Dynamic Adaptation

 The intention is to enable the mobile code to move with a small core part which is environment independent through the net and to add dynamically, i.e. without termination, environment dependent implementation for the particular environments only. The environment is assumed to be static for the runtime of the mobile code on a host.

 

Figure: Dynamic adaptation overview  

A change of the environment occurs only if the mobile code moves to a different host. The decision which implementation must be used is done without a prior knowledge of the environment of the host.

The mobile code is split into an environment-independent non-adaptable core part and an environment-dependent adaptable part. With dynamic adaptation only code is moved over the network when it is actually needed.

The input of dynamic adaptation is a set of environment dependent implementations (X,Y), a mobile core application and an environment. The result of dynamic adaptation is the selection of the right implementation for the environment and the linking of the selected implementation into the core application. This is also depicted in figure .

 

Figure: Event and object of dynamic adaptation  

The trigger for dynamic adaptation is the movement of code. If the mobile code moves to a new host, the dynamic adaptation procedure is initiated. (s. fig. ).

The objects of dynamic adaptation are code fragments, which are exchanged (s. fig. ).