By disseminating data through Vehicular Ad-hoc Networks (VANETs), vehicles are able to share relevant sensor data to acquire information about accidents, traffic, and even pollution. Data relevance is measured by a utility function which considers the contextual information that vehicles currently have about their environment. To be effective, data dissemination protocols must cope with intermittent connectivity due to the high speeds of vehicles. Problems arise when not all data can be exchanged due to the limited time available. In this paper, we explore and compare two fundamentally distinct approaches to tackling this problem. The first aims to maximize the system efficiency. In contrast, the second trades efficiency by a fair data distribution over vehicles by means of Nash Bargaining as used in game theory. By means of an extensive simulation campaign, an approach relying on fairness is shown to outperform efficiency in terms of delivery ratio, Jain's fairness index, sum of utility gains, number of hops and number of files downloaded.
展开▼
