Pub Date : 2026-01-02DOI: 10.1109/TNSM.2025.3650697
Yeryeong Cho;Sungwon Yi;Soohyun Park
This paper introduces a novel resilient algorithm designed for distributed un-crewed aerial vehicles (UAVs) in dynamic and unreliable network environments. Initially, the UAVs should be trained via multi-agent reinforcement learning (MARL) for autonomous mission-critical operations and are fundamentally grounded by centralized training and decentralized execution (CTDE) using a centralized MARL server. In this situation, it is crucial to consider the case where several UAVs cannot receive CTDE-based MARL learning parameters for resilient operations in unreliable network conditions. To tackle this issue, a communication graph is used where its edges are established when two UAVs/nodes are communicable. Then, the edge-connected UAVs can share their training data if one of the UAVs cannot be connected to the CTDE-based MARL server under unreliable network conditions. Additionally, the edge cost considers power efficiency. Based on this given communication graph, message-passing is used for electing the UAVs that can provide their MARL learning parameters to their edge-connected peers. Lastly, performance evaluations demonstrate the superiority of our proposed algorithm in terms of power efficiency and resilient UAV task management, outperforming existing benchmark algorithms.
{"title":"Joint Multi-Agent Reinforcement Learning and Message-Passing for Resilient Multi-UAV Networks","authors":"Yeryeong Cho;Sungwon Yi;Soohyun Park","doi":"10.1109/TNSM.2025.3650697","DOIUrl":"https://doi.org/10.1109/TNSM.2025.3650697","url":null,"abstract":"This paper introduces a novel resilient algorithm designed for distributed un-crewed aerial vehicles (UAVs) in dynamic and unreliable network environments. Initially, the UAVs should be trained via multi-agent reinforcement learning (MARL) for autonomous mission-critical operations and are fundamentally grounded by centralized training and decentralized execution (CTDE) using a centralized MARL server. In this situation, it is crucial to consider the case where several UAVs cannot receive CTDE-based MARL learning parameters for resilient operations in unreliable network conditions. To tackle this issue, a communication graph is used where its edges are established when two UAVs/nodes are communicable. Then, the edge-connected UAVs can share their training data if one of the UAVs cannot be connected to the CTDE-based MARL server under unreliable network conditions. Additionally, the edge cost considers power efficiency. Based on this given communication graph, message-passing is used for electing the UAVs that can provide their MARL learning parameters to their edge-connected peers. Lastly, performance evaluations demonstrate the superiority of our proposed algorithm in terms of power efficiency and resilient UAV task management, outperforming existing benchmark algorithms.","PeriodicalId":13423,"journal":{"name":"IEEE Transactions on Network and Service Management","volume":"23 ","pages":"2051-2063"},"PeriodicalIF":5.4,"publicationDate":"2026-01-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146026588","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}
Pub Date : 2025-12-30DOI: 10.1109/TNSM.2025.3649503
Surabhi Sharma;Sateesh Kumar Peddoju
Internet of Things (IoT) applications continuously generate large volumes of transient data. Delivering transient data efficiently is challenging because it is short-lived, highly dynamic, and often critical for time-sensitive services. Caching at the Edge offers a practical solution by storing frequently requested content closer to users, reducing delivery delays, and easing network congestion. However, existing caching approaches in Edge-assisted IoT networks face four significant limitations: (i) lack of freshness-aware policies, leading to outdated data, (ii) static or centralized coordination, which restricts scalability, (iii) inability to adapt to bursty and heterogeneous traffic patterns, and (iv) inefficient handling of resource-constrained Edge nodes. IoT-Cooperative Caching (IoT-C${}^{2}$ ) addresses these issues with a framework based on multi-agent reinforcement learning. Using the framework, Edge servers make decentralized, adaptive decisions that account for both user demand and data freshness. IoT-C2 introduces topic-based grouping of Edge nodes and a hierarchical state model that supports collaboration across local, group, and global states. Experiments show that IoT-C2 increases cache hit rates, reduces latency, and improves freshness compared with state-of-the-art techniques. These improvements make the proposed approach well-suited for time-critical IoT applications like smart cities, healthcare, and industrial networks.
{"title":"Cooperative Multi-Agent Strategy for Caching of Transient Data in Edge-Assisted IoT Networks","authors":"Surabhi Sharma;Sateesh Kumar Peddoju","doi":"10.1109/TNSM.2025.3649503","DOIUrl":"https://doi.org/10.1109/TNSM.2025.3649503","url":null,"abstract":"Internet of Things (IoT) applications continuously generate large volumes of transient data. Delivering transient data efficiently is challenging because it is short-lived, highly dynamic, and often critical for time-sensitive services. Caching at the Edge offers a practical solution by storing frequently requested content closer to users, reducing delivery delays, and easing network congestion. However, existing caching approaches in Edge-assisted IoT networks face four significant limitations: (i) lack of freshness-aware policies, leading to outdated data, (ii) static or centralized coordination, which restricts scalability, (iii) inability to adapt to bursty and heterogeneous traffic patterns, and (iv) inefficient handling of resource-constrained Edge nodes. IoT-Cooperative Caching (IoT-C<inline-formula> <tex-math>${}^{2}$ </tex-math></inline-formula>) addresses these issues with a framework based on multi-agent reinforcement learning. Using the framework, Edge servers make decentralized, adaptive decisions that account for both user demand and data freshness. IoT-C2 introduces topic-based grouping of Edge nodes and a hierarchical state model that supports collaboration across local, group, and global states. Experiments show that IoT-C2 increases cache hit rates, reduces latency, and improves freshness compared with state-of-the-art techniques. These improvements make the proposed approach well-suited for time-critical IoT applications like smart cities, healthcare, and industrial networks.","PeriodicalId":13423,"journal":{"name":"IEEE Transactions on Network and Service Management","volume":"23 ","pages":"1369-1380"},"PeriodicalIF":5.4,"publicationDate":"2025-12-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145929422","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}
Generalized Frequency Division Multiplexing (GFDM) is considered a strong candidate to replace Orthogonal Frequency Division Multiplexing (OFDM) in 5G MIMO networks because of its enhanced spectral utilization and design flexibility. Despite these advantages, GFDM faces the drawback of producing a relatively high Peak-to-Average Power Ratio (PAPR), which limits the efficiency of power amplifiers. To address this issue, the Partial Transmit Sequence (PTS) method is often employed for PAPR reduction. Nevertheless, the effectiveness of PTS is hindered by the intensive computational effort required for searching multiple phase factors. To overcome this challenge, we propose a method that integrates the Enhanced Squirrel Search Algorithm (ESSA) with an adaptive parameter control mechanism and a Grey Wolf Optimizer (GWO), enabling a dynamic balance between exploration and exploitation during phase factor selection. This improvement reduces the computational overhead, accelerates the convergence, and enhances the robustness of the phase sequence optimization. Simulation results show that the Hybrid PTS-ESSA-GWO-RPSM model achieves superior PAPR reduction compared to conventional ESSA-based approaches, while also providing better BER and SNR performance under varying channel conditions. The proposed method therefore offers an efficient trade-off between complexity and PAPR reduction, making it suitable for practical deployment in MIMO-GFDM-based 5G systems. The proposed scheme is evaluated against related methods by analyzing key performance indicators, including Complementary Cumulative Distribution Function (CCDF), Bit Error Rate (BER), Peak-to-Average Power Ratio (PAPR), and Signal-to-Noise Ratio (SNR).
{"title":"A Dynamic PAPR Reduction Method Using PTS-ESSA for MIMO Generalized FDM Wireless System","authors":"Surendra Kumar;Jitendra Kumar Samriya;Rajeev Tiwari;Mohit Kumar;Shilpi Harnal;Neeraj Kumar;Mohsen Guizani","doi":"10.1109/TNSM.2025.3619945","DOIUrl":"https://doi.org/10.1109/TNSM.2025.3619945","url":null,"abstract":"Generalized Frequency Division Multiplexing (GFDM) is considered a strong candidate to replace Orthogonal Frequency Division Multiplexing (OFDM) in 5G MIMO networks because of its enhanced spectral utilization and design flexibility. Despite these advantages, GFDM faces the drawback of producing a relatively high Peak-to-Average Power Ratio (PAPR), which limits the efficiency of power amplifiers. To address this issue, the Partial Transmit Sequence (PTS) method is often employed for PAPR reduction. Nevertheless, the effectiveness of PTS is hindered by the intensive computational effort required for searching multiple phase factors. To overcome this challenge, we propose a method that integrates the Enhanced Squirrel Search Algorithm (ESSA) with an adaptive parameter control mechanism and a Grey Wolf Optimizer (GWO), enabling a dynamic balance between exploration and exploitation during phase factor selection. This improvement reduces the computational overhead, accelerates the convergence, and enhances the robustness of the phase sequence optimization. Simulation results show that the Hybrid PTS-ESSA-GWO-RPSM model achieves superior PAPR reduction compared to conventional ESSA-based approaches, while also providing better BER and SNR performance under varying channel conditions. The proposed method therefore offers an efficient trade-off between complexity and PAPR reduction, making it suitable for practical deployment in MIMO-GFDM-based 5G systems. The proposed scheme is evaluated against related methods by analyzing key performance indicators, including Complementary Cumulative Distribution Function (CCDF), Bit Error Rate (BER), Peak-to-Average Power Ratio (PAPR), and Signal-to-Noise Ratio (SNR).","PeriodicalId":13423,"journal":{"name":"IEEE Transactions on Network and Service Management","volume":"23 ","pages":"1162-1175"},"PeriodicalIF":5.4,"publicationDate":"2025-12-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145852470","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}
Pub Date : 2025-12-26DOI: 10.1109/TNSM.2025.3648815
Tianzi Zhao;Xinran Liu;Zhaoxin Zhang;Dong Zhao;Ning Li;Zhichao Zhang;Xinye Wang
Fine-grained IP geolocation plays a critical role in applications such as location-based services and cybersecurity. Most existing fine-grained IP geolocation methods treat the task as a coordinate regression problem. However, due to inherent noise in the IP features, they frequently produce kilometer-scale coordinate errors, which in turn lead to inaccuracies when the predicted coordinates are mapped to geographic regions. To alleviate this issue, this paper introduces a novel IP region prediction framework HMCGeo, framing IP region prediction as a hierarchical multi-label classification problem. City administrators divide urban areas into regions at multiple granularities. The proposed framework employs residual connection-based feature extraction units to obtain IP representations at each granularity and introduces class prototype attention to predict the IP’s belonging region at the current granularity. Additionally, it adopts an output fusion strategy combined with hierarchical focal loss to further enhance region prediction performance. We evaluate HMCGeo on real-world datasets from New York, Los Angeles, and Shanghai. It significantly outperforms existing methods in region prediction across all granularities and achieves lower coordinate errors on most samples by similarity-weighted averaging of candidate region centers.
{"title":"HMCGeo: IP Region Prediction Based on Hierarchical Multi-Label Classification","authors":"Tianzi Zhao;Xinran Liu;Zhaoxin Zhang;Dong Zhao;Ning Li;Zhichao Zhang;Xinye Wang","doi":"10.1109/TNSM.2025.3648815","DOIUrl":"https://doi.org/10.1109/TNSM.2025.3648815","url":null,"abstract":"Fine-grained IP geolocation plays a critical role in applications such as location-based services and cybersecurity. Most existing fine-grained IP geolocation methods treat the task as a coordinate regression problem. However, due to inherent noise in the IP features, they frequently produce kilometer-scale coordinate errors, which in turn lead to inaccuracies when the predicted coordinates are mapped to geographic regions. To alleviate this issue, this paper introduces a novel IP region prediction framework HMCGeo, framing IP region prediction as a hierarchical multi-label classification problem. City administrators divide urban areas into regions at multiple granularities. The proposed framework employs residual connection-based feature extraction units to obtain IP representations at each granularity and introduces class prototype attention to predict the IP’s belonging region at the current granularity. Additionally, it adopts an output fusion strategy combined with hierarchical focal loss to further enhance region prediction performance. We evaluate HMCGeo on real-world datasets from New York, Los Angeles, and Shanghai. It significantly outperforms existing methods in region prediction across all granularities and achieves lower coordinate errors on most samples by similarity-weighted averaging of candidate region centers.","PeriodicalId":13423,"journal":{"name":"IEEE Transactions on Network and Service Management","volume":"23 ","pages":"1913-1926"},"PeriodicalIF":5.4,"publicationDate":"2025-12-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145982177","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}
Stochastic disruptions such as flash events arising from benign traffic bursts and switch thermal fluctuations are major contributors to intermittent service degradation in software-defined industrial networks. These events violate IEC 61850-derived quality of service requirements and user-defined service-level agreements, hindering the reliable and timely delivery of control, monitoring, and best-effort traffic in IEC 61400-25-compliant wind power plants. Failure to maintain these requirements often results in delayed or lost control signals, reduced operational efficiency, and increased risk of wind turbine generator downtime. To address these challenges, this study proposes a threshold-triggered Deep Q-Network self-healing agent that autonomically detects, analyzes, and mitigates network disruptions while adapting routing behavior and resource allocation in real time. The proposed agent was trained, validated, and tested on an emulated tri-clustered switch network deployed in a cloud-based proof-of-concept testbed. Simulation results show that the proposed agent improves disruption recovery performance by 53.84% compared to a baseline shortest-path and load-balanced routing approach, and outperforms state-of-the-art methods, including the Adaptive Network-based Fuzzy Inference System by 13.1% and the Deep Q-Network and Traffic Prediction-based Routing Optimization method by 21.5%, in a super-spine leaf data-plane architecture. Additionally, the agent maintains switch thermal stability by proactively initiating external rack cooling when required. These findings highlight the potential of deep reinforcement learning in building resilience in software-defined industrial networks deployed in mission-critical, time-sensitive application scenarios.
{"title":"A Threshold-Triggered Deep Q-Network-Based Framework for Self-Healing in Autonomic Software-Defined IIoT-Edge Networks","authors":"Agrippina Mwangi;León Navarro-Hilfiker;Lukasz Brewka;Mikkel Gryning;Elena Fumagalli;Madeleine Gibescu","doi":"10.1109/TNSM.2025.3647853","DOIUrl":"https://doi.org/10.1109/TNSM.2025.3647853","url":null,"abstract":"Stochastic disruptions such as flash events arising from benign traffic bursts and switch thermal fluctuations are major contributors to intermittent service degradation in software-defined industrial networks. These events violate IEC 61850-derived quality of service requirements and user-defined service-level agreements, hindering the reliable and timely delivery of control, monitoring, and best-effort traffic in IEC 61400-25-compliant wind power plants. Failure to maintain these requirements often results in delayed or lost control signals, reduced operational efficiency, and increased risk of wind turbine generator downtime. To address these challenges, this study proposes a threshold-triggered Deep Q-Network self-healing agent that autonomically detects, analyzes, and mitigates network disruptions while adapting routing behavior and resource allocation in real time. The proposed agent was trained, validated, and tested on an emulated tri-clustered switch network deployed in a cloud-based proof-of-concept testbed. Simulation results show that the proposed agent improves disruption recovery performance by 53.84% compared to a baseline shortest-path and load-balanced routing approach, and outperforms state-of-the-art methods, including the Adaptive Network-based Fuzzy Inference System by 13.1% and the Deep Q-Network and Traffic Prediction-based Routing Optimization method by 21.5%, in a super-spine leaf data-plane architecture. Additionally, the agent maintains switch thermal stability by proactively initiating external rack cooling when required. These findings highlight the potential of deep reinforcement learning in building resilience in software-defined industrial networks deployed in mission-critical, time-sensitive application scenarios.","PeriodicalId":13423,"journal":{"name":"IEEE Transactions on Network and Service Management","volume":"23 ","pages":"1297-1311"},"PeriodicalIF":5.4,"publicationDate":"2025-12-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=11313855","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145929613","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-12-22DOI: 10.1109/TNSM.2025.3646778
Xi Liu;Jun Liu;Weidong Li
Vehicle computing has emerged as a promising paradigm for delivering time-sensitive computing services to Internet of Things applications. Intelligent vehicles (IVs) offer onboard computing and sensing capabilities for delivering a wide range of services. In this paper, we propose a dynamic adaptability service model that leverages the swift mobility of vehicles to adjust the distribution of IVs to users’ dynamically changing locations. There are two types of areas in our model: the user area and the parking area. The former is where services are provided, while the latter serves as the preparation zone for backup IVs. IVs in the parking area are dispatched to service areas, where existing vehicle resources cannot meet user demand, and they return to the parking area after delivering the service. Multiple users share sensing resources, and our model allocates the costs among them. To ensure strategy-proofness, we introduce the concepts of no additional cost and allocation stability. We propose a strategy-proof cost-sharing mechanism for dynamic adaptability service. The proposed mechanism achieves no positive transfers, voluntary participation, individual rationality, consumer sovereignty, budget balance, no additional costs, and allocation stability. Moreover, the proposed mechanism’s approximation performance is analyzed. We further use comprehensive simulations to verify the effectiveness and efficiency of the proposed mechanism.
{"title":"Strategy-Proof Cost-Sharing Mechanism for Dynamic Adaptability Service in Vehicle Computing","authors":"Xi Liu;Jun Liu;Weidong Li","doi":"10.1109/TNSM.2025.3646778","DOIUrl":"https://doi.org/10.1109/TNSM.2025.3646778","url":null,"abstract":"Vehicle computing has emerged as a promising paradigm for delivering time-sensitive computing services to Internet of Things applications. Intelligent vehicles (IVs) offer onboard computing and sensing capabilities for delivering a wide range of services. In this paper, we propose a dynamic adaptability service model that leverages the swift mobility of vehicles to adjust the distribution of IVs to users’ dynamically changing locations. There are two types of areas in our model: the user area and the parking area. The former is where services are provided, while the latter serves as the preparation zone for backup IVs. IVs in the parking area are dispatched to service areas, where existing vehicle resources cannot meet user demand, and they return to the parking area after delivering the service. Multiple users share sensing resources, and our model allocates the costs among them. To ensure strategy-proofness, we introduce the concepts of no additional cost and allocation stability. We propose a strategy-proof cost-sharing mechanism for dynamic adaptability service. The proposed mechanism achieves no positive transfers, voluntary participation, individual rationality, consumer sovereignty, budget balance, no additional costs, and allocation stability. Moreover, the proposed mechanism’s approximation performance is analyzed. We further use comprehensive simulations to verify the effectiveness and efficiency of the proposed mechanism.","PeriodicalId":13423,"journal":{"name":"IEEE Transactions on Network and Service Management","volume":"23 ","pages":"1942-1959"},"PeriodicalIF":5.4,"publicationDate":"2025-12-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145982313","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}
This paper addresses network slicing in a large-scale Multi-Access Edge Computing (MEC)-enabled Radio Access Network (RAN) comprising heterogeneous edge nodes with varying computing and storage resource capacities. These resources are dynamically allocated to slice requests and released when the service of a slice request is completed. Our objective is to optimize the resource allocation for each admitted arriving slice request, considering its demands for computing and storage resources, to maximize the long-run average Earning Before Interest and Taxes (EBIT) of the MEC slicing system. We formulate the optimization problem as a Restless Multi-Armed Bandit (RMAB)-based resource allocation problem with a nonlinear cost rate function. To solve this, we introduce a new policy called Prioritizing-the-Future-Approximated earning per request (PFA) where for each admitted slice request, we always prioritize the allocation of the resource combination that gives the highest achievable earning, considering the future effects of this allocation. PFA is designed to be scalable and applicable to large-scale networks. We numerically demonstrate the superior performance of PFA in maximizing long-run average EBIT through simulations, comparing it with two baseline policies, at various cases of parameter values. Moreover, our findings offer insights for network operators in resource allocation policy selection.
{"title":"Network Slicing in MEC-Based RANs With Nonlinear Cost Rate Functions","authors":"Jiahe Xu;Jing Fu;Bige Yang;Zengfu Wang;Jingjin Wu;Xinyu Wang;Moshe Zukerman","doi":"10.1109/TNSM.2025.3646478","DOIUrl":"https://doi.org/10.1109/TNSM.2025.3646478","url":null,"abstract":"This paper addresses network slicing in a large-scale Multi-Access Edge Computing (MEC)-enabled Radio Access Network (RAN) comprising heterogeneous edge nodes with varying computing and storage resource capacities. These resources are dynamically allocated to slice requests and released when the service of a slice request is completed. Our objective is to optimize the resource allocation for each admitted arriving slice request, considering its demands for computing and storage resources, to maximize the long-run average Earning Before Interest and Taxes (EBIT) of the MEC slicing system. We formulate the optimization problem as a Restless Multi-Armed Bandit (RMAB)-based resource allocation problem with a nonlinear cost rate function. To solve this, we introduce a new policy called Prioritizing-the-Future-Approximated earning per request (PFA) where for each admitted slice request, we always prioritize the allocation of the resource combination that gives the highest achievable earning, considering the future effects of this allocation. PFA is designed to be scalable and applicable to large-scale networks. We numerically demonstrate the superior performance of PFA in maximizing long-run average EBIT through simulations, comparing it with two baseline policies, at various cases of parameter values. Moreover, our findings offer insights for network operators in resource allocation policy selection.","PeriodicalId":13423,"journal":{"name":"IEEE Transactions on Network and Service Management","volume":"23 ","pages":"1989-2005"},"PeriodicalIF":5.4,"publicationDate":"2025-12-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146026417","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}
Traffic analysis plays a pivotal role in network management. However, despite the prevalence of encryption, attackers are still able to deduce privacy elements such as user behavior and OS identification through advanced learning-based methods that exploit side-channel features. Existing defense strategies, which manipulate feature distribution to evade traffic analyzers, are often hampered by the need for impractical decoder deployment across all routes in symmetric framework methods. Moreover, reversing feature distribution modifications to real-time traffic, especially through dummy packet crafting or padding, is a complex task. In response to these challenges, we propose Veil, a novel and practical defender designed to protect live connections against encrypted network traffic analyzers. Leveraging an asymmetric deployment structure, Veil is capable of reconstructing live streams at the packet-block level, thereby allowing for seamless deployment on any connection node while enforcing transmission constraints. By employing a traffic-customized DQN framework, Veil not only reverses statistical feature perturbations back to the traffic space but also directs the distribution towards a target class. Extensive experiments conducted on real-world datasets validate the efficacy of Veil in efficiently evading analyzers in both targeted and untargeted modes, outperforming existing defense mechanisms. Notably, Veil addresses the key issues of impractical decoder deployment and complex real-time traffic manipulation, offering a more viable solution for network traffic privacy protection. The source code is publicly available at https://github.com/SecTeamPolaris/Veil, facilitating further research and application in the field of network security.
{"title":"Online Traffic Camouflage Against Network Analyzers via Deep Reinforcement Learning","authors":"Wenhao Li;Jie Chen;Zhaoxuan Li;Shuai Wang;Huamin Jin;Xiao-Yu Zhang","doi":"10.1109/TNSM.2025.3646259","DOIUrl":"https://doi.org/10.1109/TNSM.2025.3646259","url":null,"abstract":"Traffic analysis plays a pivotal role in network management. However, despite the prevalence of encryption, attackers are still able to deduce privacy elements such as user behavior and OS identification through advanced learning-based methods that exploit side-channel features. Existing defense strategies, which manipulate feature distribution to evade traffic analyzers, are often hampered by the need for impractical decoder deployment across all routes in symmetric framework methods. Moreover, reversing feature distribution modifications to real-time traffic, especially through dummy packet crafting or padding, is a complex task. In response to these challenges, we propose <monospace>Veil</monospace>, a novel and practical defender designed to protect live connections against encrypted network traffic analyzers. Leveraging an asymmetric deployment structure, <monospace>Veil</monospace> is capable of reconstructing live streams at the packet-block level, thereby allowing for seamless deployment on any connection node while enforcing transmission constraints. By employing a traffic-customized DQN framework, <monospace>Veil</monospace> not only reverses statistical feature perturbations back to the traffic space but also directs the distribution towards a target class. Extensive experiments conducted on real-world datasets validate the efficacy of <monospace>Veil</monospace> in efficiently evading analyzers in both targeted and untargeted modes, outperforming existing defense mechanisms. Notably, <monospace>Veil</monospace> addresses the key issues of impractical decoder deployment and complex real-time traffic manipulation, offering a more viable solution for network traffic privacy protection. The source code is publicly available at <uri>https://github.com/SecTeamPolaris/Veil</uri>, facilitating further research and application in the field of network security.","PeriodicalId":13423,"journal":{"name":"IEEE Transactions on Network and Service Management","volume":"23 ","pages":"1927-1941"},"PeriodicalIF":5.4,"publicationDate":"2025-12-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145982182","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}
Pub Date : 2025-12-18DOI: 10.1109/TNSM.2025.3645305
Lei Zhang;Wanting Su;Qin Ni;Jiawangnan Lu;Bin Chen
With the evolution of mobile networks towards Artificial Intelligence as a Service (AIaaS), generative radio maps not only need to reflect the signal strength distribution in specific areas, but also possess the capability of proactive prediction. However, due to the rapid updates in urban infrastructure and the network iterations, crafting radio maps in complex urban environments represents a substantial challenge. In this paper, a multi-output framework for generating radio maps in real multi-building scenarios is proposed, based on Reference Signal Received Power (RSRP) and Reference Signal Received Quality (RSRQ) extracted from actual urban and suburban Measurement Reports (MRs). Specifically, An image encoding method integrating environmental features and base station system information is designed, while considering the sector antenna characteristics in actual communication environments. Then, a multi-output Conditional Wasserstein Generative Adversarial Network (CWGAN) is constructed for image conversion, and the radio maps are generated by learning the mapping from environmental & system information to RSRP & RSRQ radio maps, on the basis of image encoding that incorporates the physical laws of radio propagation. By calculating the priority of communication link gains at receiving points, it provides generative networks with reliable theoretical basis and conditional information, for serving cells and first neighboring cells. Experimental results show that the root mean square errors (RMSE) of the proposed method for RSRP / RSRQ of serving and neighboring cells are 1.7821 / 2.2251 and 0.8108 / 1.5121, which demonstrates the proposed method outperforms the baseline results. Simultaneously radio maps generation endows the cellular network with a certain “prophetic” capability, significantly enhancing the live service experience.
{"title":"GAN4RM: A CWGAN-Based Framework for Radio Maps Generation in Real Cellular Networks","authors":"Lei Zhang;Wanting Su;Qin Ni;Jiawangnan Lu;Bin Chen","doi":"10.1109/TNSM.2025.3645305","DOIUrl":"https://doi.org/10.1109/TNSM.2025.3645305","url":null,"abstract":"With the evolution of mobile networks towards Artificial Intelligence as a Service (AIaaS), generative radio maps not only need to reflect the signal strength distribution in specific areas, but also possess the capability of proactive prediction. However, due to the rapid updates in urban infrastructure and the network iterations, crafting radio maps in complex urban environments represents a substantial challenge. In this paper, a multi-output framework for generating radio maps in real multi-building scenarios is proposed, based on Reference Signal Received Power (RSRP) and Reference Signal Received Quality (RSRQ) extracted from actual urban and suburban Measurement Reports (MRs). Specifically, An image encoding method integrating environmental features and base station system information is designed, while considering the sector antenna characteristics in actual communication environments. Then, a multi-output Conditional Wasserstein Generative Adversarial Network (CWGAN) is constructed for image conversion, and the radio maps are generated by learning the mapping from environmental & system information to RSRP & RSRQ radio maps, on the basis of image encoding that incorporates the physical laws of radio propagation. By calculating the priority of communication link gains at receiving points, it provides generative networks with reliable theoretical basis and conditional information, for serving cells and first neighboring cells. Experimental results show that the root mean square errors (RMSE) of the proposed method for RSRP / RSRQ of serving and neighboring cells are 1.7821 / 2.2251 and 0.8108 / 1.5121, which demonstrates the proposed method outperforms the baseline results. Simultaneously radio maps generation endows the cellular network with a certain “prophetic” capability, significantly enhancing the live service experience.","PeriodicalId":13423,"journal":{"name":"IEEE Transactions on Network and Service Management","volume":"23 ","pages":"1329-1341"},"PeriodicalIF":5.4,"publicationDate":"2025-12-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145929419","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}
Pub Date : 2025-12-18DOI: 10.1109/TNSM.2025.3645463
Shuyi Liu;Yuang Chen;Zhengze Li;Fangyu Zhang;Hancheng Lu;Xiaobo Guo;Lizhe Liu
Segment Routing over IPv6 (SRv6) gives operators explicit path control and alleviates network congestion, making it a compelling technique for traffic engineering (TE). Yet two practical hurdles slow adoption. First, a one-shot upgrade of every traditional device is prohibitively expensive, so operators must prioritize which devices to upgrade. Second, the Segment Routing Header (SRH) increases packet size; if TE algorithms ignore this overhead, they will underestimate link load and may cause congestion in practice. We address both challenges with DRL-TE, an algorithm that couples deep reinforcement learning (DRL) with a lightweight local search (LS) step to minimize the network’s maximum link utilization (MLU). DRL-TE first identifies the smallest set of critical devices whose upgrade yields the largest drop in MLU, enabling hybrid IP/SRv6 networks to approach optimal performance with minimal investment. It then computes SRH-aware routes, and the DRL agent, augmented by a fast LS refinement, rapidly reduces MLU even under traffic variation. Experiments on an 11-node hardware testbed and three larger simulated topologies show that upgrading about 30% of devices allows DRL-TE to match fully upgraded networks and reduce MLU by up to 34% compared with existing algorithms. DRL-TE also maintains high performance under link failures and traffic variations, offering a cost-effective and robust path toward incremental SRv6 deployment.
{"title":"Segment Routing Header (SRH)-Aware Traffic Engineering in Hybrid IP/SRv6 Networks With Deep Reinforcement Learning","authors":"Shuyi Liu;Yuang Chen;Zhengze Li;Fangyu Zhang;Hancheng Lu;Xiaobo Guo;Lizhe Liu","doi":"10.1109/TNSM.2025.3645463","DOIUrl":"https://doi.org/10.1109/TNSM.2025.3645463","url":null,"abstract":"Segment Routing over IPv6 (SRv6) gives operators explicit path control and alleviates network congestion, making it a compelling technique for traffic engineering (TE). Yet two practical hurdles slow adoption. First, a one-shot upgrade of every traditional device is prohibitively expensive, so operators must prioritize which devices to upgrade. Second, the Segment Routing Header (SRH) increases packet size; if TE algorithms ignore this overhead, they will underestimate link load and may cause congestion in practice. We address both challenges with DRL-TE, an algorithm that couples deep reinforcement learning (DRL) with a lightweight local search (LS) step to minimize the network’s maximum link utilization (MLU). DRL-TE first identifies the smallest set of critical devices whose upgrade yields the largest drop in MLU, enabling hybrid IP/SRv6 networks to approach optimal performance with minimal investment. It then computes SRH-aware routes, and the DRL agent, augmented by a fast LS refinement, rapidly reduces MLU even under traffic variation. Experiments on an 11-node hardware testbed and three larger simulated topologies show that upgrading about 30% of devices allows DRL-TE to match fully upgraded networks and reduce MLU by up to 34% compared with existing algorithms. DRL-TE also maintains high performance under link failures and traffic variations, offering a cost-effective and robust path toward incremental SRv6 deployment.","PeriodicalId":13423,"journal":{"name":"IEEE Transactions on Network and Service Management","volume":"23 ","pages":"1260-1275"},"PeriodicalIF":5.4,"publicationDate":"2025-12-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145929552","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}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：