This document presents a comprehensive validation and integration methodology designed to ensure the realization of the MultiX technologies using vertical oriented Proof of Concepts (PoCs). The document provides an overall description of the methodology to be used during the project execution, focusing on the roadmap and phases envisioned during the project, and the mechanisms established to monitor the progress of the integration and validation activities, and at the same time assess the completeness of the targeted tests. This document also describes the MultiX Open Labs environments, detailing the available and targeted technologies to be used during the validation activities, and the mechanisms to access the testbeds to deploy software components. Finally, this document provides an initial set of experiments and demonstrations envisioned during the project execution providing details regarding the MultiX sensing technologies to be validated and the targeted KPIs.

Federated learning (FL) enables multiple devices to collaboratively train a global model without sharing local data. However, due to limited local computing capability and communication bandwidth, FL suffers from high learning latency, especially when the model size is large. To address these issues, we propose APQ-FL, an adaptive model pruning and quantization method for wireless FL, to reduce the neural network size and improve communication efficiency. Moreover, device selection and wireless resource allocation are also integrated. We first present a convergence analysis of FL with model pruning and quantized transmission, and then jointly optimize the pruning ratio, quantization bit width, device selection, and wireless bandwidth allocation to minimize the convergence upper bound under latency and bandwidth constraints. We prove that the optimized quantization bit width can be obtained via binary search, and derive the closed-form solutions for the optimal pruning ratio and bandwidth allocation. Subsequently, we propose an efficient device selection strategy and further introduce its fairness-aware extension, APQ-FL-Fair. Experiments show that APQ-FL and APQ-FL-Fair improve test accuracy by 4.96%–15.33% while reducing 23.14%–74.45% communication overhead compared to other methods, and exhibit stable and superior performance even under stringent latency constraints.

Beyond-5G and 6G networks must support diverse QoS, from ultra-reliable low latency communications to enhanced mobile broadband. This requires joint, efficient resource management that combines network slicing and dynamic beam management. This paper addresses the complexity of optimizing resource allocation across users and slices served by a set of beams. We propose SBQF, a high-performance technique that optimizes resource allocation at both the MAC and the physical layers. SBQF is designed to maximize the overall user utility and is in particular efficient in improving the performance of users requiring latency-critical services. Validated using real-world vehicular traces, SBQF significantly improves latency, user utility, and data rates over existing solutions.

The integration of intelligent reflecting surfaces (IRS) and non-orthogonal multiple access (NOMA) within the Open Radio Access Network (O-RAN) offers significant opportunities for 6G Internet of Things (IoT) systems but also raises new challenges in security, reliability, and resource efficiency. In this paper, we propose a cross-layer secure sum-rate maximization framework that jointly addresses physical-layer secrecy and network-layer packet loss in IRS-NOMA O-RAN environments. We analytically derive closed-form expressions for secrecy rates and packet loss under a cascaded Rician fading model, and formulate a secrecy sum-rate optimization problem that accounts for IRS phase shifts, NOMA power allocation, and energy-harvesting constraints. The resulting problem is NP-hard due to non-convex coupling across layers. To overcome this, we develop SecureO-RAN-SAC, a deep reinforcement learning algorithm based on Soft Actor-Critic v2, which learns near-optimal policies in real time. Simulation results demonstrate that SecureO-RAN-SAC achieves comparable or superior performance to grid-based search (GBS) while requiring only similar to 10% of its computational cost for a 64-element IRS. These findings highlight the scalability and efficiency of our approach, establishing a new paradigm for secure, resource-aware, and ML-driven cross-layer optimization in O-RAN-enabled IoT networks.

We consider the collaborative inference acceleration task via cloud-edge-end collaboration, which involves a series of tightly coupled decision-making steps, including which DNN model to be selected, how much to compress model, how to partition model, and where to offload partitioned submodels. In practical deployments, these decisions jointly affect both fine-tuning and inference performance, and must jointly account for such aspects as the model being used, the computational resources and local datasets available at each device, as well as network latencies, which significantly increases the complexity of optimizing the problem. Yet, no existing studies focus on such joint optimization problem for these tightly coupled decisions. In this paper, we model this problem as a multi-dimensional optimization problem, jointly optimizing collaborative inference and fine-tuning by selecting the DNN model, compression level, partition strategy, and computational resource allocation, with the objective of minimizing the overall energy consumption of the learning-inference process, subject to accuracy and latency constraints. To this end, we propose an algorithmic framework called JQODI combining a time-energy tree diagram to represent the learning process, a dynamic programming solution strategy, and a data-driven theoretical approach to predict the expected total number of training epochs that meet the accuracy requirements. We prove that JQODI approximates the optimal solution with polynomial complexity. Numerical results demonstrate that JQODI surpasses state-of-the-art methods in both energy efficiency and latency.

Edge computing is a promising paradigm for deploying latency-sensitive applications (Apps) as it brings resources closer to end users. Edge Apps often adopt a microservice (MS) architecture, breaking monolithic Apps into lightweight, containerized MSs that can be dynamically and independently deployed. However, managing such Apps involves three key challenges: (i) optimizing the placement of MSs to reduce both response time and resource overhead, (ii) handling MS migration or relocation as users move while minimizing App service disruption (App downtime), and (iii) enabling MS sharing across Apps while ensuring performance guarantees. We formulate this as an optimization problem, named Multi-microservice Application Placement (MAP), prove its NP-hardness, and introduce STEP (State and Topology-aware Edge-MS Placement), a polynomial-time heuristic. STEP distinguishes itself from prior work by: (i) jointly considering stateful and stateless MS characteristics in deployment decisions, (ii) exploiting MS shareability to reduce resource usage, (iii) balancing response latency, App downtime, and resource utilization, and (iv) leveraging multiple versions of the same MS to adapt quality of service to available edge resources. Our results in a small-scale scenario show that STEP achieves near-optimal performance with only 7% higher CPU cost than the optimal solution. Large-scale real-time experiments on a Kubernetes cluster demonstrate that STEP consistently outperforms competing methods, achieving up to 50% lower deployment costs while delivering 50% gain in app quality and saving 15% in radio resources with over 90% request success rates.

Time Sensitive Networking (TSN) is fundamental for the reliable, low-latency networks that will enable the Industrial Internet of Things (IIoT). Wi-Fi has historically been considered unfit for TSN, as channel contention and collisions prevent deterministic transmission delays. However, this issue can be overcome by using Target Wake Time (TWT), which enables the access point to instruct Wi-Fi stations to wake up and transmit in non-overlapping TWT Service Periods (SPs), and sleep in the remaining time. In this paper, we first formulate the TWT Acceptance and Scheduling Problem (TASP), with the objective to schedule TWT SPs that maximize traffic throughput and energy efficiency while respecting Age of Information (AoI) constraints. Then, due to TASP being NP-hard, we propose the TASP Efficient Resolver (TASPER), a heuristic strategy to find near-optimal solutions efficiently. Using a TWT simulator based on ns-3, we compare TASPER to several baselines, including HSA, a state-of-the-art solution originally designed for WirelessHART networks. We demonstrate that TASPER obtains up to 24.97% lower mean transmission rejection cost and saves up to 14.86% more energy compared to the leading baseline, ShortestFirst, in a challenging, large-scale scenario. Additionally, when compared to HSA, TASPER also reduces the energy consumption by 34% and reduces the mean rejection cost by 26%. Furthermore, we validate TASPER on our IIoT testbed, which comprises 10 commercial TWT-compatible stations, observing that our solution admits more transmissions than the best baseline strategy, without violating any AoI deadline.

In federated learning (FL), balancing privacy protection, learning quality, and efficiency remains a challenge. Privacy protection mechanisms, such as Differential Privacy (DP), degrade learning quality, or, as in the case of Homomorphic Encryption (HE), incur substantial system overhead. To address this, we propose Alt-FL, a privacy-preserving FL framework that combines DP, HE, and synthetic data via a novel round-based interleaving strategy. Alt-FL introduces three new methods, Privacy Interleaving (PI), Synthetic Interleaving with DP (SI/DP), and Synthetic Interleaving with HE (SI/HE), that enable flexible quality-efficiency trade-offs while providing privacy protection. We systematically evaluate Alt-FL against representative reconstruction attacks, including Deep Leakage from Gradients, Inverting Gradients, When the Curious Abandon Honesty, and Robbing the Fed, using a LeNet-5 model on CIFAR-10 and Fashion-MNIST. To enable fair comparison between DP- and HE-based defenses, we introduce a new attacker-centric framework that compares empirical attack success rates across the three proposed interleaving methods. Our results show that, for the studied attacker model and dataset, PI achieves the most balanced trade-offs at high privacy protection levels, while DP-based methods are preferable at intermediate privacy requirements. We also discuss how such results can be the basis for selecting privacy-preserving FL methods under varying privacy and resource constraints.

Transmitting rich visual data in resource-constrained environments like Non-Terrestrial Networks (NTNs) poses a significant challenge. While current Semantic Communication (SC) approaches reduce bandwidth consumption, they often lack flexibility and/or compromise the perceptual fidelity of critical details. This paper first analyzes the fundamental trade-offs that exist between perceptual fidelity, semantic fidelity, and bandwidth utilization. It then introduces SPIFF, an SC-Generative AI framework that, by supporting selective fidelity, enables fine-grained control over the above trade-offs while meeting delay requirements. SPIFF features a lightweight, semantic-aware encoder performing semantic segmentation and applying a novel patch preservation strategy that retains perceptually significant regions while adapting lower-relevance areas compression to bandwidth availability. SPIFF also offloads high-complexity reconstruction tasks to a Generative AI-enabled decoder at the receiver, thus addressing asymmetric computation requirements. To support adaptation under dynamic conditions, while meeting system and application constraints, we equip SPIFF with a learning-based decision engine that is able to cope with the system non-linearities and effectively tune SPIFF’s configuration online. We evaluate SPIFF by implementing a full encoder-decoder pipeline. Results show that SPIFF fulfills perceptual reconstruction quality in scenarios where SC fails, and improves over state-of-the-art solutions both bandwidth savings (by up to 21%) and perceptual fidelity (by up to 13%).

We consider the efficient resource allocation task in the blockchain-based mobile edge computing (MEC) system that requires decentralized transaction management to validate transactions between edge servers (ESs) and mobile devices (MDs). In such task allocation process (where MDs' resources are limited and privacy-sensitive), it is a significant challenge to guarantee individual rationality with satisfactory system stability while enabling flexible task offloading under various locality constraints (e.g., communication distance, bandwidth and delay). In this paper, we formulate the target problem as a blockchain-assisted task-resource matching model, and then propose a near optimal locality-aware resource allocation mechanism over smart contract to enable automatic and efficient transactions in MEC system. More specifically, for the service agents selection, we design the preference-based selection strategy to get highest estimated profit. For the flexible task offloading, we develop the minimum delay task graph partitioning algorithm to determine the optimal task offloading solution for MD under different resource bundles. For the task-resource matching, we propose a task-resource matching game (based on potential game) with the second lowest cost strategy to determine the matching of task-resource and decide the price of resource bundle. For the transaction verification and block allocation, we propose a social welfare-driven consensus mechanism to enable verified transaction and fair block allocation in a reward-free way. Strict theoretical analysis and extensive simulations demonstrate that our mechanism guarantees individual rationality, Nash Equilibrium, and stable near optimal solution.

To address the limitations of existing wireless networks for demanding applications like brain-computer interfaces and intelligent transportation systems, we propose an advanced framework for joint resource allocation and task offloading across integrated terrestrial and non-terrestrial networks (TN-NTN). This framework utilizes multiple layers, including ground users, UAVs, HAPs, and satellites, to improve service quality and immersive experiences, particularly in scenarios like Metaverse applications. Ground users request resources, while UAVs and HAPs serve as resource providers, and satellites ensure reliable communication during emergencies. A double auction-based incentive scheme is employed in which operators control UAV and HAP resources to maximize utility, and users aim to minimize computation costs and protect data privacy. To handle the complexity of the operator-user interaction, which results in an NP-hard optimization problem, we applied a hierarchical multi-agent federated deep reinforcement learning (FeDRL) approach. Our simulation results demonstrate that the FeDRL algorithm significantly improves social welfare by 6.38%, 17.43%, and 28.73% over modified MADDPG, FRL, and DDPG algorithms, respectively.

In an edge-cloud multi-tier network, datacenters provide services to mobile users, with each service having specific latency constraints and computational requirements. Deploying such a variety of services while matching their requirements with the available computing resources is challenging. In addition, time-critical services may have to be migrated as the users move, to keep fulfilling their latency constraints. Unlike previous work relying on an orchestrator with an always-updated global view of the available resources and the users' locations, this work envisions a distributed solution to the above problems. In particular, we propose a distributed asynchronous framework for service deployment in the edge-cloud that increases the system resilience by avoiding a single point of failure, as in the case of a central orchestrator. Our solution ensures cost-efficient feasible placement of services, while using negligible bandwidth. Our results, obtained through trace-driven, large-scale simulations, show that the proposed solution provides performance very close to those obtained by state-of-the-art centralized solutions, and at the cost of a small communication overhead.

Machine learning (ML) can be often distributed, owing to the need to harness more resources and/or to preserve privacy. Accordingly, distributed learning has received significant attention from the literature; however, most works focus on the expected learning quality (e.g., loss) attained and do not consider the distribution thereof. It follows that ML models are not dependable, and may fall short of the required performance in many real-world cases. In this work, we tackle this challenge and propose DepL, a framework attaining dependable learning orchestration: DepL efficiently makes joint, near-optimal decisions concerning (i) which data to use for learning, (ii) the ML models to use - chosen within a set of full-size models and compressed versions thereof - and when to switch from one model to another, and (iii) the clusters of physical nodes to use for the learning. DepL improves over previous works by guaranteeing that the learning quality target (e.g., a minimum loss) is achieved with a target probability, while minimizing the learning (e.g., energy) cost. DepL has provably low polynomial computational complexity and a constant competitive ratio. Further, experimental results using the CIFAR-10 and GTSRB datasets show that it consistently matches the optimum and outperforms state-of-the-art approaches (30% faster learning and 40-80% lower cost).

The rapid advancements of next-generation vehicular networks require intelligent, low-latency, and efficient resource management to support heterogeneous services. In this work, we propose a Traffic-aware Dynamic Resource Allocation (TADRA) architecture for UAV-assisted vehicular O-RAN to address the challenges of dynamic traffic conditions, infrastructure failures, and stringent quality of service (QoS) requirements. Due to the dynamic mobility and flexible deployment characteristics, UAV Open Radio Units (O-RUs) in the TADRA architecture support the terrestrial infrastructure under overload or failure conditions, dynamically extending coverage, balancing traffic loads, and restoring service to maintain uninterrupted QoS across diverse and heterogeneous traffic demands. Unlike existing static or single-layer solutions, our proposed TADRA integrates RAN Intelligent Controllers (RICs) with a Hierarchical Traffic-Aware Multi-Agent Twin-Delayed (TMT) algorithm to optimize the allocation of computation and radio resources. This joint optimization problem is NP-hard, highly dynamic, and coupled across agents, making TMT a tractable and adaptive alternative. This hierarchical framework performs traffic prioritization at the upper (application) layer and resource allocation at the lower (MAC) layer, facilitating adaptive decision-making under diverse vehicular traffic patterns. Numerical results demonstrate that our solution provides substantial gains over MATD3, MADDPG, and GA, achieving 17% lower latency, 10% higher throughput, 14% lower energy consumption, and 6.5% higher reliability.

Early Exiting (EE) is an emerging paradigm in deep learning that equips Deep Neural Networks (DNNs) with intermediate classifiers, enabling a trade-off between inference accuracy and latency. In this work, we investigate the integration of EE mechanisms into edge computing architectures, focusing on a representative use case involving task execution in resource-constrained computing and communications environments for connected and automated vehicles (CAVs). We develop a detailed system model that captures the complex interplay among time-varying system components, including wireless channel coherence and the dynamic availability of computational and communication resources. Building on this model, we formulate a joint optimization problem encompassing task offloading, resource allocation, and early exit selection. We demonstrate how EE enhances system adaptability under stringent constraints, such as limited bandwidth, computing capacity, or delay requirements. To tackle the complexity of the proposed optimization, we adopt a novel solution approach based on the distributional Soft Actor-Critic (SAC) Deep Reinforcement Learning (DRL) algorithm, which quantifies the uncertainty of the learned policy. Simulation results confirm that integrating EE with edge computing significantly improves the trade-off between inference accuracy and latency, achieving up to 212% improvement in the average task completion ratio compared to edge computing systems without EE, under the considered simulation settings.

Low Earth orbit (LEO) satellite communications play a critical role in achieving global connectivity, yet they face significant challenges due to high satellite mobility and incomplete channel state information (CSI). Moreover, the integration of reconfigurable intelligent surfaces (RIS) in certain scenarios introduces additional complexities. In this paper, we propose a novel MIMO channel prediction framework tailored for LEO satellite communications involving unmanned aerial vehicle-mounted RIS (UAV-RIS), employing a spatiotemporal-attention (ST-attention) mechanism to capture both the spatial correlations among antennas and the temporal dynamics of rapidly varying channels. Furthermore, we leverage masked pretraining to enhance the model’s robustness under scenarios of severe CSI incompleteness, enabling effective reconstruction of missing channel information. Comprehensive simulations demonstrate that our approach outperforms traditional model-based predictors, whether historical CSI is fully available or only partially observed.

Deep neural networks (DNNs) with Early Exits (EE) are widely adopted at the network edge to reduce inference latency and resource consumption by allowing samples to exit the DNN early when sufficient confidence is achieved. However, their effectiveness significantly degrades under popularity drift – where the frequency of target classes shifts over time. This paper introduces ADEx, an Adaptive Drift-aware EE framework that enables dynamic and lightweight adaptation of DNNs with EEs to class popularity drift. ADEx continuously monitors EE behavior to detect popularity shifts and selectively retrains only the affected EEs, while keeping the backbone network unchanged. For self-supervised adaptation, ADEx uses the final exit as a teacher to pseudo-label new data, applying a priority-aware loss that enforces high confidence on popular classes and uncertainty on others. Also, we analyze execution strategies balancing adaptation speed, latency, and resource usage. Experimental results demonstrate that ADEx restores inference efficiency, reducing main-branch usage from a post-drift peak of 46.14% to 7.65% (near the 7.46% pre-drift level) and lowering mean latency from 3.05ms to 2.57ms. Compared to a Joint Fine-tuning baseline, ADEx achieves similar accuracy recovery while reducing peak GPU memory by 52.3% and avoiding full-model shadow copies.

Open Radio Access Network (RAN) is a key paradigm to attain unprecedented flexibility of the RAN via disaggregation and Artificial Intelligence (AI)-based applications called xApps. In dense areas with many active RAN nodes, compute resources are engineered to support potentially hundreds of xApps monitoring and controlling the RAN to achieve operator’s intents. However, such resources might become underutilized during low-traffic periods, where most cells are sleeping and, given the reduced RAN complexity, only a few xApps are needed for its control. In this paper, we propose CORMO-RAN, a data-driven orchestrator that dynamically activates compute nodes based on xApp load to save energy, and performs lossless migration of xApps from nodes to be turned off to active ones while ensuring xApp availability during migration. CORMORAN tackles the trade-off among service availability, scalability, and energy consumption while (i) preserving xApps’ internal state to prevent RAN performance degradation during migration; (ii) accounting for xApp diversity in state size and timing constraints; and (iii) implementing several migration strategies and providing guidelines on best strategies to use based on resource availability and requirements. We prototype CORMORAN as an rApp, and experimentally evaluate it on an ORAN private 5G testbed hosted on a Red Hat OpenShift cluster with commercial radio units. Results demonstrate that CORMORAN is effective in minimizing energy consumption of the RAN Intelligent Controller (RIC) cluster, yielding up to 64% energy saving when compared to existing approaches.

Federated testbeds enable collaborative research by providing access to diverse resources, including computing power, storage, and specialized hardware like GPUs, programmable switches and smart Network Interface Cards (NICs). Efficiently sharing these resources across federated institutions is challenging, particularly when resources are scarce and costly. GPUs are crucial for AI and machine learning research, but their high demand and expense make efficient management essential. Similarly, advanced experimentation on programmable data plane requires very expensive programmable switches (e.g., based on P4) and smart NICs. This paper introduces SHARY (SHaring Any Resource made easY), a dynamic reservation system that simplifies resource booking and management in federated environments. We show that SHARY can be adopted for heterogenous resources, thanks to an adaptation layer tailored for the specific resource considered. Indeed, it can be integrated with FIGO (Federated Infrastructure for GPU Orchestration), which enhances GPU availability through a demand-driven sharing model. By enabling real-time resource sharing and a flexible booking system, FIGO improves access to GPUs, reduces costs, and accelerates research progress. SHARY can be also integrated with SUP4RNET platform to reserve the access of P4 switches.

An increasing number of mobile applications leverage deep neural networks (DNN) as an essential component to adapt to the operational context at hand and provide users with an enhanced experience. It is thus of paramount importance that network systems support the execution of DNN inference tasks in an efficient and sustainable way. Matching the diverse resources available at the mobile-edge-cloud network tiers with the applications requirements and the complexity of their, while minimizing energy consumption, is however challenging. A possible approach to the problem consists in exploiting the emerging concept of dynamic DNNs, characterized by multi-branched architectures with early exits enabling sample-based adaptation of the model depth. We leverage this concept and address the problem of deploying portions of DNNs with early exits across the mobile-edge-cloud system and allocating therein the necessary network, computing, and memory resources. We do so by developing a 3-stage graph-modeling method that allows us to represent the characteristics of the system and the applications as well as the possible options for splitting the DNN over the multi-tier network nodes. Our solution, called Feasible Inference Graph (FIN), can determine the DNN split, deployment, and resource allocation that minimizes the inference energy consumption while satisfying the nodes’ constraints and the requirements of multiple, co-existing applications. FIN closely matches the optimum and leads to over 89% energy savings with respect to state-of-the-art alternatives.