Heterogeneity in contact dynamics: Helpful or harmful to forwarding algorithms in DTNs?

In this paper we focus on how the heterogeneous contact dynamics of mobile nodes impact the performance of forwarding/routing algorithms in delay/disruption-tolerant networks (DTNs). To this end, we consider two representative heterogeneous network models, each of which captures heterogeneity among node pairs (individual) and heterogeneity in underlying environment (spatial), respectively, and examine the full extent of difference in delay performances they cause on forwarding/routing algorithms through formal stochastic comparisons. We first show that these heterogeneous models correctly capture non-Poisson contact dynamics observed in real traces. Then, we consider direct forwarding and multicopy two-hop relay protocol and rigorously establish stochastic/convex ordering relationships on their delay performances under these heterogeneous models and the corresponding homogeneous model, all of which have the same average inter-contact time over all node pairs. We show that heterogeneous models predict an entirely opposite ordering relationship in the delay performances depending on which of the two heterogeneities is captured. This suggests that merely capturing non-Poisson contact dynamics - even if the entire distribution of aggregated inter-contact time is precisely matched, is not enough and that one should carefully evaluate the performance of forwarding/routing algorithms under a properly chosen heterogeneous network setting. Our results will also be useful in correctly exploiting the underlying heterogeneity structure so as to achieve better performance in DTNs.