Gene Kim

In previous years, Rosalind Radcliffe shared the career journey that led her to becoming a technical fellow at IBM, the highest rank of individual contributors at IBM, of which there are only 84 active right now.

What is astonishing particularly about her elevation is a business problem that was big enough to warrant Rosalind becoming a technical fellow. And she's the first to be made in the finance and operations group instead of the R&D group, which basically is every other technical fellow. And this is because there was an urgent business need that arrived because IBM was spinning out Kyndryl.

So let me just present the potential business context that has been written up in the press about why IBM did this.

Kyndryl was formerly known as IBM Global Technology Services. Many of you may know these people, right? This is the long-term business outsourcing division. It's the largest outsourcer in the industry by far, with over 80,000 employees in a portfolio of over 4,000 customers, 75% of the Fortune 100. So they manage infrastructure services, including cloud computing, data center management, security, and so forth.

Forrester wrote that this managed services market was getting more and more competitive. So IBM CEO Arvind Krishna described how the spinout would allow IBM to focus on its more profitable software and consulting businesses.

So here's the problem: IBM had all of its critical infrastructure being managed by Kyndryl, and for a variety of reasons, IBM wanted to bring a thousand systems back in-house, and they had to do it in two years.

IBM had systems as part of IBM Cloud, but as you will see in the story, a large part of the critical systems were running on Z, and more specifically z/OS mainframes, which would not work in IBM Cloud.

So this is something they had to get right the first time. If they didn't successfully transition these systems and business processes, one can imagine that Kyndryl would have IBM right where they wanted them and could insist on another long-term outsourcing contract. Does that resonate with anyone here at all?

How many people here have had that experience with long-term outsourcing contracts? By round of applause, right? Okay.

So Rosalind was asked to oversee this astounding project of staffing up an internal capability that would take over these applications and, in some cases, build out new data centers where they would run. This is an incredible mission that mattered, where a deadline could not be moved and would have astounding downside consequences if they failed. And she will describe some of the incredible benefits that will result from this as well.

So here to tell that story is Rosalind.

Rosalind Radcliffe

Thank you, Gene, for that wonderful introduction. I'm happy to be here again to be sharing my story.

And my story is of an individual contributor. I've been in IBM only 36 years, and I say "only" for a reason. There are plenty of people who've been there longer than I have. And the reason I've been there is because I've done so many different things. I haven't had one job for 36 years. I've had, well, this chart shows you some of the various different opportunities that I've had.

I started early in my career on mainframes, but I had an opportunity to do industry standards work. And Control-C, Control-V, Control-X, that's there because of the standard. I worked on little things throughout my career.

So my opportunity to work with clients, my opportunity to work in services, all those things gave me a challenge, gave me an opportunity to do different work.

In my last role before this one, I was working in Z systems, responsible for modernizing the environment, truly making Z DevOps-able, as I like to say. And I was speaking here about that.

But then, as was mentioned, in order to get to IBM Fellow, you not only have to have accomplished great things, you have to have something new to accomplish that is great and is a miracle.

And so I came into the CIO organization. And just to give you a perspective, this is what the CIO runs. We provide all the machines, all the internal IT infrastructure. We support the hybrid cloud. We don't run IBM Cloud, but we run all the other internal IT systems: payroll, all the things that you can imagine. And it's a large challenge.

And 48% of our financially significant applications actually run on Z.

So I come into this organization, and here we are. We have 11 data centers, which we no longer own. We have a thousand-plus systems managed as pets. Oh, isn't this fun?

We have separated this business, and now we have to figure out what we're going to do. We're going to end up being in four data centers, except we only have two.

And I discovered in January of last year, or no, December, we only have two planned data centers. I need four because I need two in Europe, right? Logical. This makes sense.

So in January, we start the process to build out data centers in Europe. Now, I don't build a building. I find cage space. But we have to get cage space, we have to order all the hardware. And you do remember COVID kind of screwed up the supply chain, right?

So imagine having to build out two data centers with Z, Power, and Intel hardware, and all the networking equipment and everything else. And oh, by the way, be live now.

So we've done it. I have all four data centers up and running. We have begun to move workload there. And if this were one month from now, I'd be standing up here and saying we're done. But it's not, so we're still working.

Even this morning, I had that wonderful message: "Rosalind, we need your help."

And this is where, when you think about it with my history, I was an ideal candidate to help this transition. I could easily identify what the challenges were and help explain what the risk to doing something would be, or the risk to not doing something.

So today's question is all about risk. It was all about, should we allow this or not? Well, I have the understanding and the knowledge, so I could do that.

The other thing that was a challenge in this is we had to build up an entirely new team. So yes, between then and now, we've staffed an entirely new team, including z/OS system programmers, Z system people.

Everybody tells me there's a skills shortage. What skills shortage? I found them.

Now, okay, I admit, I got a lot of friends in the industry. So I found some friends who are willing to come help. And I hired a lot of new people, because if we treat infrastructure as code, if we use modern SRE practices, then it doesn't matter whether or not it's Intel, Power, or Z. I can do things in a consistent way so I can train up new skills without a problem.

I have new college grads who are doing system programming. In effect, they're building automation for my systems.

We will have completed the top one: the four data centers moved all of our workload and built an entire new team by fourth quarter 2023. So next month.

The standardized systems, fully automated, we aren't quite there yet. I had to make a decision: what was I going to do? Exit on time or be perfect? We are exiting on time. We're not missing that deadline.

But we still created new world. We created the new environment and began to move workload to it. Already, our new workload is running in the hybrid cloud environment. That is truly an integrated environment.

Whether or not you are running containers in our OpenShift environment, or you're running in a z/OS application, you have the same developer experience, the same pipeline.

We've brought into our new world pervasive encryption. We're working to consolidate our workloads to simplify that set of pets down to truly standardized, automated infrastructure.

And one of the key reasons we're doing this work is because we want to be a showcase. We want to show what is possible when you truly use the systems together appropriately.

We intend to get the new machines. I have my new data centers. I got two new z16s. I was really happy. Latest Telum processor. When we have the next one, whatever Z Next is, I'll bring that in right away so that I can show and demonstrate before we GA and showcase what's capable.

The other advantage: if I'm client zero, I get to beat it up first before we ever make it available.

So we ran a watsonx challenge actually inside IBM, so that IBM and all 250,000-plus IBMers could beat up on technology before we make it available. So we find out how it scales before we ship it out.

You'll notice in this picture we've got IBM Cloud and we have our CIO private cloud. We have four data centers for our CIO private cloud because there are a set of workloads that really logically would run in your own data centers. And so we want to be representative of that.

We don't just run Z and Intel. We run Power, we run i, we run manufacturing. And so IBM i is part of this as well. So it really is a fully integrated environment to make it simple, make it easy for people to work with.

And as that individual contributor, I could help do this. I'm not the manager leading the team. I'm the person who can help answer those questions when there's an issue. How should we do this? Help the team design and build the new world infrastructure, help them understand how to do the work through this challenge.

One of the things that we run into is people not necessarily wanting to be helpful. Yeah, you imagine you have this two-year deadline, and sometimes other people have priorities, for some reason. So generally they refer to me as the hammer, sometimes.

My hammer is an interesting hammer, though. Yes, I have the weight of, "We have to get done," behind me, but I also have the knowledge of how things work.

So when we had a problem moving systems, I could have a conversation with the technical people about what we were trying to do, why we were doing it, how it needed to be done, and could work through the challenge.

And yes, it started with a hammer that said, "Oh, by the way, we have to be done." But I could work with them in a way that, if I had just been management saying, "Thou must," I wouldn't have gotten as friendly an answer. I wouldn't have gotten that collaboration that I could get by being a person they knew, they understood, and I had worked with them through the years.

And we have all sorts of challenges in this migration. We have to literally move systems that may have been running for the last, I don't know, 40, 50 years.

I have applications that were written for the moon launch. And I am not joking, for the moon launch. Yes, we're still running.

Now, really and truly, when I look at our applications, yes, we have some applications that were written for the moon launch and haven't had enough updates to them. But we also have plenty of applications that actually have modernized throughout the years. So yes, they started 40 years ago, or 50, or 60, or whatever it is. But they have, over the years, modernized their practices. They've modernized their applications and remained being services.

But when you modernize your application, but your infrastructure is being managed like the 1980s, literally, because it looks just like when I started, you don't get very far. It's a lot harder to do your work.

And so, not only in this process, we've worked on simplifying our architecture and adding the automated pipeline for zero-touch operations.

I loved Topo Pal when he talked about access to production and break glass. My intention is zero. No one touches production. Am I getting that way with that yet? No. Am I going to? Hopefully, at least not for write access.

I don't want anyone in operations touching production to make a change. Will I have to have a break-glass ID so that they can go see what's going on? Yes, probably. But not make changes.

The simplified architecture and the advanced self-healing. One of the things, I have a huge advantage: I have a whole lot of systems. They're running a whole lot of workload. And one of the things that's really nice about AI, if you can train it on a whole bunch of data, it gets a lot better.

I've got a lot of data to help train this AI, to help improve our self-healing and improve the way we can run our systems.

And by doing the work we've done, we're decoupling infrastructure entirely from the application such that I can roll my updates in the infrastructure without affecting the application uptime.

Our intention with this, in being client zero and being the showcase, we intend to demonstrate eight nines. And yes, I just said eight nines, and I mean eight nines.

It is important that we can showcase that systems really can run this way using modern SRE practices. We can have the concept of fully automated system and infrastructure, and treat z/OS and our Z systems just like the hybrid cloud. It's just another cloud system.

Developers don't have to worry about where their applications run. It's handled through the pipeline and through the capabilities.

Changing the way, getting this done, required a huge change in ways of working and ways of thinking. Our application teams, I'll have to admit, still have a long way to go. But our infrastructure team, since we started from scratch, we hired a new set of people and started with a dynamic operating model.

And so we move people around in a sense, so that we can ensure we get the functions that need to happen when they need to happen.

Right now, almost everybody is focused on this final stage of exit, building out our systems and moving the last pieces of workload. But when they're not doing that, then there's opportunity to work on new world z/OS. We can ensure we have the right resources focused on run versus focused on change.

And this transition allows a better way of working. It allows people to learn, transition, and grow their own careers, and grow even more individual contributors within our organization.

Now, hopefully I told you a little bit about our wonderful story. We will be done. We will be done one month from now. We will be standing up celebrating that we finished the exit. And by the end of next year, we will have finished the migration to new world z/OS and a fully intelligent workload placement system for our entire environment.

So we still have work to do. It's never over. This first step will be done and we'll keep moving.

But it's really important. And the help I need, and the help I want to ask for, is the recognition that individual contributors, though we're not managers, I have no authority, actually, sort of in some ways. But I have a lot of authority because I know I know what I'm doing.

I've built that credibility through the history of my working in the company. And so people will follow what I say because I really do actually understand the challenges that were going on.

Now, do I know everything? Absolutely not. I work with all of the other people in my organization to make sure we're doing the right thing at the right time with the least risk. We're doing a lot of risky things right now, but we're doing the least risky that we can to still make the date.

But it's important to recognize this. Recognize that new-collar talent matters. People don't have to have a college degree. I do, but that doesn't mean that's the reason I can do what I'm doing.

You don't have to have a college degree. IBM has had this idea of new-collar talent to bring in people from different backgrounds, different experiences, to help grow our industry and to help grow diversity and drive innovation.

We would not have been successful in what we have done if we had tried to insist on having everybody of a particular mold. We would not have succeeded without having people challenging the status quo. We would not have done the amazing things that we've done without recognizing everybody has a skill to contribute.

The other thing that's really important is to think about rotation programs. Rotation programs help drive skill. They help improve your internal people.

Now, we have a rotation program with the development organization. It's absolutely wonderful. People who build products don't usually run them, so it's really nice to bring them into an organization where they have to actually run it.

All of a sudden, things change. They understand logging, they understand a whole bunch of other challenges. So it really is important, this rotation program, to help drive innovation, to drive knowledge and continuous learning.

We have a huge challenge. We're getting through it because we recognize and value people. We value our individual contributors. We value ideas that people come up with. We appreciate what they can contribute to us.

This has been a fun year. It'll be even better in November when I can take a vacation. But it's great. It's been a great opportunity and a new challenge.

And my next challenge will be build new world z/OS and move all my workload into that. And then I'll go on to the next things.

So thank you. Happy to meet with you and talk to you afterwards. Have a great rest of the conference.