Abstract:
In previous work the authors have described an approach for building distributed selfâ¿¿healing systems â¿¿ referred to as EmbryoWare â¿¿ that, in analogy to Embryonics in hardware, is inspired by cellular development and differentiation processes. The approach uses â¿ artificial stem cellsâ¿ that autonomously differentiate into the node types needed to obtain the desired systemâ¿¿level behaviour. Each node has a genome that contains the full service specification, as well as rules for the differentiation process. This approach has inherent self-healing behaviours that naturally give rise to fault tolerance. Previous evaluations of this fault tolerance have however focused on individual node failures. A more systemic fault modality arises when the nodes become mobile, leading to regular changes in the network topology and hence the potential introduction of local node type faults. In this paper we evaluate the extent to which the existing fault tolerance copes with the class of faults arising from node mobility and associated network topology changes. We present simulation results that demonstrate a significant relationship between network stability, node speed, and node sensing rates.
