Identity comes first

If you're concerned enough about web services to want to manage them, then you'll be even more concerned to control who gets access to them. That's the rationale advanced by Oblix for its acquisition of Confluent — as reported in this week's Loosely Coupled feature article, Beating a path from identity to services — and I'm sure it's an accurate analysis. But does it justify the purchase?

Oblix has observed that customers who buy web services management software already have identity management in place. To be honest, this is not exactly surprising, considering that most enterprise applications and networks are at the very least password-protected, and that directory services to help manage all of that have been around for quite a long time. So any enterprise that's sophisticated enough in its use of IT to be thinking about managing its web services has probably long since taken steps to get on top of identity management.

Having done so, will it then want to use the same vendor for web services management? This I think is where the flaw lies in Oblix's thinking. If a company is of the type that prefers to get all its software from a single source, then it won't have become a customer of Oblix in the first place. It will have gone to IBM, or Computer Associates, or whichever other vendor already supplies the bulk of its infrastructure management software.

The type of company that is going to select a specialist vendor like Oblix for its identity management needs is far more likely to select another specialist when it considers its web services management needs. It's not as if web services is just an add-on to identity management, after all. To say that identity management is a prerequisite for operating a services-oriented architecture does not make it an all-encompassing framework for everything else.

Identity does come first, because in a true services environment there needs to be an independent function that can confirm the credentials of each participant to everyone's satisfaction. But the smooth operation of the infrastructure depends on a lot more besides.

From a purist point of view, then, web services management is a separate skill from identity management — in fact, you can probably subdivide it further into a cluster of separate skills, each as complex as identity management. Companies that want to take full advantage of all these capabilities right from the start won't be buying services management from Oblix. But such companies are exceptions. The majority of enterprises take a more pragmatic approach, and will be happy to work within an existing supplier relationship. If Oblix has enough of these companies in its customer base then it has the opportunity to do well out of acquiring Confluent.

posted by Phil Wainewright 6:39 PM (GMT) | comments | link