Preliminary Design of Hybrid Simulation Framework for Workload Analysis in Automotive Edge Computing

Many vehicular tasks shall be completed in a timely manner as they move. In automotive edge computing, a vehicle can offload such tasks to an edge server in proximity to process them faster than using the cloud computing model. However, such an edge server has limited communication coverage and relatively smaller computing power compared to the general cloud server, it is important to distribute the offloaded workloads appropriately. One needs to understand such workload patterns and quantitatively analyze them for such workload distribution. We propose a preliminary hybrid simulation framework that generates compute workloads via a combination of the virtual vehicular traffic and the physical edge-server platforms.