OCTANE: A Joint Computation Offloading and Resource Allocation Scheme for MEC Assisted 5G NR Vehicular Networks

New vehicular applications like Augmented Reality (AR), Virtual Reality (VR), and High Definition Map (HD Map) have computational intensive and latency-sensitive traits and require collaboration among nearby vehicles. Computational offloading is used to improve the accuracy and performance of these applications, as it allows computational jobs to be processed on MEC servers at the cell edge. Here, the challenge is how to effectively take offloading decisions at the MEC server by considering wireless transmission delay and computational delay in the presence of time varying channel conditions due to vehicular mobility. In this work, we aim to maximize the number of jobs offloaded to the MEC server under application's deadline constraints while ensuring fairness among vehicles. First, we formulate the computational offloading as an integer linear programming (ILP) problem where both the transmission delay of 5G NR and MEC computational resources are taken into account. Then, we propose an online heuristic for joint computational offloading and resource allocation, OCTANE, that jointly takes 5G NR radio resources and computational resources into consideration while taking offloading decisions. Further, to provide fairness among vehicles, Transport Block Size (TBS) based Medium Access Layer (MAC) strategy is proposed for allocation of TDMA symbols in the 5G NR uplink. Finally, extensive simulations are performed in the NS-3 5G NR module with mobility traces taken from SUMO using OpenStreetMap to evaluate OCTANE and ILP model. Simulation results show that the proposed OCTANE scheme performs better than a state-of-the-art solution and is close to the ILP model in terms of offloading success rate.