What are IBM's Hybrid Cloud and AI Capabilities?

Machine-generated from the episode audio and not hand-corrected, so names and technical terms may be imperfect. The audio is authoritative.

All right, as you can tell by the B-roll, we're in snowy Boston, just perfect timing. We pulled the Airstream out of storage and drove it halfway across the country just to do a video interview remotely with Roger Primo, General Manager of Strategy and Ventures at IBM. Roger, welcome to the program. Good to be here. Thanks, Keith. Sorry about the weather. Can't control it. You know what? I'm from Chicago, so I can relate. I apologize for our weather half the year, so there you go.

Roger, I think we're going to start out with the hybrid cloud, a conversation that I think both me and you could have predicted two or three years ago. It's all the rage now, but hybrid cloud is the present and the future of enterprise IT. I want you to take a victory lap. Talk about why hybrid is the obvious future of enterprise IT. Well, I think you just start with, yeah, obviously, it's thinking that's part of the Red Hat acquisition has been thinking that's been brewing inside of IBM for five, six, seven years, but it was counterintuitive at the time, which is just this notion that technology is heterogeneous.

I think a lot of folks assume everything would land on a singular public cloud, and it hasn't. There's a lot of cost reasons. There's a lot of risk diversification. There's a lot of regulations. There's a lot of reasons that you need data local to where it's used. a. hybrid and multi-cloud. And I think that's the enduring state that practically every large company in the world is going to be in. And increasingly, we're seeing it with software innovators, too, because their client base is saying, I need my workload to be in this cloud or that cloud or behind my firewall.

So even with the SaaS companies of the world, we're increasingly seeing that trend as well. So we'll talk about how IBM Cloud enables that heterogeneity a little bit later, but I want to dig into kind of the drivers behind hybrid cloud. We talked with David Lithicum, who's the chief cloud strategy officer over at Deloitte last week during our virtual event, and he talked about the value that enterprise IT brings back to the business, which is the return value, the number one objective. That value can be an agility or that value can be in cost savings.

And a lot of the industry is focused a lot on cost savings. I want you to talk about both. What is the agility that hybrid cloud can bring? And then talk to me about the potential of cost savings. So I'd say, first off, so we talked about hybrid as that heterogeneity statement. IBM talks about its hybrid cloud platform with Red Hat. That is the solution to the heterogeneity statement. So, you know, Google gave us this gift in 2014 when they open sourced Kubernetes.

So it gave us an open source project that brought cloud capabilities as software. And obviously, over the past nearly a decade, right, all overnight successes really take a decade. The open source communities rallied around it. So it can really do this ambitious thing, which is bring cloud native capabilities, the same or similar cloud native capabilities to each and every environment in which you run technology. And what that does is it means that and David's really articulate on this. Right. The cost of the cost that really penalized business was that complexity.

The thing that slowed down innovation was the complexity. Because if you do everything in a bespoke way in each different environment you serve, you say, how do I take innovation on one side of my business and bring it to the other? How do I think about skill populations of developers, data scientists, et cetera, make sure their skills can bring can work in whatever part of the business I need innovation. That complexity was slowing down the enterprise and creating friction costs, security costs, operational costs, fragmenting developer skill pools.

There's a number of reasons that that really slowed down the enterprise. If you go to that level of standardization to get to a singular platform, it drives massive productivity for innovation, for your developers, because you can have a common way to do DevOps across your environments. It simplifies a lot of the operating and security complexity. It lowers the number of security gaps in your architecture. There's all sorts of both innovation acceleration and cost reduction opportunities that come out of that. Now, David's also articulate on a couple of things also, which is, one, there's no such thing, you don't want to go to architectural purity.

It doesn't mean one size fits all. And there's also trade-offs in that. When you go to a standard, there are some ways in which each individual cloud or particular technology stack has differentiation. Every enterprise has to think about the balance of that overall value prop I just talked to versus specific technical value props they may get from a particular service. A smart architecture isn't 100% in one way or the other, but we think this move towards standardization, leveraging what the open source community has really brought to life with containers and Kubernetes, is the direction of travel that makes that value statement come to life.

So let's talk about some of this nuance. We spent 80% of our IT budget keeping the lights on, very consistent. I have this saying that nothing ever dies in enterprise IT. I'm going to have everything from legacy Unix on my floor, mainframes, Linux, Windows, et cetera. And when I look at a company like IBM, I'm missing the fabulous Z because of the weather. I'm missing the fabulous conference on the Z systems. I see a company capable of the hybrid spreading across the entire control plane, if you will, the ability to manage these systems.

When you're talking to customers, how are they approaching you about, you know, applying DevOps strategies or operating models to something like a Z system, enabling a Z system for the hybrid cloud? Yeah, Keith, it's a good question. And I think we have to it's a conversation where we look at that traditional application estate, maybe on Z on other platforms. And we have a conversation about modernization. First of all, we talked about Red Hat and, you know, that the Red Hat distribution of containers, Kubernetes, and the surrounding capabilities is OpenShift.

And so we run OpenShift on Z. So you can take the performance, security, stability characteristics of a Z and still do cloud native development on top of it. Right. And so when we look at traditional applications, we take a modernization approach, which is taking these applications often been built, you know, in a monolithic way years ago and really decomposing them into microservices. Some of those may land on a public cloud. None of them may be modernized and sit on top of the Z.

But it is the way that we look at, you know, first getting the most out of that Z footprint, because there are real, you know, if it's averaging only three seconds a year of downtime, there's some real characteristics of the Z that make it really differentiated for the core capabilities of the business. But it's also being realistic that there's a lot of applications that were built around that Z that do get modernized, put into containers, run as microservices and land on top of a public cloud or land on top of a different hardware platform.

So we try and go with a modernization with ultimately the client value around that application being the main thing we're optimizing to. And that helps us figure out which workloads to place in what location through that modernization journey. So let's talk about the other extreme. We talked about Z systems and kind of modernizing in place and using Red Hat to do that, bringing DevOps type of practices to mainframe workloads that run on a mainframe like system. But what about AWS? When I think of IBM hybrid cloud, I don't necessarily always think about the other cloud providers.

I think about IBM's cloud. What if I'm developing on AWS? Where's the value prop when it comes to IBM hybrid cloud? So first off, so I'm the executive that runs leads our relationship with AWS. I see this in first full color kind of on a daily basis. So, first of all, I talked about Red Hat and OpenShift. So on AWS, we call it Rosa. It's Red Hat OpenShift on AWS is the offering, but it is AWS managed OpenShift for customers.

So you can build the right and it's really run by by AWS as if it was one of their own services. So you can get that capability delivered as a service inside of AWS, similar offerings available from Azure and others. So you can if you're a cloud developer, you can get that value prop. Let's say first and foremost, I mean, OpenShift is a very is a best practice way to do DevOps on top of containers and Kubernetes. And so the way they've built the surrounding open source projects to deliver that kind of DevSecOps value proposition and platform.

First of all, as a standalone, we think it's it's best in class. But obviously there are equivalent or similar offerings from those clouds. But especially if you find out that that innovation that starts on the cloud may land on other clouds or may land on prem. We do think starting on that Rosa starting point, even for cloud development, is a smart decision for the enterprise because it gives them that flexibility and that set of business value propositions. We talked about before of having a common platform layer.

So, again, we're going to talk to Red Hat Stu Miniman specifically about Rosa tomorrow. But I want to extract out to this larger conversation of. Hector, I hate this word heterogeneity. I bet I've. Pause. Kid, I record knock out that and replace with this question. So we're going to talk to Stu Miniman tomorrow about Rosa specifically and some of the technical details. But I want to zoom out because I want the same developer experience on these systems. I want the same developer experience in the public clouds.

I want the same developer experience on my on prem hardware. And what sounds like you're saying is that OpenShift, because of open source, has powered my ability as a CTO to give my developers what they want, which is basically an API to the hybrid cloud. That's right. That's right. I mean, I'd say first and foremost, I separate it from IBM for a second. I think just thinking about the ways to create standards and how you develop and operate your application environment is good advice.

We also think if you look at interrogate that in open source, you'll see containers and Kubernetes and the surrounding projects as kind of the standard for that. And then we think then get to that technical architecture and then we can talk about vendors. And we do think Red Hat has the leading distribution of containers, Kubernetes for the hybrid for the hybrid cloud kind of being able to run across those environments. But absolutely, we think it really does deliver that value proposition of of, you know, that business value abstracted across those environments.

And so for us, that's, you know, IBM believed it enough that we bought Red Hat to kind of gain that capability because we thought it was so critical to the future of enterprise IT. So you've been very gracious in talking about open source and the power of open source, what Red Hat and OpenShift brings to it and even ROSA. I have to ask you the IBM lens. What is the thing that IBM, the IP that IBM brings to not just open source and Kubernetes, but this whole hybrid cloud conversation?

I could get Kubernetes elsewhere. I can even go and buy OpenShift, you know, just via the open market. When I'm thinking about IBM cloud, IBM big, where is the value prop for IBM? Yeah, I'd say a few things. I think, first of all, moving our business to where it creates value for clients has to be, you know, how you evolve a strategy for a company, right? So if you say we create value, we talked about that before, and then Red Hat is kind of the core of us capturing value.

But then we've woven that hybrid, that hybrid strategy throughout all of our businesses. So if you think about our consulting business, we have the biggest Red Hat practice in the world. We invested to build that out. We have the most skills on hybrid cloud inside IBM consulting, more so than a lot of our other systems integration partners. We've also invested a lot in building out capabilities around AWS, Azure, Salesforce, etc. All the ecosystem partners that a client would use to pull together hybrid applications and modernize their IT in a hybrid world.

So consulting, we've made that a differentiator for that business and they help our clients realize the value of that hybrid architecture inside their business. Inside our software portfolio, we do things like data and AI, security, sustainability, IT process automation, business process automation. Actually, before we bought Red Hat, we'd already made the technical calling on containers and Kubernetes. So we made some technical bets at the time, which have really borne out. And so we re-platformed all of our software onto OpenShift to get that same value prop.

And that's how we're doing multi-cloud SaaS. It's how we're able to innovate for our software portfolio. So all of our software both benefits from that value proposition and extends that value proposition in the domains that it serves. And then we talked about it before, but in our infrastructure portfolio, we also run OpenShift on top of that portfolio. And it does things like saying, yes, you can take a cloud native developer and have them develop on top of a Z mainframe. That's kind of a counterintuitive thought to a lot of customers.

But it really modernizes a lot of the value proposition of our infrastructure portfolio as well. So, you know, you see it kind of delivering value across all of IBM. And then we did, you know, over my tenure, we've done some exits and spins and things like that. There are parts of IBM that didn't fit that value proposition. And we've really, you know, exited those parts of the business to really focus on the place where that value proposition kind of really comes together.

So, Roger, I know we've talked mainly about hybrid cloud, but I don't think we can talk about hybrid cloud without talking about large language, large learning module link. I don't think we can talk about hybrid cloud without talking about AI in the lift it takes for these large language models, LLM, and how IBM is enabling this. I've talked about use cases with IBM AI in the past, but how's IBM engaging customers around AI? So I'd say a few things. I mean, I think we're pretty proud that, you know, IBM helped introduce AI to the world a decade ago with Watson.

I think obviously now it is really hitting the mainstream with what we've seen with these foundation models, long, large language models and chat GPT. But I say first, before we get to large language models, you know, obviously we've been delivering the Watson portfolio, Watson data, you know, the Watson data studio capabilities to help the enterprise use more traditional types of AI for a long time. We also use AI to run, you know, it is a hybrid cloud and AI strategy. And we also use AI to run that hybrid cloud.

So if you think of what we do in AI ops, that heterogeneous environment is getting so complex, like a single human being can't manage that. You can't do a hundred thousand storage optimizations on the cloud one at a time to save a half a million bucks. It's just not economic to do with people. You need AI. So that's a big piece of our strategy. But then specifically on the foundation model topics and large language models are, you know, obviously the maybe the hottest subset of foundation models.

There's a lot that core technical stack is also with it, you know, with OpenShift. That is where these large language models are both trained and executed on containers and Kubernetes. And so we deliver that in our portfolio as well. The ability to have a scalable infrastructure so you can do the training of these models. We've extended a lot of core Kubernetes and it's going back into open source to put in the scale and duration of some of these jobs, because often the training of these models takes weeks or months.

And then if you look at the inference of AI, that is also done inside of containers and Kubernetes, because you want to wrap an AI algorithm with the environment inside a container. So you get repeatability because if you just deploy kind of a naked algorithm into a different environment, you can get same data in different inside out if you don't control that. So it very much plays to, you know, the technical standard that we're building to. And that's where that is the technical standard that, you know, it might not be the Red Hat variant of it, but you see so many of these large companies, you know, building on that same technical stack.

We're doing we've been working on large language models inside of IBM for several years now. So we're bringing that to the enterprise on top of that same Red Hat stack as well. Yeah, these two questions are both topics I addressed in my light board video talking about IBM's hybrid cloud and capability from the software packs in AWS to orchestration to AI, AI ops. IBM has led the way, honestly, in these areas for a very long time. So if you want to see more content, CTO advisor content on this topic, visit the Web site, the CTO advisor dot com.

And you'll find both the video that I talked about, IBM cloud and the content that we created at our virtual event. Did I miss something? Should I have asked Roger a topic that we didn't cover? We covered AI, hybrid roles, AWS. You know, we we've we've hit the things. But if I didn't in my traditional way, ask a tough enough question. You can DM me at CTO advisor on Twitter and I'll email Roger. We'll try and get that answer to you until then.

Catch you on the next CTO advisor CTO dose.