Geographic validity aware content retention in vehicular networks

Abstract

In Vehicle-to-Vehicle (V2V) communication, interest-based content dissemination carries equal importance. For example, a group of vehicles involved in a safari may communicate with each other about the location of rare animals. These messages are useful only within a certain geography and within a certain timespan. Hence, messages injected into the V2V network should be retained within these boundaries regardless of the highly dynamic nature of the underlying V2V network. To ensure that the content is retained within the V2V network both efficiently and with high certainty, it is important to address problems such as how and when to disseminate content, how to maintain order and honor priorities of content, how many replicas to maintain, and when to evict the content. However, if the content message is passed every time a pair of vehicles comes into each other's range, it will lead to message implosion while suboptimally utilizing the wireless links, power, and content storage. Therefore, to ensure that all the vehicles get the message without high certainty and efficiency, it is imperative to identify with what probability a message should be disseminated. In this research, we identify this probability value that could lead to successful retention of the message within the network given the parameters such as the valid geographical boundary, time span, and vehicle arrival rate. We developed a model that estimates the minimum probability that needs to be maintained to ensure that the specific content is seeded among other nodes. The model was developed for straight roads, T-junctions, and four-way intersections by varying other parameters such as the valid geographic area, vehicle speed, and density. Simulation bases analysis shows that the proposed model could reasonably estimate the minimum probability that needs to be met for the message to be replicated in other nodes.