Dynamic Virtual Network Embedding Leveraging Neighborhood and Preceding Mappings Information

Future transportation systems are primarily based on the concept of the Internet of Vehicles (IoV). However, to fully unleash the potential of IoV, Network Virtualization (NV) is regarded as one of the key enablers. With NV, heterogeneous service requests can be deployed quickly, cost-effectively, and on-demand on a shareable infrastructure to meet stringent resource requirements. Virtual Network Embedding (VNE), one of the main challenges in NV, has been extensively investigated in the data-center paradigm, in which the network topology is static. Some existing VNE solutions have tackled the VNE problem in data-center networks while considering IoV demands, but very few have directly solved the problem considering vehicle mobility. Therefore, solving the online VNE problem in dynamic IoV environments, where connected and moving vehicles function as physical nodes to handle network service requests, still remains at an early stage. Towards that end, this paper proposes a novel heuristic algorithm that efficiently ranks available moving vehicles based on multiple network attributes, their neighborhood information, and the correlation of preceding mappings to tackle the online VNE problem in IoV. Moreover, we investigated the performance of several VNE algorithms using the Random Waypoint mobility model on different sizes of the Substrate Network (SN). We also introduce additional performance metrics to demonstrate the impact of vehicle mobility. Extensive simulation results indicate that the proposed algorithm performs better than state-of-the-art VNE algorithms in multiple performance metrics.