A Realistic Analytical Methodology for Mobility-assisted Routing

Traditionally, ad hoc networks have been viewed as a connected graph over which end-to-end routing paths had to be established. Mobility was considered a necessary evil that invalidates paths and needs to be overcome in an intelligent way to allow for seamless communication between nodes. However, it has recently been recognized that mobility can be turned into a useful ally, by making nodes carry data around the network instead of transmitting them. This model of routing departs from the traditional paradigm and requires new theoretical tools to model its performance. A mobility-assisted protocol forwards data only when appropriate relays encounter each other. To be able to evaluate the performance of any mobility-assisted routing scheme, it is necessary to know the statistics of various quantities related to node encounters, such as meeting times and contact duration. These quantities depend on the specific mobility model in hand. In this paper, we present an analytical methodology to calculate a number of useful encounter-related statistics for a general class of mobility models. We first apply our methodology to derive accurate closed form expressions for popular mobility models like Random Direction or Random Waypoint. To validate the usefulness of our analytical methodology, we then apply it to a more sophisticated mobility model that better captures some important mobility characteristics observed in real traces. Finally, we show how our results fit into a general theoretical framework that can be used to analyze the performance of mobility-assisted routing schemes or other processes based on node encounters (e.g. wireless virus spread). We demonstrate that derivative results concerning the delay of various routing schemes are very accurate, under all the mobility models examined. Thus, this work can help in better understanding how various algorithms and processes that depend on mobility perform in different settings, and can facilitate the design of more robust protocols.