[ Home ] [ Up ] [ Remote Objects ] [ Remote Method Calls ] [ Components ] [ The RMI Registry ] [ A Simple Example ] [ Listing Remote Objects ] [ Extended Example ] [ Caveats ] [ Passing Objects in RMI ] [ Remote Callbacks ]

How Does RMI Perform a Remote Method Call?

There are a number of components and actions involved in performing a remote method call:

• In order for a remote method call to appear as if it is local, there has to be a local object which 'stands in for' the remote object -- that is, it presents the same interface to the client as the remote object does.  
  • The local object is called the client stub.
  • The interface that both the stub and the remote object present is known as the remote interface
• When a call is made to one of the methods in the client stub, the stub is responsible for somehow:
  • Invoking the corresponding method on the remote object
  • Passing any parameters provided on the local call to the remote object's method
  • Eventually, obtaining any return values from the remote method (output parameters, or the return value from the method) and passing them back to the local caller.
• The mechanism used to pass values back and forth between the client stub and the remote object is called marshalling:
  • Data to be passed over the connection from the client to the server is first marshalled into a form that may be conveniently transported over the 'wire'.
  • On the server side, the data is received, and must be unmarshalled into a form that can be used by, and supplied to, the remote object.
  • Data to be returned from the remote object to the client stub must similarly be marshalled on the server side, and subsequently unmarshalled on the client side.
  • Remember that code can cause exceptions to occur;  exceptions can be [un]marshalled, too.