5G mobile network defines a set of use cases with very distinct requirements with regard to throughput, low latency, and high reliability. Network slicing is essential for addressing these end-to-end needs, breaking away from the monolithic architectures characteristic of previous mobile network generations. The Open RAN movement promotes the disaggregation of infrastructure components, allowing for contributions from various vendors and enhancing flexibility and openness within the ecosystem. On the other hand, this paradigm shift introduces potential security threats and vulnerabilities that challenges Slice reliability in Open RAN domains. This work presents a comprehensive investigation through a systematic literature review, examining the current research landscape regarding the security aspects of Open RAN slicing based on O-RAN standards, including studies focused on security metrics and indicators. Additionally, it evaluates the effectiveness of proposed solutions and identifies critical gaps that require further research. By providing a detailed overview of existing research and potential security strategies, this work aims to contribute to the development of secure and robust 5 G networks.
As Artificial Intelligence (AI) adoption accelerates across industries, the need for robust real-time cybersecurity solutions becomes increasingly critical. Traditional security frameworks that rely on static rules, signature-based detection, and threshold alerts struggle to detect sophisticated identity-based attacks that exploit behavioral anomalies rather than explicit policy violations. These limitations, combined with the growing volume of telemetry data and the latency of conventional detection methods, leave organizations vulnerable to prolonged breaches and data exfiltration. This paper presents a use case with a scalable, AI-driven approach to sensitive data leak detection using NVIDIA Morpheus, a GPU-accelerated cybersecurity framework. We demonstrate how Morpheus, in conjunction with NVIDIA BlueField DPUs and DOCA, enables high-throughput, low-latency inspection of real-time network telemetry without impacting system performance. The solution leverages a pre-trained NLP-based Sensitive Information Detection model capable of identifying ten categories of sensitive data within packet payloads, including credentials, PII, and cryptographic keys. Network-captured PCAP data serialized in JSONlines format passes through a linear, feed-forward Morpheus pipeline, configured via CLI and powered by Triton Inference Server with dynamic batching for optimized inference. The final output provides binary classification results that indicate the presence or absence of sensitive information. This case study illustrates how AI-enhanced telemetry analysis can significantly reduce detection and response times, offering a proactive defense mechanism against modern cyber threats.
Wi-Fi mesh networks are an excellent complementary solution to extend the reach of 5G infrastructure, particularly by providing local coverage or backhaul connectivity in hard-to-reach areas. However, the deployment of audio and video services in mesh network is challenging due to unstable wireless link quality. In this context, the choice of transport protocol can play an important role in the resulting Quality-of-Service (QoS) of applications and the perceived Quality-of-Experience (QoE) of users. In this paper, we evaluate video traffic performance over the UDP and TCP transport protocols and the audio traffic over UDP. We also examine the impact of competing traffic on audio performance in a tactical mesh network deployment for the scenario of firefighting intervention. The experimental results show that in small networks, where TCP can quickly retransmit lost packets, TCP-based video streaming can greatly outperform UDP. The competing video traffic can decrease the instantaneous audio similarity score.
This work evaluates the use of Wi-Fi 7 with Multi-Link Operation (MLO) for indoor positioning based on RSSI, seeking to reduce the density of Reference Points (RPs) and simplify the infrastructure. Using only 0.7656 RPs/m(2) and a single access point operating simultaneously in the 2.4 GHz, 5 GHz and 6 GHz bands, simulations were performed in GNU Octave. The results demonstrate that the SVM model achieved 92% accuracy with errors below 3 meters, while the MLP model attained 65% accuracy with errors below 1.71 m. The Random Forest model excelled in stability across varying RP densities, achieving 88% accuracy with errors below 3 m and 64% accuracy with errors below 1.71 m These findings highlight the efficient use of MLO in maintaining high accuracy with reduced infrastructure, positioning Wi-Fi 7 as a scalable and cost-effective solution for indoor positioning.
We propose a correlation-based detection method using Support Vector Machine (SVM) for ambient backscatter based tags. The proposed detector is evaluated against the Minimum Mean Square Error (MMSE) detector in various scenarios, including low Signal-to-Noise Ratio (SNR) conditions and multi-transmitter environments. Moreover, we compare our approach to an energy detector in terms of Bit Error Rate (BER) performance. The reader employs beamforming to steer the beams towards tags and RF source, improving the strength of the tag signal, attenuating the interferences and improving the estimation of the source signal, which is used in the correlation technique to build features for the SVM detector. Simulation results show that our correlation-based detector consistently out-performs the MMSE and energy detection benchmarks, especially in low SNR conditions and can support detection in the presence of multiple RF sources.
