To help make sure its research investments have a major impact on its huge services business, IBM (IBM) has assigned some research executives to work within IBM Global Services. One is David McQueeney, vice-president for technology assets. He recently spoke with BusinessWeek Senior Writer Steve Hamm. Here are edited excerpts from their discussion:
Q: You have coined a term -- the science of services. What does that mean?
A: A new science is being invented, which is an outgrowth of computer science but also an outgrowth of research and project management and a bunch of other things. In a few years we'll look back and say this was the beginning of the scientific and technical base underneath services.
I talk to the university community, people like Dan Huttenlocher at Cornell. He's a professor of computer science and business. He's part of a new breed of academics that are starting to populate this space. He's interested in working with companies like ours -- asking, "What's the new curriculum I need to develop to train people in this new science, however it evolves?" We're trying to build partnerships with the university community to take seriously the question of how do we develop a new science.
Q: Where does this lead 10 years from now? What does this do for the world?
A: The great good of it is pretty simple. Clients always want the maximum value from their investment in information technology with the minimum need to be IT experts. It used to be they had to have huge inside IT staffs. They had to know every detail of all the plumbing. But over time as we as an industry get better and move up the stack of technology abstractions, we can give them the power to change their business and require them to be less and less the experts -- and focus their energy on the things that are higher in the food chain.
This is a big step in making all the things we offer to our customers more coherent, more clear in terms of its value, and its impact on the business. Our customers are telling us to solve more of the details, and to do it for them in a way that creates the end result without them having a lot of expertise.
Q: You talk about seeing a stack of services emerging that's analogous to the stack of software used by corporations. What do you mean by that?
A: I'm seeing patterns in the evolution of the service businesses. Services is the youngest of all the IT businesses, compared to the server business and the software products business. Looking at hardware and software and how they have matured, I see a lot of parallels to what's happening as the services business matures.
With this as a backdrop, look at IT services. In particular, look at the work we're doing in basic infrastructure operations -- computers, storage, and networking in a data center. These are the things that are part of a typical outsourcing contract. To make this efficient and scalable, you have to create a set of best practices.
You have a standard server configuration, so you can bring servers on line efficiently, and you can use fewer people to do troubleshooting and maintenance. You set standard configurations. You start seeing patterns. You start to capture some of that in software -- the way to do a standard server configuration with Windows NT or Linux or our mainframe software. It's broad enough to satisfy complexity but repeatable enough across a broad set of customers to gain the efficiencies.
You start to see it in IBM Business Consulting Services, in the work they're doing on the component business model and Web services. We can capture the business-process intellectual property into software. Our friends in BCS are working very hard on seeing how far they can push it and how well they can start capturing elements of a business process in software.
The goal is not to replace people. The people are where the creativity comes from. But you take the repeatable, standardizeable pieces and put them in software that we deliver either as products or as specialized services. The humans can apply more of their creativity to the piece that only humans can do.
I believe this is the beginning of creating interfaces in the services business that will look a lot like the software stack that has evolved over the past 20 or 30 years.... I think XML and Web services are the key technologies that are allowing us to actually start to make progress on part of the stack where the consultants work, and at the data-center level, which is lower in the stack. We're going to start to see a services stack emerge.
Q: What are the pieces of the services stack?
A: Low down in the stack is basic infrastructure. You have the data center, and we're getting it to work really efficiently in a regularized, repeatable way. Being low in the stack doesn't mean low value.
Universal Management Infrastructure [UMI] is our stake in the ground. It's the technology we use to build a repeatable, regularizeable way of running some pieces of data-center operations. Many of the things you had to do in the past with human hands now can be done with tools and software and with less human involvement.
We're defining the way the data center works not by writing code but by setting policies, like what kind of response time to do we want and what kind of uptime do we want, and how fast can we respond to capacity changes, and how quickly can we install new abilities.
Once you do that you can start going up the stack of abstractions. Instead of applying all of my energy to getting the next upgrade of the operating system, I'm going to set the policies for a data center in financial services, or a tax agency in government, or a telecommunications company. If I have done a good job of designing my infrastructure, the infrastructure will be the same and the policies will be different.
Q: How are the polices different for different industries?
A: If you look today at a best-practice data-center operation for a telecom company, an insurance company, and a bank, you'll find they made different choices in how much emphasis they put on computing power vs. storage -- how much bandwidth they need per unit of application. An airline might have a lot of high-performance computing capability doing optimizations on their deployment of aircraft and crews. Another enterprise might have less of that. You describe these things with different policies.
If you put in the right bottom layer of this services stack, and have a set of interfaces that allow the data center to work differently by changing parameters, rather than reimplementing it, then I can take the intellectual horsepower of our team and our customers' team and apply it to optimizing that data center for my industry rather than optimizing it around a fixed point design.
Q: What's the next step up -- some sort of industry-specific middleware for services?
A: I don't have enough clarity yet to know what to call it. I can see some clarity low in the stack, where I can tune operations more to business value. And I can see things happening at the top of the stack in consulting, where we're learning how to capture business processes in software. My view of the middle of the stack isn't clear yet. An industry-specific version of middleware for services could be a way to complete the middle of the stack, but I'm not sure I can say that yet.
Q: Web services is the intersection of software and business processes, right?
A: That's exactly right. The emergence of Web services is the technological base and the lingua franca that makes the idea of a services stack possible. You have to somehow bridge the world of software to the world of business processes.