International audience

Fifth-generation (5 G) networks have the potential to enhance a wide range of services through improved Quality of Service (QoS) and efficient service differentiation (Network Slicing). However, for many projects such as FerroMobile (revitalizing secondary railway lines using hybrid road-rail vehicles), before deploying new 5G-based solutions, it is essential to evaluate real-world performance using emulators under realistic conditions: geographical locations, travel speeds, varying traffic loads, etc. This paper aims to fill the mobility support gap in current open-source 5G emulators, notably srsRAN and OAI. While these platforms offer robust 5 G emulation features, they do not natively support mobility scenarios in 5 G networks. To address this, we introduce and validate a new software component that integrates srsRAN with GNU Radio and Open5GS. Our solution can be extended to other emulators, particularly OAI, and includes automated scenario generation to streamline experimentation. By detailing a real-world application in the FerroMobile context, we demonstrate this software brick's practicality and adaptability.

In today’s intelligent transportation systems, emerging vehicular applications, such as multi-vehicle data fusion, platoon management, and high-definition video processing, demand a new generation of communication architectures to meet stringent processing and latency requirements. NOMA-based Vehicular Edge Computing (VEC) architectures have gained prominence as a solution, offering efficient spectrum reuse and reduced processing delays by leveraging distributed edge resources. While decentralized management of these architectures promises real-time, adaptive task offloading, existing approaches are limited by their reliance on local observations and insufficient integration of comprehensive global network state information. To address these challenges, we propose ROLAM, a novel framework that fuses LSTM-based state estimation with Multi-Agent Deep Deterministic Policy Gradient (MADDPG) to optimize resource allocation under dynamic conditions. Our evaluation demonstrates that ROLAM significantly improves energy efficiency and overall system performance, underscoring its potential to robustly meet the evolving needs of next-generation vehicular applications.

The railway industry is rapidly transitioning toward full automation and enhanced safety, driven by the need to reduce greenhouse gas emissions and revitalize secondary lines. While 5G promises improved latency, reliability, and throughput, its high deployment costs and limited coverage hinder universal adoption. Emerging research advocates the transition to 6G, supported by a vision of Networks of Networks that would allow the transparent use of different access networks (e.g. satellite, IoT). This paper extends that paradigm by introducing a broader vision centred on multi-operator resource federation, hybrid classical/quantum communications, and distributed artificial intelligence. We propose a five-layer architecture that integrates these pillars by federating infrastructure across operators for seamless coverage and combining quantum-secure channels with conventional links for robust security. Additionally, our design leverages federated artificial intelligence at the edge for real-time decision-making. Finally, we outline the primary technical challenges and research directions necessary to realize resilient, adaptive, and secure railway connectivity in the 6G era.

The Future Railway Mobile Communication System (FRMCS), based on 5G technology, is being developed to support high-speed train mobility and increasing data-rate demands within the limited$5-10 \text{MHz}$frequency bandwidth allocated to railway operators. The fluctuating channel conditions and limited radio resources present a significant challenge for the Radio Access Network (RAN). Railway communications are structured around application-specific requirements, categorized into three classes: critical (ultra-reliable train control and safety operations), performance (bandwidth-intensive services like video surveillance and infotainment), and business (on-board Wi-Fi). Existing RAN slicing solutions, are designed for public 5G Networks, assume wideband carriers or static quotas at slice level or user level, so they cannot guarantee application-level requirements or adapt to dynamic channel-quality indicator (CQI) swings in railways. That is why we propose HSRS (Distributed Hierarchical Slack-Reduction Slicing), which formulates this NPhard allocation problem as an integer linear program and applies a two-phase heuristic: 1) minimize SLA slack for critical; 2) maximize weighted throughput for performance and business slices. Simulations using SNR traces captured on real trains traveling up to$350 ~\text{km} / \mathrm{h}$demonstrate the improved application-level SLA satisfaction compared to baseline methods.

Multi-access Edge Computing (MEC) promises ultra-low end-to-end latency for$5 \mathrm{G} / 6 \mathrm{G}$services, but each additional edge site drives CAPEX and OPEX, making accurate dimensioning essential. Existing studies either optimize runtime placement on a pre-deployed edge tier or size a single layer under synthetic traffic, leaving operators without a principled way to decide where to build servers or how many. We propose a fourphase, measurement-driven workflow that (i) records live radio and transport delays, (ii) converts real applications into machinereadable profiles, (iii) derives tier-specific latency budgets, and (iv) greedily assigns every service to the cheapest network layer that still meets its Service Level Agreement. Applied to FerroMobile, an autonomous railway system with URLLC and eMBB workloads, evaluations demonstrate the relevance of such a framework and the ability of a multi-layer infrastructure to reduce costs while guaranteeing performance.

The advent of cooperative intelligent transport systems (C-ITS) and connected vehicles has brought with it a raft of novel services, including autonomous driving and dynamic traffic management, which necessitate both real-time processing and substantial data processing capacity. These technological advancements have imposed significant network constraints, requiring the development of robust network infrastructures to guarantee the reliable transmission of critical information in real time. In consequence, numerous studies have concentrated on NOMA-assisted Multi-Access Edge Computing (NOMA-MEC) to facilitate the provision of new services and, in particular, on their management through the utilisation of Multi-Agent Deep Reinforcement Learning (MADRL) algorithms. Nevertheless, the absence of a unified framework for the rigorous evaluation and comparison of these different approaches represents a significant issue. For this reason, we have established a unified, adaptable, and extensible framework, NedgeSIM-RL, with the objective of evaluating decentralised offloading strategies within the context of NOMA-VEC architectures. In this paper, we illustrate the functionality of the framework, the potential configurations, and the areas of research that can be pursued. We also present a basic simulation that can be executed on the framework to manage decentralised offloading via NOMA on a single MEC server by multiple vehicles.

To allow the large-scale deployment of automated and connected vehicles, their safety must be ensured. The Operational Design Domain (ODD) aims to define under which conditions an Automated Driving System (ADS) can operate safely: speed, type of road, weather conditions, etc. Clearly identifying the characteristics and boundaries of the ODD is an important issue today for ADS. To this end, the design and use of an ODD taxonomy seems to be a relevant approach considered by both the industrial and academic worlds. Therefore, in this paper, we propose an analysis and comparison of the main existing taxonomies. We also define a new generic taxonomy, combining the different approaches proposed in the literature, applicable to both vehicles and road sections. Then, this taxonomy is applied to a specific use case (Bus Station Automated Service). Finally, we identify potential directions for an extended ODD taxonomy.

Data processing architectures are currently evolving to enable the deployment of critical real-time applications, particularly in the railway environment. Indeed, these new computing architectures could contribute to the definition of innovative services: platooning, remote driving, autonomous trains, etc. This is why many studies today aim at setting up optimal Edge Computing architectures that would allow critical services to be deployed as close as possible to the end user. It is therefore necessary to design a powerful simulation/emulation environment that could be used to validate the proposed solutions. In this paper we demonstrate how the platform we have implemented, Emu5GNet, could be used to enable the rapid design and evaluation of new Edge Computing solutions. We present the potential applications of Emu5GNet and a reusable use case comparing Edge and Cloud deployment as well as different strategies for Edge resources management.

Edge Computing promise a bright future for Internet of Things (IoT). Edge Computing servers can be geographically distributed to be closer to users and ensure low latency communications and real-time data processing for the third-party applications using this infrastructure. However, user mobility and limited Edge servers capabilities (CPU, memory, bandwidth) may cause many services placement failures. That's why in this article we propose a new strategy aimed at enabling a fair sharing of available Edge resources through a dynamic and real-time IoT Service placement. Our strategy uses Fuzzy Logic to enable 1) quick and low-cost deployment of the solution and 2) real-time modification of the policies defined by the Edge operator. It takes advantage, among others, of a microservice decomposition to optimize the use of the Edge architecture. Experiments demonstrate the benefits of our approach in terms of placement reliability, execution time and resources utilization.

MQTT is currently one of the most widely used protocols at the application layer for the Internet of Things (IoT). To manage an increasing number of IoT devices and optimize the transmission of information, the distribution of the MQTT broker (MQTT cluster) and the integration of the SDN technology in the MQTT architecture are nowadays strongly promoted ideas. Unfortunately, existing proposals have several limitations: 1) they are incompatible with currently deployed solutions, 2) they induce significant delays for information exchange between MQTT brokers within the same cluster, and 3) they do not take into account the distribution of data to the end users (subscribers). That is why, in this paper, we propose a new SDN-based Real-Time Distributed MQTT Broker: SoD-MQTT. The architecture and protocols described aim to minimize communication delays between brokers and to support very low latency applications (e-health, transportation). By leveraging the advantages of SDN, subscription, and publishing can also be efficiently managed. The conducted evaluations demonstrate the relevance of the solution in terms of latency and network utilization optimization compared to existing SDN-based MQTT brokers.

Software Defined Vehicular Networks (SDVN) are increasingly being considered for vehicular communication sys-tems as they could provide both flexibility, programmability and improved performance. The SDVN control plane should be distributed between a set of SDN controllers managing specific geographical areas and network functions (load balancing, fire-wall, etc.) to ensure a high level of scalability. In this environment, the Blockchain technology could be used to guarantee the coordination of the different SDVN controllers. Therefore, many works have demonstrated the benefits of Blockchain-based SDVN architectures. However, they did not investigate the optimization of the underlying Blockchain network. That is why, in this paper, we propose a high-performance multi-level Blockchain architecture which could enable SDVNs to scale up. We also introduce a Fuzzy logic-based mechanism allowing the real-time reconfiguration of the Blockchain network depending on the actual SDVN requirements: traffic increase, integration of a new controller, etc. We finally demonstrate, through experimentation, the benefits of our approach compared to existing solutions in terms of latency, CPU, throughput and scalability.

Network emulators are a key component for the implementation and evaluation of new applications. A network emulator designed for 5G networks should support the deployment of a complete 5G architecture (core and access network) as well as the deployment of edge data centers (edge computing) and the implementation of flexible SDN (Software Defined Networks)/NFV (Network Function Virtualization) architecture. To date, there is no environment enabling this combination of features. In this paper, we introduce a new network emulator named Emu5GNet, allowing the deployment of applications in complex 5G networks enabling both emulation of the network and data processing. Then, we present the architecture set up to achieve this objective and identify some of the applications that could be deployed in this environment: network performance, mobility management, etc. We also propose a quick evaluation to show how our emulator could potentially be used.

Software Defined Vehicular Networking (SDVN) could be the future of the vehicular networks, enabling interoperability between heterogeneous networks and mobility management. Thus, the deployment of large SDVN is considered. However, SDVN is facing major security issues, in particular, authentication and access control issues. Indeed, an unauthorized SDN controller could modify the behavior of switches (packet redirection, packet drops) and an unauthorized switch could disrupt the operation of the network (reconnaissance attack, malicious feedback). Due to the SDVN features (decentralization, mobility) and the SDVN requirements (flexibility, scalability), the Blockchain technology appears to be an efficient way to solve these authentication and access control issues. Therefore, many Blockchain-based approaches have already been proposed. However, two key challenges have not been addressed: authentication and access control for SDN controllers and high scalability for the underlying Blockchain network. That is why in this paper we propose an innovative and scalable architecture, based on a set of interconnected Blockchain sub-networks. Moreover, an efficient access control mechanism and a cross-sub-networks authentication/revocation mechanism are proposed for all SDVN devices (vehicles, roadside equipment, SDN controllers). To demonstrate the benefits of our approach, its performances are compared with existing solutions in terms of throughput, latency, CPU usage and read/write access to the Blockchain ledger. In addition, we determine an optimal number of Blockchain sub-networks according to different parameters such as the number of certificates to store and the number of requests to process.

The vehicular networks are evolving towards a new paradigm, the Internet of Vehicles. Different types of services, with different QoS requirements, should be enabled: road safety and traffic efficiency, infotainment and advertising, etc. For many of these services, a Publish/Subscribe (Pub/Sub) approach is an efficient way to enable a scalable and real-time data dissemination. That is why different Pub/Sub middlewares, in particular Software Defined Networking (SDN)-based Pub/Sub middlewares, have already been proposed. Indeed, thanks to the high level of flexibility and programmability of SDN, different improvements of the existing Pub/Sub middlewares could be enabled: QoS awareness, mobility management, etc. However, for many vehicular applications, data dissemination is closely linked to the geographic position of the emitter. To the best of our knowledge, the idea of a location-aware SDN-based Pub/Sub middleware has not been proposed yet. Thus, in this paper, we aim to demonstrate that, using the Openflow protocol and the SDN benefits, an efficient and flexible location-aware and QoS-aware middleware can be designed and implemented. To do so, this middleware, its main components and an algorithm enabling an efficient SDN-based QoS-aware geographic data dissemination are described. Moreover, the performances of the proposed architecture and mechanism are evaluated through experiments (latency, bandwidth efficiency, control overhead, etc.).