This paper proposed the Predicted Utility-centric Path Switching (PUPS) method for k-hop Vehicle-To-Vehicle (V2V) data offloading based on the Software Defined Network (SDN) Controller inside the Multi-Access Edge Computing (MEC) server's architecture. Let two moving vehicles Va and Vb regularly use Base Stations (BSs) of 4G/5G cellular network to communicate with each other and periodically report their contexts to the MEC server. The SDN Controller inside the MEC server can calculate and predict the k-hop offloading paths that are able to exist potentially for Va and Vb based on the reported contexts that are periodically received from all of the vehicles administered by the MEC server, i.e., the control plane of the k-hop V2V offloading is handled by the SDN Controller inside the MEC server. When a feasible k-hop V2V path can be found and used for offloading the communicated data between Va and Vb from 4G/5G cellular network, the data plane, i.e., forwarding data between Va and Vb, of the k-hop V2V offloading is handled by the constituent vehicles of the k-hop V2V offloading path. Additionally, after the V2V offloading session is launched, the MEC server can notify Va and Vb to switch to the k-hop V2V offloading path X having the better utility than the currently used one on the time point that path X was predicted to exist. The performance evaluation results indicated that the proposed method has the higher data offloading fraction, the lower data loss rate and the longer offloading service time in the situations of the longer limit of hop count and the low/medium/high vehicle density.
All Science Journal Classification (ASJC) codes
- Automotive Engineering
- Electrical and Electronic Engineering