Asynchronous modeling - how it works

The approach described for synchronous tools, cannot cope with asynchronous modeling. In these tools, it is the subsystem itself that decides when a call will actually take place - the VM is a callee in this case, not a caller. Not only that, but the asynchronous call can't afford to wait for the call to return (and therefore, output parameters are meaningless). The ``bridge'' mechanisms must therefore be different for asynchronous modeling tools.