Optimizing energy efficiency in MEC networks: a deep learning approach with Cybertwin-driven resource allocation

Umesh Kumar Lilhore, Sarita Simaiya, Surjeet Dalal, Neetu Faujdar, Roobaea Alroobaea, Majed Alsafyani, Abdullah M. Baqasah, Sultan Algarni
{"title":"Optimizing energy efficiency in MEC networks: a deep learning approach with Cybertwin-driven resource allocation","authors":"Umesh Kumar Lilhore, Sarita Simaiya, Surjeet Dalal, Neetu Faujdar, Roobaea Alroobaea, Majed Alsafyani, Abdullah M. Baqasah, Sultan Algarni","doi":"10.1186/s13677-024-00688-8","DOIUrl":null,"url":null,"abstract":"Cybertwin (CT) is an innovative network structure that digitally simulates humans and items in a virtual environment, significantly influencing Cybertwin instances more than regular VMs. Cybertwin-driven networks, combined with Mobile Edge Computing (MEC), provide practical options for transmitting IoT-enabled data. This research introduces a hybrid methodology integrating deep learning with Cybertwin-driven resource allocation to enhance energy-efficient workload offloading and resource management in MEC networks. Offloading work is essential in MEC networks since several applications require significant resources. The Cybertwin-driven approach considers user mobility, virtualization, processing power, load migrations, and resource demand as crucial elements in the decision-making process for offloading. The model optimizes job allocation between on-premises and distant execution using a task-offloading strategy to reduce the operating burden on the MEC network. The model uses a hybrid partitioning approach and a cost function to optimize resource allocation efficiently. This cost function accounts for energy consumption and service delays associated with job assignment, execution, and fulfilment. The model calculates the cost of several segmentation and offloading procedures and chooses the lowest cost to enhance energy efficiency and performance. The approach employs a deep learning architecture called “CNN-LSTM-TL” to accomplish energy-efficient task offloading, utilizing pre-trained transfer learning models. Batch normalization is used to speed up model training and improve its robustness. The model is trained and assessed using an extensive mobile edge computing public dataset. The experimental findings confirm the efficacy of the proposed methodology, indicating a 20% decrease in energy usage compared to conventional methods while achieving comparable or superior performance levels. Simulation studies emphasize the advantages of incorporating Cybertwin-driven insights into resource allocation and workload-offloading techniques. This research enhances energy-efficient and resource-aware MEC networks by incorporating Cybertwin-driven techniques.","PeriodicalId":501257,"journal":{"name":"Journal of Cloud Computing","volume":null,"pages":null},"PeriodicalIF":0.0000,"publicationDate":"2024-08-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Journal of Cloud Computing","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1186/s13677-024-00688-8","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
引用次数: 0

Abstract

Cybertwin (CT) is an innovative network structure that digitally simulates humans and items in a virtual environment, significantly influencing Cybertwin instances more than regular VMs. Cybertwin-driven networks, combined with Mobile Edge Computing (MEC), provide practical options for transmitting IoT-enabled data. This research introduces a hybrid methodology integrating deep learning with Cybertwin-driven resource allocation to enhance energy-efficient workload offloading and resource management in MEC networks. Offloading work is essential in MEC networks since several applications require significant resources. The Cybertwin-driven approach considers user mobility, virtualization, processing power, load migrations, and resource demand as crucial elements in the decision-making process for offloading. The model optimizes job allocation between on-premises and distant execution using a task-offloading strategy to reduce the operating burden on the MEC network. The model uses a hybrid partitioning approach and a cost function to optimize resource allocation efficiently. This cost function accounts for energy consumption and service delays associated with job assignment, execution, and fulfilment. The model calculates the cost of several segmentation and offloading procedures and chooses the lowest cost to enhance energy efficiency and performance. The approach employs a deep learning architecture called “CNN-LSTM-TL” to accomplish energy-efficient task offloading, utilizing pre-trained transfer learning models. Batch normalization is used to speed up model training and improve its robustness. The model is trained and assessed using an extensive mobile edge computing public dataset. The experimental findings confirm the efficacy of the proposed methodology, indicating a 20% decrease in energy usage compared to conventional methods while achieving comparable or superior performance levels. Simulation studies emphasize the advantages of incorporating Cybertwin-driven insights into resource allocation and workload-offloading techniques. This research enhances energy-efficient and resource-aware MEC networks by incorporating Cybertwin-driven techniques.
优化 MEC 网络的能效:采用赛博特文驱动的资源分配的深度学习方法
Cybertwin(CT)是一种创新的网络结构,它以数字方式模拟虚拟环境中的人和物品,对 Cybertwin 实例的影响远远超过普通虚拟机。Cybertwin驱动的网络与移动边缘计算(MEC)相结合,为传输物联网数据提供了实用的选择。本研究介绍了一种混合方法,该方法将深度学习与赛伯特云驱动的资源分配相结合,以增强 MEC 网络中的节能工作负载卸载和资源管理。卸载工作在 MEC 网络中至关重要,因为一些应用需要大量资源。Cybertwin驱动方法将用户移动性、虚拟化、处理能力、负载迁移和资源需求作为卸载决策过程中的关键要素。该模型采用任务卸载策略,在本地执行和远程执行之间优化任务分配,以减轻 MEC 网络的运行负担。该模型采用混合分区方法和成本函数来有效优化资源分配。该成本函数考虑了能耗以及与任务分配、执行和完成相关的服务延迟。该模型会计算多个分区和卸载程序的成本,并选择成本最低的程序来提高能效和性能。该方法采用了一种名为 "CNN-LSTM-TL "的深度学习架构,利用预先训练好的迁移学习模型来完成高能效的任务卸载。批量规范化用于加快模型训练并提高其鲁棒性。利用广泛的移动边缘计算公共数据集对该模型进行了训练和评估。实验结果证实了所提方法的有效性,表明与传统方法相比,该方法的能耗降低了 20%,同时达到了相当或更优的性能水平。仿真研究强调了将赛博特资讯驱动的见解纳入资源分配和工作负载卸载技术的优势。这项研究通过采用赛伯特赢驱动技术,提高了 MEC 网络的能效和资源感知能力。
本文章由计算机程序翻译,如有差异,请以英文原文为准。
求助全文
约1分钟内获得全文 求助全文
来源期刊
自引率
0.00%
发文量
0
×
引用
GB/T 7714-2015
复制
MLA
复制
APA
复制
导出至
BibTeX EndNote RefMan NoteFirst NoteExpress
×
提示
您的信息不完整,为了账户安全,请先补充。
现在去补充
×
提示
您因"违规操作"
具体请查看互助需知
我知道了
×
提示
确定
请完成安全验证×
copy
已复制链接
快去分享给好友吧!
我知道了
右上角分享
点击右上角分享
0
联系我们:info@booksci.cn Book学术提供免费学术资源搜索服务,方便国内外学者检索中英文文献。致力于提供最便捷和优质的服务体验。 Copyright © 2023 布克学术 All rights reserved.
京ICP备2023020795号-1
ghs 京公网安备 11010802042870号
Book学术文献互助
Book学术文献互助群
群 号:481959085
Book学术官方微信