
Research Article
Joint Collaborative Task Offloading for Cost-Efficient Applications in Edge Computing
@INPROCEEDINGS{10.1007/978-3-030-41114-5_7, author={Chaochen Ma and Zhida Qin and Xiaoying Gan and Luoyi Fu}, title={Joint Collaborative Task Offloading for Cost-Efficient Applications in Edge Computing}, proceedings={Communications and Networking. 14th EAI International Conference, ChinaCom 2019, Shanghai, China, November 29 -- December 1, 2019, Proceedings, Part I}, proceedings_a={CHINACOM}, year={2020}, month={2}, keywords={Edge computing Task offloading Quality of service Cost-efficiency}, doi={10.1007/978-3-030-41114-5_7} }
- Chaochen Ma
Zhida Qin
Xiaoying Gan
Luoyi Fu
Year: 2020
Joint Collaborative Task Offloading for Cost-Efficient Applications in Edge Computing
CHINACOM
Springer
DOI: 10.1007/978-3-030-41114-5_7
Abstract
Edge computing is a new network model providing low-latency service with low bandwidth cost for the users by nearby edge servers. Due to the limited computational capacity of edge servers and devices, some edge servers need to offload some tasks to other servers in the edge network. Although offloading task to other edge servers may improve the service quality, the offloading process will be charged by the operator. In this paper, the goal is to determine the task offloading decisions of all the edge servers in the network. A model is designed with different types of cost in edge computing, where the overall cost of the system reflects the performance of the network. We formulate a cost minimization problem which is NP-hard. To solve the NP-hard problem, we propose a Joint Collaborative Task Offloading algorithm by adopting the optimization process in nearby edge servers. In our algorithm, an edge server can only offload its tasks to other edge servers within a neighborhood range. Based on the real-world data set, an adequate range is determined for the edge computing network. In cases of different density of tasks, the evaluations demonstrate that our algorithm has a good performance in term of overall cost, which outperforms an algorithm without considering the influence of neighborhood range.