Walt Johnson is an IT planner at California Independent System Operator (CalISO), the not-for-profit operator of the state’s wholesale power grid. I met him at InfoWorld’s SOA Executive Forum last week, where he described CalISO’s transition to service-oriented architecture.
Now entering its fourth generation, CalISO’s electronic marketplace uses Web services on an enterprise service bus to receive bids and send notifications. None of this is visible to the customers drawing power from the grid. They just assume that there’s a competitive market and — if they think about it at all — that suppliers are interchangeable.
Several attendees asked when consumers of outsourced business services will enjoy the same kind of substitutability. To be honest, we have a long way to go. If you rely on multiple sources for credit checks today, for example, and you want to plug them in to your SOA, you’ll have to normalize the interfaces yourself.
According to one of the credit bureaus’ customers, RouteOne’s Technology Director T.N. Subramaniam, that’s true because some of these companies are making money hand over fist doing things their own way. There’s no real pressure yet to present SOAP and WSDL interfaces.
Several of our panelists pointed out that customers should beat up on vendors whose SOA products aren’t interchangeable at the level of basic standards. The same holds true for services. QoS is the only acceptable lock-in. Providers that don’t make their customers’ data available in a common XML format, or whose services can’t easily be second-sourced, will meet growing resistance.
When SOA-style services finally become commodities in b-to-b markets, there are reasons to hope that those markets could work quite well. Self-describing data flows will lend themselves to rigorous analysis. Formal interface contracts will enable equivalent services to substitute for one another. Service consumers will be able to easily check quality metrics published by service providers. Call me an optimist, but I really believe that the transparency made possible by SOA could influence the b-to-b landscape in profoundly positive ways.
Of course I’m a realist too. Information silos don’t crumble just because XML formats arrive on the scene. Customers must light up the dark corners of service interoperability. How? My guess is that new forms of social software will be our flashlights.
Nowadays on the Web, people are cataloging everything in sight. At LibraryThing, for example, you can maintain a personal inventory of books that interest you. You can use it to organize books — either formally by ISBN, Dewey Decimal, and Library of Congress numbers, or informally using del.icio.us-style tags — then search along these and other axes.
What makes LibraryThing more than a pastime for lonely obsessives, however, is its social dimension. The service finds collections like yours and immediately puts you in contact with people who share your interests, whose knowledge of the books in your collection complements your own, and whose experiences with them can be helpful to you.
Commodity services today are few and far between, and substitution of one for another is a rare event. But all this will change, and when it does the mechanisms we use to collectively catalog and evaluate everything else under the sun will make us better consumers of services, too.