Peer-to-peer data structures for cooperative traffic information systems

• Published in: Pervasive and mobile computing Vol. 8; no. 2; pp. 194 - 209
• Main Authors: Rybicki, Jedrzej, Scheuermann, Björn, Mauve, Martin
• Format: Journal Article
• Language: English
• Published: Elsevier B.V 01.04.2012
• Subjects: Inter-vehicular communication; Peer-to-peer; Traffic information system; Traffic information system; Peer-to-peer; Inter-vehicular communication
• ISSN: 1574-1192; 1873-1589
• DOI: 10.1016/j.pmcj.2011.06.002

Cooperative traffic information systems support the driver of a car in selecting a route, based on traffic information collected by other cars. We propose to use a peer-to-peer network based on Internet access via cellular networks to distribute traffic information between the participants of such a system. This approach avoids the well-known limitations of VANET-based communication. Since the data maintained in a cooperative traffic information system has a very specific structure, it is particularly profitable—in terms of bandwidth consumption and latency—to tailor the system to this specific application domain instead of re-using generic peer-to-peer approaches. This realization led us to the development of GraphTIS—a peer-to-peer network specifically designed to manage traffic information. In this paper, we derive, step-by-step, the core mechanisms of GraphTIS, starting with a standard peer-to-peer system, outlining a first solution—named PeerTIS—which is based on a modification of this standard DHT, and then presenting GraphTIS, a novel peer-to-peer system that has been specifically designed to support traffic information systems. ► We examine a cellular network-based P2P system for sharing road traffic information. ► This approach avoids the capacity limitations of VANET communication. ► We show that for this application it is vital to preserve the structure of the data. ► We compare approaches based on geographical proximity and graph-based relations. ► Both approaches achieve low latencies, but differ in terms of the load distribution.