Bob Sutor predicts that in three years, your business will depend on Web services—software that helps companies connect their systems across networks. As IBM’s WebSphere Infrastructure Software director—the person who drives the company’s key Web services product line—Sutor might be expected to unleash a hard sell. But given the opportunity, he allows that Web services are a lot like plumbing: a largely invisible technology that’s at its best when it’s out of sight and out of mind. Sutor refers to Web services as a “stealth technology,” a description that belies the overt impact it will have on business. In this edited interview with HBR’s Gardiner Morse, Sutor discusses how and when Web services will change the ways companies do their work.
Just the thought of Web services seems to make a lot of managers anxious. Why is that?
There’s a lot of hype and confusion about Web services. Many businesspeople don’t really understand what they are, but they sense there’s an IT revolution going on, and they’re worried they’ll get left behind. Actually, we’re in an evolution, not a revolution.
If you think about the ways that businesses have connected their distinct software in the past—say, for placing orders at one end and invoicing and shipping at the other—the problem has been that there have been so many different ways of doing it. If I have one software solution for connecting with you, and another solution for connecting with another partner, and so on with many partners, you can see that my IT department is spending a lot of time and money just keeping an increasingly complex system up and running. A Web service application is simply a piece of software that sits between my partners and me and allows all these disparate systems to communicate more easily. So if we can reduce the complexity of connecting systems together, we can either reduce our IT resources or put them to better use to make companies more efficient and competitive.
What’s a real-world example of Web services changing how a company does business?
Bekins is a major shipping company. One of its units specializes in delivering high-value consumer goods, like large-screen TVs, from retailers to homes and businesses. To do this, Bekins uses a network of 1,600 agents, who own the trucks. In the past, when Bekins received a shipping order, if the centrally managed fleet couldn’t handle it, the company used the phone and fax to place the job with a particular agent. The process wasn’t always efficient, and it could be inequitable, since many agents had overlapping territories. The question was, what’s a better way to connect orders and agents, and automate the process, given that they all used different systems and software?
Bekins built a Web-services-based system that essentially created a virtual marketplace in which agents could select jobs. When Bekins got a shipping order, the company would tender it via Web services simultaneously to all the agents who’d signed up for the system. Any agent could accept the job, and once accepted by an agent, it would become unavailable to the others. The result has been increased efficiency, faster response time, less idle time for trucks, and more satisfied retailers. And because of the system’s efficiency, Bekins is also able to accept lower-margin jobs that it would have passed on before. The system is expected to increase shipment volumes and deliver increased revenue to Bekins by as much as $75 million annually.
Many companies are developing Web services software—Oracle, Microsoft, IBM, and Sun, among others. If I’m a company considering using Web services, shouldn’t I wait to see who will become the dominant player?
I don’t believe there will be a dominant player in the long run. To be honest, Web services are like plumbing. Houses have standardized pipes; they’re all designed to connect, and there are rules about how you connect them. Web services are like these standardized pipes. There isn’t one single pipe supplier—there are many, and their pipes are all compatible. The pipes are less interesting than the fixtures at the ends. The fixtures—the software that Web services technology connects—is where the value is going to be, and that’s where we think IBM and our customers will come out ahead. At the risk of forcing the analogy: Once the plumbing is installed and it works, you don’t think about it. Web services will be the same way. Much of it will become invisible, and managers will focus on what Web services allow them to do—things like searching across multiple suppliers simultaneously for the best price, outsourcing operations, and connecting with an acquired company’s systems. I’m not saying you can’t do those things somehow, some way, right now. What Web services can do is standardize and simplify these activities.
With so many disparate systems being connected through Web services, shouldn’t companies be concerned about security?
Security is a hot topic right now, for very good reasons. It’s a major area of Web services development, and it needs to be a lot more sophisticated than the security you use to send credit card data over the Web. For example, imagine I have a business that keeps all employee information in-house in an ERP system. My employees can sit down at our intranet portal, enter their serial numbers and passwords, and get full access to their job and salary histories, 401(k) accounts, and so on. Security is provided in some way so that only the appropriate people can view and update HR data. Now, suppose I want to outsource all these HR functions to a specialized company that can provide greater value and additional features. This company has many clients and its own security system. I am not willing to change my intranet security infrastructure to match the HR company’s, since my infrastructure is used elsewhere in my enterprise and, after all, I’m paying for the outsourcing. Can we somehow bridge this divide so that I can outsource securely and my employees can still seamlessly access their data? The IT industry is now working on standards to provide these kinds of security features. There’s a good road map showing what needs to be done, and you should see standards-compliant Web services security products from several vendors by the end of this year.
Are we past the point of early adoption?
Web services are about three years old, so we’re past the time when the earliest adopters decided to take a risk on an unproven technology. It’s not at all a wild frontier out there now. I’d say we’re in the early mainstream period. There will be continued standardization of components that make up Web services, and that process should be complete around the end of 2005. For many companies, downstream it’s not going to be a very active decision to use Web services because the services will be built in semiautomatically by the software-development tools that are or will shortly be available. So, over time, as companies build applications, Web services will sneak in there. It will be kind of a stealth technology.
