Future Information and Telecommunication Systems are expected to be pervasive and ubiquitous solutions, able to consider users’ context and to automatically adapt to their environments. Traditional configuration and management tools are not adapted. The richness, the heterogeneity and the complexity of the upcoming systems require automated solutions able to gather contextual information, to reason on them and to make the appropriate adaptation decisions. The representation and the sharing of contextual information is a key issue. In this thesis, we proposed and used a methodology to conceive « Ubiquity-Ont », a generic ontology dedicated to Information and Telecommunication Systems. Contextual information are the described through semantic concepts, instances and relations. We then proposed an overlay architecture, composed of two virtualization layers that can integrate a semantic context management framework over existing networking environments. This architecture is able (a) to hide any heterogeneity among the system components and (b) to augment the different entities with additional capacities for context gathering, reasoning and sharing operations. The proposed solutions were then implemented and tested in Lab for two applications. The fisrt is related to mobility management over heterogeneous Wireless Networks and the second aims to power optimization on mobile terminals. These two case studies helped in proving and enhancing the proposed solutions
