Providing for change through Adaptive Object Models and Autonomous Computing techniques

This paper explores how autonomic behaviour might be supported by software that is constructed using adaptive object modelling (AOM). AOMs provide a meta-layer architecture through which controlled (i.e. domain meaningful and correct) transformations of base-level computation can be achieved. Our work is derived from a model driven approach to systems construction, in which architectural modelling artefacts of a development process are used directly to construct a system AOM. Modification of modelling is explicitly mapped to AOM adjustment, and thus to system modification. We propose an extension to this approach to capture dynamic and autonomic concerns. We explore how resulting systems models might readily accommodate unforeseen changes in the environment and react appropriately without the need for programmer intervention