Calling time on ESB
For some time now, I’ve been holding back from letting a growing frustration boil over. Specifically, my frustration is with the creation of new terminology to represent concepts which are already well-understood, with the purpose of creating new product opportunities; and with the twisting of existing terminology to represent concepts other than those which were originally framed. But with the growing interest in the “Enterprise Service Bus” (ESB), I’m afraid I just can’t control myself any longer.
First, let me say that I’m convinced that the concept of an ESB is absolutely valid in the context of implementing a service-oriented approach to application delivery. Indeed Sonic Software (often credited with the creation of the term, supported by
Roy Schulte) provides a sensible definition: “An ESB is software infrastructure that simplifies the integration and flexible reuse of business components using a service-oriented architecture. An ESB makes it easy to dynamically connect, mediate and control services and their interactions.” The
Wikipedia entry for ESB is good, too.
What’s getting me really uppity is the way that some vendors (you know who you are) are attempting to make it very clear that “old-school” technology vendors offering this technology are not really worth consideration, because either (a) their technology isn’t built from the ground up based on Web Services; or (b) they don’t have a product called an ESB.
Let’s be honest – this is a cynical marketing move. The ESB product category is a complete fiction, invented by software companies (abetted by some of my fellow analysts) intent on convincing the world that their technology is materially different from that which has been delivered by the “old-school” players for 10+ years. What’s really incredible, is that these old-school vendors are now coming under pressure from their salespeople because prospects are saying “you don’t have an ESB product, so I can’t work with you”. Somewhere along the line these prospects become convinced that only a product with ESB on the box, is going to do the job. This interview with IBM’s
Steve Mills is just a symptom of what’s going on.
But the truth is that if you’re an IBM customer with a significant investment in WebSphere’s integration technology, there's a good chance you already have an ESB. Same goes for SeeBeyond (now
part of Sun). And TIBCO. And webMethods.
The “ESB product vendors” try to convince prospects that because products from the old-school vendors don’t always provide the same features as they do out of the box, they’re not doing the job. But customers of the old-school vendors already know that implementing enterprise-wide application infrastructure isn’t something you can do by just installing software from a DVD.
So, come on – let’s hear it. Am I
tilting at windmills, or
naming the elephant?
