Publish-Subscribe framework for disruption tolerant networks

Disruption Tolerant Networks (DTNs) are characterized by limited end-to-end connectivity between nodes. As a result, nodes exchange messages through interactions that exploit opportunistic connectivity and propagate data through a store-carry-forward model. DTNs have been envisioned for numerous applications that require collaborative communication. Publish-Subscribe framework is one way to address the challenges involved in collaborative communication for DTNs. There are three important components to any publish-subscribe framework: event dispatcher, event matching, and event forwarding. This dissertation specifically focuses on addressing challenges involved in developing protocols for event forwarding and event dispatching in DTNs.;The contributions of this thesis are threefold. First, it focuses on reliable multicasting for one to many data collaboration. This thesis proposes a scalable, distributed multicasting scheme that deterministically guarantees message delivery to all multicast receivers in a group. Unlike, traditional reliable multicasting, the proposed scheme does not utilize tree construction or flooding and instead makes use of opportunistic connectivity between the nodes. Performance evaluations show that the proposed scheme is better in terms of latency and overhead as compared to other traditional schemes. Second, this thesis proposes an anonymous communication and key management scheme to ensure privacy of network nodes in the context of reliable multicasting. Simulations and experimentation on a mobile platform are used to characterize the overhead of the proposed scheme. Third, this thesis develops schemes that trade off delivery guarantee for scalability. These schemes encourage intermediate nodes to act as forwarding agents by providing them with an incentive for cooperation. Simulation and experimentation on a mobile platform demonstrate the efficiency and feasibility of such an approach. Collectively, these three contributions create an efficient framework for data collaboration among network nodes in DTN.