Soft-VAN: Mobility-Aware Task Offloading in Software-Defined Vehicular Network

摘要

In this paper, we propose a mobility-aware task offloading scheme, named as Soft-VAN, with an aim to minimize task computation delay in a software-defined vehicular network. The proposed scheme consists of two phases - fog node selection and task offloading. In the first phase, we formulate an integer linear program (ILP), and solve the problem to get optimal number of fog nodes required for a given network. In the task offloading phase, we formulate an optimization problem to minimize overall delay in task computation, while considering associated constraints. As finding optimal solution to the problem is NP-hard, we propose a greedy heuristic approach in two phases - task offloading and computed task downloading - to solve it in polynomial time. The greedy solution for offloading takes into account network delay, flow-rule capacity, and link utilization. On the other hand, the solution for computed task downloading considers vehicle's mobility in addition to the parameters associated with the offloading decisions. Experimental results show that the proposed scheme, Soft-VAN, is capable of enhancing the performance approximately by 30%, 45%, and 50% in terms of delay compared to state-of-the-art schemes - Detour, DAGP, and SD2O, respectively.