Abstract
The concept of Delay Tolerant Network (DTN) is a relatively new research field, and one for which many applications have been identified. It is an overlay network working over heterogonous networks in a challenging environment, where the links are less reliable than in traditional networks and the delays are expected to be very long. Due to the difficult working conditions of DTN, all available methods of data transmission are utilized to ensure a better delivery of messages, including physical transportation methods such as buses and ferries. In this thesis, I have proposed the use of aircraft on scheduled routes for data transportation as a new method of DTN data transport. Aircraft fly daily routes and pass over remote locations where the communication infrastructure is limited. They can be used to relay users’ messages to their destinations using reasonably simple radio devices. The thesis presents an analysis of aircraft routes and possible scenarios for the realization of this DTN concept, supported by various simulation results. The notion of Quality of Service (QoS) in DTN is difficult, due to the challenging nature of DTN and the fact that it may suffer from intermittent connections and long delays. I have explained the meaning of QoS as applied to DTN and have suggested how to improve it in such a challenging environment. After defining the QoS notion in DTN, I have presented the requirements for QoS in DTN and identified the DTN flow characteristics and QoS performance metrics, showing how the control and management of these metrics enhance the DTN performance. Furthermore, since DTN networks have scarce resources, it is important to control admission to those resources and to best share them among the users, according to each user’s privilege. Finally, I have proposed the Fair Allocation of Resource Model (FARM) for DTN applications. This model is based on the local information of the node resources and function to avoid network congestion and enhance DTN performance. The model controls admission to DTN resources, based on an evaluation process of each user’s past admission attempts, class of service and their time request order. Furthermore, I show the implementation of FARM in the Opportunistic Network Environment (ONE) simulator, which is a platform for DTN application simulations. The simulator produced good results, and indicates the advantages of using FARM in a DTN environment.