Geocasting is the delivery of a message to nodes within a geographical region. With geocast, new services and applications are feasible, such as finding friends who are nearby, geographic advertising, and accident or wrong-way driver warning on a motorway. In this article we present a survey on geocast routing protocols. The protocols mainly differ in whether they are based on flooding, directed flooding, or on routing without flooding, and whether they are suitable for ad hoc networks or for infrastructure networks. Based on these criteria we propose a classification of geocast protocols. Our protocol comparison includes message and memory complexity, robustness, and the ability to deliver geocast packets in partially partitioned networks. Finally, we present simulations to compare the approaches based on flooding, directed flooding, and routing without flooding.

Inter-vehicle communication is regarded as one of the major applications of mobile ad hoc networks (MANETs). Compared to other MANETs, these so called vehicular ad hoc networks (VANETs) have special requirements in terms of node mobility and position-dependent applications, which are well met by geographic routing protocols. Functional research on geographic routing has already reached a considerable level, whereas security aspects have been vastly neglected so far. Since position dissemination is crucial for geographic routing, forged position information has severe impact regarding both performance and security.In order to lessen this problem, we propose a detection mechanism that is capable of recognizing nodes cheating about their position in beacons (periodic position dissemination in most single-path geographic routing protocols, e.g. GPSR). Unlike other proposals described in the literature, our detection does not rely on additional hardware or special nodes, which contradicts the ad hoc approach. Instead, this mechanism uses a number of different independent sensors to quickly give an estimation of the trustworthiness of other nodes' position claims without using dedicated infrastructure or specialized hardware.The simulative evaluation proves that our position verification system successfully discloses nodes disseminating false positions and thereby widely prevents attacks using position cheating.

SummaryInter-vehicle communication is regarded as one of the major applications of mobile ad hoc networks (MANETs). Compared to MANETs or wireless sensor networks (WSNs), these so-called vehicular ad hoc networks (VANETs) have unique requirements on network protocols. The requirements result mainly from node mobility and the demands of position-dependent applications. On the routing layer, those requirements are well met by geographic routing protocols. Functional research on geographic routing has already reached a considerable level, whereas security aspects have only been recently taken into account. Position information dissemination has been identified as being crucial for geographic routing since forged position information has severe impact regarding both performance and security.In this work, we first summarize the problems that arise from falsified position data. We then propose a framework that contains different detection mechanisms in order to mitigate or lessen these problems. Our developed mechanisms are capable of recognizing nodes cheating about their position in beacons (periodic position dissemination in most single-path geographic routing protocols, e.g., GPSR). Unlike other proposals described in the literature, our detection system does not rely on additional hardware or special nodes, which would contradict the ad hoc approach. Instead, we use a number of different independent sensors to quickly give an estimation of the trustworthiness of other nodes' position claims. The different sensors run either autonomously on every single node, or they require cooperation between neighboring nodes.The simulation evaluation proves that the combination of autonomous and cooperative position verification mechanisms successfully discloses most nodes disseminating false position information, and thereby widely prevents attacks using position cheating.

Abstract. There has been a lot of effort in the research on routing in mobile ad hoc networks in the last years. Promising applications of MANETs, e.g. in the automotive domain, are the drive for the design of inter-vehicle networks. So far, several projects in this field have chosen geographic routing approaches because of their outstanding performance and the possibility to support location-based applications like traffic warning functions. Having reached a reasonable functional level, a next step will be a deeper study of safety and security issues. With this paper, we dive into that area by assuming defective or malicious nodes that disseminate wrong position data. First, we have a look at the local problems that may arise from falsified position data, then we show the global effects on the routing performance by simulating malicious nodes. Simulation results show that the overall ratio of successfully delivered messages decreases, depending on the number of maliciously acting nodes, even up to approximately 30%. We conclude from this result that future work should take these threats into account in order to design more robust routing protocols.

No abstract