Research Article
D2D-Based Multi-relay-Assisted Computation Offloading in Edge Computing Network
@INPROCEEDINGS{10.1007/978-3-030-92638-0_3, author={Xuan Zhao and Song Zhang and Bowen Liu and Xutong Jiang and Wanchun Dou}, title={D2D-Based Multi-relay-Assisted Computation Offloading in Edge Computing Network}, proceedings={Collaborative Computing: Networking, Applications and Worksharing. 17th EAI International Conference, CollaborateCom 2021, Virtual Event, October 16-18, 2021, Proceedings, Part II}, proceedings_a={COLLABORATECOM PART 2}, year={2022}, month={1}, keywords={Compute-intensive task Computation offloading Edge computing D2D Multi-relay}, doi={10.1007/978-3-030-92638-0_3} }- Xuan Zhao
Song Zhang
Bowen Liu
Xutong Jiang
Wanchun Dou
Year: 2022
D2D-Based Multi-relay-Assisted Computation Offloading in Edge Computing Network
COLLABORATECOM PART 2
Springer
DOI: 10.1007/978-3-030-92638-0_3
Abstract
With the rapid development of edge computing, a large number of compute-intensive tasks are offloaded to the edge server, and computation offloading strategy has become a hot research topic. Because the deployment of edge servers is still in its infancy, there would be some service blind area. Through D2D technology, mobile devices in the blind area can obtain edge services with the help of relays. However, the existing methods usually select single relay to transmit tasks, and seldom consider the sensitivity of processing delay for compute-intensive tasks. When the available bandwidth of single relay is not enough to complete the data transmission in a limited time, it will produce a large delay, which seriously affects the quality of user experience. In view of this challenge, a D2D-based multi-relay-assisted computation offloading method is proposed. Technically, multiple mobile devices in the available area of edge service are used as the relays. Based on D2D technology, mobile devices in the blind area use the relays to transmit the computing task to the edge server to improve the application computing efficiency. A large number of experiments with real-world datasets have proved the feasibility and effectiveness of our method.
