Discovering Web Services

Discovering Web Services

Three things happen during the life cycle of Web service client development and deployment: The first is discovering the Web service you will be using. Next is examining the description of that service and using it to build a client. Third is the execution—the sending of messages.

This book is working through this list backwards: We started with SOAP and HTTP, protocols that are part of the execution phase. After that, we looked at WSDL, which is used to describe Web services. This description phase shouldn't be downplayed, as it is the area of greatest confusion, particularly in terms of interoperability. This chapter covers the very first phase: discovery. If you haven't noticed this missing piece of the puzzle before now, you aren't alone. Ad hoc discovery (simply writing a custom application for registering and finding Web services, such as those found on Xmethods.net) is still very popular. In fact, some would argue that discovery is an application-specific task.

I disagree. Ad hoc discovery can be useful, particularly for very informal development efforts. But discovery is not merely a design-time query for the service with which you are interacting. For example, it also can be a runtime query used to find new service endpoints during failure rollover. And, at design time, there are still many reasons to standardize how your company publishes both internal and external Web services. Figure illustrates how this process of discovery, description, and messaging can occur.