A blockchain-based framework to secure vehicular social networks

Abstract

Vehicular social network is emerging as a new promising concept, combin-ing two types of network paradigms, namely, vehicular networks and socialnetworks. In order to manage efficiently the security and the control of the net-work, this paper proposes a new framework based on the emerging concepts ofsoftware-defined vehicular network (SDVN) and blockchain. Using the SDVNmakes the network more programmable, virtualized, and partitionable. How-ever, on the other hand, it also creates a well-known vulnerability of a singlepoint of failure. Hence, we propose to introduce the blockchain paradigm thatwill enable the certification of transactions and ensure data anonymity in afully distributed manner. To this end, three levels of controllers are needed: aprincipal controller (PC), roadside units (RSUs), and a local controller. In orderto dynamically select miners, a distributed miners connected dominating setalgorithm (DM-CDS) has been proposed. The DM-CDS is a single-phase dis-tributed algorithm that supports a dynamic topology based on a trust model andsome other network parameters, such as the connectivity degree, the averagelink quality indicator, and the rank. The performance of the proposed DM-CDSis evaluated throughout multiple scenarios using different parameters, such astrust metric, node density, node mobility, and radio range. The obtained resultshighlight the importance of such proposed architecture, especially in terms ofnumber of required miners. For instance, when the density of nodes increases,the number of selected miners increases similarly to when the network lengthincreases. The node mobility impacts also on the stability of the selected miners,in terms of withdrawing and joining, showing a variation between 0% and 10%.The trust metric has also an important impact on the selection of miners, as onlynodes with a higher trust level are selected to endorse the roles of miners.