This topic describes two brief scenarios that illustrate the role that Web Services Invocation Framework (WSIF) plays in the emerging Web services environment.
Scenario: Redevelopment and redeployment
If you are implementing Web services today, you are probably working with simple prototypes. As your Web services move into production, you need to reimplement and redeploy them. WSIF uses the same API calls with different underlying technologies. If you use WSIF you can reimplement and redeploy your services without changing the client code, and you can use existing highly reliable and high-performance infrastructures like RMI-IIOP without sacrificing the location-independence that the Web service model offers.
Scenario: Service Flow composition
A service flow typically invokes a Web service, then passes the response from one Web service into the next Web service, perhaps performing some transformation in the middle.
There are two key aspects to this that WSIF provides:
For example, imagine that you build a meta-service that uses a number of services to build a process. Initially several of those services are simple JavaBean prototypes that are written and exposed through Simple Object Access Protocol (SOAP), but you plan to reimplement some of them as EJB components, and to out-source others.
If you use SOAP, it ties up multiple threads for every onward invocation, as they pass through the webserver and servlet engine into the SOAP router. If you use WSIF to call the beans directly, you get much better performance compared to SOAP, and you don't lose access or location transparency. Using WSIF, you can move some of the Web services from local implementations to external SOAP services you just update the WSDL.