An edge server placement based on graph clustering in mobile edge computing

Abstract With the exponential growth of mobile devices and data traffic, mobile edge computing has become a promising technology, and the placement of edge servers plays a key role in providing efficient and low-latency services. In this paper, we investigate the issue of edge server placement and u...

Full description

Saved in:
Bibliographic Details
Main Authors: Shanshan Zhang, Jiong Yu, Mingjian Hu
Format: Article
Language:English
Published: Nature Portfolio 2024-12-01
Series:Scientific Reports
Subjects:
Online Access:https://doi.org/10.1038/s41598-024-81684-5
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1846137082396803072
author Shanshan Zhang
Jiong Yu
Mingjian Hu
author_facet Shanshan Zhang
Jiong Yu
Mingjian Hu
author_sort Shanshan Zhang
collection DOAJ
description Abstract With the exponential growth of mobile devices and data traffic, mobile edge computing has become a promising technology, and the placement of edge servers plays a key role in providing efficient and low-latency services. In this paper, we investigate the issue of edge server placement and user allocation to reduce transmission delay between base stations and servers, and balance the workload of individual servers. To this end, we propose a graph clustering-based edge server placement model by fully considering the constraints such as the distance, coverage area and number of channels of base stations. The model mainly consists of a two-layer graph convolutional network (GCN) component and a differentiable version of K-means clustering component, which transforms the server placement problem into an end-to-end learning optimization problem on a graph. It trains the GCN network to achieve the best clustering results with the expectation of average delay and load balancing as the loss function to obtain the edge server placement and user assignment scheme. We conducted experiments based on the Shanghai Telecom dataset, and the results show the effectiveness of our approach in both latency reduction and load balancing.
format Article
id doaj-art-1f8d6e72af5f4d8db0619c7b6009e9b6
institution Kabale University
issn 2045-2322
language English
publishDate 2024-12-01
publisher Nature Portfolio
record_format Article
series Scientific Reports
spelling doaj-art-1f8d6e72af5f4d8db0619c7b6009e9b62024-12-08T12:29:22ZengNature PortfolioScientific Reports2045-23222024-12-0114111910.1038/s41598-024-81684-5An edge server placement based on graph clustering in mobile edge computingShanshan Zhang0Jiong Yu1Mingjian Hu2School of Computer Science and Technology, Xinjiang UniversitySchool of Computer Science and Technology, Xinjiang UniversityXinjiang Petroleum Engineering Co., LtdAbstract With the exponential growth of mobile devices and data traffic, mobile edge computing has become a promising technology, and the placement of edge servers plays a key role in providing efficient and low-latency services. In this paper, we investigate the issue of edge server placement and user allocation to reduce transmission delay between base stations and servers, and balance the workload of individual servers. To this end, we propose a graph clustering-based edge server placement model by fully considering the constraints such as the distance, coverage area and number of channels of base stations. The model mainly consists of a two-layer graph convolutional network (GCN) component and a differentiable version of K-means clustering component, which transforms the server placement problem into an end-to-end learning optimization problem on a graph. It trains the GCN network to achieve the best clustering results with the expectation of average delay and load balancing as the loss function to obtain the edge server placement and user assignment scheme. We conducted experiments based on the Shanghai Telecom dataset, and the results show the effectiveness of our approach in both latency reduction and load balancing.https://doi.org/10.1038/s41598-024-81684-5Edge server placementMobile edge computingGraph clusteringGraph convolutional network
spellingShingle Shanshan Zhang
Jiong Yu
Mingjian Hu
An edge server placement based on graph clustering in mobile edge computing
Scientific Reports
Edge server placement
Mobile edge computing
Graph clustering
Graph convolutional network
title An edge server placement based on graph clustering in mobile edge computing
title_full An edge server placement based on graph clustering in mobile edge computing
title_fullStr An edge server placement based on graph clustering in mobile edge computing
title_full_unstemmed An edge server placement based on graph clustering in mobile edge computing
title_short An edge server placement based on graph clustering in mobile edge computing
title_sort edge server placement based on graph clustering in mobile edge computing
topic Edge server placement
Mobile edge computing
Graph clustering
Graph convolutional network
url https://doi.org/10.1038/s41598-024-81684-5
work_keys_str_mv AT shanshanzhang anedgeserverplacementbasedongraphclusteringinmobileedgecomputing
AT jiongyu anedgeserverplacementbasedongraphclusteringinmobileedgecomputing
AT mingjianhu anedgeserverplacementbasedongraphclusteringinmobileedgecomputing
AT shanshanzhang edgeserverplacementbasedongraphclusteringinmobileedgecomputing
AT jiongyu edgeserverplacementbasedongraphclusteringinmobileedgecomputing
AT mingjianhu edgeserverplacementbasedongraphclusteringinmobileedgecomputing