Study of Time Constrained Task off loads for Traveling Vehicles in Multi-Edge Computing Environment
With the rapid advancement in vehicle communications and related applications, the computing resource shall be properly arranged for time constraint services due to the limited computing power of the vehicle. Although the cloud server can provide powerful computing resource, the cloud server is far away and it always results in excessively long transmission delay. The edge server, which is located beside the base station, is quite straightforward to be the computing resource for the offloading of the time critical tasks. And the execution result shall be transmitted back from the server after the task is completed. As the vehicle is moving, the execution results may need to be forward to the base station, which the vehicle is currently connected to, for the delivery to the vehicle. This paper studies the allocation of computing resource for task offloading in multiple edge-servers environment. Both the service migration, which means to transfer the residual work of the executing offloaded task from current server to the other server, and the initial task offloading are studied in the proposed scheme. The proposed adaptive migration scheme analyzes the computing resource of each edge server and effectively arrange the accepting of new offloaded task and the service migration of existing offloaded tasks to maximize the resource utilization. Exhaustive simulations were performed to investigate the performance of the proposed scheme. The simulations considered the moving vehicles with time critical tasks for offloading, and both balanced and unbalanced traffic models were applied. The simulation results demonstrate that the proposed scheme achieves lower blocking ratio of the offloading requests especially in non-uniform traffic condition.
Keywords - Internet of Vehicles, Task off loading, Service Migration, Edge Computing.