Task Offloading and Collaborative Backhaul System based on Multi-level Edge Computing in the Internet of Vehicles

Main Article Content

Jin Lin
ZeQin Li
Ruofei Wang
RuYue Gong
HongJing Wu

Abstract

With the development of 5G and the Internet of Vehicles, diverse in-vehicle services continue to emerge. Computation-intensive and delay-sensitive in-vehicle tasks pose significant challenges to in-vehicle devices and represent one of the bottlenecks limiting the development of Internet of Vehicles technology. This paper proposes a Speed-Sensitive Offloading (SSO) and collaborative backhaul solution to address the problem of task offloading and result backhaul failure caused by vehicle movement, including a multi-level MEC architecture solution, speed sensitive task offloading and an MEC collaboration-based task return scheme (SSCOM). Through preliminary experimental verification, as the number of vehicles increases, the average task offloading time of all schemes shows an upward trend, but the SSCOM scheme has the smallest increase; compared with the schemes of random offloading, speed-prioritized and data-volume-prioritized offloading, the present scheme can significantly reduce the average task offloading time; collaborative backhaul can also solve the problem of result backhaul failure caused by vehicles driving out of the coverage area, etc., can improve the task backhaul success rate and MEC resource utilization rate by at least 5%. 

Article Details

Section
Special Issue - Efficient Scalable Computing based on IoT and Cloud Computing