Computer Networks最新文献

筛选
英文 中文
Performance modeling and comparison of URLLC and eMBB coexistence strategies in 5G new radio systems 5G 新无线电系统中 URLLC 和 eMBB 共存策略的性能建模与比较
IF 4.4 2区 计算机科学
Computer Networks Pub Date : 2024-11-14 DOI: 10.1016/j.comnet.2024.110904
Daria Ivanova , Elena Zhbankova , Ekaterina Markova , Yuliya Gaidamaka , Konstantin Samouylov
{"title":"Performance modeling and comparison of URLLC and eMBB coexistence strategies in 5G new radio systems","authors":"Daria Ivanova ,&nbsp;Elena Zhbankova ,&nbsp;Ekaterina Markova ,&nbsp;Yuliya Gaidamaka ,&nbsp;Konstantin Samouylov","doi":"10.1016/j.comnet.2024.110904","DOIUrl":"10.1016/j.comnet.2024.110904","url":null,"abstract":"<div><div>Industrial deployments of 5G New Radio (NR) communications are expected to be the first ones, where time-critical communications implemented via ultra-reliable low-latency service (URLLC) will coexist with enhanced mobile broadband (eMBB) service on the same radio interface. However, owing to the demanding nature of the former service and principally different traffic patterns, the conventional coexistence techniques based on a complete reservation or complete sharing either result in inefficient usage of radio resources or inability to deliver the required service level. Motivated by the lack of comparison-based studies for URLLC and eMBB coexistence at the air interface that does not affect the current 5G NR multiple access scheme, in this paper, we compare several resource allocation strategies for URLLC and eMBB traffic. To this aim, by utilizing directional antenna array and propagation models, we formulate a general performance assessment framework for a partial sharing coexistence strategy with preemptive and non-preemptive priorities. As this strategy includes several coexistence schemes as its special cases, we then proceed to assess their performance utilizing user- and operator-centric key performance indicators including the session drop and preemption probabilities, and system resource utilization. Our numerical results indicate that utilizing preemption service is vital for efficient coexistence between URLLC and eMBB traffic. Out of all the considered schemes, one of the simplest ones — preemptive priority without resource reservation is characterized by the minimum resource requirements. This scheme allows to achieve 50% gain in terms of the amount of required resources as compared to the strategies without preemption or with complete reservations.</div></div>","PeriodicalId":50637,"journal":{"name":"Computer Networks","volume":"255 ","pages":"Article 110904"},"PeriodicalIF":4.4,"publicationDate":"2024-11-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142662330","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
引用次数: 0
Integrating Unmanned Aerial Vehicles (UAVs) with Vehicular Ad-hoc NETworks (VANETs): Architectures, applications, opportunities 将无人驾驶飞行器 (UAV) 与车载 Ad-hoc 网络 (VANET) 相结合:架构、应用和机遇
IF 4.4 2区 计算机科学
Computer Networks Pub Date : 2024-11-13 DOI: 10.1016/j.comnet.2024.110873
Muhammad Mansoor Ashraf , Saadi Boudjit , Sherali Zeadally , Nour El Houda Bahloul , Nouman Bashir
{"title":"Integrating Unmanned Aerial Vehicles (UAVs) with Vehicular Ad-hoc NETworks (VANETs): Architectures, applications, opportunities","authors":"Muhammad Mansoor Ashraf ,&nbsp;Saadi Boudjit ,&nbsp;Sherali Zeadally ,&nbsp;Nour El Houda Bahloul ,&nbsp;Nouman Bashir","doi":"10.1016/j.comnet.2024.110873","DOIUrl":"10.1016/j.comnet.2024.110873","url":null,"abstract":"<div><div>Unmanned Aerial Vehicles (UAVs) and Vehicular Ad Hoc NETworks (VANETs) have evolved significantly in the last decade, providing a wide range of advanced options for transportation networks. Despite these advancements, both networks may suffer from frequent link failures due to their dynamic nature, variable network density, and limited transmission range. To address these challenges, researchers have recently shifted their focus on how UAV–VANET integration can improve the functionality and performance of both UAVs and VANETs in Intelligent Transportation Systems (ITS). However, there is a lack of comprehensive study that thoroughly examines the existing literature and identifies the open challenges of an integrated network of UAVs and VANETs In this survey, first, we discuss the advantages and applications of the integrated network and identify communication layer challenges that must be addressed for seamless, efficient, and robust UAV–VANET integration. Secondly, we present a comparative analysis of the recent advancements in UAV–VANET integration solutions and discuss their advantages and limitations. To the best of our knowledge, this is the first comprehensive study that reviews the various recent integration approaches of UAVs and VANETs. This study will help researchers working on the development of future UAV–VANET architectures, applications, and protocols.</div></div>","PeriodicalId":50637,"journal":{"name":"Computer Networks","volume":"255 ","pages":"Article 110873"},"PeriodicalIF":4.4,"publicationDate":"2024-11-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142662350","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
引用次数: 0
Deep reinforcement learning for autonomous SideLink radio resource management in platoon-based C-V2X networks: An overview 基于排的 C-V2X 网络中自主 SideLink 无线电资源管理的深度强化学习:概述
IF 4.4 2区 计算机科学
Computer Networks Pub Date : 2024-11-12 DOI: 10.1016/j.comnet.2024.110901
Nessrine Trabelsi , Lamia Chaari Fourati , Wael Jaafar
{"title":"Deep reinforcement learning for autonomous SideLink radio resource management in platoon-based C-V2X networks: An overview","authors":"Nessrine Trabelsi ,&nbsp;Lamia Chaari Fourati ,&nbsp;Wael Jaafar","doi":"10.1016/j.comnet.2024.110901","DOIUrl":"10.1016/j.comnet.2024.110901","url":null,"abstract":"<div><div>Dynamic and autonomous SideLink (SL) Radio Resource Management (RRM) is essential for platoon-based cellular vehicular networks. However, this task is challenging due to several factors. These include the limited spectrum below 6 GHz, stringent vehicle-to-everything (V2X) communications requirements, uncertain and dynamic environments, limited vehicle sensing capabilities, and inherent distributed operation. These limitations often lead to resource collisions, data packet loss, and increased latency. Current standardized approaches in Long-Term Evolution-V2X (LTE-V2X) and New Radio-V2X (NR-V2X) rely on random resource selection, limiting their efficiency. Moreover, RRM is inherently a complex combinatorial optimization problem. It may involve conflicting objectives and constraints, making traditional approaches inadequate. Platoon-based communication necessitates careful resource allocation to support a diverse mix of communication types. These include safety–critical control messaging within platoons, less time-sensitive traffic management information between platoons, and even infotainment services like media streaming. Optimizing resource sharing inter- and intra-platoons is crucial to avoid excessive interference and ensure overall network performance. Deep Reinforcement Learning (DRL), combining Deep Learning (DL) and Reinforcement Learning (RL), has recently been investigated for network resource management. It offers a potential solution for these challenges. A DRL agent, represented by deep neural networks, interacts with the environment and learns optimal decision-making through trial and error. This paper overviews proposed DRL-based methods for autonomous SL RRM in single and multi-agent platoon-based C-V2X networks. It considers both intra- and inter-platoon communications with their specific requirements. We discuss the components of Markov Decision Processes (MDP) used to model the sequential decision-making of RRM. We then detail the DRL algorithms, training paradigms, and insights on the achieved results. Finally, we highlight challenges in existing works and suggest strategies for addressing them.</div></div>","PeriodicalId":50637,"journal":{"name":"Computer Networks","volume":"255 ","pages":"Article 110901"},"PeriodicalIF":4.4,"publicationDate":"2024-11-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142662351","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
引用次数: 0
Privacy-preserving local clustering coefficient query on structured encrypted graphs 结构加密图上的隐私保护局部聚类系数查询
IF 4.4 2区 计算机科学
Computer Networks Pub Date : 2024-11-10 DOI: 10.1016/j.comnet.2024.110895
Yingying Pan , Lanxiang Chen , Gaolin Chen
{"title":"Privacy-preserving local clustering coefficient query on structured encrypted graphs","authors":"Yingying Pan ,&nbsp;Lanxiang Chen ,&nbsp;Gaolin Chen","doi":"10.1016/j.comnet.2024.110895","DOIUrl":"10.1016/j.comnet.2024.110895","url":null,"abstract":"<div><div>Graphs and graph databases serve as the fundamental building blocks for various network structures. In real-world network scenarios, nodes often aggregate due to their approximate organizational associations with each other. The local clustering coefficient, which evaluates the proximity of nodes within a graph, plays an important role in quantifying the structural properties of graphs in scrutinizing network robustness and understanding its intricate dynamics. Despite the growing popularity of easily accessible cloud services among small and medium-sized enterprises as well as individuals, the potential risk of data privacy disclosure when outsourcing large graphs to third-party servers is increasing. It is vital to explore a technique for executing queries on encrypted graph data. In this paper, we propose a <u>st</u>ructured <u>e</u>ncryption scheme to achieve privacy-preserving local <u>c</u>lustering <u>c</u>oefficient query (<span><math><mrow><mi>STE</mi><mtext>-</mtext><mi>CC</mi></mrow></math></span>) on the outsourced encrypted graphs. To calculate the clustering coefficient, we design the <span><math><msub><mrow><mi>PSI</mi></mrow><mrow><mi>sum</mi></mrow></msub></math></span> protocol to sum the number of intersections, in which the basic private set intersection (PSI) protocol combines Bloom filter (BF) and garbled Bloom filter (GBF) to perform the private matching for counting the number of common neighbors. When configured with appropriate parameters, it can achieve no false negatives and negligible false positives. Finally, the security analysis and experimental evaluation on real-world graph data substantiate the effectiveness and efficiency of our approach.</div></div>","PeriodicalId":50637,"journal":{"name":"Computer Networks","volume":"255 ","pages":"Article 110895"},"PeriodicalIF":4.4,"publicationDate":"2024-11-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142662354","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
引用次数: 0
Robust and energy-efficient RPL optimization algorithm with scalable deep reinforcement learning for IIoT 采用可扩展深度强化学习的鲁棒节能 RPL 优化算法,适用于物联网
IF 4.4 2区 计算机科学
Computer Networks Pub Date : 2024-11-10 DOI: 10.1016/j.comnet.2024.110894
Ying Wang, Yuanyuan Li, Jianjun Lei, Fengjun Shang
{"title":"Robust and energy-efficient RPL optimization algorithm with scalable deep reinforcement learning for IIoT","authors":"Ying Wang,&nbsp;Yuanyuan Li,&nbsp;Jianjun Lei,&nbsp;Fengjun Shang","doi":"10.1016/j.comnet.2024.110894","DOIUrl":"10.1016/j.comnet.2024.110894","url":null,"abstract":"<div><div>The increasing complexity and quantity of the Industrial Internet of Things (IIoT) pose new challenges to the traditional routing protocol for low-power and lossy networks (RPL) in terms of dynamic management, data transmission reliability, and energy efficiency optimization. This paper proposes a scalable deep reinforcement learning (DRL) algorithm with a multi-attention actor double critic model for routing optimization (MADC) to meet the requirements of IIoT for efficient and intelligent routing decisions while improving data transmission reliability and energy efficiency. Specifically, MADC employs the centralized training and decentralized execution (CTDE) learning paradigm to decouple the model’s training and inference tasks, which reduces the difficulty and computational cost of model learning and improves the training efficiency. In addition, a lightweight actor network based on multi-scale convolutional attention mechanism is designed in MADC, which can provide intelligent and real-time decision-making capabilities for resource-constrained nodes with low computational and storage complexities. Moreover, a scalable critic network utilizing multiple attention mechanisms is proposed. It is not only suitable for dynamic and changing network environments but also can more comprehensively and accurately evaluate local observation states, providing more accurate and efficient guidance for model optimization. Furthermore, MADC incorporates a double critic network architecture to mitigate potential overestimation issues during training, thereby ensuring the model’s robustness and reliability. Simulation results demonstrate that MADC outperforms existing RPL optimization algorithms in terms of energy efficiency, data transmission reliability, and adaptability.</div></div>","PeriodicalId":50637,"journal":{"name":"Computer Networks","volume":"255 ","pages":"Article 110894"},"PeriodicalIF":4.4,"publicationDate":"2024-11-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142662328","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
引用次数: 0
Secure fuzzy retrieval protocol for multiple datasets 多数据集安全模糊检索协议
IF 4.4 2区 计算机科学
Computer Networks Pub Date : 2024-11-09 DOI: 10.1016/j.comnet.2024.110891
Jie Zhou , Jiao Deng , Shengke Zeng , Mingxing He , Xingwei Liu
{"title":"Secure fuzzy retrieval protocol for multiple datasets","authors":"Jie Zhou ,&nbsp;Jiao Deng ,&nbsp;Shengke Zeng ,&nbsp;Mingxing He ,&nbsp;Xingwei Liu","doi":"10.1016/j.comnet.2024.110891","DOIUrl":"10.1016/j.comnet.2024.110891","url":null,"abstract":"<div><div>With the diversification of data sources and the massive growth of datasets, data retrieval has become increasingly complex and time-consuming. In the traditional retrieval method, if a user wants to query multiple datasets, the general approach is to retrieve them one by one in order, which may lead to duplication of work and waste of resources. Private set intersection is a specific issue in secure multi-party computation. It allows several participants, each holding different sets, to jointly calculate the intersection of their sets without revealing any information other than the intersection. This method is naturally suitable for data fusion. In this work, we propose a secure fuzzy retrieval protocol for multiple datasets. First, we use private set intersection technology to fuse multiple datasets. Then, we perform secure retrieval based on this fused dataset, effectively avoiding the waste of resources caused by separate retrievals, thereby maximizing resource efficiency. It is worth mentioning that the protocol proposed in this paper can also be used for fuzzy retrieval to improve the user’s search experience. More importantly, the protocol can maximize privacy protection during the retrieval process, including strict protection of sensitive information such as retrieval keywords, ensuring that user data and query intentions will not be leaked during the entire retrieval process. Finally, we provide a rigorous security proof and demonstrate the effectiveness of the protocol through simulation experiments.</div></div>","PeriodicalId":50637,"journal":{"name":"Computer Networks","volume":"255 ","pages":"Article 110891"},"PeriodicalIF":4.4,"publicationDate":"2024-11-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142662329","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
引用次数: 0
Complex-network based model for SMS spam filtering 基于复杂网络的垃圾短信过滤模型
IF 4.4 2区 计算机科学
Computer Networks Pub Date : 2024-11-09 DOI: 10.1016/j.comnet.2024.110892
Shaghayegh Hosseinpour, Hadi Shakibian
{"title":"Complex-network based model for SMS spam filtering","authors":"Shaghayegh Hosseinpour,&nbsp;Hadi Shakibian","doi":"10.1016/j.comnet.2024.110892","DOIUrl":"10.1016/j.comnet.2024.110892","url":null,"abstract":"<div><div>With the advancement of technology and the widespread use of mobile phones and wireless communication, SMS has become the most popular texting method due to its high response rate, affordability, and no internet connection requirement. A survey found that 3.5 billion users, or 80% of active users worldwide, use SMS for communication. SMS, however, has also attracted spammers, resulting in an explosion in spam messages, especially in Asia. Users are annoyed, lose money, and waste their time by receiving spam messages intended to serve various purposes, such as advertising, adult content, smishing, and fraud. Spam messages are a problem for users and providers, which calls for a mechanism to identify and filter them out. With supervised machine learning techniques, we propose a novel approach to classify spam and ham messages based on complex network theory. The proposed approach integrates complex network based features with statistical TF-IDF and grammatical rules features. Also, an under-sampling method has been employed in order to cope with the imbalanced data issue. We evaluated the performance of several supervised learners in terms of accuracy, precision, recall, F1-score, and AUC. In our experiments, Random Forest successfully classified spam messages more accurate than statistical methods that only extracted TF-IDF features.</div></div>","PeriodicalId":50637,"journal":{"name":"Computer Networks","volume":"255 ","pages":"Article 110892"},"PeriodicalIF":4.4,"publicationDate":"2024-11-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142662355","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
引用次数: 0
SARM: A network State-Aware Adaptive Routing Mutation method for power IoT SARM:面向电力物联网的网络状态感知自适应路由突变方法
IF 4.4 2区 计算机科学
Computer Networks Pub Date : 2024-11-09 DOI: 10.1016/j.comnet.2024.110889
Tianshuai Zheng , Jinglei Tan , Xuesong Wu , Ruiqin Hu , Qifang Chen , Zhiquan Liu , Ye Du
{"title":"SARM: A network State-Aware Adaptive Routing Mutation method for power IoT","authors":"Tianshuai Zheng ,&nbsp;Jinglei Tan ,&nbsp;Xuesong Wu ,&nbsp;Ruiqin Hu ,&nbsp;Qifang Chen ,&nbsp;Zhiquan Liu ,&nbsp;Ye Du","doi":"10.1016/j.comnet.2024.110889","DOIUrl":"10.1016/j.comnet.2024.110889","url":null,"abstract":"<div><div>With the rapid development of Power Internet of Things (PIoT), the application of Internet of Things technology in smart grids is becoming increasingly widespread, but it also enlarges the attack surface of the system. Against this backdrop, traditional defense methods are limited by the asymmetry of network attack and defense, which makes it difficult to effectively resist the evolving sniffing attacks and link flooding attacks. In order to improve the security of PIoT, this paper proposes an Adaptive Route Mutation based on Network State Awareness (SARM). It generates a route mutation space using a path state matrix and optimizes the time complexity of mutation space generation with a backtracking method. Furthermore, the SARM can dynamically adjust the route mutation strategy according to the real-time network state to realize self-adaptive defense. In conclusion, SARM is evaluated through simulations conducted with Mininet. Compared to Random Routing Mutation (RRM), it enhances defense against Sniffing and Distributed Denial of Service attacks by approximately 30% and 35% respectively. Additionally, in various example topologies, SARM consistently outperforms RRM.</div></div>","PeriodicalId":50637,"journal":{"name":"Computer Networks","volume":"255 ","pages":"Article 110889"},"PeriodicalIF":4.4,"publicationDate":"2024-11-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142662353","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
引用次数: 0
Inventory of Load Balancing Parameters in MPTCP Schedulers MPTCP 调度器中的负载平衡参数盘点
IF 4.4 2区 计算机科学
Computer Networks Pub Date : 2024-11-09 DOI: 10.1016/j.comnet.2024.110880
Mohamed Rabie Naimi , Chakib Zouaoui , Mohamed Elbahri , Abdenacer Bounoua
{"title":"Inventory of Load Balancing Parameters in MPTCP Schedulers","authors":"Mohamed Rabie Naimi ,&nbsp;Chakib Zouaoui ,&nbsp;Mohamed Elbahri ,&nbsp;Abdenacer Bounoua","doi":"10.1016/j.comnet.2024.110880","DOIUrl":"10.1016/j.comnet.2024.110880","url":null,"abstract":"<div><div>In the era of ubiquitous connectivity, where multiple wired and wireless interfaces connect internet users, we can now ensure the reliability of online services. Multi-Path TCP (MPTCP) significantly improves data transport efficiency by aggregating multiple network paths into a single session to optimize network resources use. MPTCP's effectiveness depends on its scheduling decisions, which directly affect data flow, throughput, and service quality. This work presents a first labeled dataset detailing critical MPTCP scheduling parameters—congestion window sizes, unacknowledged transmitted segments, and latency—from the most popular classic schedulers: RR (RoundRobin), BLEST (Blocking Estimation-based MPTCP Scheduler), and ECF(Earliest Completion First). It allows an in-depth study of different load-balancing policies, thus clarifying the complexity of schedulers and the impact of parameters used in each load-balancing policy on data transfer. With <strong>80033271</strong> rows extracted across <strong>15 scenarios</strong>, providing usable <strong>21</strong> numeric values per row, the objective of this dataset is to inventory all the decisive parameters in schedulers' decision-making. This dataset will not only facilitate the selection of optimal parameters for load-balancing policies, but also serve as a foundation for the development of numerous novel, supervised machine learning methods specifically tailored for scheduling tasks.</div></div>","PeriodicalId":50637,"journal":{"name":"Computer Networks","volume":"255 ","pages":"Article 110880"},"PeriodicalIF":4.4,"publicationDate":"2024-11-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142662352","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
引用次数: 0
Model Parameter Prediction Method for Accelerating Distributed DNN Training 加速分布式 DNN 训练的模型参数预测方法
IF 4.4 2区 计算机科学
Computer Networks Pub Date : 2024-11-08 DOI: 10.1016/j.comnet.2024.110883
Wai-xi Liu , Dao-xiao Chen , Miao-quan Tan , Kong-yang Chen , Yue Yin , Wen-Li Shang , Jin Li , Jun Cai
{"title":"Model Parameter Prediction Method for Accelerating Distributed DNN Training","authors":"Wai-xi Liu ,&nbsp;Dao-xiao Chen ,&nbsp;Miao-quan Tan ,&nbsp;Kong-yang Chen ,&nbsp;Yue Yin ,&nbsp;Wen-Li Shang ,&nbsp;Jin Li ,&nbsp;Jun Cai","doi":"10.1016/j.comnet.2024.110883","DOIUrl":"10.1016/j.comnet.2024.110883","url":null,"abstract":"<div><div>As the size of deep neural network (DNN) models and datasets increases, distributed training becomes popular to reduce the training time. However, a severe communication bottleneck in distributed training limits its scalability. Many methods aim to address this communication bottleneck by reducing communication traffic, such as gradient sparsification and quantization. However, these methods either are at the expense of losing model accuracy or introducing lots of computing overhead. We have observed that the data distribution between layers of neural network models is similar. Thus, we propose a model parameter prediction method (MP<sup>2</sup>) to accelerate distributed DNN training under parameter server (PS) framework, where workers push only a subset of model parameters to the PS, and residual model parameters are locally predicted by an already-trained deep neural network model on the PS. We address several key challenges in this approach. First, we build a hierarchical parameters dataset by randomly sampling a subset of model from normal distributed trainings. Second, we design a neural network model with the structure of “convolution + channel attention + Max pooling” for predicting model parameters by using a prediction result-based evaluation method. For VGGNet, ResNet, and AlexNet models on CIFAR10 and CIFAR100 datasets, compared with Baseline, Top-k, deep gradient compression (DGC), and weight nowcaster network (WNN), MP<sup>2</sup> can reduce traffic by up to 88.98%; and accelerates the training by up to 47.32% while not losing the model accuracy. MP<sup>2</sup> has shown good generalization.</div></div>","PeriodicalId":50637,"journal":{"name":"Computer Networks","volume":"255 ","pages":"Article 110883"},"PeriodicalIF":4.4,"publicationDate":"2024-11-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142662349","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
引用次数: 0
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×
相关产品
×
本文献相关产品
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