Infrastructure Based Chord Structure for P2P File Sharing over Vehicular Network

Abstract

Vehicular network is different from wired network due to its network environment changes rapidly. The connection lifetime between vehicles is usually short because vehicles move in high speeds. Hence, deploy peer-to-peer (p2p) applications over vehicular network is a challenging research issue. There are many problems confronted in p2p file sharing over vehicular network. Therefore, how to search files effectively, how to share files with distant vehicles, how to fully utilize the limited bandwidth, etc. have became important issues. In this paper, we propose an infrastructure-based Urban Multi-Layered Chord (UML-Chord) architecture for p2p file sharing over an urban vehicular network. In this architecture, vehicles are grouped into clusters. A multi-layer chord is employed to manage both files information and clustered vehicles locations. Connection lifetime between vehicles is estimated on a traffic aware basis. Target vehicle selection considers connection lifetime, bandwidth, and vehicle transmission status. Simulations results show that UML-Chord outperforms Chord, clustered organization and overlay based approaches in terms of number of hops (per query), control message overhead, query delay time, average download time, and file complete ratio.