New GM Wants To Push IBM Power With Hybrid Cloud And AI
November 13, 2023 Timothy Prickett Morgan
Three months ago, the Power division within IBM’s Infrastructure group got a new general manager. This changing of the guard happens every couple of years as Big Blue moves executives around so they can get managerial experience across different lines of business, and this time around, it is Tom McPherson who has been tapped for the top job in the division that is the home of the IBM i platform.
McPherson got his bachelor’s degree in electrical and electronics engineering from Rutgers University in 1990 and got his master’s degree in computer engineering from Syracuse University in 1994. We don’t know what McPherson did immediately after graduation, but he worked on design for many generations of IBM Z processors, and also the Power6 processor, and eventually became director of IBM Z processor development in June 2008 and then was vice president of IBM Z development until late 2017, when he then became vice president of platform services for IBM Cloud infrastructure. In January 2021, McPherson went back to the IBM Z division to be its vice president of the IBM Z software stack, and in April 2022, McPherson did a lateral move to become vice president of IBM Power products under general manager Ken King, who took over the position that McPherson now holds in June 2021 after running the OpenPower organization for more than seven years.
It is tradition for us to give the new general manager of the IBM midrange and then Power business a chance to settle in before hopping on the phone and having a chat, and last week was a sort of coming out party for McPherson as he did a blog post on the IBM site talking about the renewed focus the IBM Power division would have in the coming year – something we were unaware of until after we did an interview with him and special guest Steve Sibley, vice president and global offering manager for IBM Power.
McPherson is taking over the running of the Power division as it is on the rebound, with SAP HANA in memory databases, OpenShift, application and database modernization, and Power Virtual Server all combining to drive growth as the AIX and IBM i businesses are stable or growing, depending on the quarter and depending on how far we are into the Power9 and Power10 upgrade cycles and looking at how the Power11 upgrade cycle that starts in 2025 might go.
Timothy Prickett Morgan: You have a long history that weaves together Power and Z, which seems like an asset to me given that these machines are the closest thing to a traditional integrated stack that is left in the datacenter.
Tom McPherson: When I joined IBM Power 18 months ago, I was leading product management. And that was when Steve and I and others, including Ken King, general manager of IBM Power, and Ric Lewis, senior vice president of Infrastructure, were working on the strategy pivots and driving the business that led to me becoming general manager of IBM Power back in August.
Before coming to Power, I was running IBM Z software for a year and a half – so I had responsibility for CICS, IMS, TPF, and compilers. Before that I was in IBM Cloud for three years leading the development of the virtual private cloud infrastructure, which is the foundation of IBM Cloud for Financial Services, or FS Cloud. And then before that, I spent almost twenty years in IBM Z development roles and was responsible for the z12, z13, and z14 generations of IBM Z from a system development point of view. I did another five generations of IBM Z processor development for IBM Z before that, where I was actually a designer within the processor team and then became director of the team. There was a period of time where I had responsibility for Power microprocessors in the Power6 timeframe together with the IBM Z processors.
TPM: Oh, that was back in the ECLipz generation, where people got the wrong impression that IBM was going to converge to a single processor and didn’t. . . . I thought that was an interesting idea, but someone would have to be running in emulation mode so I didn’t see it really happening.
Tom McPherson: We put a lot of the building blocks together that are shared across IBM Power and IBM Z back then for development efficiency. Because of all of this, I have a lot of background in development and product management, cloud, software, and this has all lined up well for the responsibilities I had in Power and lead to the role I’m in now.
TPM: You have been in both of these systems divisions throughout your career at IBM. What makes them different?
Tom McPherson: A few things stuck out to me when I came back into Power. I always knew the team talent level is extremely high, just like it is with IBM Z. But the thing that’s really different between Power and Z is the number of clients and the amount of engagement with partners – and how much business goes through partners – with IBM Power.
And knowing this, we architected some strategy changes. We took a position where we’re not just going to preside over the trend in IBM Power over the past few years, but rather actively drive a new focus with our strategy. You can think of it as: Preserve the core, stimulate progress. We’re focusing on growth drivers while continuing to preserve the core through continuous innovation in AIX and IBM i – driving double digit growth in SAP, driving double digit growth in modernization and OpenShift and AI, and driving triple digit growth in as a service with Power Virtual Server.
The organization structure that Steve, Ken, and I designed was to align these growth drivers, from transistors on up through the ecosystem, in terms of a product management focus. They’re almost pillars inside Power. We’re doing more frequent releases on our roadmap on a quarterly basis, which is more aligned to our strategy segments. And we have more engagement with clients so we can respond where we have to pivot. We have probably done three or four pivots in SAP already, for instance, just based on the dynamics in the market. In modernization, we’re driving more pivots to AI and watsonx so we’re really in a sense and respond mode.
TPM: Talk about growth in IBM Power because I have to make my own estimates, and I have a sneaking suspicion that the IBM Power business is growing faster than I think it is, but I can’t prove it. From what I can see, it looks like it is going to be three years of growth. It’s not a huge growth, mind you, but we have had fifteen years of decline before that. So this feels pretty good compared to that. So my question is this: Is this modest growth sustainable or can you accelerate it even more?
Tom McPherson: I read most of what you put out, and I definitely read each quarterly earnings story you have done – I don’t know for how many years you’ve done it . . . .
TPM: It is now year 35, if you can believe it. The white patch in the center of my beard, that comes from IBM. And my kids are the cause of the white in my sideburns. [Laughter]
Tom McPherson: Fair enough. [Laughter] You’re in the right zip code with the growth you are showing in your model for IBM Power.
TPM: My gut has been telling me it is a little low. But I don’t have any proof. All I know is Steve is smiling more than he usually does – and mind you, he smiles a lot as it is.
Tom McPherson: Our strategy is designed to get the business in a mode where we are mitigating the cycle dynamics. So when you put the fuel in from our growth drivers, such as Power VS, which is growing at a high double digit rate this year and hopefully a triple digit rate next year, even though it is a small part of the business now, as it grows into the tens of millions of dollars, it starts to fill in part of the valley in the system cycle.
TPM: Ya know, I had not realized I should allocate Power VS revenue to IBM Power, so there you go.
Tom McPherson: But we also have SAP HANA driving double digit growth. We’re going to continue to drive the modernization part.
We are focusing a lot in the banking area going forward. There are thousands of clients on Power in banking, and they are the tier two banks that do not overlap with the IBM Z customers. And there is a natural effect of growing as the banking and financial services industries are growing – kind of in the high single digit growth range. But in banking, as you know, if you’re not modernizing and moving into a hybrid cloud architecture, increasing your productivity, driving automation, you’re just going to lose out to the competition.
So, there’s a lot of modernization happening in the banking industry. And we’re benefiting from that because our modernization play of surrounding the core AIX workloads with OpenShift, leveraging Power10 for AI inferencing for fraud detection, putting more AIops automation around the banking estate – that’s driving double digit growth for us. But we also have this trend of clients who are adopting OpenShift on Power and are growing the AIX core at a much higher rate than clients that are not, so this is another cycle mitigator where we’re driving growth on OpenShift and hybrid cloud architecture and it helps lift the core business.
TPM: That certainly started working for the IBM Z business with low-cost Linux engines and Linux applications back in the early 2000s right up through today. That mainframe Linux drove direct MIPS consumption on mainframes and then drove secondary consumption as Linux applications accessed z/OS applications and databases. We have not seen that same kind of Linux uplift with IBM Power, but I have been writing about how OpenShift can be the new PASE for IBM i. If you can make OpenShift invisible, so integrated into IBM i that customers don’t know and don’t care that they are using it, then IBM i customers can do all these modern things without having to go through all the grief of setting up a Linux cluster on some X86 servers over there and networking them to their production IBM i systems. So I think there’s great potential there.
Steve Sibley: Merlin is clearly an example of how you start to move in that direction. You and I talked a little bit about this — that making OpenShift invisible is a challenge at the moment, but we’re working in that direction to simplify and to integrate OpenShift environments.
TPM: OK, new topic. In the past couple of weeks, you have seen me getting on a high horse – or maybe it’s a reasonably large pony – about having to do a Power10+ upgrade in 2024 because it is a long time to wait for Power11 in 2025 or so. Is this stupid?
Tom McPherson: Now, Tim, you know that we don’t talk about unannounced products. [Laughter] All I will say is that we are working on stuff all the time.
Steve Sibley: We are working with Samsung to drive some improved performance of transistors and doing a number of things in the software stack to improve performance. But one of the other things we’re really focused on, as we get feedback from clients, is how do we continue to drive a more continuously available environment and more automation skills. These are the key things that we continue to hear from customers. And it’s not just skills you need for IBM Power. Companies just can’t invest as much to maintain their infrastructure, which is part of why people are moving to cloud. So autonomic computing is one of our key focus areas, as is delivering more capabilities as a service. Both take burdens off of customers.
Tom McPherson: We have more of a stack level view than we have had in the past. In our client councils, we’re doing design engagements with the clients, trying to align more around making sure we’re developing in the areas that resonate the most and drive the most value with them. You can say, well, you’ve always been doing that. But what I’m calling out here is that we’re doing next level, full stack engagement with the clients around our next generation system. So for example, we are looking at frictionless hybrid cloud, where if a client buys a next generation system, they will have hybrid cloud capacity credits that will enable them in a frictionless way to have the same experience in IBM Cloud as their on prem datacenter. And we are looking at delivering on premises Power as a service, too.
The other advantage we’re going to have is that clients can move from their on prem environment into Power VS without major IT refactoring. The bridge to get to Power VS is a low friction bridge, so you can come into PowerVS without having to do a massive IT transformation. Then, you could go at your pace for transforming for cloud native or expanding cloud services after you’re on Power VS. And I think that’s another key differentiator because if clients are looking at a full cloud native transformation, and replatforming, Power VS will win every time from a cost-velocity-business outcome perspective. So, I think that’s a good position that we are fortifying around.
TPM: At the turn of the last century, Linux was the hot thing. Now, the cloud is the hot thing. And I think it’s far more important for us to think about how much of the capacity of across IBM i and AIX customers is going to be on a cloud versus on premise and see how those numbers change than it is for us to worry about Linux on Power.
Where do you think it will balance out? I’ve joked – at least in the broader IT market – that it’ll be one third on premise, one third in the cloud, and one third in co-location facilities where people halfback. You know, they go to the cloud and see it is too expensive and then they rent iron under a cloud pricing model and put it into a co-lo to get something that is priced like cloud for less money but gives them control like on premises.
How do you think this will all shake out in the long term?
Tom McPherson: I follow l the trends from the major consultants where they say that by the end of next year, we could be at a point where 50 percent of the purchases are done “as a service” under flexible consumption models. What we are trying to position for is that we want both deployments. So we’re going to have Power VS in the IBM Cloud and then a local deployment of Power VS, too. That local deployment could be in a client’s datacenter in an Equinix or other co-lo datacenter. And I think the ”as a service” portion could be two thirds of the business in the range of five years or so. We’re moving more to flexible consumption, but we’re not going to eliminate the transactional stuff we have now. So I think we’ll just set the dial and work it as we go.