Configuration Management for Proactive Adaptation in Pervasive Environments

Abstract

Pervasive computing applications have to adapt in order to cope with changes in their environment. Proactive adaptation allows to change the application and/ or the context based on prediction of context and user behavior, in order to reduce adaption delay and provide a seamless service to the user. Thus, such applications are self-organizing systems. Ideally, self-organizing systems adapt by changing their structure or behavior without requiring the user's intervention. A prerequisite to that is the knowledge of the possible configurations and their order with respect to the suitability. We introduce a comprehensive framework based on an application model with suitability and cost metrics. Based on the application model, we construct a CSP and develop an algorithm with two heuristics that finds all configurations. We rate the configurations depending on their expected instantiation using novel utility and cost functions. In the evaluation, we show the feasibility of our approach.