Message routing in mobile delay tolerant networks inherently relies on the cooperation between nodes. In most existing routing protocols, the participation of nodes in the routing process is taken as granted. However, in reality, nodes can be unwilling to participate. We first show in this paper the impact of the unwillingness of nodes to participate in existing routing protocols through a set of experiments. Results show that in the presence of even a small proportion of nodes that do not forward messages, performance is heavily degraded. We then analyze two major reasons of the unwillingness of nodes to participate, i.e., their rational behavior (also called selfishness) and their wariness of disclosing private mobility information.Our main contribution in this paper is to survey the existing related research works that overcome these two issues. We provide a classification of the existing approaches for protocols that deal with selfish behavior. We then conduct experiments to compare the performance of these strategies for preventing different types of selfish behavior. For protocols that preserve the privacy of users, we classify the existing approaches and provide an analytical comparison of their security guarantees.

The increasing pervasiveness of mobile devices with networking capabilities has led to the emergence of Mobile Delay Tolerant Networks (MDTNs). The characteristics of MDTNs, which include frequent and long-term partitions, make message routing a major challenge in these networks. Most of the existing routing protocols either allocate an unlimited number of message copies or use a fixed number of message copies to route a message towards its destination. While the first approach unnecessarily floods the network, the rigidity of the second approach makes it inefficient from the viewpoint of message replication. Hence, the question that we address in this paper is: "how to dynamically allocate message copies in order to strike a balance between the delay and cost of message delivery?". We present a novel adaptive multi-step routing protocol for MDTNs. In each routing step, our protocol reasons on the remaining time-tolive of the message in order to allocate the minimum number of copies necessary to achieve a given delivery probability. Experiment results demonstrate that our protocol has a higher delivery ratio and a lower delivery cost compared to the state-of-the-art Spray-and-Wait and Bubble protocols.

No abstract