When you call a long lived process you will see that you have a couple of additional methods to the web service. In the short lived process you only have one method (invoke). So calling the long lived process you woudl use the invoke method and it will return a process id. Now that you have the process id you can make subsequent calls to get the status as well as the result of th eweb service call. The secondary call requires that you pass in the process id that you recieved from the 1st step.
Remember that the long lived process is set up that way because it will take time to get the result as opposed to the short lived process which will return the result in a short period of time.