Computer Networks最新文献

{"title":"Federated learning-blockchain synergy for multi-objective replication optimization in dynamic cloud-edge networks","authors":"Peng Xiao,&nbsp;Saifeng Zeng","doi":"10.1016/j.comnet.2026.112014","DOIUrl":"10.1016/j.comnet.2026.112014","url":null,"abstract":"<div><div>The evolution of cloud-edge computing has intensified the need for efficient data replication strategies to balance latency, resource efficiency, and adaptability in dynamic environments. Existing approaches often prioritize isolated objectives, neglect privacy risks, and lack scalability under fluctuating network conditions and node churn. This paper introduces RPMO, a multi-objective replication algorithm integrating Lagrangian Relaxation and adaptive bound adjustments to optimize latency, energy consumption, and coverage systematically. The framework is further enhanced by FL-MORP, a federated learning model that embeds multi-objective optimization into distributed training and blockchain-based trust management, enabling proactive replication while preserving data privacy and auditability with minimal overhead. Experimental evaluations on a hybrid cloud-edge testbed demonstrate that FL-MORP reduces worst-case latency by 18–50 %, energy consumption by 13–27 %, and convergence time by 26 % compared to baselines.</div></div>","PeriodicalId":50637,"journal":{"name":"Computer Networks","volume":"277 ","pages":"Article 112014"},"PeriodicalIF":4.6,"publicationDate":"2026-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145981820","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}

{"title":"V-FedMM: Dynamic sample selection for efficient multimodal federated learning over vehicular networks","authors":"Haoyu Tu ,&nbsp;Wen Wu ,&nbsp;Liang Li ,&nbsp;Yongguang Lu ,&nbsp;Lin Chen ,&nbsp;Xu Chen","doi":"10.1016/j.comnet.2026.112020","DOIUrl":"10.1016/j.comnet.2026.112020","url":null,"abstract":"<div><div>In vehicular networks, federated learning (FL) using multimodal sensory data from connected vehicles is critical for emerging autonomous driving applications, such as trajectory prediction, object detection, and semantic segmentation. However, imbalanced learning contributions across modalities and heterogeneous computation/communication capabilities of vehicles pose fundamental challenges, leading to biased model training, slow convergence, and inefficient resource utilization. In this paper, we propose V-FedMM, a novel framework for <u>v</u>ehicular <u>fed</u>erated learning with <u>m</u>ulti<u>m</u>odal data that optimizes both sample and vehicle selection. Particularly, sample selection follows a two-stage mechanism: (i) modality-level selection, where the server determines vehicle participation and per-modality sample counts based on onboard data distribution and modality-specific training contributions; and (ii) sample-level selection, where each vehicle selects specific multimodal data samples by balancing sample importance and utilization to strategically incentivize the participation of low-contribution data. We theoretically analyze the impact of varying modality contributions on training performance by deriving an upper bound for the loss function. To determine the optimal per-modality sample counts for each vehicle under a strict per-round delay constraint, we formulate a joint vehicle selection and per-modality sample count optimization problem that maximizes the contribution from all selected samples. We solve optimization problem by first selecting vehicles guided by a theoretical property, then solving the remaining sample allocation problem as an integer linear programming (ILP) problem with the branch and bound algorithm. Extensive simulation results demonstrate the superior performance of V-FedMM, achieving nearly 5% higher accuracy compared to conventional vehicular FL approaches and saving 16.8% computation overhead compared to the conventional sampling algorithm.</div></div>","PeriodicalId":50637,"journal":{"name":"Computer Networks","volume":"277 ","pages":"Article 112020"},"PeriodicalIF":4.6,"publicationDate":"2026-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146039704","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}

{"title":"Energy minimization for cellular-connected UAV-aided inspection systems: A dual radio map-driven hierarchical optimization algorithm","authors":"Fahui Wu ,&nbsp;Shi Peng ,&nbsp;Wei Wang ,&nbsp;Jiangling Cao ,&nbsp;Dingcheng Yang ,&nbsp;Sihan Fu ,&nbsp;Xiaoli Ye","doi":"10.1016/j.comnet.2026.112027","DOIUrl":"10.1016/j.comnet.2026.112027","url":null,"abstract":"<div><div>The burgeoning development of low-altitude economy networks (LAEN) relies heavily on the efficient and sustainable operation of cellular-connected unmanned aerial vehicles (UAVs) for critical services, particularly in infrastructure inspection. While UAV-aided inspection offers significant advantages over traditional inspection methods, the challenge of minimizing total system energy consumption remains paramount, constrained by both limited onboard battery capacity and highly dynamic low-altitude wireless networks (LAWN). To address this issue, a cellular-connected UAV-aided inspection system is investigated, and a dual radio map-driven hierarchical optimization algorithm is proposed to enhance the system´s energy efficiency. Initially, urban map data and relevant transmitter information are utilized to construct a dual radio map, which comprises the communication outage probability (COP) map and the reference signal received power (RSRP) map. Subsequently, leveraging this a priori channel knowledge, the A star (A*) algorithm is employed to optimize the UAV’s trajectory between any two inspection points, thereby stabilizing the communication link. Finally, the optimized trajectory information is fed into the particle swarm optimization (PSO) algorithm to determine the optimal task assignment strategy for the energy-constrained UAV, achieving the overall energy minimization objective. The numerical results validate that our proposed algorithm achieves lower energy consumption compared to benchmark schemes, while simultaneously guaranteeing the required communication quality for the low-altitude UAV.</div></div>","PeriodicalId":50637,"journal":{"name":"Computer Networks","volume":"277 ","pages":"Article 112027"},"PeriodicalIF":4.6,"publicationDate":"2026-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146039284","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}

{"title":"ControlShard: A cross-shard transaction processing method based on congestion control strategy","authors":"Junfeng Tian,&nbsp;Yuyang Zhao,&nbsp;Jin Tian","doi":"10.1016/j.comnet.2026.112028","DOIUrl":"10.1016/j.comnet.2026.112028","url":null,"abstract":"<div><div>Sharding technology significantly enhances blockchain performance by dividing nodes into multiple groups to propose blocks in parallel. However, cross-shard transactions require timely communication between shards to ensure information consistency, which has a great impact on the performance of sharded blockchains. To address this challenge, we propose ControlShard, which builds shards specifically for cross-shard transactions and allows users to have different account status in different shards, further reducing communication overhead. ControlShard allows multiple blocks to be committed in a single consensus round and employs a congestion control strategy to determine the maximum number of blocks suitable for the current network conditions. Additionally, ControlShard employs a pipelining strategy to further enhance parallel processing capabilities. We implemented a system prototype and evaluated its performance using real Ethereum transactions. Under the experimental conditions specified in this paper, ControlShard outperforms existing advanced solutions in terms of throughput and transaction confirmation latency, achieving a throughput up to 2.2 times that of BrokerChain and 4.1 times that of Monoxide.</div></div>","PeriodicalId":50637,"journal":{"name":"Computer Networks","volume":"277 ","pages":"Article 112028"},"PeriodicalIF":4.6,"publicationDate":"2026-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146039289","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}

{"title":"JUST: Cost-efficient joint cluster upgrade sequencing and task scheduling for containerized edge computing","authors":"Yuting Hu ,&nbsp;Xuan Xie ,&nbsp;Zhiqing Tang ,&nbsp;Wenmian Yang ,&nbsp;Jianxiong Guo ,&nbsp;Tian Wang ,&nbsp;Weijia Jia","doi":"10.1016/j.comnet.2026.112065","DOIUrl":"10.1016/j.comnet.2026.112065","url":null,"abstract":"<div><div>The rise of Edge Computing (EC) has led to an increased deployment of applications on edge servers using containers, which rely on multi-layer image files that often need to be upgraded. While these upgrades are essential, traditional cloud upgrade strategies are not suitable for EC due to resource constraints, geographical distribution, and latency sensitivity. Consequently, efficiently updating containers in EC is a significant challenge. In this paper, we present the online layer-aware cost-efficient <u>J</u>oint cluster <u>U</u>pgrade <u>S</u>equencing and <u>T</u>ask scheduling (JUST) algorithm to quickly and efficiently address the upgrades in EC. Our JUST algorithm comprises two main mechanisms. First, we model the edge cluster upgrade problem by evaluating both short-term and long-term node upgrade costs along with task priorities, establishing a cluster upgrade sequencing mechanism to minimize disruptions to ongoing tasks. Second, we introduce a layer-aware task scheduling mechanism that reduces initial scheduling and upgrade-related rescheduling latency by optimizing layer transmission, downloads, and computation time. This approach minimizes upgrade disruptions by integrating node resources, layer distribution, and task characteristics through reinforcement learning and adaptive factorization networks. Finally, experiments with simulated and real datasets show that our JUST algorithm outperforms baseline methods, achieving a 47% improvement in total latency.</div></div>","PeriodicalId":50637,"journal":{"name":"Computer Networks","volume":"277 ","pages":"Article 112065"},"PeriodicalIF":4.6,"publicationDate":"2026-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147398083","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}

{"title":"AoI-minimal trajectory planning and data collection in UAVs crowdsensing by graph reinforcement learning","authors":"Jiang Tong ,&nbsp;Yubo Wang ,&nbsp;Tiancong Huang ,&nbsp;Xu Zhao ,&nbsp;Guojin Liu","doi":"10.1016/j.comnet.2026.112056","DOIUrl":"10.1016/j.comnet.2026.112056","url":null,"abstract":"<div><div>In multi-Unmanned Aerial Vehicles (UAVs) data collection scenarios, it remains a critical challenge to jointly optimize time-sensitive responsiveness and maximize the volume of collected data. To address this issue, a graph reinforcement learning (RL) framework termed STIG-RL is proposed. The framework integrates an attention-based spatio-temporal interactive graph (ASTIG) network to extract efficient graph representations of the UAV-user state. Additionally, a hybrid decision-making strategy that combines model-based and model-free RL is employed to enhance policy efficiency. To further improve learning performance and reduce reliance on real-world experience, a simulated trajectory generation module (SRM) with a state prediction mechanism is incorporated. Extensive experimental evaluations demonstrate that the proposed approach outperforms five baseline algorithms in terms of data collection efficiency and responsiveness.</div></div>","PeriodicalId":50637,"journal":{"name":"Computer Networks","volume":"277 ","pages":"Article 112056"},"PeriodicalIF":4.6,"publicationDate":"2026-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147398084","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}

{"title":"Attribute-based searchable encryption for secure data sharing with blockchain in VANETs","authors":"Zhenqi Wang ,&nbsp;Peng Yang ,&nbsp;Angze Du ,&nbsp;Dongmei Yang ,&nbsp;Xiaoyan Liu","doi":"10.1016/j.comnet.2026.112026","DOIUrl":"10.1016/j.comnet.2026.112026","url":null,"abstract":"<div><div>Vehicular Ad Hoc Networks (VANETs) are critical for enhancing road safety and optimizing traffic flow. Attribute-based searchable encryption (ABSE) with blockchain supports one-to-many encryption, ciphertext search, and decentralized storage, providing a solution to enable data sharing in VANETs. However, existing VANETs schemes face many limitations: blockchain miner nodes are at risk of stealing vehicle attribute keys during key verification; the lack of permission control may also lead to unauthorized downloads; and multi-keyword search, encryption, and decryption require a lot of computing power. To address these issues, we propose an attribute-based searchable encryption scheme for secure data sharing with blockchain (ABSE-VANETs). Specifically, we introduce multiple authorization authorities to manage attributes and generate obfuscated key parameters via bilinear pairing to prevent the leakage of vehicle attribute keys. We combine linear integer secret sharing (LISS) to bind the ciphertext download permission and the vehicle attribute set to prevent unauthorized ciphertext download. Moreover, we optimize the multi-keyword index generation algorithm and trapdoor structure and construct a multi-keyword index with the help of blockchain, which achieves the fast positioning of the ciphertext address on the chain. Finally, we utilize the edge cloud server to realize proxy decryption of ciphertexts, which further reduces the computation pressure on the vehicle terminal. Formal security analysis and performance comparison experiments show that our scheme is highly secure under the DBDH assumption and performs well in terms of communication and computation efficiency.</div></div>","PeriodicalId":50637,"journal":{"name":"Computer Networks","volume":"277 ","pages":"Article 112026"},"PeriodicalIF":4.6,"publicationDate":"2026-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146039291","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}

{"title":"Dynamic task offloading and resource allocation with emergency/general task coexistence in vehicle edge computing","authors":"Yaoping Zeng,&nbsp;Shisen Chen,&nbsp;Yimeng Ge","doi":"10.1016/j.comnet.2026.112023","DOIUrl":"10.1016/j.comnet.2026.112023","url":null,"abstract":"<div><div>Vehicle edge computing (VEC) meets the growing computational demands of vehicles by bringing computing resources closer to the vehicle through roadside units (RSUs). However, the types of tasks to be handled in different vehicular applications are different, such as streaming media, where tasks to be handled are energy-sensitive, termed by general tasks (GTs), while in autonomous driving, tasks are latency-sensitive, called by emergency tasks (ETs). Therefore, how to effectively utilize system resources to satisfy these heterogeneous demands has become a current challenge. In this paper, we propose a novel algorithm aiming to minimize the average total cost, and establish a joint optimization objective function for the energy cost of GTs and the processing delay of ETs. Considering that the proposed problem is a long-term stochastic optimization problem, we leverage the Lyapunov optimization to transform it into two deterministic online optimization subproblems based on variable types. In the subproblem related to discrete variables, namely transmit power control and channel selection, we innovatively establish it as a potential game model by utilizing the linear combination of exact potential game and ordinary potential game, and prove the existence of the Nash equilibrium. Another subproblem involves continuous variables, i.e., RSU computing resource allocation, and we obtain their optimal solutions using convex optimization theory. Finally, numerical results show that the proposed algorithm outperforms other baseline algorithms.</div></div>","PeriodicalId":50637,"journal":{"name":"Computer Networks","volume":"277 ","pages":"Article 112023"},"PeriodicalIF":4.6,"publicationDate":"2026-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146039283","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}

{"title":"Decentralized federated learning with vertical model parallelism and topology construction","authors":"Yang Xu ,&nbsp;Zuan Xie ,&nbsp;Hongli Xu ,&nbsp;Yunming Liao ,&nbsp;Zhiwei Yao","doi":"10.1016/j.comnet.2026.112015","DOIUrl":"10.1016/j.comnet.2026.112015","url":null,"abstract":"<div><div>In order to avoid the possible bottleneck of conventional parameter server architecture, the decentralized federated learning (DFL) is developed on the peer-to-peer (P2P) communication. However, due to the intrinsic features of edge computing (EC), such as limited and heterogeneous computing capability, constrained bandwidth and non-IID data, DFL with a unified model structure and fixed communication topology results in a slow convergence rate. To address these challenges, we adopt the vertical model parallelism and propose an efficient model training framework called SWIFT. Specifically, we divide the complete model into multiple non-overlapping columns, and allow each client to train a sub-model with different length of column group according to its computing capacity. We theoretically analyze the convergence upper bound of SWIFT, and reveal that the convergence rate will be impacted by the consensus distance among different clients. To reduce the consensus distance and accelerate training, we propose a dynamic topology construction algorithm by selecting appropriate neighbors for each client before model aggregation. Experimental results demonstrate that, compared with the baseline methods, SWIFT can approximately achieve a <strong>2.3 × </strong> speedup and save <strong>46%</strong> communication traffic.</div></div>","PeriodicalId":50637,"journal":{"name":"Computer Networks","volume":"277 ","pages":"Article 112015"},"PeriodicalIF":4.6,"publicationDate":"2026-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146039286","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}

{"title":"FederNet: A network and device-aware emulation platform for federated learning benchmarking","authors":"Antonio Boiano ,&nbsp;Marta Avanzini ,&nbsp;Mattia Brambilla ,&nbsp;Monica Nicoli ,&nbsp;Alessandro E.C. Redondi","doi":"10.1016/j.comnet.2026.111995","DOIUrl":"10.1016/j.comnet.2026.111995","url":null,"abstract":"<div><div>Federated Learning (FL) has emerged as a pivotal privacy-preserving machine learning paradigm, enabling collaborative model train across distributed data sources. A main issue, however, is the lack of comprehensive testing environments that can accurately emulate real-world conditions at a large scale, particularly the impact of network dynamics and device capabilities on FL algorithm performance. To this end, we introduce FederNet, a novel platform designed to facilitate the development and testing of FL algorithms with realistic network and device emulation. We show how the proposed system provides a versatile platform for researchers to evaluate the performance, robustness, and scalability of FL algorithms under diverse and configurable scenarios. We describe the FederNet architecture, detail its network and device emulation capabilities, and outline potential use cases that demonstrate its utility in advancing FL research. By bridging the gap between algorithmic development and practical deployment challenges, FederNet aims to accelerate the innovation and adoption of FL technologies.</div></div>","PeriodicalId":50637,"journal":{"name":"Computer Networks","volume":"277 ","pages":"Article 111995"},"PeriodicalIF":4.6,"publicationDate":"2026-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145981827","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}