A Propagation Based Model for Communication in Vehicular Cloud

Abstract

Vehicular Cloud, a new concept that merges Cloud Computing and Vehicular Network is an example of pervasive computing. This idea is based on of benefiting from the unused vehicles parked in a specific location for a period of time by running a computational operation. Because of the embedded access point in vehicles, Vehicular Clouds are main part of Internet of Things. One of the challenges in Vehicular Cloud is the communication between the vehicles. Because of dynamic nature of Vehicular Clouds, communication links may emerge and fail randomly which leads to connection failure in the entire network. Here we use the underlying network structure to propose a new strategy to mitigate this failure of service. We use label propagation method in order to cluster the network. In short, the idea is to connect a computation node with nodes within the community rather than nodes outside the community. This model increases the locality which is the main factor in performance optimization.