 Okay, welcome back everyone. This is theCUBE, our flagship program. We go out to the events, extract the signal from the noise. I'm John Furrier, the founder of SiliconANGLE, I'm joined by my co-host, Paul Gillam, with SiliconANGLE, my co-host. And we're here live at IBM Impact, and our next guest is the Baylaw GM of the Power Systems Group at IBM, CUBE alum that we spoke with him just a few weeks ago at IBM Pulse. A lot's changed. Well first, welcome back. It's great to be back in Vegas. It feels like many times already this year. Lots changed. At Pulse, we were talking a lot about some of the power systems and kind of how that HPC, mainframe, all that software enablement, but a lot's changed. Big announcements to everyone. Cisco, just give us an update quickly. What's transpired since Pulse to today? Yeah, so thanks John. Thanks John and Paul for having me here today. As you said, a lot has changed. We've been moving very fast with our power platform to connect it to the whole trends we see in the marketplace around cloud analytics, mobile and social. At the end of the day, it's really around data. And what we see from the new power-each systems is they are squarely designed to focus on solving the big data problem. So as these new systems come to market, they're not just addressing the traditional sort of relational database structure data. They're working on solving client problems around just scale of data, but all kinds of data. Unstructured, mobile data, right? MongoDB, right? No SQL, the list goes on and on. So data is a big trend. The other is economics for the cloud. These are really our first systems designed for scale out computing. Driving a differentiated price point for the cloud. And the third is the whole aspect around openness. We have fundamentally opened up the power platform. Paul and I were talking in the intro, IBM's obviously 50 years of innovation to the mainframe celebrating. That's kind of like Apple Pie and hot dogs and baseball. You know, it's great, but it's not a strategic, it's a cash cow, but it's not a strategic thing. But we were talking in Paul's just moved to the cloud. And in that sense, it's like a mainframe concept. So share with the folks out there that the tea leaves, read the tea leaves around, okay, mainframe's great, great legacy, you know, the motherhood of the mainframe's great, but now it's moving to cloud. What similarities are you seeing from those mainframe growth days where it was a real force and how is cloud becoming that new, in my words, the new kind of mainframe for lack of a better description? And how does that software component play into it? Yeah, certainly from a cloud standpoint, what we see is clients with on-premise infrastructure looking to find ways to rapidly and flexibly sort of move to a hybrid infrastructure. Yeah, so perhaps the cloud is a place they do their mobile application, they compose the objects together, they do their web enterprise applications, quickly connecting to that backend infrastructure. We saw some great examples of that here in Marie's section just moments ago on stage. So this hybrid notion, I think is really where we'll see enterprises going, leveraging their power infrastructure or mainframe infrastructure on-premise and connecting that to the cloud. Could be the IBM cloud or some other cloud. In the IBM cloud case, it could be running power here with our announcement at Pulse of power as part of the software infrastructure. One of the things you've done, Doug, is you created the Open Power Foundation now really to re-emphasize the idea that power is an open platform and you're looking to bring other partners into the fold. How are you, what are you doing to convince companies who may be accustomed to thinking of IBM as proprietary, as all or nothing to have a platform that you really are open, you really do want to involve them? Well, one of the interesting trends we see, sort of as we deal with this sort of composable application, composable enterprise model is every company is looking for the best technology to solve the issues they have. But the best technology doesn't necessarily mean it comes from one company. And they're all incredibly uncomfortable with one company setting the innovation agenda. So with our Open Power Foundation, we've said, you know what, we completely agree. IBM cannot be the company that thinks we can do it all. It's going to take a community of innovators to bring together the best pieces to solve different problems. So that started with an idea with Google, NVIDIA, Melanox, and TyAnne, the founding five members. It now has blossomed up to 26 members and growing daily. In fact, in San Francisco last week, we were signing new partners such as Canonical, which brings a bun through the market, and Rice University, who's working with the Texas Medical Center into the Open Power Foundation all around innovation and solving real problems with new approaches. And nevertheless, when you look at the members of the foundation, there's one company that stands out and that is Google. I thought you were going to say IBM. How important is Google's endorsement and they're really taking this ball and running with it to the success of this initiative. Yeah, so Gordon McKean is from Google. He's responsible for their infrastructure. He's also the chairman of the Open Power Foundation. So they've had a very strong hand in setting the direction for where the Open Power Foundation goes so far. And obviously, with Gordon as the chairman, he will continue to do that. They play a strong role in not only what they're going to do with it, and I mean, they've contributed developers, they've contributed software to the Open Power Foundation. Obviously, we've contributed the Powerade architecture in a licensing model, but also reference design. So, and I think that's what's exciting people to join. If you look at the memory companies joining the accelerator companies joining, and now really the new trend is end users joining. And I think that's creating sort of this whole global movement around open this versus closed. What does an end user have to gain by joining a hardware design coalition? Yeah, and we actually see it not as a hardware, although that's certainly where it started, Paul. So I acknowledge that. It's really sort of this end to end ecosystem around innovation. What I see end users looking for, back to this point of flexibility, speed of development, rapid, composable, fit my need is what they're seeing, fit my need. I don't want to feel like somebody else is defining what I need. I want to be able to take the best pieces and fit them together. Some of great examples would be the ability to use Powerade systems with NVIDIA, right? And bringing that sort of hybrid compute model of NVIDIA for GPU acceleration, say for Java workloads, great example of 10x acceleration with Java workloads with NVIDIA plus Powerade. We've got another example down on the floor of, again, composing together these systems to go do Monte Carlo simulation for the finance market. What is a magnitude improvement in performance? So we're seeing, we've got some great demos down on the floor of these sort of, couldn't you imagine? What if you could put together these pieces in a composable system and bring that to new use cases? I've got some questions out in the Twitter sphere in the social crowd chat. If you're watching this, you want to ask Doug a question, go to crowdchat.net slash IBM impact. We'll be fielding questions there. We've got the first question coming in. IBM Power sounds so IBM centric. How is IBM changing that perception to be broader? And how are you tracking non-IBM customers to come along with this new offering? Well, certainly Power is the IBM brand for our core processor, the Powerade processor, but then the systems we build on top of it. One of the aspects of Open Power is there will be other server companies in the market who will bring this power technology to their clients. So as I think about Open Power, I sort of describe it this way. It is about bringing power technology to new buyers and new markets, new buyers potentially being clients like Google and others in the sort of the hyperscale data center, and new markets like China. Certainly China is undergoing a tremendous change in the way in which they think about the domestic IT market. It's also about creating this innovation pipeline so that this community of innovators who are coming up with creative ideas to solve different problems I've described are able then to sort of bring this innovation to the power platform. So we're absolutely opening it up and I think of no better way to attract new buyers than to talk about openness. And certainly Google will be here. Google's involved in the openness of this. Talk about Google's role involved in power. Yeah, I can't, obviously I can't talk for Google. That's only Google for Google. No, that wouldn't be appropriate for me. I can't talk about the role, as I mentioned a minute ago, in terms of a chairman of the Open Power Foundation. I do understand here, down on the demo floor, Google has brought their own Powerade motherboard and I think Gordon McKean will be showing that off down there if anybody wants to stop by and see it. You announced a pulse that you'd be moving power to SoftLayer and lots of Watson on top of SoftLayer by the autumn, by the fall. Is that schedule still on track? Actually, if you saw a minute ago, what Mike wrote and was talking about in terms of Watson available in the cloud, that's actually running on Power this quarter. Is that for developers only at this point for business partners or is that for end customers as well? I think if I remember from what Mike described, it's really across the spectrum of potential users, developers and users as well. It's a great question for Mike, sort of the ultimate consumption of it is in his space. But from a rollout for Power in the SoftLayer, now it's interesting, we've sort of been talking about these things you mentioned, Paulson, obviously over the last couple of quarters and months. Underneath it all has been Powerade, we just never said that. When we talked about our commitment to KVN and Linux, the little Indian Linux, by the way, for ease of application porting, it was all about Powerade. When we talked about Watson as a service, it was about Powerade. We talked about Watson and SoftLayer, Powerade. We're sort of now here, so we've been sort of building up to this over the last nine months. So I want to ask you, Doug, one of the things that comes off, see IBM is so good about the messaging. You guys are all, you're right on the fault line of all the killer megatrends as the place shift and the tectonic shifts in the industry. One of them is real-time analytics. The other one that we were talking on the intro is this whole trend towards this maker movement. And I was commenting, I was at the Macintosh's 30th anniversary celebration of the Mac, and then, you know, and the open compute was now on this the next week. This movement towards tinkering and hacking on hardware with open source software really put a big trend. So you're seeing some things with Raspberry Pi here. Developers are really geeking out on this maker movement. It's really spurring creativity. So I got to ask you, how does that power, you hit a nerve there with power. So how is a developer who really wants to get their hands dirty and build software, how does the power system help them? What kinds of things can they do? Is it analytics, is there more sexier applications? What's your take, share some color on that? Yeah, so as we think about the sort of where power plays best in the marketplace. Certainly, from a design point, we see power really, as I said earlier, addressing all the types of data. And the good news is, there's no lack of data right now, right, and it's all kind of different kind of data that we see going on in the marketplace. So, you know, we talked about as an example, IBM Cloud, clouded a no SQL database for SaaS. So think about the benefits one could see of no SQL data on power and software, which is something we're actually working on, right, from moving power into that sort of no SQL space. So I think any aspect around data is a good fit for the power architecture. Now to your tinkerer question, one of the things we've really focused on here with this launch is making application and tinkering simple on power. And we've done that by completely supporting an open stack of software from Linux, not big Indian Linux, if you want to get a little geeky for a second, it's really the little Indian. What does that mean? It's the bitmaps are the same way as they are on x86. So now with little Indian Linux, KBM as a hypervisor, open stack for management and then the smart cloud portfolio on top, you've got a completely open stack of software that makes picking up a higher level language application, you can pick it up and drop it over on power without a single change to it. Think about minutes versus hours or days of what it used to take. The same is true of CC++. It's a simple recompile and go for 95% of the app. So we truly have addressed this sort of speed and development of applications on power. Doug, IBM sold off its x86 business to Lenovo and now you set up the OpenPower Foundation. Are you directly in conflict with Intel at this point? Are you competing directly with Intel? I think for the targeted workloads, I'm talking about data-centric workloads, enterprise applications based on Java, workloads such as mobile that are enablement to the core, absolutely yes. In fact, the economic point I made earlier, we see the economics of our scale off platform which delivers great performance at up to 20 to 30% less acquisition price. So it's a direct play for the infrastructure market. How important is it to you? You currently don't have any partners who are actually fabbing the PowerAge chips. How important is it to you to get a partner in the door actually making PowerAge? Yeah, the fab part isn't the most important part. What's important to me from an ecosystem is having chip development companies with me that are doing derivatives of the power architecture like Suzo PowerCore. It's about having IO members, memory members, software partners doing software around it and around end users. Building out that ecosystem now at 26, as I mentioned in growing, is really the focus of the open power foundation. If somebody wants to join from a fab, we're completely open to that, but I don't think that's as super important as the other areas. There have been previous efforts by companies to create sort of standards around risk architectures, PA risk, sun risk, even IBM many years ago. What did you learn from those efforts, most of which didn't go very far. What did you learn from those efforts that you're applying to power foundation? Yeah, I mean it always starts with, and it's going to sound kind of weak, but it always starts with listening to the marketplace. And the big thing that we did this time, and I would give Google credit for us for grabbing us by the shoulders virtually and saying, guys, get on with the Linux play. Get on with it in a big way and make it a little Indian so it's easy for applications to port to. And Buntu, we'll talk later here today with me about the way in which they've taken thousands of applications now within their ecosystem and brought them to the power architecture. They couldn't have done that before. So that was sort of the big aha moment we had based on the open power relationships. This get on with it from a Linux client, so the market's chosen around an open stack of software for new application development. You've mentioned to Buntu a couple of times now, are they sort of the preferred Linux platform for power? No, I would say they're the most recent to join. We continue to have great relationship to Red Hat and Suse. And Buntu happens to be the new one we're announcing today, is sort of the third member of the distro. There also happened to be the first bringing little Indian Linux to the platform. The others have plans in the future, but that's partly why I mentioned them. They're new, and then the first with little Indian Linux. You were, you headed up IBM's mainframe division before you moved to this role. That's right. You were trying to tell a Linux story to those, to that group of course, ZOS focused very much a different kind of migration play there. What did that experience teach you as you're trying to promote Linux as really the platform going forward for the future? Yeah, I mean Linux on the mainframe has been incredibly successful. If you look at the capacity now we ship each quarter on the mainframe, the majority of that capacity is now for Linux applications on the mainframe versus sort of that core infrastructure that's existed for years. So it has been a great enabler for new application capture on the mainframe. What did I learn? Get on with it. Back to the point of the application developers have decided OpenStack of Software is the right answer. So me trying to convince them to do new application development on my core infrastructure of AX and IBMI, it's a much harder play. That remains super important though. And this is the question just like the mainframe I get all the time. AX, IBMI, super important to my strategy going forward. It's where my tens of thousands of clients are today. I need to augment that though with a new application capture engine and that's Linux. That's what I learned. Talk about the investment you're making in the power. Obviously IBM is all in on Linux. You see OpenStack big part of the blue mix, all the cloud stuff, blue acceleration, a lot of stuff in the portfolio. So talk about how this fits in with the existing portfolio within IBM. And also how does this tie into some of the key applications around analytics and then the cloud. What is it? I mean, are you guys not giving up the hardware per se? You're opening it up. You're seeing that being more of an enabling strategy with openness. So how does that all reconcile? I mean, you got big dollars going into the market with Linux. Take us through that next generation roadmap. Where does it lead to? Yeah, great question. So as we've sort of been sharing here on announce day, I mean, we didn't just start this move to power aid here just in the last nine months. In fact, the power aid processor, really what we see as the first processor design for big data, that started over three years ago. And we saw the problem coming in the market of a true need for a system and processor that was going to help clients with the big data problem. We've spent, I think, our best estimate is over two and a half billion dollars in those three years developing power aid processors and the technology around it. So pretty significant investment. Recently, as you mentioned, John, we announced our second billion dollar commitment to Linux. This time, Linux on power. And it's really about the ecosystem. It's about the porting centers. It's about the IS fees. It's about the go-to-market and obviously the products as well too. So- I mean, a billion dollars is not that much money these days. How about 10 billion? I mean, you know, I mean, you know, I don't know. 10 billion is the new one billion. Ah, come on. I think we'd all agree. I'd take a billion anyway, right? A billion is still an important number, right? For a company the size of IBM as well. So sizable investments we've made. Now to your point earlier in terms of, so where is this play out? You know, we're not getting out of the power of hardware business. This is incredibly important. Our chairman, Ginny Romendi, has been very clear on the importance of, you know, what we call high-end, high-value systems and the role power plays in that. I actually think with the divestiture that's still underway, this helps sort of get clarity of the portfolio where we're placing our bets from the server. We need to be more specific. And where we can bring value and innovation. Yeah, I mean, you guys have a lot of value. I mean, if you pick a customer base vertically, you guys are pretty much there, but I want to get more specific around HPC because HPC market's changing, right? It used to be throw a box at it and you're done. Right, right. Now you got cloud as a big part of the HPC equation. You're talking about compute, a lot of unlimited compute theoretically, but big data is kind of the things where we're going to talk about internet of things today. Right. There's all kinds of stuff going on in the cloud. So it's not so much the big data pipe, it's the pilot data. So if someone wants to do a computation, is this kind of where you see it going? Is that the focus? Is it the, is HPC the new, is HPC, what is the new HPC? Yeah, so I say the new HPC is much more analytics driven than it is pure compute, right? If you think about the scenes, the scene HPC was all just about compute part, right? To be able to run fast, that's still important, but what we see, and to no surprise again, is many of these, whether it be government oriented, utilities, whatever it happens to be, are very much focused on the same problem commercial clients are, data. And so HPC evolves to, sometimes we call it HPA, high performance analytics, taking advantage of this, lots of cores, lots of threads, large memory working space, and yet very big IO bandwidth to move the data around. So I got to ask Paul, because we're a throwback here. It's not throwback Thursday yet, but are we going back to MIS departments and data processing departments? I mean, you're basically talking about data processing. This is data scientists, right? I mean, we see kind of this role evolving around folks who sort of focus every day on the science of data, getting business value out of this gobs of data they have. You've got a lot of partners working with the PowerAid reference spec right now. Can you give us an idea preview when we might see chips coming out that would include, say, an IBM Power with NVIDIA or Melanox? You know, I can obviously, again, speak for them and announce their own plans. What would you like to see? You know, I think the faster we create innovation, the better for the marketplace. I know one thing that is being shown down on the floor and was publicly stated by NVIDIA last week, so I'll repeat it, is this sort of marriage between NVIDIA GPUs and Power on a single system. I think you can look for that later this year. We got some commentary from the crowd again. Tim Crawford, shout out to Tim. My point of view, IBM would do well to play down their hardware position, not build on it. IBM's hardware play is better suited for cloud providers than enterprises moving forward. What's your comment on that, on his take on that? Yeah, I think part of it is, so we're not going to play it down, obviously we're loud and proud, right? So I'm putting that part aside for a second. I do agree with them on the role the cloud plays. You know, I think what we see going forward is this incredibly important role around sort of what we call sort of the hyperscale clients. Those are building out the Internet data center that we talk about at sort of the dinner table and at home every day. But then there's the managed service providers, those who are sort of much more mid-sized and have drawn a circle on a map and said, you know what, I'm going to be the service provider for this city, this state, this country, and we're bringing power technology to them, sort of married with PureFlex, by the way, for the mid-range service provider. And that's got to come with a different economic model by the way in which they acquire that infrastructure. So I completely agree with the second part, power for the cloud is a very important play for us. So in the last minute I want to just get you to take on what you're expecting over the next year, obviously the customer activity here, give us some taste of the meat on the bone with customers and some deployments. Where is the action happening and what do you see projecting out of the next 12 months? Just kind of a high-level execution. Yeah, certainly we are looking to power each systems to be a catalyst for improving our business performance around the power platform. Our CFO, Martin Schroder, has been very clear on as I have been that powering and all we've done to reposition the platform, new applications with Linux, the target for the cloud and the role power plays for big data are where the market's going or gone and is where power's going. So I think from a business perspective we are now very much aligned with the market and based on that we will look forward to strong success and adoption. So next year or next event at Pulse are here. What's the big success, top three things you want to check off to have a good year for power? What's the big objectives? Yeah, I think with anything where you're sort of shifting to a new space for ourselves, although a space that's large, I'm absolutely looking forward to sharing sort of proof points and references around the role power is playing for big data solutions to clients, right? Second part would be the success of open power, continuing to grow, but also a bit more than just growth of members, continuing to bring innovation to the market and continuing to demonstrate as we are today on the show for true innovation that benefits clients in the end. As people start to instrument the business, this is my last question and we'll break and then I'll let you get the last word in, but as customers start instrumenting their business, big data, obviously a big part of it. Yes. Now you're starting to see the convergence of physical spaces, the work environment. It could be in the hospital, machines, to other physical attributes. How do you see that vision of the internet of things playing out with regard to power? Yeah, I think it comes back to, clients want to consume IT in the way in which they want to consume it. It could be infrastructure as a service, it could be platform as a service, it could be software as a service. My goal is to bring power to however they want to consume it, right? It could be on-premise, it could be on-premise in the cloud, it could be as a way in which we deliver phenomenal benefits under a blue mix composable set of applications. So however they want to consume IT, power will be there. Does that include embeddable devices? You know, less focused on embeddable. I mean, we've had power in the past focused on embeddable. I'm really focused with power and open power on the server market. Sometimes that server is more of an appliance based. That's fine, but try putting in the cars and so forth, that's not our focus. Doug, I'll let you get the final word in. Tell the folks out there why this show is so important. What is the main message to your customers and IBM across the portfolio? Yeah, so I think the reason we're at this show announcing the new power systems is back to the point of the changes we see in the marketplace. This, every mobile transaction, right? Every reach to the cloud, every social post, it drives data. And data ultimately does depend on the infrastructure. And the infrastructure we believe certainly matters. That's why we're investing in new power systems focused on, back to what I said before, data, cloud and being the most open server platform in the marketplace. Doug, the general manager of the power systems group at IBM. I mean, I'm the main creator of 50th anniversary. Like I was saying earlier, data processing is back in vogue the whole term. But that's what's happening. It's data science, data processing. This is theCUBE. We're processing the data and sharing it with you. We'll be right back with our next guest after this short break.