A risk-reward competitive analysis for online routing algorithms in delay tolerant networks

Abstract

Delay Tolerant Networks (DTNs) suffer from frequent disruptions and partitioning. Because of dynamic environment and lack of knowledge about the whole network topology, routing is quiet challenging in DTNs. Prediction based algorithms in DTNs are able to predict nodes future contacts in the case where each node has a movement model. In this paper we suppose that a new node is added to the network which is about to send messages to a destination node and has no knowledge about the upcoming contacts. Deciding about the next node to forward the messages to, is an online problem. We are going to analyze a greedy forwarding strategy using competitive analysis and also propose a risk-reward analysis framework for the risky version of the algorithm and obtain risk and reward values for different forecasts. Risk-reward analysis can be used to develop better forwarding strategies which decide about the next hops more wisely. © 2012 Springer-Verlag.