The Internet of things (IoT) has opened new dimensions of novel services and computing power for modern living standards by introducing innovative and smart solutions. Due to the extensive usage of these services, IoT has spanned numerous devices and communication entities, which makes the management of the network a complex challenge. Hence it is urgently needed to redefine the management of the IoT network. Software-defined networking (SDN) intrinsic programmability and centralization features simplify network management, facilitate network abstraction, ease network evolution, has the potential to manage the IoT network. SDN’s centralized control plane promotes efficient network resource management by separating the control and data plane and providing a global picture of the underlying network topology. Apart from the inherent benefits, the centralized SDN architecture also brings serious security threats such as spoofing, sniffing, brute force, API exploitation, and denial of service, and requires significant attention to guarantee a secured network. Among these security threats, Distributed Denial of Service (DDoS) and its variant Low-Rate DDoS (LR-DDoS), is one of the most challenging as the fraudulent user generates malicious traffic at a low rate which is extremely difficult to detect and defend. Machine Learning (ML), especially Federated Learning (FL), has shown remarkable success in detecting and defending against such attacks. In this paper, we adopted Weighted Federated Learning (WFL) to detect Low-Rate DDoS (LR-DDoS) attacks. The extensive MATLAB experimentation and evaluation revealed that the proposed work ignites the LR-DDoS detection accuracy compared with the individual Neural Networks (ANN) training algorithms, existing packet analysis-based, and machine learning approaches.

To efficiently allocate communication and computation resources among enhanced mobile broadband (eMBB) and ultrareliable low-latency communication (uRLLC) applications, this article proposes a joint communication and computation resource allocation scheme in sixth generation (6G) information-centric networks (ICNs) named: CCR-ICN. To achieve efficient communication and computation resources sharing, the CCR-ICN scheme designs a microservice-centric interest naming structure to enable named-based communication. Furthermore, the CCR-ICN scheme designs a robust forwarding strategy specifically tailored to efficiently allocate communication resources, such as bandwidth among eMBB and uRLLC application requests. In addition, for the allocation of computation resources, the proposed scheme designs a priority-driven computation resources allocation mechanism ensuring effective allocation of multiaccess edge computing computation resources. To evaluate the proposed scheme, we performed extensive simulations in a widely used NS-3-based ndnSIM simulator and the results reveal the CCR-ICN achieves 66.82% higher communication resource allocation, and a 17.51% lower communication overhead compared to the TS-6G scheme. Moreover, the CCR-ICN performs computation request satisfaction at 42.03% and 12.72% higher than the TS-6G and extended weighted fair queueing with latency constraint (EWFQ/LC) schemes, respectively. Furthermore, the CCR-ICN achieves computation resource utilization of 35.54% and 21.63% lower than the EWFQ/LC and TS-6G schemes, respectively.

The Internet of Things (IoT) has shown remarkable competence in uniting devices and internet connectivity to make modern life’s applications a reality. These networked devices, empowered with computing and communications, usually exhibit an exponential increment in network traffic to enhance the end-user quality of experience (QoE). The exponential increase in network traffic makes the traditional communication system congested. To reduce the burden of network traffic, the data could be offloaded to device-to-device (D2D) based infrastructure. This offloading scheme results in an innovative, eco-friendly, and sustainable data dissemination approach known as social networking and is mostly done in opportunistic ways. In this paper, we propose a scheme to explore the performance of the existing routing protocols with the movements and spatial closeness relations among ‘things’, termed as Internet of Moving Things (IoMT). The spatial closeness known as traces will help to make optimal data offloading decisions while considering the mobility traces of humans in the social network. The proposed mechanism is designed to optimally utilize moving physical objects to be involved as a third party in the data communication process between the sources and the destinations rather than using the traditional infrastructure while considering the quality of service (QoS). The simulation results have corroborated the proposed scheme and showed the remarkable performance of routing algorithms while incorporating mobility traces.

Autonomous vehicles are revolutionizing the future of intelligent transportation systems by integrating smart and intelligent onboard units (OBUs) that minimize human intervention. These vehicles can communicate with their environment and one another, sharing critical information such as emergency alerts or media content. However, this communication infrastructure is susceptible to cyber-attacks, necessitating robust mechanisms for detection and defense. Among these, the most critical threat is the denial-of-service (DoS) attack, which can target any entity within the system that communicates with autonomous vehicles, including roadside units (RSUs), or other autonomous vehicles. Such attacks can lead to devastating consequences, including the disruption or complete cessation of service provision by the infrastructure or the autonomous vehicle itself. In this paper, we propose a system capable of detecting DoS attacks in autonomous vehicles across two scenarios: an infrastructure-based scenario and an infrastructureless scenario, corresponding to vehicle-to-everything communication (V2X) Mode 3 and Mode 4, respectively. For Mode 3, we propose an ensemble learning (EL) approach, while for the Mode 4 environment, we introduce a gossip learning (GL)-based approach. The gossip and ensemble learning approaches demonstrate remarkable achievements in detecting DoS attacks on the UNSW-NB15 dataset, with efficiencies of 98.82% and 99.16%, respectively. Moreover, these methods exhibit superior performance compared to existing schemes.

Technological advancements have significantly improved cattle farming, particularly in sensor-based activity monitoring for health management, estrus detection, and overall herd supervision. However, such a sensor-based monitoring framework often illustrates several issues, such as high cost, animal discomfort, and susceptibility to false measurement. This study introduces a vision-based cattle activity monitoring approach deployed in a commercial Nestlé dairy farm, specifically one that is estrus-focused, where overhead cameras capture unconstrained herd behavior under variable lighting, occlusions, and crowding. A custom dataset of 2956 Images are collected and then annotated into four fine-grained behaviors—standing, lying, grazing, and estrus—enabling detailed analysis beyond coarse activity categories commonly used in prior livestock monitoring studies. Furthermore, computer vision-based deep learning algorithms are deployed on this dataset to classify the aforementioned classes. A comparative analysis of YOLOv8 and YOLOv9 is provided, which clearly illustrates that YOLOv8-L achieved a mAP of 91.11%, whereas YOLOv9-E achieved a mAP of 90.23%.

The rapid advancement of deepfake generation techniques poses significant challenges to the credibility of digital content by producing highly realistic manipulated content. This study explores two primary scenarios: the first is model configuration, and the second is training strategies. In the model configuration scenario, we combine deep learning-based feature extraction with machine learning-based classification, where deep learning serves as the encoder and machine learning serves as the classifier. Additionally, under training strategies,we investigate howdifferent components of the model, whether fixed or learnable, affect detection performance. Specifically, we explore three learning modes: Learnable–Learnable (LL), Learnable–Fixed (LF), and Fixed–Learnable (FL), each of which affects the model’s adaptability and efficiency in detecting deepfakes. Through extensive experimentation on publicly available benchmark datasets, including DFDC, Celeb-DF, and FaceForensics++, our framework demonstrates how learning dynamics influence both detection accuracy and computational efficiency. Furthermore, we assess the robustness of our approach against adversarial attacks to ensure the model’s resilience in real-world scenarios. Additionally, the integration of Explainable AI (XAI) techniques enhances model interpretability by identifying the critical features that drive the model’s predictions. The performance analysis across these datasets, combined with robustness testing, provides valuable insights for designing scalable, efficient, and explainable deepfake detection systems suitable for real-world deployment.