Predicting Queueing Delays In Delay Tolerant Networks With Application In Space

In this paper we present a study of queueing delays experienced in Delay Tolerant Networks with topology based on deterministic contact plan schedules. We examine a generic scenario and propose a sampling procedure that extracts measurements of queueing rates and queue lengths. Sampling queueing information is transmitted to network nodes, which then form time series and can be used to forecast future queueing rates. Through simulations we show that the introduced method can be useful for DTNs with predetermined contact schedules, such as the Interplanetary Internet, providing accurate end-to-end delivery delay predictions.