EDiPSo: An Efficient Scalable Topology Discovery Protocol for Software-Defined Vehicular Networks

Abstract

With the vast data loads produced by numerous applications and services, network softwarization became a vital addition to the design of intelligent heterogeneous vehicular networks. Smart Software-Defined Vehicular Networks (S-SDVN) are promising solutions that support various applications and protocols such as network mobility management, routing, traffic control and provides a centralized innovative decision-making protocol. SDVNs remove the limitations imposed by traditional hardware-based network devices to offer a global view of the networks’ status and give a standard framework for heterogeneous wireless technologies (WiMAX, WiFi/6, 5G, 6G). However, to capture the network’s topology and manage data traffic, an efficient discovery protocol is needed to mitigate the challenges under vehicles’ rapid mobility. In this article, we present an efficient and scalable topology discovery protocol for distributed software-defined vehicular networks, which we refer to as EDiPSo. The proposed protocol reduces the overhead and time complexity of the discovery process based on the centrality score measurements and vehicle traffic densities over switch-enabled roadside units. EDiPSo protocol is evaluated and compared to existing benchmark solutions using realistic mobility traces of Shenzhen City. The results showed that the proposed distributed discovery process significantly reduces the overhead compared to other protocols and decreases the time required to complete a discovery cycle as the number of locally distributed controllers increases.