Motorola Inc. is an anomaly when it comes to Web services. With more than 100 in production, the global communications company is in elite company with, by its estimates, fewer than 5% of companies worldwide having developed as many Web services.
The company has built Web services that go into devices, corporate services that go across the firewall for commercial transactions and some at the business management layer connecting data and applications. These services, built on J2EE, .NET and from third parties, are part of a successful service-oriented architecture (SOA) implemented by vice president of IT strategy, architecture and e-business Toby Redshaw.
Redshaw was hired to bring efficiency to Motorola's IT operations. He battled some internal reticence from developers over SOA, but the payoff has been enormous -- 5x to 10x performance and efficiency gains, he estimates.
"If you've done the Web services management and directory layers properly, you'll get what the manufacturing industries calls 'component reuse'," Redshaw said. "That's a huge win."
Effective management was imperative, especially once Motorola had more than 40 Web services in production.
"Once you get toward 100 or 150, you can poison the well if you don't get it right," Redshaw said. "If you think of doing Web services as a new product for IT, you can tarnish its brand if you don't manage it properly. You will blow it up and blow up parts of your business, not just IT. Business people hate that."
Redshaw is managing Motorola's Web services with products from Oakland, Calif.-based AmberPoint Inc., including AmberPoint Management Foundation, AmberPoint Service Level Manager and AmberPoint Exception Manager. The software products track Web service performance and availability, detect and resolve faults, provide access control and integrate with third-party UDDI registries.
Effective management makes for easier buy-in from C-level executives, Redshaw said.
"Web services make you think in terms of process flows, and that's how business people think," Redshaw said. "When you bring this to the business people, they say 'Thank God, IT is speaking our language. We've been waiting 20 years for this.' IT, meanwhile, says 'this is the fastest, best prototype we've ever done.' Being able to extend this on an enterprise level is huge."
Buy-in from management is a breeze compared to finagling buy-in from developers who are often reticent to reuse code developed elsewhere in an organization.
"You build once in a Web services/SOA model and never have to code that again," Redshaw said.
Developers, however, don't always buy into this thinking.
"You've got three types of people. The first, all they care about is the next best thing. They want to be able to tell their buddies they're working on cool stuff like Web services. The next type is the guy who has always been a COBOL guy and will always be a COBOL guy. You have to do the training and benefit selling with these folks and bring them along. Finally, you've got the group that has figured out that not only is this a great step forward, but they are not going to need as many people any more. Instead of coding something seven times, do it once. It's threatening to them," Redshaw said. "The solution is that you have to have an intervention with them and be dictatorial -- 'This is the way we're going and there is no debate.' If they don't like it, they can play somewhere else."