Computer Networks最新文献

{"title":"ESTK-JC: Encrypted malicious traffic detection fusing spatio-temporal features and JointCloud entity knowledge for JointCloud environment","authors":"Rongwei Yu ,&nbsp;Zijian Zhou ,&nbsp;Yuhao Zhang ,&nbsp;Lina Wang ,&nbsp;Qiyun Shao","doi":"10.1016/j.comnet.2025.111495","DOIUrl":"10.1016/j.comnet.2025.111495","url":null,"abstract":"<div><div>JointCloud computing is a new cloud computing paradigm that enables interconnection between clouds. Compared with other network environments, large-scale diverse data interactions and data collaboration have become the norm in the JointCloud environment based on JointCloud computing. The complex behaviors between entities in the JointCloud environment are realized through the interaction of network traffic. Attackers often mix malicious traffic with benign network traffic to break the JointCloud ecosystem. Among them, encrypted malicious traffic poses a huge security risk to JointCloud environments due to its strong invisibility and fast propagation speeds. Currently, there is a gap in researches on malicious traffic detection in the JointCloud environment, especially encrypted malicious traffic detection. To the best of our knowledge, this paper is the first research work to conduct encrypted malicious traffic detection for JointCloud environment. Specifically, to cope with the complex network traffic data in the JointCloud environment, this paper proposes a spatio-temporal feature extraction method for encrypted traffic to enrich the features of the original traffic. Subsequently, we propose a method of mining JointCloud entity knowledge based on the characteristics of traffic distribution in the JointCloud environment, which can improve detection performance. Finally, we construct an encrypted malicious traffic detection model fusing spatio-temporal features and JointCloud entity knowledge for JointCloud environment (ESTK-JC). In experiments in a simulated JointCloud environment, ESTK-JC exhibits detection performance superior to current state-of-the-art models.</div></div>","PeriodicalId":50637,"journal":{"name":"Computer Networks","volume":"270 ","pages":"Article 111495"},"PeriodicalIF":4.4,"publicationDate":"2025-07-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144534289","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":"nnWeb: Towards efficient WebGPU-based DNN inference via automatic collaborative offloading","authors":"Jiawei Liu ,&nbsp;Bing Dong ,&nbsp;Weilong Wang ,&nbsp;Muhan Yuan ,&nbsp;Borui Li ,&nbsp;Zhao-Dong Xu ,&nbsp;Shuai Wang","doi":"10.1016/j.comnet.2025.111489","DOIUrl":"10.1016/j.comnet.2025.111489","url":null,"abstract":"<div><div>In-browser neural network inference offers the promise of cross-platform AI applications, but faces severe latency and energy challenges on resource-constrained devices. In this paper, we present nnWeb, a WebGPU-based in-browser neural network inference framework with optimized latency and energy efficiency. nnWeb dynamically partitions neural network and facilitates the collaborative offloading between client browser and server. nnWeb operates in two phases: (1) <em>layer-wise isolation-based profiling</em>, which is used to predict per-layer execution latency and energy on heterogeneous hardware; and (2) <em>asynchronous execution-based DNN partitioning,</em> which continuously monitors network bandwidth and device load to select the optimal partition point using WebGPU’s native pipeline parallelism, minimizing total latency or energy consumption by solving a closed-form optimization at runtime. Extensive evaluation on various in-browser AI models and networking conditions shows that nnWeb achieves an average reduction of 30% to 52% in total inference latency compared with static partitioning. Moreover, nnWeb realizes energy savings ranging from 11.3% to 44.0% in contrast to standalone browser inference.</div></div>","PeriodicalId":50637,"journal":{"name":"Computer Networks","volume":"270 ","pages":"Article 111489"},"PeriodicalIF":4.4,"publicationDate":"2025-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144534290","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":"Cache-assisted task offloading strategy based on multi-agent deep reinforcement learning","authors":"Mande Xie,&nbsp;Longchen Li,&nbsp;Xueping Ni","doi":"10.1016/j.comnet.2025.111483","DOIUrl":"10.1016/j.comnet.2025.111483","url":null,"abstract":"<div><div>In the context of 5G architecture, mobile edge computing (MEC) has emerged as a promising computing paradigm within this framework. It involves deploying servers at the edge of the network with computing and storage resources to meet the low-latency and high-energy requirements of emerging applications. While many studies have focused on task offloading decisions in MEC, the performance improvement brought by content caching has often been neglected. In this paper, the joint optimization problem of task offloading and content caching in a multi-user, multi-server MEC system is investigated. An offloading method based on multi-agent deep reinforcement learning is proposed, which integrates self-attention with multi-agent deep deterministic policy gradient (MADDPG) to determine task offloading policies. To further enhance the performance of offloading tasks, a dynamic cache decision optimization algorithm (DCDOA) is developed for more effective content caching decisions. The simulation results demonstrate that the proposed method outperforms other baseline algorithms in different scenarios. Specifically, the proposed method reduces the long-term average user cost by 45.4%, 22.2%, 10.6%, and 8%, respectively, compared to the all-cloud execution method, the random offloading method, the DDPG method, and the MADDPG method.</div></div>","PeriodicalId":50637,"journal":{"name":"Computer Networks","volume":"270 ","pages":"Article 111483"},"PeriodicalIF":4.4,"publicationDate":"2025-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144549689","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":"Joint resource allocation for high-mobility HCNs with D2D communications","authors":"Jingya Ma,&nbsp;Hongyuan Gao","doi":"10.1016/j.comnet.2025.111491","DOIUrl":"10.1016/j.comnet.2025.111491","url":null,"abstract":"<div><div>Heterogeneous cellular networks (HCNs) with high-mobility, as an important part of future mobile communication systems, can effectively integrate multiple different types of network nodes and provide more flexible and efficient communication services. However, the high-mobility and heterogeneity will bring about frequent network dynamics variation and severe interference, impacting resource utilization efficiency and Quality of Service (QoS), particularly in scenarios involving Device-to-Device (D2D) communications. Consequently, optimizing the total energy efficiency (EE) of both uplink and downlink while ensuring QoS becomes a critical challenge. We aim to optimize the total EE of uplink and downlink in high-mobility HCNs with D2D communications to meet the resource allocation (RA) demands in dynamic mobile environments. Through rigorous theoretical modeling, we formulate an energy-efficient joint RA issue that incorporates minimum data rate requirements for both uplink and downlink, as well as cross-tier interference and power constraints, making it an NP-hard hybrid optimization challenge. To address this, a novel evolution algorithm named quantum-inspired differential evolution algorithm (QDEA) is designed to achieve synchronized configuration of time-division duplexing (TDD), spectrum, and power resources. Simulation results reveal that the proposed QDEA significantly enhances EE while maintaining QoS in high-mobility HCN scenarios, showing a high adaptability under various network conditions.</div></div>","PeriodicalId":50637,"journal":{"name":"Computer Networks","volume":"270 ","pages":"Article 111491"},"PeriodicalIF":4.4,"publicationDate":"2025-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144571251","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":"A Management Data Analytics Function for ethical 6G networks","authors":"Milad Akbari ,&nbsp;Raffaele Bolla ,&nbsp;Roberto Bruschi ,&nbsp;Chiara Lombardo ,&nbsp;Nicole Simone Martinelli ,&nbsp;Beatrice Siccardi","doi":"10.1016/j.comnet.2025.111487","DOIUrl":"10.1016/j.comnet.2025.111487","url":null,"abstract":"<div><div>In order to satisfy the ethical requirements that are expected from upcoming 6G technologies, the knowledge of the power consumption ascribable to the applications and network functions is crucial to enforce a sense of responsibility and joint efforts towards value-driven sustainability. In this respect, this paper presents a modular, energy-focused prototype of the Management Data Analytics Function (MDAF), that represents the cornerstone of the Observability framework recently developed by the 6Green Project. Its main goal is to provide management data to upper layers (i.e., network and end-users/verticals). A 5G network was tested on both the User-Plane and Control-Plane; at the same time, said network was monitored by both the proposed MDAF and a common power monitoring solution: Scaphandre. Results show that the MDAF measures a higher power consumption than Scaphandre; the difference lies between 50% (when idle) and 4% (at the maximum offered load). This difference corresponds to the ”indirect” power consumption that the MDAF is able to ascribe to the containers/Virtual Machines (VMs) (i.e., the Network Functions (NFs)). The more accurate power consumption measurements of the single NFs is a first step towards the need to spread energy/carbon awareness to all the involved stakeholders.</div></div>","PeriodicalId":50637,"journal":{"name":"Computer Networks","volume":"270 ","pages":"Article 111487"},"PeriodicalIF":4.4,"publicationDate":"2025-06-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144563960","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":"DPoW: A decentralized proof-of-work consensus mechanism for blockchain system","authors":"Shasha Yu,&nbsp;Yanan Qiao,&nbsp;Fan Yang,&nbsp;Junge Bo","doi":"10.1016/j.comnet.2025.111490","DOIUrl":"10.1016/j.comnet.2025.111490","url":null,"abstract":"<div><div>Proof of Work (PoW) serves as a consensus mechanism in blockchain that verifies and selects the next node through solving computationally intensive puzzles. As competition in computation has intensified and mining difficulty has increased, individual miners have tended to join forces by forming mining pools. This allows them to generate blocks more quickly and receive a consistent portion of the block reward, rather than randomly. However, this collaboration could potentially introduce a centralization risk, as a few pools gain significant control over the network’s hash power. To alleviate the centralization in PoW mining, this research proposes an incentive mechanism named Decentralized Proof of Work (DPoW), where a dynamic reward allocation algorithm is designed and the mining rewards are dynamically allocated between the first and second block proposer. Then a dynamic evolutionary game model based on node behavior is constructed to simulate the pool evolution process. This model considers different pool selection preference towards original PoW and the proposed DPoW. Solo miners can dynamically adjust their strategies based on the expected payoff and revenue stability. Furthermore, the equilibrium stability of the model in DPoW is compared with that in original PoW. Through theoretically probing miner preferences in pool selection and assessing decentralization metrics, this research demonstrates DPoW enables the blockchain network to enjoy decentralization superiority compared to traditional PoW.</div></div>","PeriodicalId":50637,"journal":{"name":"Computer Networks","volume":"270 ","pages":"Article 111490"},"PeriodicalIF":4.4,"publicationDate":"2025-06-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144571269","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":"Device and data Heterogeneity Aware SplitFed Learning for Digital Twin empowered Industrial Internet of Things","authors":"Himel Saha ,&nbsp;Md Nur Ahmed ,&nbsp;Palash Roy ,&nbsp;Md. Abdur Razzaque ,&nbsp;Nafis Fuad Tanvir ,&nbsp;Mohammad Mehedi Hassan ,&nbsp;Md Zia Uddin","doi":"10.1016/j.comnet.2025.111478","DOIUrl":"10.1016/j.comnet.2025.111478","url":null,"abstract":"<div><div>Distributed learning methods in the Industrial Internet of Things (IIoT) face challenges due to their dynamically changing learning environment. This is mainly caused by two facts. Firstly, the complex nature of statistically heterogeneous data, often non-independent and identically distributed (non-IID), from geographically scattered clients hampers the accuracy of the training model. Secondly, the heterogeneity in computational and communication resources among IIoT devices frequently introduces instability in the training process, resulting in a delay referred to as the straggler. Existing literature addresses only the unilateral side of the heterogeneity issue but lacks comprehensive efforts to tackle the joint problem of device and data heterogeneities for the resource-constrained IIoT. To address these multifaceted challenges, in this article, we have introduced device and data Heterogeneity Aware SplitFed Learning, namely Het-SFL, a distributed learning framework designed for deployment within dynamic IIoT networks. The developed Het-SFL framework optimizes resource utilization and reduces the computational burden by dynamically assessing the optimal split point of the training model for the resource-limited devices, considering factors such as training time, device energy consumption, and model accuracy. A clustering mechanism is employed to mitigate the straggler effect and to reduce the solution space to obtain the optimal split point within a significantly shortened deadline. Furthermore, the Het-SFL framework leverages emerging Digital Twin (DT) technology to facilitate real-time analysis of heterogeneous data, thereby improving the performance of the training model in non-IID contexts. The numeric performance analysis reveals that the Het-SFL framework improves training performance in terms of accuracy, training time, and device energy consumption by up to 30%, 40%, and 55%, respectively, compared to other state-of-the-art works.</div></div>","PeriodicalId":50637,"journal":{"name":"Computer Networks","volume":"270 ","pages":"Article 111478"},"PeriodicalIF":4.4,"publicationDate":"2025-06-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144522115","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":"LAS: Lightweight Aggregate Signcryption for federated learning with blockchain in IoT","authors":"Chen Yang,&nbsp;Lei Tian,&nbsp;Xiang Wang,&nbsp;Feilong Lin,&nbsp;Riheng Jia,&nbsp;Zhonglong Zheng,&nbsp;Minglu Li","doi":"10.1016/j.comnet.2025.111480","DOIUrl":"10.1016/j.comnet.2025.111480","url":null,"abstract":"<div><div>Federated learning enables edge servers to share models instead of exchanging data, achieving high-quality model training. Nevertheless, the unreliability of communication environments presents security challenges for model transmission between edge servers. Signcryption can protect the security of the model during transmission, but existing schemes lack efficient pseudonym verification and revocation of signcryption permissions. To address these challenges, this paper proposes Lightweight Aggregate Signcryption for federated learning (LAS). LAS leverages blockchain technology for trusted storage and verification, while incorporating a pseudonym verification mechanism based on relevant proofs instead of repeated verification requests. Furthermore, we introduce a signcryption permission revocation mechanism based on the Chinese Remainder Theorem, ensuring that once a edge server is flagged as malicious, it can no longer generate valid ciphertexts. LAS also supports ciphertext aggregation and batch verification. Finally, we theoretically prove that LAS achieves IND-CCA2 and EUF-CMA security. Extensive experimental results demonstrate its feasibility and advantages in terms of computational overhead, communication overhead, and functionality.</div></div>","PeriodicalId":50637,"journal":{"name":"Computer Networks","volume":"270 ","pages":"Article 111480"},"PeriodicalIF":4.4,"publicationDate":"2025-06-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144534288","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":"Leakage-resilient certificateless public key encryption with equality test resistant to side-channel attacks","authors":"Tung-Tso Tsai,&nbsp;Jung-Hao Yang","doi":"10.1016/j.comnet.2025.111485","DOIUrl":"10.1016/j.comnet.2025.111485","url":null,"abstract":"<div><div>Public key encryption (PKE) has revolutionized modern cryptography, providing robust security for communication and data protection. However, traditional PKE schemes rely heavily on certificates, resulting in complex and inefficient certificates management problem. To address this problem, certificateless public key encryption (CL-PKE) was introduced to offer a more flexible and efficient approach for key distribution and management. While CL-PKE provides advancements in key management, it is unable to carry out equality tests on ciphertexts, which is essential for tasks such as verification of encrypted data (ciphertexts). To tackle this limitation, the integration of an equality test into CL-PKE, namely certificateless public key encryption with equality test (CL-PKEET), has been proposed. Recent discoveries have highlighted the insecurity of public-key systems due to side-channel attacks. To the best of our knowledge, no CL-PKEET scheme has been proposed to withstand such attacks. Hence, our objective is to present the <em>first</em> CL-PKEET scheme capable of withstanding side-channel attacks, which we call the leakage-resilient CL-PKEET (LR-CL-PKEET) scheme. The proposed scheme offers IND-CCA security and OW-CCA security. Furthermore, the scheme ensures robust security even in the presence of side-channel attacks.</div></div>","PeriodicalId":50637,"journal":{"name":"Computer Networks","volume":"270 ","pages":"Article 111485"},"PeriodicalIF":4.4,"publicationDate":"2025-06-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144517246","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":"Joint multidimensional features for LoRa reception in burst traffic","authors":"Kai Sun ,&nbsp;Bin Hu ,&nbsp;Zhimeng Yin ,&nbsp;Shuai Wang ,&nbsp;Shuai Wang ,&nbsp;Zhuqing Xu ,&nbsp;Tian He","doi":"10.1016/j.comnet.2025.111476","DOIUrl":"10.1016/j.comnet.2025.111476","url":null,"abstract":"<div><div>LoRa has gained significant attention as a promising communication technology in the IoT field. However, with the widespread use of LoRa, network performance faces challenges due to signal collisions at base stations during concurrent transmissions. Traditional methods rely on signal characteristics like frequency to separate colliding packets but have limitations in burst traffic scenarios. These methods fail to accurately separate signals due to unstable and insufficiently detailed signal features. In this paper, we propose a novel physical layer approach called SCLoRa, which can decode overlapping LoRa signals that have collided. SCLoRa utilizes cumulative spectral coefficients, combining frequency and power spectral density, to identify symbols in overlapping signals. This approach takes into account practical factors such as channel fading, symbol boundary alignment, and spectral leakage, which are crucial for accurate signal separation. Enhanced Dynamic-Window and Reuse-Window designs are introduced to further improve decoding reliability and reduce the computational cost. We implement SCLoRa on USRP B210 base stations and standard LoRa nodes (e.g., SX1278). Experiments across various scenarios and radio parameter configurations show that SCLoRa achieves a <span><math><mrow><mn>3</mn><mo>×</mo></mrow></math></span> throughput improvement compared to state-of-the-art methods.</div></div>","PeriodicalId":50637,"journal":{"name":"Computer Networks","volume":"270 ","pages":"Article 111476"},"PeriodicalIF":4.4,"publicationDate":"2025-06-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144549691","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}