Lately I've heard the term Enterprise Service (ESB) Bus tossed around so much that I question if some people realize how big a deal it would be to put one in place. However, it is not as though a person can run out and try playing with one on their desktop because it looks as though the average price to purchase the software for one is $10,000 per cpu. This is not an option for most people, including me, so I'll cut them some slack.

I am however able to read about the products and the ideas behind them. Most of what I read is grand talk about how these products will solve most problems in the enterprise by connecting all of our applications and providing the ability of universal access and organization. At the most basic level that I can understand the ESB is simply a registry for web services and then a layer of abstraction that clients of those services can interact with to exercise a desired function. Also, I understand that more than web services can connect to an ESB but I don't care about that aspect of things.

So I'm going to take a leap of faith and say that the majority of the applications that would connect to an ESB would be doing so via web services. I will guess again that the majority of the existing applications in a typical "enterprise" don't currently expose their functionality through web services. Therefore exposing functionality in this way will require a lot of planning and an enormous amount effort.

In my opinion organizations should target specific applications and more specifically functions within those applications to pilot this type of effort. Simply saying "we are going to implement an Enterprise Service Bus" should stop most people in their tracks. Doing a pilot can help blaze a trail for those that will follow and set internal best practices which others can use later on.