arXiv subjects

Fabrizio Granelli

Publications and source records attributed to Fabrizio Granelli.

14 recordsLinked to original sources

arXiv2026

Hybrid spatial-temporal graph neural network Powered NDTs:Towards Next-Gen Smart Infrastructure Twins

Network Digital Twins (NDTs) enable proactive network management and optimization by predicting system behavior before control actions are applied to live infrastructures, supporting critical operations in Internet Service Provider (ISP) networks and wide-area networks (WANs). However, to anchor the superior performance NDTs promise to provide, key enabler techniques are required. Given that mobile networks are modeled as graphs, graph-based architectures such as graph neural networks (GNNs) have shown promising performance in modeling network behavior. This work proposes a novel Hybrid Spatial-Temporal Graph Neural Network (HSTGNN) architecture. Unlike single-branch GNN approaches, we propose a multi-scale design that combines three complementary message-passing paradigms: local neighborhood aggregation, spectral filtering, and learnable attention-based weighting. When benchmarked against other approaches, the proposed HSTGNN achieved superior performance delivering a coefficient of determination score of approximately 0.8816, 17.5\% better than the best baseline ChebNet. Furthermore, HSTGNN achieved the lowest Mean Absolute Error (MAE) of 0.0300, and Root Mean Squared Error (RMSE) of 0.0458, significantly outperforming baseline frameworks and certifying the proposed framework's capability in enabling NDTs.

cs.ET↗

arXiv2026

Lightweight PID-Based Drift Mitigation for Cellular Traffic Forecasting

As mobile networks transition from Beyond 5G (B5G) towards 6G, accurate traffic forecasting is a prerequisite for improving network management. However, with increasing heterogeneity and a massive surge in connected devices, combined with dynamically evolving traffic patterns, accurate forecasting is a persistent bottleneck. Existing frameworks, while generally effective, often lack efficiency and degrade under drift, thus requiring costly model retraining to restore performance. In this paper, we propose a lightweight error correction framework that improves forecasting accuracy by integrating a Proportional-Integral-Derivative (PID) controller as a correction layer enhancing Hierarchical Spatio-temporal Models (HiSTM). Unlike retraining-based model adaptation, our framework performs online error correction without modifying the model parameters. Results from the proposed framework, evaluated across drift scenarios and cell-level analysis, demonstrate reduced Mean Absolute Error (MAE) and Root Mean Squared Error (RMSE), achieving an average drift mitigation of up to 30.18\% in MAE and 26.68\% in RMSE, thereby validating the robustness of the PID framework as a drift mitigation mechanism for network traffic forecasting.

cs.ET↗

arXiv2026

NET4EXA: Pioneering the Future of Interconnects for Supercomputing and AI

NET4EXA aims to develop a next-generation high-performance interconnect for HPC and AI systems, addressing the increasing demands of large-scale infrastructures, such as those required for training Large Language Models. Building upon the proven BXI (Bull eXascale Interconnect) European technology used in TOP15 supercomputers, NET4EXA will deliver the new BXI release, BXIv3, a complete hardware and software interconnect solution, including switch and network interface components. The project will integrate a fully functional pilot system at TRL 8, ready for deployment into upcoming exascale and post-exascale systems from 2025 onward. Leveraging prior research from European initiatives like RED-SEA, the previous achievements of consortium partners and over 20 years of expertise from BULL, NET4EXA also lays the groundwork for the future generation of BXI, BXIv4, providing analysis and preliminary design. The project will use a hybrid development and co-design approach, combining commercial switch technology with custom IP and FPGA-based NICs. Performances of NET4EXA BXIv3 interconnect will be evaluated using a broad portfolio of benchmarks, scientific scalable applications, and AI workloads.

cs.NI↗

arXiv2025

AI-Enabled Digital Twins for Next-Generation Networks: Forecasting Traffic and Resource Management in 5G/6G

As 5G and future 6G mobile networks become increasingly more sophisticated, the requirements for agility, scalability, resilience, and precision in real-time service provisioning cannot be met using traditional and heuristic-based resource management techniques, just like any advancing technology. With the aim of overcoming such limitations, network operators are foreseeing Digital Twins (DTs) as key enablers, which are designed as dynamic and virtual replicas of network infrastructure, allowing operators to model, analyze, and optimize various operations without any risk of affecting the live network. However, for Digital Twin Networks (DTNs) to meet the challenges faced by operators especially in line with resource management, a driving engine is needed. In this paper, an AI (Artificial Intelligence)-driven approach is presented by integrating a Long Short-Term Memory (LSTM) neural network into the DT framework, aimed at forecasting network traffic patterns and proactively managing resource allocation. Through analytical experiments, the AI-Enabled DT framework demonstrates superior performance benchmarked against baseline methods. Our study concludes that embedding AI capabilities within DTs paves the way for fully autonomous, adaptive, and high-performance network management in future mobile networks.

cs.NI↗

arXiv2025

Building Network Digital Twins Part II: Real-Time Adaptive PID for Enhanced State Synchronization

As we evolve towards more heterogeneous and cutting-edge mobile networks, Network Digital Twins (NDTs) are proving to be a promising paradigm in solving challenges faced by network operators, as they give a possibility of replicating the physical network operations and testing scenarios separately without interfering with the live network. However, with mobile networks becoming increasingly dynamic and heterogeneous due to massive device connectivity, replicating traffic and having NDTs synchronized in real-time with the physical network remains a challenge, thus necessitating the need to develop real-time adaptive mechanisms to bridge this gap. In this part II of our work, we implement a novel framework that integrates an adaptive Proportional-Integral-Derivative (PID) controller to dynamically improve synchronization. Additionally, through an interactive user interface, results of our enhanced approach demonstrate an improvement in real-time traffic synchronization.

cs.ET↗

arXiv2025

A Network Digital Twin of a 5G Private Network: Designing a Proof-of-Concept from Theory to Practice

Network Digital Twins represent a key technology in future networks, expected to provide the capability to perform accurate analysis and predictions about the behaviour of 6G mobile networks. However, despite the availability of several theoretical works on the subject, still very few examples of actual implementations of Network Digital Twin are available. This paper provides a detailed description about the characteristics of Network Digital Twin and provides a practical example about real deployment of the technology. The considered network infrastructure is a real 5G private network running in a lab. The Network Digital Twin is built based on open source network emulation software and is available to the community as open source. Measurements on both the physical infrastructure and the related Digital Twin demonstrate a high accuracy in reproducing the state and behavior of the actual 5G system.

cs.NI↗

arXiv2025

Interference-Aware PMI selection for MIMO systems in an O-RAN scenario

The optimization of Precoding Matrix Indicators (PMIs) is crucial for enhancing the performance of 5G networks, particularly in dense deployments where inter-cell interference is a significant challenge. Some approaches have leveraged AI/ML techniques for beamforming and beam selection, however, these methods often overlook the multi-objective nature of PMI selection, which requires balancing spectral efficiency (SE) and interference reduction. This paper proposes an interference-aware PMI selection method using an Advantage Actor-Critic (A2C) reinforcement learning model, designed for deployment within an O-RAN framework as an xApp. The proposed model prioritizes user equipment (UE) based on a novel strategy and adjusts PMI values accordingly, with interference management and efficient resource utilization. Experimental results in an O-RAN environment demonstrate the approach's effectiveness in improving network performance metrics, including SE and interference mitigation.

cs.NI↗

arXiv2025

PPO-EPO: Energy and Performance Optimization for O-RAN Using Reinforcement Learning

Energy consumption in mobile communication networks has become a significant challenge due to its direct impact on Capital Expenditure (CAPEX) and Operational Expenditure (OPEX). The introduction of Open RAN (O-RAN) enables telecommunication providers to leverage network intelligence to optimize energy efficiency while maintaining Quality of Service (QoS). One promising approach involves traffic-aware cell shutdown strategies, where underutilized cells are selectively deactivated without compromising overall network performance. However, achieving this balance requires precise traffic steering mechanisms that account for throughput performance, power efficiency, and network interference constraints. This work proposes a reinforcement learning (RL) model based on the Proximal Policy Optimization (PPO) algorithm to optimize traffic steering and energy efficiency. The objective is to maximize energy efficiency and performance gains while strategically shutting down underutilized cells. The proposed RL model learns adaptive policies to make optimal shutdown decisions by considering throughput degradation constraints, interference thresholds, and PRB utilization balance. Experimental validation using TeraVM Viavi RIC tester data demonstrates that our method significantly improves the network's energy efficiency and downlink throughput.

cs.NI↗

arXiv2025

XAI-Driven Client Selection for Federated Learning in Scalable 6G Network Slicing

In recent years, network slicing has embraced artificial intelligence (AI) models to manage the growing complexity of communication networks. In such a situation, AI-driven zero-touch network automation should present a high degree of flexibility and viability, especially when deployed in live production networks. However, centralized controllers suffer from high data communication overhead due to the vast amount of user data, and most network slices are reluctant to share private data. In federated learning systems, selecting trustworthy clients to participate in training is critical for ensuring system performance and reliability. The present paper proposes a new approach to client selection by leveraging an XAI method to guarantee scalable and fast operation of federated learning based analytic engines that implement slice-level resource provisioning at the RAN-Edge in a non-IID scenario. Attributions from XAI are used to guide the selection of devices participating in training. This approach enhances network trustworthiness for users and addresses the black-box nature of neural network models. The simulations conducted outperformed the standard approach in terms of both convergence time and computational cost, while also demonstrating high scalability.

cs.IT↗

arXiv2025

A measurement-based approach to analyze the power consumption of the softwarized 5G core

In light of the ever growing energy needs of the ICT sector, a value that is becoming increasingly important for a mobile network is its power consumption. However, the transition away from legacy network deployments tightly coupled with the underlying hardware and the adoption of the Network Function Virtualization (NFV) paradigm has made more difficult to accurately evaluate their energy and carbon footprint. In this paper, we propose and validate a measurement-based approach to analyze the power consumption of a virtualized 5G core network (5GC) deployment. We design an experimental testbed using commercial off-the-shelf (COTS) hardware and open-source software as a sample architecture simulating an edge computing node and supporting three different virtualization options. We make use of both hardware-based and software-based power meters to investigate the power consumption trends associated with increasing levels of traffic and multiple 5GC deployment types. The results show the feasibility of a real-time power monitoring system and highlight how deployment choices, such as virtualization framework and 5GC software, can significantly impact on the power consumption of the network.

cs.NI↗

arXiv2024

Optimizing Intrusion Detection System Performance Through Synergistic Hyperparameter Tuning and Advanced Data Processing

Intrusion detection is vital for securing computer networks against malicious activities. Traditional methods struggle to detect complex patterns and anomalies in network traffic effectively. To address this issue, we propose a system combining deep learning, data balancing (K-means + SMOTE), high-dimensional reduction (PCA and FCBF), and hyperparameter optimization (Extra Trees and BO-TPE) to enhance intrusion detection performance. By training on extensive datasets like CIC IDS 2018 and CIC IDS 2017, our models demonstrate robust performance and generalization. Notably, the ensemble model "VGG19" consistently achieves remarkable accuracy (99.26% on CIC-IDS2017 and 99.22% on CSE-CIC-IDS2018), outperforming other models.

cs.CR↗

arXiv2022

A New Agent-Based Intelligent Network Architecture

The advent of 5G and the design of its architecture has become possible because of the previous individual scientific works and standardization efforts on cloud computing and network softwarization. Software-defined Networking and Network Function Virtualization started separately to find their convolution into 5G network architecture. Then, the ongoing design of the future beyond 5G (B5G) and 6G network architecture cannot overlook the pivotal inputs of different independent standardization efforts about autonomic networking, service-based communication systems, and multi-access edge computing. This article provides the design and the characteristics of an agent-based, softwarized, and intelligent architecture, which coherently condenses and merges the independent proposed architectural works by different standardization working groups and bodies. This novel work is a helpful means for the design and standardization process of the futureB5G and 6G network architecture.

cs.NI↗

arXiv2022

An Analytical Study on Functional Split in Martian 3D Networks

As space agencies are planning manned missions to reach Mars, researchers need to pave the way for supporting astronauts during their sojourn. This will also be achieved by providing broadband and low-latency connectivity through wireless network infrastructures. In such a framework, we propose a Martian deployment of a 3-Dimensional (3D) network acting as Cloud Radio Access Network (C-RAN). The scenario consists, mostly, of unmanned aerial vehicles (UAVs) and nanosatellites. Thanks to the thin Martian atmosphere, CubeSats can stably orbit at very-low-altitude. This allows to meet strict delay requirements to split functions of the baseband processing between drones and CubeSats. The detailed analytical study, presented in this paper, confirmed the viability of the proposed 3D architecture, under some constraints and trade-off concerning the involved Space communication infrastructures, that are discussed in detail.

cs.NI↗

arXiv2014

Electric Power Allocation in a Network of Fast Charging Stations

In order to increase the penetration of electric vehicles, a network of fast charging stations that can provide drivers with a certain level of quality of service (QoS) is needed. However, given the strain that such a network can exert on the power grid, and the mobility of loads represented by electric vehicles, operating it efficiently is a challenging problem. In this paper, we examine a network of charging stations equipped with an energy storage device and propose a scheme that allocates power to them from the grid, as well as routes customers. We examine three scenarios, gradually increasing their complexity. In the first one, all stations have identical charging capabilities and energy storage devices, draw constant power from the grid and no routing decisions of customers are considered. It represents the current state of affairs and serves as a baseline for evaluating the performance of the proposed scheme. In the second scenario, power to the stations is allocated in an optimal manner from the grid and in addition a certain percentage of customers can be routed to nearby stations. In the final scenario, optimal allocation of both power from the grid and customers to stations is considered. The three scenarios are evaluated using real traffic traces corresponding to weekday rush hour from a large metropolitan area in the US. The results indicate that the proposed scheme offers substantial improvements of performance compared to the current mode of operation; namely, more customers can be served with the same amount of power, thus enabling the station operators to increase their profitability. Further, the scheme provides guarantees to customers in terms of the probability of being blocked by the closest charging station. Overall, the paper addresses key issues related to the efficient operation of a network of charging stations.

math.OC↗