Theoretically you can use (almost) any transport type for web services, like email, JMS, FTP, HTTP or pidgeon, but I think that HTTP is the only one officially supported at the moment.
If you look at the JAX-WS web page https://jax-ws.dev.java.net/ there is a link to JAX-WS over JMS.
The best place to look for information is the JAX-WS standards document: http://jcp.org/en/jsr/detail?id=224 Best wishes!
Thanks Ivan for the reply. I had a look at the links, they are not providing the information which I seeking. I want to know does JAX-WS web service, uses Java RMI or JAVA RPC feature or something else.
Further to Ulf's explanation, to answer your specific question, Web Services can publish their services (and methods / contracts) using UDDI. This contains information like the Organization hosting the services, the service names, method names, description of services and the WSDL URLs. The clients who want to access this have to query the UDDI registry (clients of course do need to have knowledge of the registry itself) and get to the required service.
That said, this is not at all compulsory. Many time, what happens is that once the service is found out (by physically locating it OR by sharing it across teams) clients stubs are generated and the method calls are given - All that matters then is the WSDL / URL to the WSDL.
Hi, no it does not use RMI. The previous standard JAX-RPC used it.
In rmi you do a lookup and the concept is somewhat similar in JAX-WS, i.e. to locate a service, in your code you use specific objects that work with the information in the wsdl which specifies where each service is located. Once the url is found you just call the functions in the local stubs and the framework takes care of the rest. The wsdl, you either know where to find it or you do a more formal look up as specified by Shailesh