Design, implementation and evaluation of a structured P2P algorithm for vehicular communication systems

Distributed information management is particularly powerful in vehicular networks. For example, it enables vehicular communication systems to persistently provide up-to-the-minute information about accessible parking spaces, existing hazards or traffic densities and to keep these items up to date. The Vehicular Information Space framework (VIS) implements distributed information management. Vehicles running the VIS form a kind of distributed database. They are able to provide information items in a location-aware and fully distributed manner. Moreover, they can retrieve, modify and delete these items. For this purpose, the VIS divides the vehicular network into separate, interacting segments each running the VIS Structured P2P Algorithm (VSPA). This article focuses on the VSPA's design, implementation and evaluation. The algorithm is based on features derived from existing structured P2P algorithms and extended to suit the specific characteristics of vehicular networks. Simulation studies show that the VSPA overlay state maintains consistent given long-run simulations. On inconsistencies, the overlay stabilizes rapidly. This is evaluated using an urban traffic environment implying few messages for network adjustments. Also, it is examined on a motorway given high vehicular density and high data load to investigate scalability. Finally, we show that the VSPA is able to rebuild an overlay in case of a network breakdown.