There is a growing interest in Delay Tolerant Networking (DTN). This research field aims at providing communication means to extend the current Internet architecture for the support of challenged networks. These networks are mainly characterized by the fact that connectivity between entities suffers from disruptions. We examine in this thesis the problem of routing by using knowledge about network connectivity. In this thesis, we make several contributions to DTN routing. We investigate particularly scenarios where network entities are mobile (e.g., mobile phones, PDAs) and carried by people sharing social relationships. First, we show, with the analysis of real traces, that there is heterogeneity in interactions between participants of such networks and we demonstrate that this heterogeneity could be taken into account to propose efficient routing schemes. Second, moving in sich direction, we propose routing algorithms based on the use of a high-dimensional Euclidean space, that we call MobySpace, constructed upon nodes' mobility patterns. We have shown, through the replay of real mobility traces, that MobySpace-based routing schemes can applied to DTNs and that it can bring benefits in terms of enhanced bundle delivery and reduced communication costs. Finally, to contribute to the on-going data collection effort, we present an analysis of contact traces that we collected in an experiment we conducted in Cambridge, UK. This experiment allowed us to study the feasibility of a city-wide content distribution architecture composed of short range wireless access points.
