A Fog-Assisted System to Defend Against Sybils in Vehicular Crowdsourcing
Technological advancements in vehicular transportation systems have led to the growth of novel paradigms, in which vehicles and infrastructures collaborate to infer high-level knowledge about phenomena of interest. Vehicular Social Network (VSN) is one such paradigm in which vehicular network entities are considered as part of an Online Social Network (OSN), paving the way for new services derived from social context. Although vehicular crowdsourcing has tremendous benefits, its deployment in real systems requires to solve important challenges including defense against Sybil attacks. This paper proposes a novel fog-assisted system that uses SybilDriver to minimize the presence of Sybil entities in VSN-based crowdsourcing applications. The proposed system exploits the characteristics of Vehicular Ad-hoc NETworks (VANETs) and OSNs to effectively recognize Sybils, and the adoption of fog computing helps reduce the overall network overhead by processing data closer to the vehicles. We perform detailed experiments on real-world publicly available datasets primarily to assess the effectiveness of SybilDriver against different Sybil attack strategies. Our experimental results show that SybilDriver detects Sybils with higher performance than state-of-the-art techniques under different settings. Furthermore, an evaluation of the fog architecture in terms of message complexity demonstrates low impact on the network overhead.
F. Concone et al., "A Fog-Assisted System to Defend Against Sybils in Vehicular Crowdsourcing," Pervasive and Mobile Computing, vol. 83, article no. 101612, Elsevier, Jul 2022.
The definitive version is available at https://doi.org/10.1016/j.pmcj.2022.101612
Keywords and Phrases
Crowdsourcing; Fog Computing; Sybil Detection; Truthfulness; Vehicular Social Network
International Standard Serial Number (ISSN)
Article - Journal
© 2022 Elsevier, All rights reserved.
01 Jul 2022
This work was supported by the National Science Foundation, Grant CNS-1818942.