Exploiting Spatial Locality for Content Placement in Roadside-Unit Caching with Delay Constraint

To cope with the growing demands for mobile traffic in Vehicular ad hoc networks (VANET), Roadside-Unit (RSU) caching is considered as an effective way to reduce the downloading latency. In this paper, RSUs are grouped into disjoint clusters based on their spatial locality. Then in-cluster cache space is considered as an entity by means to increase local content diversity. We first use Markov model to derive the expected download data for delay downloading service under a given content placement. Then we propose two clustering algorithms to perform content allocation which repeatedly update the content placement with a lower download cost. We also explore the impact on the download cost under different simulation parameters, such as the RSU density, the number of contents and the length of download period.