Cost-Aware Node Ranking Algorithm for Embedding Virtual Networks in Internet of Vehicles

Internet of Vehicles (IoV), a subset of the Internet of Things (IoT), has been commonly considered as a primary paradigm for the anticipated success of the intelligent transportation. Network Virtualization (NV) enables flexible, cost-effective and on-demand services over the deployments of heterogeneous network service requests on a shared physical infrastructure. The most challenging problem of NV is Virtual Network Embedding (VNE) which involves embedding Virtual Network Requests (VNRs) into the substrate network efficiently and effectively, meeting several rigid resource constraints. In fact, the conventional VNE problem has been extensively investigated in the datacenter architecture in which the network topology is always fixed. Although recent studies have addressed the VNE problem considering IoV demands in datacenter networks, the development of VNE in IoV contexts, where connected and autonomous vehicles operate as substrate network nodes to handle incoming VNRs, is still in its early stages. This paper proposes a dual ranking-value and cost-aware heuristic algorithm, called CARA, for dealing with the online VNE problem in IoV. By considering vehicle mobility, our solution guarantees that the selected vehicles will remain within the preferable radius of the VNR while serving it. A thorough evaluation of our proposed VNE algorithm under the Random Waypoint (RWP) mobility model reveals that it accepts more than 40% VNRs while maintaining a drop-out ratio of almost zero and an execution time that is very practical.