 I liked about that. So my next guest is Michael McInerney from the BCS. From BCS. BCS, okay, welcome to theCUBE. Business Critical Systems. Okay, I'm John Furrier with SiliconANGLE.com. We're here in Las Vegas, Nevada. HP Discover 2011 where we're on the ground with theCUBE, our flagship telecast. We talk to the smartest guys around. Executives, thought leaders, bloggers, anyone who wants to share their knowledge with us. We're going to have a conversation and go deep, expand those ideas. So Michael, welcome to theCUBE. So one, give me an update on what's going on. We've had some of your folks in talking about some of the changes in the products. We had Intel on, talking about the chip. Okay. Give us a quick update from your perspective on what's going on in your world today. Sure, no. So I'm responsible for building, essentially the mission critical products at HP. So all of the servers, integrity-based, and also some of the large-scale x86 servers. So Ben, as you've been mentioning and talking about, you know, a lot of the effort has been, excuse me, has been around driving the converged infrastructure, driving the mission critical systems in the converged infrastructure. So we've been investing a lot of time into that, so you'll see we have a common architecture from x86 all the way up through our sort of flagship Superdome, all built out on a common blade architecture. So our customers have a very seamless experience as they're deploying out their architectures, not get rid of the sort of islands of isolation in their data center. So a lot of people don't know this, but I actually worked at HP for nine years, from 1988 to 1997, and HP was an eight billion dollar company when I joined in the late 80s, and you know, there are hundreds and hundreds of billions of dollars, 26 billion roughly revenue. Give or take a few tens of billions here and there. So they're known for high performance computing. It's not a new DNA for HP. No, absolutely not, right? So we're out there running billions of credit cards, hundreds of millions of dollars of transactions on these mission critical systems, stuff that you'd never see, but you know, how many of you used your cell phone today, likely that you were actually touching one of our integrity systems or one of our high end, high availability systems. So we continue to see that out. We see some interesting things happening out in some of the emerging economies where they're out trying to deploy this infrastructure that here in the US were sort of, we expect, like we have credit cards, we have banking infrastructure, we have customers coming to us and saying, you know, I need to go out and deploy 7,000 new banks over the next two years. You know, how do I do that? And we say, ah, we have an idea, all right? And so, you know, we have that sort of trusted mission critical architectures out there that we deliver in the marketplace. Yeah, obviously the press, I mean, mostly in the blog's blog side, in which we do have a blog, but technically we're bloggers, we consider press, but even on the mainstream press, they love to talk about the death of something, right? The death of, you know, the death of the mainframe, death of the mini computer, death of the land, death of client server. And so cloud brings in that conversation, the death of this, and you know, when we talk to the Intel, they're like, oh, Intel's going to go down, Moore's Law, people of virtualization's going to kill Intel. If they invest in cloud, their business is growing, there's great investment, so the world really never changes. It's kind of the same pattern and movie all the time. So we still have a need, even though there might be federated clouds and, you know, distributed network out there, it's still the same distributed network. Computing and high performance computing still is a big part of that. So just for the folks out there that aren't on top of the whole, you know, high end, I don't want to say mainframe class, as a kind of mainframe is kind of old, but there is that high performance, it's kind of shrunk a little bit in size, but horsepower wise, there's still that massive nodes out there of high performance, and that's not going to change. So just share the folks, with the folks out there, that dynamic. Yeah, I think there's two things that go on there, right? As a lot of it is, there's some very hardened infrastructure that's out there that's deployed, right? So if you're out there and you're, you know, a large manufacturing company, and you have an accounts receivable program that's out there and running, and or a billing program that's out there from a telco that's out there providing your monthly cell bill, you know, what happens if that didn't work, right? So, well, that'd be a big problem, right? Collecting money is an important part of business. So a lot of these applications and these infrastructures have been hardened over time, and people just are not going to replace those. The risk reward of going in there and saying, I'm going to change my billing application out because I think it'll be fun, isn't happening in the environment. So you continue to see that, and we continue to see customers looking at, you know, where is it that, you know, I have my cloud-based architectures, I have my sort of more general-purpose computing platforms, but there's some places in my enterprise that I still need better than general-purpose, right? I still need that, I need higher levels of reliability. What's changing though, I mean, so we still have the need for high performance. So a lot of people in the cloud side say, hey, there's an abundance of compute. So we've done a lot about the storage, talked about three bar a year ago. Storage has always been, is now sexy in the middle of the conversation as a subsystem, but there's been a conversation that people have been saying, oh, we have an abundance of compute, and we just, the old grid argument, I don't remember those grid days, but you know, I mean, is that true, is it false, is it kind of true? Can I just talk about these issues? No, it's a great, I mean, it's a great question, right? So I think the idea here is, you know, what is performance? Right, so when we start talking about mission-critical environments, what does performance mean to you? Is performance, so I'm a hardware guy, right, and so I want to talk about processors and cores and frequencies, but when you go out there and look in the mission-critical environments, performance to those customers is not necessarily how many cores do I have, it's what's the response time on my application? How often do I need to tear down my application a year to upgrade things? How do I, how many unplanned failures do I have in my architecture? And this is where minutes matter, right? If my billing application for texting goes down for a half hour, how much money does a large telco lose? It's real dollars, that's a real dollar. Real dollars. So the production, we talk about SAP with the software side, same thing, you've got production systems out there, since the, you know, how do you change the air plane out, midair, blah, blah, blah, blah, all that argument, but you know, you're talking about computing, you're talking about mission-critical, how do you guys fit that asset or capability into these new dynamic environments like cloud and big data? Absolutely, I think that cloud and big data are just great examples of sort of an implementation, right? So I can go out and say, can I do these things on the cloud, right? Well, a lot of times for the mission-critical pieces, these will be the last pieces you'll see sort of go out to the cloud, not because of the technology, but more because of the sensitivity of the information, right, if it's in healthcare. This is actually my, you know, my provider data or my customer data, I just can't share that. Financial services, telecommunications, this is my core data, this is my company. You know, I can do some of these things on the cloud, but these other things I'm going to keep in. So you'll start seeing us talk about hybrid clouds and building clouds that are internal and external. That'll definitely happen in this marketplace. And then the question just becomes is, do you have those mission-critical clouds, right? Do you have the high availability? Do you have the scalability that you need within those cloud architectures? And that'll be the interesting thing I think you'll see develop out there. We're here with Michael McNerney with HP, runs the group Hardware for the Big Super Dome. I mean, this is the high-end performance, computing the devices, talking about all the things and why it's relevant to have all those high-performance dedicated machines from HP. Kind of a philosophical question for you, Michael, around systems, you're a systems guy. I can imagine, you talk about, you know, you love cores, you love frequencies. So, you know, when you design a system, a motherboard or whatnot, there's a lot of engineering involved. And you know, it's packaged up into a hardware machine. But now, we've been talking about the systems view and Dave Donatelli and I were chatting about that, yes, on Monday night around the cloud, in a way, as an operating system, or its own system. It might not be encased in a machine, but there's system elements involved. So, can you share with us the dynamics or kind of high-level philosophy around architecture around how your machines, your hardware, as a component of a new meta, higher-level system? Sure. It certainly has been true, and I think definitely, where do you define that layer of abstraction? Where do you define that layer of systems-ness? When we look at the mission-critical systems, this is what we're talking about, right? It's like, how do you build reliability in not just at the CPU level? So, my favorite story here is with the integrity systems, we have un, what are they called? Unconsumed and consumed errors. So, there's consumed errors in the processor. This means that a processor has a fault. But it doesn't mean that the system fails because the rest of the system is able to sort of solve that. So, actually the processor has failed, but the system hasn't failed. It's been able to back it out. It's been able to sort of either identify the application that that processor or that memory fault was associated with, and kill just that process or something like that. So, you see that sort of systems-level things are built in, and this is really where we drive things. So, if you look at like virtualization capabilities, having those tightly integrated into management so that then we can provision new servers, do capacity on demand. You know, I want to fail over a server. I want to move a bunch of workloads on it. I want to dynamically add capacity to that new server. I don't want to just have that sitting there. You know, how do you tie all these things together? How do you build out that system? That's really the key of what we work on. Talk about virtualization from your perspective. Obviously, virtualization's been the heart of a lot of the conversations. Paul Maritz, who's speaking, I saw on Twitter, actually, I think he's speaking here, talked at VMworld last year. We had theCUBE there. He said, first time in history, virtual machines, server shipments, surpassed the physical machines. Sure. That scares people, but it's a good technology advancement. Has virtualization helped your business on your end, or is it just you deal with it? So, two answers, really, and this is maybe one is sort of just showing my age is, you know, Unix systems have been virtualizing workloads since the beginning of time, right? We used to call this time sharing, and now we moved into, so virtualization is definitely sort of alive and well in the mission critical environment. A lot of things we do here is, we will do things like electrically isolated partitions. So what this means is I've virtualized the system, I've sliced it up, and nothing that happens into the system down to, you know, you putting a bullet through it, is gonna impact the blade sitting next to it, right? It's completely electrically isolated. So when we start talking about mission critical virtualization, you start talking about this higher level. How do we do things in a hardware level to guarantee complete isolation? How do we partition workloads versus sort of virtualizing with a hypervisor, you know, that a lot of people can't take the performance hits on, I've virtualized IO, now all of a sudden there's this intermediary in my IO causing delays, I need raw performance on my IO. I think most people, I would say to the folks out there, it helps your business because it puts an emphasis on some of those mission critical virtualization. I guess what I would say is, you guys are in essence in the mission critical business. Yes, absolutely. Whether it's virtualization or... Yeah, and so there's this sort of notion of a hardened virtualization. What does it mean in a mission critical environment? And I've had customers sort of have that conversation with me where it's like, you know, what do you think of virtualization and a little bit of a broken English conversation and they're sort of like, no, no, no, we want the UNIX partitioning, we want these electrically isolated partitions, not just virtualization with the operating environment. You know, I've been watching, obviously, working at HP going back to the day, actually met Bill and Dave when they were still alive in the late 80s when I joined the company when they first would welcome guests. HP's story, history around HP UX and the servers, when even the old mini-computer days, it's legendary, we all know that. But what's interesting that people might not know about is that HP is really the only mini-computer company to survive, you know, deck, you know, kind of she rolled into compact, which ended up ending up into HP, data general went to EMC and so, you guys have transformed and reinvented yourself. That was helped with the commodity server side of the business within, et cetera. So you guys have successfully always innovated on that. And when we talked to Intel, we asked them a question about their innovation challenges when a year or two years ago, and Cloud was an early investment for them. So the question, and it's been paying off for Intel. So my question for you is around investments. How are you, what are you guys investing in now that you see paying dividends down the road on the roadmap? So I think, you know, we continue to invest in, you know, it's not super exciting, right? It's continue to invest in availability, continue to invest in scalability, so. The boring stuff that keeps the lights on and keeps the lights on. Keep the lights on. And yeah, and so people, yeah. We sit around in marketing meetings for a long time and come up with what's our value proposition and it comes back to RAZ like every time. And sometimes we, you know, we change the letters around and try to create some new acronym. But ideally it really is the scalability to system. So today, in this show, we're announcing we're gonna scale Superdome to 32 socket. So we're gonna take 16 sockets, essentially two chassis, wire them together, turn it into a 32 socket system. You know, sort of really sort of upping that level of scalability. You'll see us continue to drive that with Pulse and we'll go up to like 256 cores on that system. You know, single image system or, as you mentioned, a lot of people partition that up into different workloads to drive that. So we'll see a lot of the scalability, a lot of the availability and then just keeping all of the virtualization capabilities and other pieces sort of built out. So what's the price tag for a 250 core machine, just ballpark or something? You know, I don't know the number off the top of my head. It's not cheap. I mean, it's not 5,000. It's not 20,000. I mean, it's like, I mean, it's in the, it's big numbers. It is, absolutely. What solar system is it in? I think the, I think it's hundreds of thousands of dollars. Absolutely. And I think when you look at those costs, right, what's really interesting about that is what is that compared to your total solution cost? So if you're talking about deploying a large scale SAP environment with, you know, help from outside, essentially services, consulting, you look at the software costs, you look at all these things, the actual hardware cost is fairly minimal in that equation. So a lot of times I think one of the conversations I have with people is, you know, how much does it cost for a Superdome? Well, how much does it cost you not to have a Superdome, right? So what does it mean when your APOs, SAP APOs goes down? It's a big purchase because of the mission critical. I suppose it's going to get a feel for the, for the device component and see the services wrapped around it. And I think you also have to look at what is the cost of building that equivalent architecture, right? So you sit there and say, I have Superdome. It means I have this cluster of servers built into this fabric. Can I do that with an x86 environment? Can I take x86 boxes, put them together, with let's say Infiniband, cluster them, do all of these things and put that whole package together. You know, what does that cost me? What does that cost me to actually operate and run? And I will tell you, we have those conversations with customers who are saying, I'm going to go build this out in a x86 architecture. And we say, okay, what is your sort of core value? Where do you want to invest as a company? Do you want to be a server vendor? Do you want to be a bank, right? And so Superdome can come in and build you a turnkey mission critical environment. And when you start looking at those costs, you start looking at the TCOs, you know, that will still come out on the positive side. And you're dealing with a very sophisticated, very savvy audience because you've been selling to these environments. It's just your standard big investment houses and they're big companies, multinational companies, right? It's not like, it's not SMB as everyone would think, hundreds of thousands of dollars. If they want to buy a Superdome, we're glad to sell it to them. Law firm, buy Superdome, you know? I guess my final question would be around where the action is right now. I'll see cloud big data, all that trends going on here. Where's the action for you guys in the marketplace and use cases in particular? Because it's a high-end machine, it fits into the architecture. You got the conversion networking group with all the Lego blocks from entry level all the way scaling up to the low end. We just had Jeff Miller on from, or Paul Miller on from the ESSN group talking about the Lego blocks and appliances. What's the use case that's sexy right now or hot for you? I think, you know, in the mission critical environment, what you tend to see is, you know, you talk about big data, right? Adding all this extra data, having the ability to analyze and drive all of these data. In most enterprises, everything sort of always comes back to a couple of core systems, right? So you still can do a lot of analysis, you can still have a lot of opportunity, but it's that core that you have to make sure is going to scale to meet this demand. That core has to be flexible. So how do I take my core data? How do I take my core business applications? That it's great to have these new capabilities like big data and being able to do analysis where you never could before, but a lot of times you're still going to be pinging on those core systems or still dependent on those core systems. And so as an IT manager, how do you make sure those aren't going to become the bottleneck, right? How am I going to make sure that core is going to scale when I need it to, expand when I need it to, to meet all of these tremendous opportunities out there? So basically what you're saying is, no matter what people talk about, the game is still the same. Whatever architectures they roll to, cloud, big data, core systems are at the heart of any infrastructure. I just think it's a, you want to make sure, for our customers, a lot of them are focused on how do I make sure my core data doesn't become the bottleneck, right? I can go out and do all these things with big data and make all these things happen, but I still have to be able to get to customer, these five facts about my customer in sub millisecond time. Okay, how do I do that? And that's where they build out these mission critical architecture for. Final question, I'll let you go. For Superdome and the super monster machines that you have there, horsepower integrated in, what's the future for you guys? What can people expect from Superdome and some of the new advances you guys are coming out? Sure, no, it's a great question. So again, boring is hack, right? We're going to get bigger, we're going to get faster, we're going to get more reliable. Betcha you talked to. Yeah, exactly. And you know, this is what we do, is we'll eat for lunch, you know, breakfast, dinner, and that's what we're going to see from us is just continuing to drive that envelope. We think the mission critical requirements of today are going to pale to what we need to do to deliver that for 2020. So how do we keep building and driving that new limit? Okay, okay, Michael McInerney, thanks for coming inside the Cube, Superdome BCS, which stands for? Business Critical Systems. There you go. Okay, thank you so much for coming on the Cube. Thank you so much. Thank you.