server, the storage array, the network? There are two different levels. There's a level of management that's required for active data, and there's a level of management that's required for archiving and other sorts of data like that. So, and they can be in two different places. But it's clearly that if big data is the way that things are going IO-centric, the data, the architecture will be top down. You'll develop the applications together in the servers. You'll bring together that data. You'll bring it down so that the metadata and the indices can be provided in real time. And the management of that has to be much more server-centric than is storage-centric, which has been the way it's gone for the last 20 years. It's now the pendulum's moving back. So I strongly believe that the management of that will have to come more from the server level than it will from the storage level. For the archiving data and the backup data, I think that can be much more of a storage-led environment. But you cannot manage things from a slower place to a faster place. You have to be top down. You have to manage top down. There's a quote here on Twitter. I'm reading a comment. Just checking out the Twitter stream while you're elaborating on the architecture. You have to be careful not to make the servers too smart. We don't want to be out of a job. So, I mean, that always brings, first of all, we've been talking about how the economy has turned around. For the past decade, IT's been a living hell. You know, budget cuts, budget cuts, get the OPEX number down. But now we're seeing this shift towards this new architecture IO-centric among other things. But we've talked in the past, the three of us on theCUBE here, about top-line growth, where these apps are generating business value, right? So there's that top-line revenue growth. So it's a zero-sum game, in my opinion. What's your view on all this? To me, it's not a zero-sum game. I think the value, the business value of big data applications, as they come on board, as they're redesigned, as IO is taken as a constraint, is taken away as a constraint. The value of those applications is going to be enormous. And I see an increased spend in terms of new applications, new ways of doing things that will give very, very high returns, particularly to those that are the leaders. So answer the guy's comment. I mean, he's obviously being tongue-in-cheek, but I mean, it is a concern. People always worry about their job, and we're hearing great things about the channel here, and HP Storage is opening 100% selling through the channel, and new clients and storage, and so there's money-making and services, the services angles we say, yeah. So what do you have a job? If you're a disk monkey, and you are replacing disk, it was a variation of a tape monkey, then yes, your job is in danger, but there's so much opportunity by being able to... So when I say zero-sum game, I mean, okay, we talk about DBAs being shifting to data science. The server guys, managed in the racks, they're going to become more analysts. They're going to become partners with the applications on delivering the lowest cost, highest availability performance of systems. That's where they need to be, is helping out. So they won't be data center monkeys. No, they'll be application. You know, John, I just wrote a piece on the way out here. That's a blog post title. So I got a piece that's up on the homepage of Wikibon. How storage admins can stay relevant? Yeah. What's your advice to server admins? How can they stay relevant? Answer the question, what can I do to help applications run better, faster, better response time, less downtime? If you're answering that question and you're relevant to the applications, that is the way that you stay in your job. Is that now, I mean, that's always, we've talked about this, the foreign culture of DevOps, developers versus operations, and some people say ops and devs, and so operations, just a different culture than applications, is that an issue? Sure it is. I mean, if you try and stay in the past and you want to keep things exactly as they are, that's going to be a challenge for you. You've got to be out there helping the Node.js of the world, put their applications in and get them working, even if it's in a way that you don't, is foreign to you. We are here live at theCUBE, John Furrier at SiliconANGLE.com, SiliconANGLE.tv at theCUBE, our flagship telecast, we go out and broadcast the signal from the noise, interview analysts, experts, CEOs, startups, you name it, we'll find it, we'll talk to you. Here are David Fleurer and Dave Vellante from Wikibon doing the deep dive analysis and commentary on the ULT Packard Gen 8 launch, which is the ProLiant new servers. Quick highlights here, and then I'm going to ask you guys to grade them, okay? So HP, Gen 8 solution, two year effort, $300 million effort total, eliminates manual tasks, proactive insight architecture, basically application performance increase, and then 80% less energy, huge power issue, we didn't even get to that yet, this whole thermal sensing thing is cool, whatever they called it, inside thermal something. Reduced downtime by cost by 40%, 6x performance on the smart array controllers, the new array controllers, and then management tools, 3x productivity, location discovery, science and services, this whole power thing. This is scaling beyond compute. You know, the joke we were saying earlier, servers are the new peripheral, right? So servers just compute is now integrated. So that's kind of the highlights from the ULT Packard Gen 8. So guys, on this launch, I want you to grade them from a scale of one to 10, one being low, 10 being the highest, okay? In terms of relevance to the marketplace, product, technology, business model, and benefits to customers. So market alignment with the market trends, product technology, business model, and then benefits to customers. I give them an eight on market alignment. I think they're not tapping into the big data story in a way that will differentiate them. I think they will need to catch up with that part of it for that one. The second one was... Product technology. Product technology. Within what they're offering, what they've addressed, I think they have, they're a nine on that. It's really excellent development of the products themselves. Business model. That one, I think, David, you could probably comment on better, but I was very impressed with their ability to go through the channel and provide these services to the channel themselves. You gotta give them a 10 on that one. I think so. I mean, the Russian judge would give them a 10. Yeah, and I think just to back up a little bit, I would agree with you with relevance. I had them at a seven. I thought technology and product at 10. I think business models would make them. Hold on, hold on, hold on. What's your score? So you wanna give them a 10? I was giving them an eight for the first one. I gave them a seven. You gave them a nine. Oh, hold on, let me see this. So market eight, product to technology nine, business model. Business model, I think a 10. 10, okay, benefits to IT. Benefits to the IT. Remember, there are a lot of other problems in IT as well, in the development of other things like that. So in terms of overall relevance. Is this an aspirin or a game changer? It's more than an aspirin and it's not a game changer. So it's somewhere in between. Okay, all right, David's lawyer, you're on the books. We've got that recorded. Dave, now your grades. So my grades are, so I would give them seven on relevance. I think very high on product and technology. I give them a 10 actually to David's nine. I think business model also very high on nine or a 10. I think the whole channel thrust is really impressive. HP's getting serious about the channel. There's a big land grab going on with the channel. And then I think benefits to IT, I would agree. I mean, it's not a total revolution here. I mean, it's really great for the server admin. I mean, really great news for the server admin now. Where's the server admin in the grand scheme of things? It's, to me, it's you got to touch application development and you've got to align with the business. And I think this helps. It's great. It's good stuff. And the other thing is competition. I think that it's from HP, HP continues to lead in servers that clearly are number one, particularly in this ISS space, industry standards. Commodity space. So let's run down your scores. So give me the numbers again. 10, nine. So relevance? No, you were. Oh, your score? I was an eight. No, what's your score? I was seven. Seven. Product technology? Product technology, I was a 10. I was the nine. Business model, I was a nine and a half. I was 10. And then benefits. Benefits to IT, I think, I was going to give us five or six, you know. I mean, overall IT. To server admin, you know, higher. Okay. So my scores are on the market relevance. I would have given them a five or six, mainly because of what you were saying about the big data story. They didn't really knock that out of the park. And I think the SSD story didn't come out true. But I'm going to bump them up to an eight and a half, mainly because I love the power and energy story. That thing is just, that is so relevant. And that is a game changer. So I give them an eight and a half on relevance. Product, I give them a 10. Business model, I give them a seven and a half. Because the channel, although great, you know, you have the EDS and you have direct sales force. Not sure how I was going to pay that. Only, they'll give them a seven and a half, only because I don't know that yet. Benefits to the customers, I give them a nine. Mainly because I think the energy piece and this idea of automating some of the stuff on the configuration really will make a big difference. You know, how about marketing? I think marketing wise, and I think again, the story, back to the stories, I think they could have really amped up more of the relevance around big data. They did talk about data in their announcement. It was a key theme in there. And I think if you look at the success of Splunk, for example. They didn't address the application issue. They didn't go after the application and say, how are we going to help? Okay, so that's the roundup drill down from the HP Gen 8 launch. John Furrier with our research team from Wikibon, Dave Vellante and David Fleuer. We're going to be right back to hear from the leader of HP's enterprise storage servers and networking, David Donatelli, in a few minutes. So watch this great ad by HP. Thanks for support for HP, allowing us to do our amazing editorial cube. We love doing this here. So watch this from HP and we'll be right back with David Donatelli.