Chaoxian Zhang

Chaoxian Zhang

Auto-generated profile, based on publicly available information

Chaoxian Zhang

School of Information Science and Engineering, Xiamen University Tan Kah Kee College, Xiamen, China

Research Focus & Metrics

Academic Career Length

6Years (2021-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
24 March 2021 in IEEE Internet of Things Journal
Verified
Journal Article
Closed access
YGYixiao Gu
Yixiao Gu
YYYao Yao
Yao Yao
CLCheng Li
Cheng Li
BXBin Xia
Bin Xia
DXDingjie Xu
Dingjie Xu
CZChaoxian Zhang
Chaoxian Zhang
To realize the vision of the Internet of Things (IoT), mobile-edge computing (MEC) has recently emerged as a promising paradigm to meet the computation demand from mobile users (MUs). In this article, we study the network performance in large-scale stochastic MEC wireless networks, where the tasks can be computed locally by the local computation capabilities (LCCs) or be offloaded to MEC servers for edge computing. To this end, a MEC network is modeled featuring random node distribution, dynamic task requests, orthogonal frequency-division multiple access, task retransmission, and parallel computing in MEC servers. Given the model, a 2-D discrete-time Markov chain is first adopted to characterize the task execution process, including local computing and task offloading. Based on the coupling between communication and computing, the average outage probability of the task transmission and the average MEC computation load are derived by integrating the stochastic geometry and queuing theory. Furthermore, by jointly analyzing the local computation latency, transmission latency, and edge computation latency, we derive the average end-to-end latency of the task execution. Our results show that the LCCs in MUs can improve the network performance, including communication and computation performance, in stochastic MEC networks. In addition, useful guidelines for MEC network provisioning and planning are provided to avoid either the local computing or the task offloading being the latency performance bottleneck.