Decentralized Configuration Protocols for Low-Cost Offloading From Multiple Edges to Multiple Vehicular Fogs
Li‐Hsing Yen, Jui-Chung Hu, Ying–Dar Lin, Binayak Kar
Abstract
A vehicular-fog (VF) system as an emerging platform consists of electric vehicles with computing resources that are mostly under-utilized. This paper considers a two-tier federated Edge and Vehicular-Fog (EVF) system, where edge systems may partially offload user traffic to nearby VFs for potential cost reduction. Offloading configuration is to determine the ratios and targets of offloading traffic for maximal cost reduction, which is formulated as a mixed integer programming problem in this paper. We first present a decentralized offloading configuration protocol (DOCP) for an individual edge system to set up its own offloading configuration. We then propose a matching protocol among multiple edge systems to resolve resource contention when they simultaneously request resources from the same VF. Simulation results show that the proposed approach can leverage the heterogeneity of cost and capacity between edge systems and VFs. The proposed protocol outperforms greedy approaches by at most 40% and is comparable to a centralized off-line approach that is based on Particle Swarm Optimization.