Skip to Main Content
For many applications in mobile wireless ad hoc networks (MANETs), forming an end-to-end data path is not always necessary; instead, the primary routing goal is often data collection or dissemination where only a data source is known. Routing algorithms must be carefully chosen in order to suit the needs of applications employing them. Our focus is on data collection applications in MANETs where limited mobility information is required to route data in a scalable manner. To address this goal, we employ the concepts of delay-tolerant networking (DTN) in which data makes progress toward a destination with high latency expectations and little knowledge of routing topology. Specifically, we present time-to-network (TTN) forwarding, a method of forwarding data generated by mobile nodes to a network endpoint in such a way that delivery latency is lowered without high networking cost. By segmenting mobility patterns into trips, we are able to apply TTN to a vehicular network using only an estimated destination arrival time for each vehicle. We evaluate TTN using mobility data from the TRANSIMS simulator for a real road network. Results show that our algorithm produces collection-to-network latencies similar to more generic algorithms but at a lower cost and with higher efficiency. Furthermore, we establish a lower bound for delivery latency in our experiments and compare it to TTN. This also helps normalize the interpretation of results specific to our mobility model.
Date of Conference: 14-17 June 2010