XOR Network Coding for Data Mule Delay Tolerant Networks

Abstract

We propose a simple yet efficient scalable scheme for improving the performance of Delay Tolerant Networks (DTNs) with data mules by using XOR network coding. We carry out a theoretical analysis based on a model abstracted from the Village Communication Networks (VCNs), beginning with two villages and then extending to N villages. We also examine how the delivery probability is affected by the different overlapping intervals of two data mules. The theoretical analysis indicates that the maximum delivery probability increases by 50% and our simulation results illustrate this point, showing that the overhead ratio and average delay are reduced as well. Finally, our scheme is applied to a real network, the Toulouse public transportation network. We analyze the dataset, calculate the overlapping intervals of inter-vehicles and the amount of data that transit vehicles can exchange in one day, showing a 54:4% improvement in throughput.