TNSM Searchable Index

Author(s)	Title	Year	Publication	Keywords
Abdalla Hussein, Patrick Mitran, Catherine Rosenberg	Spectrum and RAN Sharing: How to Avoid Cross-Subsidization While Taking Full Advantage of Massive MU-MIMO?	2025	Early Access	Benchmark testing Resource management Graph neural networks	Motivated by the need to use spectrum more efficiently, this paper investigates fine grained spectrum sharing (FGSS) in Multi-User massive MIMO (MU-mMIMO) systems where a neutral host enables users from different operators to share the same resource blocks. To be accepted by operators, FGSS must i) guarantee isolation so that the load of one operator does not impact the performance of another, and ii) avoid crosssubsidization whereby one operator gains more from sharing than another. We first formulate and solve an offline problem to assess the potential performance gains of FGSS with respect to the static spectrum sharing case, where operators have fixed separate subbands, and find that the gains can be significant, motivating the development for online solutions for FGSS. Transitioning from an offline to an online study presents unique challenges, including the lack of apriori knowledge regarding the performance of the fixed sharing case that is required to ensure isolation and cross-subsidization avoidance. We overcome these challenges and propose an online algorithm that is fast and significantly outperforms the static case. The main finding is that FGSS for a MU-mMIMO downlink system is doable in a way that is“safe” to operators and brings large gains in spectrum efficiency (e.g., for 4 operators, a gain above 60% is seen in many cases).	10.1109/TNSM.2025.3612425
Ziwang Wang, Huili Yan, Zhize Wu	Efficient Cross-Shard Blockchain Atomic Submission Scheme Based on Pledge Transactions	2025	Early Access	Blockchains Security Synchronization	To address the substantial coordination overhead and communication latency inherent in cross-shard transaction commits within contemporary multi-shard blockchain architectures, this paper presents an efficient cross-shard atomic commit scheme (PledgeACS), grounded in the use of pledge transactions. The proposed scheme introduces a pledge transaction mechanism that employs a null recipient address and synchronizes these transactions across shards to an auxiliary chain through a global consensus protocol. Additionally, a cross-shard transaction protocol is developed, securing recipient funds via pledge transactions within the global consensus framework. Furthermore, a batch pledge transaction record and settlement protocol tailored for shard blockchains is designed, followed by rigorous feasibility analysis and performance evaluation. Experimental results indicate that the proposed scheme markedly decreases the user-perceived latency in cross-shard transactions and enhances security relative to current solutions, providing an innovative approach for achieving high throughput and scalability in blockchain systems.	10.1109/TNSM.2025.3607349
Rifat Al Mamun Rudro, Sultanul Arifieen Hamim, Md. Hamid Uddin, Md Masuduzzaman, Md. Manzurul Hasan	Waris-Chain: The Blockchain Driven Transformation of Inheritance Solutions	2025	Early Access	Blockchains Smart contracts Fraud	In the current digital era, managing inheritance presents a critical challenge, necessitating a balance of effectiveness, security, and transparency. Traditional processes are often complex, time-consuming, and susceptible to fraud and disputes. This paper introduces the successor chain model called Waris-Chain, a blockchain-based solution designed to streamline and secure inheritance management. Waris-Chain integrates smart contracts and Non-Fungible Tokens (NFTs) to automate and verify inheritance processes, ensuring accuracy and reducing manual intervention. Developed using the Ethereum blockchain, ERC-1155 tokens, and MetaMask for authentication, Waris-Chain offers a comprehensive, adaptable, and secure platform. Performance evaluation shows that Waris-Chain achieves a high throughput of 477.36 transactions per hour, a low transaction latency of 7.54 seconds, with a 99.42% accuracy rate and a 0.58% error rate. Despite these advancements, challenges such as blockchain adoption, legal integration, and system scalability remain, suggesting avenues for future research to fully realize blockchain’s potential in inheritance management.	10.1109/TNSM.2025.3608088
Meihui Liu, Fangmin Xu, Shihui Duan, Jinyu Zhu, Wenlong Ma, Ruoyu Ji, Chenglin Zhao	Efficient SRv6 Based Multi-Path Transmission Strategy for Resilient Communication in Deterministic Computing Power Network	2025	Early Access	Reliability Processor scheduling Computer architecture	The computing power network (CPN) serves as a key infrastructure for future networks, facilitating the connection of ubiquitous computing resources distributed across various locations. The continuous emergence of computation-intensive and delay-sensitive applications highlights the crucial need to fully utilize limited computing resources and the importance of building a resilient communication network. Primary-backup (PB) based transmission is a commonly used technique to enhance network reliability. However, implementing this approach in CPN with a consideration of load balancing introduces significant complexity and has received limited research attention. In this paper, we designed a deterministic computing power network (Det-CPN) architecture based on segment routing over IPv6 (SRv6). On top of the above architecture, we proposed a best computing node selection method based on a comprehensive index calculation and ranking (CICR) algorithm to determine the optimal computing node for task transmission. Subsequently, we developed a bandwidth sharing-based multi-path transmission (BSMT) algorithm to realize the maximization of the system efficiency. Simulation results demonstrate that in adverse network conditions (overloaded with a failure rate of 0.02), compared to the traditional dual-path redundant forwarding mechanism, the proposed solution achieves an average reduction of 22.3% in transmission latency, an average improvement of 39.94% in task success rate, a decrease of 19.4% in bandwidth occupation rate, and an increase of 31.05% in computing resource utilization rate.	10.1109/TNSM.2025.3608200
Wei-Kun Chen, Ya-Feng Liu, Yu-Hong Dai, Zhi-Quan Luo	QoS-Aware and Routing-Flexible Network Slicing for Service-Oriented Networks	2025	Early Access	Reliability Delays Routing	In this paper, we consider the network slicing () problem which aims to map multiple customized virtual network requests (also called services) to a common shared network infrastructure and manage network resources to meet diverse quality of service (QoS) requirements. We propose a mixed-integer nonlinear programming (MINLP) formulation for the considered NS problem that can flexibly route the traffic flow of the services on multiple paths and provide end-to-end delay and reliability guarantees for all services. To overcome the computational difficulty due to the intrinsic nonlinearity in the MINLP formulation, we transform the formulation into an equivalent mixed-integer linear programming () formulation and further show that their continuous relaxations are equivalent. In sharp contrast to the continuous relaxation of the formulation which is a nonconvex nonlinear programming problem, the continuous relaxation of the formulation is a polynomial-time solvable linear programming problem, which significantly facilitates the algorithmic design. Based on the newly proposed formulation, we develop a customized column generation () algorithm for solving the problem. The proposed algorithm is a decomposition-based algorithm and is particularly suitable for solving large-scale problems. Numerical results demonstrate the efficacy of the proposed formulations and the proposed algorithm.	10.1109/TNSM.2025.3608074
M. Wasim Abbas Ashraf, Shivanshu Shrivastava, Om Jee Pandey, Arvind R. Singh, Arif Raza	DRL-Driven Optimal User Association and Load Balancing in Hybrid RF/LiFi Based IoT Systems	2025	Early Access	Radio frequency Internet of Things Light fidelity	The proliferation of Internet of Things (IoT) devices has raised considerable difficulties in the identification of users, optimization of power, and load balancing in hybrid RF/LiFi networks. As interconnection among devices increases, ensuring optimal performance while managing network resources efficiently becomes quite complex. This complexity arises due to accommodating the possibly diverse users’ needs, fluctuating channel conditions, and varying interference levels, all necessitating sophisticated management solutions to provide seamless connectivity and dependable communication. To tackle these issues, a deep joint hybrid system (DJHS) technique is presented, which employs proximal policy optimization (PPO), a cutting-edge deep reinforcement learning (DRL) algorithm. DJHS aims to effectively handle the intricate problems surrounding user association and load balancing while optimizing power usage in dynamic contexts. DJHS continuously updates its approach based on real-time network data through adaptive learning methods, allowing it to make intelligent decisions that improve overall system performance regarding data throughput and power optimization. Simulation results demonstrate that DJHS outperforms existing approaches such as sac, a2c, td3, and trpo regarding crucial metrics, including data rate and power transmission. Notably, DJHS’s ability to adjust to variations in signal-to-interference-plus-noise ratio (SINR) allows for enhanced resource allocation and network stability. This flexibility ensures that users receive optimal service even in changing conditions, enhancing the overall user experience.	10.1109/TNSM.2025.3607433
Anna Karanika, Rui Yang, Xiaojuan Ma, Jiangran Wang, Shalni Sundram, Indranil Gupta	There is More Control in Egalitarian Edge IoT Meshes	2025	Early Access	Internet of Things Smart devices Intelligent sensors	While mesh networking for edge settings (e.g., smart buildings, farms, battlefields, etc.) has received much attention, the layer of control over such meshes remains largely centralized and cloud-based. This paper focuses on applications with commonplace sense-trigger-actuate (STA) workloads—like the abstraction of routines popular now in smart homes, but applied to larger-scale edge IoT deployments. We present CoMesh, which tackles the challenge of building a decentralized mesh-based control plane for local, non-cloud, and hubless management of sense-trigger-actuate applications. CoMesh builds atop an abstraction called the coterie, which spreads STA load in a finegrained way both across space and across time. A coterie uses a novel combination of techniques such as zero-message-exchange protocols (for fast proactive member selection), quorum-based agreement, and locality-sensitive hashing. We analyze and theoretically prove safety and liveness properties of CoMesh. Our evaluation with both a Raspberry Pi-4 deployment and largerscale simulations, using real building maps and real routine workloads, shows that CoMesh is load-balanced, fast, faulttolerant, and scalable.	10.1109/TNSM.2025.3608796
Anna Volkova, Julian Schmidhuber, Hermann de Meer, Jacek Rak	Design of Weather-Resilient Satellite-Terrestrial ICT Networks for Power Grid Communications	2025	Early Access	Power grids Satellites Meteorology	Hybrid satellite-terrestrial communication networks can enhance the resilience of power grid communications. Recent advancements in low-Earth orbit (LEO) satellite technologies have improved their ability to meet the communication requirements of power grid applications. However, the dynamic nature of LEO networks necessitates frequent routing updates, which can potentially disrupt the transmission of critical power grid monitoring and control data. Additionally, extreme weather events, such as severe rainfall, can impair both terrestrial and satellite communication links, posing risks to the operation of the power grid. This paper presents a two-phase methodology for reducing the need for frequent routing updates by identifying stable low-latency configurations of hybrid satellite-terrestrial communication networks for power grid applications. In the proactive phase, the deterministic dynamics of LEO satellite constellations are considered to generate a sequence of stable network configurations using fine-grained temporal snapshots and graph aggregation. The adaptive phase incorporates a dynamic regional weather model to update link capacities. A minimum-delay multi-commodity flow problem is solved to determine the best traffic distribution under given conditions. Simulation results show that hybrid networks with stable configurations can reduce network reconfiguration frequency by 92%. Compared to terrestrial-only networks, the hybrid network improves end-to-end delay by 65.5% and maintains approximately 80% connectivity even under extreme rainfall conditions.	10.1109/TNSM.2025.3608855
Mounir Bensalem, Admela Jukan	Signaling Rate and Performance of RIS Reconfiguration and Handover Management in Next Generation Mobile Networks	2025	Early Access	Handover Protocols Analytical models	We consider the problem of signaling rate and performance for control and management of reconfigurable intelligent surfaces (RISs) in next-generation mobile networks. To this end, we first analytically determine the rates of RIS reconfigurations and handover using a stochastic geometry network model. We derive closed-form expressions of these rates, while taking into account static obstacles (both known and unknown), self-blockage, RIS location density, and variations in the angle and direction of user mobility. Based on the derived rates, we analyze the signaling rates of a sample novel signaling protocol, which we propose as an extension of the current handover signaling protocol. We evaluate the signaling overhead due to RIS reconfigurations and the related energy consumption. We also provide a capacity planning analysis of the related RIS control plane server for its dimensioning in the network management system. The results quantify the impact of known and unknown obstacles on the RIS reconfiguration rate and the handover rate as a function of device density and mobility. We evaluate the scalability of the model, the related signaling overhead, energy efficiency, and server capacity in the control plane. To the best of our knowledge, this is the first analytical model to derive the closed form expressions of RIS reconfiguration rates, along with handover rates, and relate its statistical properties to the signaling rate and performance in next-generation mobile networks.	10.1109/TNSM.2025.3608077
Zhuolun Li, Srijoni Majumdar, Evangelos Pournaras	Send Message to the Future? Blockchain-Based Time Machines for Decentralized Reveal of Locked Information	2025	Early Access	Cryptography Proposals Proof of Work	Conditional information reveal systems automate the release of information upon meeting specific predefined conditions, such as a designated time in the future. This paper presents a new practical timed-release cryptography system that “sends messages in the future" with highly accurate decryption times. The core of the proposed system is a novel secret sharing scheme with verifiable information reveal, and a data sharing system is devised on smart contracts. This paper also introduces a breakthrough in the understanding, design, and application of conditional information reveal systems that are highly secure and decentralized. A complete evaluation portfolio is provided to this pioneering paradigm, including analytical results, a validation of its robustness in the Tamarin Prover and a performance evaluation of a real-world, open-source system prototype deployed across the globe. Using real-world election data, we also demonstrate the applicability of this innovative system in e-voting, illustrating its capacity to secure and ensure fair elections.	10.1109/TNSM.2025.3604833
Yan Wang, Sheng Cao, Jingwei Li, Xiaosong Zhang	A Unified Framework for Hybrid Network Intrusion Detection	2025	Early Access	Detectors Biological system modeling Feature extraction	Lately, hybrid network intrusion detection systems (HNIDSs) have progressed significantly. Through the cascade or ensemble of multiple machine learning models, HNIDS benefits from each model and achieves better performance. A widely adopted framework for designing HNIDS consists of two models: a misuse detector and an anomaly detector. However, (1) benign traffic must be analyzed by both models, reducing inference speed; (2) the misuse detector performs dual functionalities, leading to suboptimal accuracy; (3) deploying the misuse and anomaly detectors on two devices introduces substantial latency and restricts distributed deployment. In this paper, we propose a unified framework called AUF. To solve (1), we deploy the anomaly detector in the first stage rather than the second, which improves inference speed. To solve (2), we employ two independent models to implement the misuse detector’s functionality, enhancing overall accuracy. To solve (3), we ensure that the different models operate independently, supporting distributed deployment. To demonstrate the effectiveness of the AUF framework, we implement XGBoost for detection and classification and propose an adaptive k-nearest neighborhood-based approach to achieve accurate discrimination. We also introduce zero-shot learning to showcase the framework’s customized model. Extensive experiments validate the effectiveness of the AUF framework and methods. Our code is available at https://github.com/wangyann2000/A-Unified-Framework-for-Hybrid-Network-Intrusion-Detection.	10.1109/TNSM.2025.3609854
Ayman Younis, Chuanneng Sun, Dario Pompili	Communication-Efficient Disaggregated and Distributed Federated Learning in NG-RANs	2025	Early Access	Training Servers Privacy	Next Generation Radio Access Networks (NG-RANs) are a promising paradigm for meeting 6G and future application requirements. However, the practical implementation of NG-RAN systems faces significant challenges due to novel technologies, network densification, and more complex applications. Specifically, the limited capacity of front-haul links and privacy concerns have posed severe constraints that must be addressed. To overcome these obstacles, we present a novel approach, called FedBNG, which is a disaggregated and distributed Federated Learning (FL)-based algorithm for NG-RAN. This algorithm enables collaboration between User Equipment (UEs) and the NG-RAN infrastructure through a learning process and shared prediction models, ultimately improving privacy and alleviating the burden on the front-haul interface. Using a shared predictive model, our proposed approach facilitates cooperative learning between Radio Units (RUs) and Distributed Units (DUs). To accomplish this, we initially used the first-phase training models of RUs and DUs as input for local training. Subsequently, the suboptimal DU models are uploaded to the Central Unit (CU) for the next phase of global training. We present numerical results to evaluate the efficacy of our proposed approach in terms of accuracy, service latency, and traffic volume. Our algorithm’s convergence properties demonstrate that it outperforms the current state-of-the-art solution based on FedAvg.	10.1109/TNSM.2025.3611235
Ori Mazor, Ori Rottenstreich	Understanding the Blockchain Interoperability Graph Based on Cryptocurrency Price Correlation	2025	Early Access	Blockchains Cryptocurrency Correlation	Cryptocurrencies have gained high popularity in recent years, with over 9000 of them, including major ones such as Bitcoin and Ether. Each cryptocurrency is implemented on one blockchain or over several such networks. Recently, various technologies known as blockchain interoperability have been developed to connect these different blockchains and create an interconnected blockchain ecosystem. This paper aims to provide insights on the blockchain ecosystem and the connection between blockchains that we refer to as the interoperability graph. Our approach is based on the analysis of the correlation between cryptocurrencies implemented over the different blockchains. We examine over 4800 cryptocurrencies implemented on 76 blockchains and their daily prices over a year. This experimental study has potential implications for decentralized finance (DeFi), including portfolio investment strategies and risk management.	10.1109/TNSM.2025.3611309
Shixuan Xian, Tuanfa Qin, Ming Yan, Wenhao Guo, Zhanyong Zhang, Weiyu Gu, Junjiang Chen, Yongle Hu	Energy-Aware Deployment of Parallelized SFCs With Delay and Reliability Constraints For Smart Firefighting	2025	Early Access	Delays Reliability Servers	Smart firefighting utilizes advanced information technologies to enhance disaster prevention and emergency response capabilities. The deployment of Service Function Chains (SFCs) in smart firefighting involves orchestrating a series of Virtual Network Functions (VNFs) as software instances on edge servers, enabling flexible and efficient service provisioning for firefighting applications. Security-enhancing VNFs, such as data encryption modules and firewalls, are deployed to strengthen the security of the smart firefighting system. A significant challenge in smart firefighting SFCs is ensuring strict low-delay and reliability constraints while optimizing energy efficiency. Traditional sequential SFCs introduce substantial end-to-end delays, rendering them unsuitable for delay-sensitive applications. Parallelized SFC addresses this issue by multiple independent VNFs in an SFC to run in parallel. In this paper, the deployment of parallelized SFCs for smart firefighting is formulated as an Integer Linear Programming (ILP) model. Since the NP-hard nature of the ILP model, we proposed a heuristic scheme named EADRC to minimize energy consumption while satisfying reliability and end-to-end delay constraints. Extensive simulations demonstrate the effectiveness of EADRC, achieving significant reductions in energy consumption, as well as improved SFC acceptance ratio and reliability compared to baseline approaches.	10.1109/TNSM.2025.3611491
Zhijie Wang, Zhen Zhang, Tengjiao He, Hao Xie	LPCD: A Parallel Candidate Deployment Strategy in Stateful Serverless Computing with Low Latency	2025	Early Access	Serverless computing Parallel processing Costs	Serverless computing has been widely regarded as an ideal computing paradigm, enabling edge servers to host serverless functions. Due to its high scalability and usage-based pricing model, it provides efficient services across various applications. However, in the deployment process of serverless applications, past works lack considerations for the parallel relationships between stateful functions, which increases end to end latency. To leverage the parallel dependencies between functions, we propose a strategy for dependent function parallelization deployment, named LPCD (Low latency Parallel Candidate Deployment strategy). By partitioning the problem into inter-layer function deployment and analyzing optimal substructures, a heuristic algorithm is introduced to determine candidate deployment strategies for each layer of the users, which aims at identifying the optimal edge server for each function instance during deployment to enhance user satisfaction. Through simulation experiments, we evaluate the performance of the strategy. The experiments results indicate that the average latency was reduced by at least 41% compared with the state-of-the-art strategies.	10.1109/TNSM.2025.3611445
Xianchao Zhang, Qinghua Zhang, Jia Chen, Deyun Gao, Shuxiao Ye, Hongke Zhang	Joint Optimization of Task Planning and Service Function Chain Scheduling in the UAVs Networks	2025	Early Access	Processor scheduling Disasters Delays	Natural disasters pose a significant threat to human life. In these extreme conditions, terrestrial networks frequently become incapacitated, hindering the provision of essential communication and computing services required for emergency response efforts. In recent years, the rapid advancement of drone technology, coupled with the maturation of lightweight communication and computing equipment, has led to the emergence of unmanned aerial vehicle (UAV) networks as a crucial asset in disaster rescue missions. These networks provide significant advantages, including rapid response times, flexible deployment capabilities, and heightened resilience to complex terrains, showcasing considerable potential for further development. UAV networks exemplify resource-constrained systems where efficient scheduling of computing resources is vital. Especially in emergency rescue scenarios, this complexity is exacerbated by the diverse range of tasks, varying demands, and stringent real-time requirements. Effectively managing and allocating the computational resources of drones is essential for maximizing their operational efficiency in response to the intricate dynamics of disaster situations. To improve the computational service efficiency of the network, this paper proposes an emergency rescue UAVs network architecture. Additionally, we investigate a joint optimization approach for task planning and SFC scheduling. Current research on SFC scheduling primarily focuses on ground data center networks, with comparatively limited investigation into UAV networks. Fully considering the mobility of computing nodes, as well as the wireless transmission modes within the aerial environment, we establish a joint optimization model for task planning and SFC scheduling aiming at minimizing the total weighted end-to-end delay. Then we design the A3C based algorithm to learn the optimization strategy. Simulation results are presented to demonstrate the superiority of the proposed approach in the aspect of total weighted end to end delay and training time against other benchmark algorithms.	10.1109/TNSM.2025.3611921
Sankalp Mittal, Praveen Tammana	Efficient In-Network Traffic Classification Using Programmable Switches With AdaFlow	2025	Early Access	Accuracy Pipelines Memory management	In-network ML-based traffic classification using programmable switches has enabled faster decisions and reduced the cost of the security infrastructure and management overheads. However, due to constraints on per-packet operations and limited stateful memory in the switch data plane, there is a fundamental tradeoff between traffic classification accuracy and switch memory requirements. Existing works fall short of accurately classifying traffic with diverse flow characteristics while keeping the memory footprint low. In this paper, we propose AdaFlow, a system that aims to address this gap by incorporating traffic-specific heuristics while designing the in-network classifier. We evaluate the AdaFlow prototype via simulations and also on a testbed with an Intel Barefoot Tofino switch. Compared to the state-of-the-art, AdaFlow improves accuracy up to 7% for various use-cases while keeping the memory overheads similar to or lower than those of the existing systems.	10.1109/TNSM.2025.3607406
Liang Feng, Cunqing Hua, Lingya Liu, Jianan Hong	Quality and Diversity Balanced Neighbor Selection against Eclipse Attack in Blockchain System	2025	Early Access	Peer-to-peer computing Blockchains Bitcoin	Blockchain technology has gained widespread adoption across diverse applications; however, its peer-to-peer network architecture remains susceptible to eclipse attacks via malicious neighbor infiltration. Existing defense mechanisms typically rely either on historical data to detect attacks post hoc or on diverse neighbor selection to prevent them. These approaches, however, exhibit critical limitations: detection-based strategies are inherently reactive, while diversity-based selection lacks rigorous quantitative models to characterize the differences between neighbors. To bridge this gap, this paper proposes integrating data analysis techniques directly into the neighbor selection process. Specifically, the proposed method dynamically evaluates peers’ block propagation performance and quantifies inter-peer differences using Wasserstein distance metrics. This enables the application of determinantal point processes (DPPs) to select an optimal set of high-performing and diverse peers as neighbors. Empirical evaluation utilizing Bitcoin network data demonstrates that the proposed scheme simultaneously achieves adaptive neighbor selection and robust protection against eclipse attacks.	10.1109/TNSM.2025.3612386
Indukuri Mani Varma, Neetesh Kumar	Blockchain-Based SDN-Enabled Lightweight Authentication Protocol for IoV Using zk-SNARK	2025	Early Access	Authentication Protocols Handover	Software-defined networking (SDN), an emerging networking paradigm, has been realized in the internet of vehicles to effectively manage the dynamic nature of vehicles and provide diverse vehicle-to-everything (V2X) services and security. Nevertheless, the lack of authentication mechanisms and malicious behavior of vehicle users has seriously threatened security and privacy concerns in the SDN-based vehicular network (SDVN). Furthermore, in a distributed SDVN, the SDN controllers (SDNCs) are prone to single point of failure (SPoF) due to various potential attack vectors. Such attacks can result in the loss of periodically exchanged vehicular information among SDNCs. To address the aforementioned issues, a novel privacy-preserving lightweight zk-SNARK-based authentication protocol in blockchain-enabled distributed SDVN is presented. To address the SPoF of SDNCs, the proposed network model removes the dependency of a controller on neighboring SDNCs as the authenticated information is validated and stored in the blockchain. As an integral part of the protocol, the vehicles are authenticated by providing succinct and constant size proofs during authentication across SDVN domains. The security and simulation performance analyses confirm the superior performance of the proposed protocol, surpassing state-of-the-art schemes in terms of authentication latency, computation time, and energy efficiency. The proposed protocol has achieved notable improvements of more than 55% in all the metrics considered during the protocol simulation when compared with other protocols.	10.1109/TNSM.2025.3613415
Mengyao Li, Noah Ploch, Sebastian Troia, Carlo Spatocco, Wolfgang Kellerer, Guido Maier	On the Optimization of Model Aggregation for Federated Learning at the Network Edge	2025	Early Access	Computational modeling Training Wide area networks	The rapid increase in connected devices has significantly intensified the computational and communication demands on modern telecommunication networks. To address these challenges, integrating advanced Machine Learning (ML) techniques like Federated Learning (FL) with emerging paradigms such as Multi-access Edge Computing (MEC) and Software-Defined Wide Area Networks (SD-WANs) is crucial. This paper introduces online resource management strategies specifically designed for FL model aggregation, utilizing intermediate aggregation at edge nodes. Our analysis highlights the benefits of incorporating edge aggregators to reduce network link congestion and maximize the potential of edge computing nodes. However, the risk of network congestion persists. To mitigate this, we propose a novel aggregation approach that deploys an aggregator overlay network. We present an Integer Linear Programming (ILP) model and a heuristic algorithm to optimize the routing within this overlay network. Our solution demonstrates improved adaptability to network resource utilization, significantly reducing FL training round failure rates by up to 15% while also alleviating cloud link congestion.	10.1109/TNSM.2025.3613772