 Live from San Francisco, it's theCUBE. Covering Red Hat Summit 2016. Brought to you by Red Hat. Now, here are your hosts, Stu Miniman and Brian Gracely. Welcome back to theCUBE, SiliconANGLE Media's flagship program here at Red Hat Summit 2016. Sitting here in the lobby at street level of Moscone West, happy to have on the program fresh off the afternoon, day three keynote, Randy Meyer, who is the president and general manager of Mission Critical Systems at Hewlett Packard Enterprise. Randy, thanks so much for joining us. Well, thank you for having me here. All right, so it's interesting times in technology, the mashing together of things that you're like, wait, how does that go together? So today, Mission Critical and what's happening with open source, bring us up to speed as to how those go together. Well, and I talked about it in the keynote. It's one of the fastest growing things we've got. That part of my business, scale up x86, resilience, open source software is growing 60% year over year for me. And we're watching, a couple of things happen. You're watching increased scale, increased capability. That's being driven by the open source community. Red Hat's certainly fostering that in their partnership with us. You're also looking at this trend towards in-memory. People are moving data in memory. SAP HANA's the poster child for that, but it happens with Oracle databases, happens with DB2 databases even. And you start to see when the data's that close to the processing, now analytics and transactions really start to meld and you can do some really cool things. You know, imagine your IoT data is now married to your transaction data and in real time you can make some decisions about that. That's changing the way people think about what they're doing. I've got a customer in Germany. They get information off of high-end air compressors, believe it or not. They have thousands of sensors in these things. It all goes into a predictive analytics database. That lets them ship parts somewhere in advance so that the air compressors there to service the Lufthansa jet that's sitting at the Frankfurt Airport so it doesn't go down. So amazing analytical capabilities and then they marry that into their transaction stream. That's one of the big confluence areas we're seeing. Yeah, we hear that a lot. We're starting to hear more. Things like predictive maintenance are one of the killer IoT applications and so forth. I mean, it really is and in order to do that you have to realize that it's not just collecting the data. You have to know how do you marry it in a smart way because the worst part about that is if you say, wow, I get the analytic data and I know that part's going to fail and I have no way to get it there in time. You got to marry it all the way back up your supply chain and in many cases that's got to go not only to your shop but to a partner where you say, I got to make my part plus somebody else's part arrive in the right place at the same time. That's the magic in this. You said scale up computing is growing. We hear lots of buzz around scale out and distributed system. What are we not seeing in the news and the headlines about scale up computing that people need to know about? Well, so there's certain problems that scale out really well and I want to go spray data across and I want to shard the database, do things. It turns out there's other kinds of problems that really don't distribute very well. Think about, I'm trying to do a massive billing application. I run 20 million telecom subscribers that all got to get a bill every month. It turns out it doesn't scale out very well. It wants to scale up. It wants to have the account database. It wants to have the call database and it wants to go map all that stuff all at once. I talked to my keynote about pillow windows. Great example. There's a billion different ways to configure a custom window and you can't spread it out because it's got to all integrate in there. So there's a certain set of applications and a lot of these for the longest time have lived in the Unix world or the mainframe world. We're now seeing the capabilities in the open source world that say I can get enough compute power going at it. I can put a 24 terabyte memory footprint all accessible by a kind of this sea of compute coming at it and I can solve those scale up problems. So what's interesting is a lot of the times that gets married to scale out environments where you say, oh, all those transactions are coming in across a web forum and I'm going to add data to them as they come across. Great. There's your scale out model that says add it right into the transaction stream, do what you have to do in the scale up model and then spit it back out that way as well. So Randy and Nukino, you said that technology today isn't about data centers, infrastructure or cloud. It's about people. Can you unpack that for our audience? Oh, you bet. My favorite concept. Ultimately, it's not about running a system or running an OS or running a database. I want to go down the street to Macy's and buy something. I want to know that it's there and I know that my credit card is going to work and guess what? What I really want to do is say I'm going to order it on my phone, I'm going to walk down the street and it's going to be there in a shopping bag for me when I pick it up. If I do that and say, okay, I've been gone a couple of days, I'm going to bring my wife a gift on the way home because I happen to live in the Bay Area. I order it, I show up there and they say, oh sorry, we only had it in the wrong size or you only had it in green even though you wanted pink. Is that okay? No, it's not okay. And you promised me that it would be there and ready. At the end of the day, that's a person trying to do something. It's somebody trying to get something done, whether that's a sales rep trying to sell something, a customer trying to pick something up, somebody trying to make an airline reservation, you trying to get data on your phone, whatever it might be. There's a person at the end of that business process and every interaction, if you can't make it work, somebody thinks you failed. And what's interesting in today's world, we're all connected. You guys are carrying at least three different devices I bet each, we don't think about it when it goes right. We assume it's gonna go right, but as soon as it goes wrong, that's a failure in your eyes. And it's not the company failure, you go, I'm personally disappointed. That's it. What do we have to do as IT people? How do I deliver on that promise every single day? How do I figure out to do that? And by the way, there's not infinite costs. There's not infinite number of resources. There's not infinite amount of processing power. Inside of all those envelopes, we all have to deal with as business people and technologists, how do I satisfy that customer 100% of the time that they're going to try to do something with me? Yeah, like you just talked about, there's different ways to access things these days. Mobile is an experience, tablets an experience, just being around is an experience. Being able to deliver that for customers, there's a cost to having access to all those experiences. What is the cost change that's going on between those Unix systems, those mainframe systems, and now moving it on to x86 and open source operating systems? Well, and that's one of the things you're getting to. The front end costs are multiplying like crazy, because you're right. The watch experience is different than the phone experience is different than the tablet experience, and I have to optimize for that. One of the things you want to do on the back end is say to the extent that I can take out the cost, but have a consistent experience there, so manage the front end the way you want to, but the back end's going to be clean and more effective. So we were talking about this earlier, there's a banking conglomerate based out of Austria, and they used to run on mainframes and IBM Power Linux systems. They went to our Superdome X product, run on Red Hat Linux for their core banking environment, which, again, you get through all these different channels. They took 78% out of their maintenance cost stream, which was massive, millions and millions of dollars a year reduced by three quarters. That's huge, and then you can go reinvest that in all the other front end, in the experience pieces of it. And they're able to maintain the same performance and availability they expected before? Same performance, same service level agreements, and the kicker was that new function, new features that would take them weeks to deploy, they're now doing overnight. The whole DevOps model because of all the tools that you get to do that is just easier for them. Yeah, if a customer sees this story, they're watching, they see the story, so Randy, what are the steps that I need to do? How can HPE help me get from that mainframe application to that environment that you're talking about? So the first thing that we want to talk to you about is explain what you're trying to do. Talk about the workload, what are your SLAs, where are you trying to go, and what do you have today? Because not very many of us walk in and say, okay, yeah, we'll just build everything new from scratch, everything will change. There's business processes that are running, you got to go do that. We look for opportunities where you say, ah, there's a piece of infrastructure that if you use a new model, this one can go quickly versus some over here you say, you know, the best model for that is gonna be connected into the new infrastructure. So our terminology, we'd say that you wanna build a hybrid infrastructure. There's gonna be some things that run in traditional IT computing models because they're either hard to move or there's no value in moving them. There's other things that you can now move today because the capabilities there, Linux has evolved, systems have evolved. We can move that and get value for you immediately. And then there's things you say, you know, it might wanna live in a public cloud. We have to say, how do you optimize that? So that's one of the things we do. We have technical professionals, we've got solution centers. You have to step back and look at the whole environment. It's not by a box. A box doesn't solve anything. For this set of workloads, here's how they have to communicate and how do you go do that? So it really is, to the extent you can talk about your business, your business processes, we can go in and say, there's the place. There's the win. Connect this in, work that on the mobile front, and that's gonna save you costs and it's gonna make you faster. Talk about it earlier. Time is the new currency. If you try to do everything big bang, it doesn't work. Solve this problem really fast and then go to the next one. So we definitely agree, we talk to customers. It's not one solution fits all. I guess one of the challenges is if I've got some big iron, I've got some public cloud, how does the management and orchestration fit across all of that? So that becomes critical and that's where software defined is the new mantra. And by the way, there's software defined elements for some of your big iron that's there. That's where as people are moving to, whether it's puppet chef, open stack, you leverage the capabilities there. I mean, I'm doing some pretty cool things. I have customers that have UNIX applications that they depend on that are unlikely to be able to move easily to something else. But I've taken and done some work, so I've got a prototype of my UNIX environment running in Docker. Okay, maintain the applications. You don't have to port those, but now I can leverage the new infrastructure. My nonstop systems, things that are running, banking networks and payments and telcos and that. I've got that running in my labs, full nonstop environment using KVM and Red Hat Linux as its underpinnings. So not too distant future, you want to put nonstop in a cloud based environment because that's where you go. I can do that. So you have to orchestrate across all those things and you got to leverage open stack, you got to leverage all the key things that are there to orchestrate that. And then underneath, you change them as you get value out of them. Which sometimes might be never. Yeah, Randy, one of the challenges we find is something like your mission critical system. Do you think nonstop Superdome? People think back, it's like, oh, I run the Superdome from 12, 15 years ago in what it does. If, you know, what message you want to give people to give them a fresh look at where your portfolio is today and how they should think about it going forward? Well, so first and foremost, the capabilities of doing what we've done for years and years and years, leveraging new platforms, X86 engines, open source software. If you can do a lot more than you thought you could today and there's clustering capabilities and there's that. There's also you got to walk away from thinking about, oh, used to have just purely specialized hardware and all this. Those are software choices that you're making, right? My nonstop business is a software choice you're making. Unix, it's actually a software choice. To the point that you can leverage that and say, know what you're doing, why you're doing it. You can create value and there's five or six different paths to success that are all going to let you reduce cost, get more agile, take advantage of the community tools that are out there, but at the same time, you can't break what you're running because your customers expect it to work all day, every day, year round. So if you're going to do one thing, have a plan, right? And if you haven't thought about it, where do you go in the future? That's where we can go help you and say, step back, look at all the different things you're doing and what is the plan? Because at the end of the day, there's a lot of opportunity out there. All right, Randy Meyer, really appreciate the updates. We'll be back with more coverage of Red Hat Summit 2016. You're watching theCUBE.