Vicinity Resource Cartography for Delay-Tolerant Networks: a Holistic Perspective

In this paper we propose a distributed algorithm allowing nodes in delay-tolerant networks to construct vicinity resource maps based on past encounters. Our algorithm estimates the availability of the following resources: energy, buffer space, and bandwidth. We then show how a store-carry-forward scheme may benefit from having access to these resource estimates. While knowledge about resources available in the vicinity allows nodes to implement meaningful custodian election strategies, an evaluation of available bandwidth and contact time allows them to carry out adaptive queue management strategies. Thus, such strategies can be approached from a holistic perspective based on the availability of the three resources under consideration in node proximity.We validate our model separately in time-varying and space-varying environments. In addition to synthetic mobility models (random waypoint) we validate our resource estimation in a disaster area mobility model (using Bonn motion traces). We show that by using this information a routing protocol may dramatically improve its delivery rate and energy overhead while keeping delivery latency almost constant.