NET4 – Network Softwarisation
Thursday, 4 June 2026, 16:30-18:00, room Sala Dirección de Certámenes (1st floor)
Session Chair: Ramon Aguero (Univ. Cantabria, ES)
Enabling near Real-Time Slicing in Open-Source O-RAN Testbed Using Dynamic Resource Allocation
Iker Hernández and Zaloa Fernández (Fundación Vicomtech, Basque Research and Technology Alliance (BRTA), Spain); Alfonso Gómez, Andrea Campos and Juan Jose Fernández (Gamma Solutions, Spain)
The Open Radio Access Network (O-RAN) architecture is a fundamental paradigm for 5G and Beyond, promising programmable, service-driven resource management. This flexibility is essential to satisfy diverse Quality of Service (QoS) requirements through network slicing. However, realizing efficient closed-loop control in O-RAN infrastructures remains an open research challenge. Conventional implementations often lack the granular scheduler interfaces required to map application-layer demands to radio resources in real-time. To address this gap, this paper introduces a device-centric dynamic slicing framework specifically designed for private 5G Non-Public Networks (NPNs). This framework is built upon an extended E2 Radio Access Network Control (RC) Service Model (SM) and dedicated xApps. This mechanism augments the scheduler’s capabilities, enabling dynamic Physical Resource Block (PRB) limit allocation at the Transmission Time Interval (TTI) level. This logic has been integrated into an end-to-end 5G Standalone (SA) testbed combining OpenAirInterface (OAI), FlexRIC, real-time RAN telemetry via a monitoring xApp, and UE-embedded Deep Packet Inspection (DPI) for traffic classification. The proposed solution remains compatible with O-RAN specifications and is experimentally validated, demonstrating effective Service Level Agreement (SLA) enforcement with deterministic algorithmic processing (approx. 30 us) and minimal scheduling overhead (< 0.2% of the TTI budget), confirming its viability for latency-sensitive industrial environments and demonstrating significant isolation improvements over conventional Proportional Fair scheduling.
Towards End-to-End Network Slicing: Slice-Aware 5G Backhaul with Isolation and QoS Guarantees
Xhulio Limani (University of Antwerp, Belgium & Imec, The Netherlands); Lorenzo Riccardi (DISI Bologna, Italy); Armir Bujari, Alessandro Calvio and Luca Foschini (University of Bologna, Italy); Miguel Camelo Botero (University of Antwerp – imec, Belgium); Johann Marquez-Barja (University of Antwerpen & IMEC, Belgium); Nina Slamnik-Krijestorac (University of Antwerp-IMEC, Belgium)
Fifth-generation (5G) network slicing enables multiple logical networks to share the same infrastructure, each designed to support distinct Service Level Agreements (SLAs). While network slicing is well supported in the Radio Access Network (RAN) and 5G Core (5GC), the Transport Network (TN) backhaul is often treated as a slice-agnostic forwarding substrate. As a result, resource contention on shared links can significantly degrade network performance, such as latency and throughput, breaking end-to-end SLAs.This paper addresses this gap by making the TN slice-aware through a lightweight framework that translates slice requirements into concrete backhaul policies. The framework (i) associates each slice with TN identifiers at domain boundaries, (ii) turns slice requirements into per-slice forwarding and scheduling policies (e.g., priority and queue selection, bandwidth bounds, and optional congestion signals), and (iii) exports per-slice measurements for monitoring and closed-loop control. We deploy a proof of concept on a programmable switch and evaluate it on a real-life 5G testbed. Our results show that, under link saturation, slices designed for mission-critical services preserve their SLAs, while best-effort traffic degrades predictably.
ORACLE: Open RAN Arbitration via Consensus Ledger Enforcement
Joss Armstrong (LM Ericsson, Ireland)
Open RAN promises innovation through vendor diversity, but realising this vision requires conflict arbitration that no single vendor controls. When rApps from competing vendors propose conflicting network configurations, operators need assurance that arbitration is impartial yet current approaches rely on centralised coordination that vendors cannot independently verify. This paper presents ORACLE, a vendor neutral arbitration mechanism for Open RAN rApp conflicts, implemented using a permissioned distributed ledger. This ledger records proposals, predictions, and outcomes as immutable shared state, while smart contracts execute decision functions deterministically. All vendors follow a uniform workflow, and no party, including the infrastructure operator, can manipulate inputs or selectively omit outcomes. We evaluate ORACLE through live integration with an ns- 3 O-RAN simulation, demonstrating practical decision latency within Non-RT RIC timing constraints and linear gas scaling with participant count. The enforcement layer provides the transparency and auditability that multi-vendor deployments require.
RAG-Enabled Intent Reasoning for Application-Network Interaction
Salwa Said Hamed Mostafa (University of Oulu, Finland); Mehdi Bennis (Centre of Wireless Communications, University of Oulu, Finland); Mohammed Elbamby (Telefónica Research, Spain); Mohamed K. Abdel-Aziz (Nokia Bell Labs, Finland)
Intent-based network (IBN) is a promising solution to automate network operation and management. IBN aims to offer human-tailored network interaction, allowing the network to communicate in a way that aligns with the network users’ language, rather than requiring the network users to understand the technical language of the network/devices. Nowadays, different applications interact with the network, each with its own specialized needs and domain language. Creating semantic languages (i.e., ontology-based languages) and associating them with each application to facilitate intent translation lacks technical expertise and is neither practical nor scalable. To tackle the aforementioned problem, we propose a context-aware AI framework that utilizes machine reasoning (MR), retrieval augmented generation (RAG), and generative AI technolo- gies to interpret intents from different applications and generate structured network intents. The proposed framework allows for generalized/domain-specific intent expression and overcomes the drawbacks of large language models (LLMs) and vanilla-RAG framework. The experimental results show that our proposed intent-RAG framework outperforms the LLM and vanilla-RAG framework in intent translation.
An Open-Source CAMARA UE Location API Implementation over 3GPP NEF for AI-Native Crowd Assessment in 6G Networks
Stefanos Plastras (National Center for Scientific Research Demokritos, Greece); Panagiotis Pavlidis and Spyridon Georgoulas (National Centre for Scientific Research Demokritos, Greece); George Makropoulos (NCSR Demokritos, Greece & National and Kapodistrian University of Athens, Greece); Vasilis Pitsilis and Harilaos Koumaras (NCSR Demokritos, Greece)
Network Application Programming Interfaces (APIs) and exposure capabilities are emerging as key enablers for 5G and 6G applications’ ecosystems, allowing third-party developers to consume fine-grained network information in a standardized way. Building on this premise, this paper presents a proof-of-concept (PoC) implementation that exposes 5G Core (5GC) User Equipment (UE) location information to application developers by combining the Common API Marketplace and Repository Architecture (CAMARA) Device Location – Location Retrieval API, the open-source implementation for the Common API Framework (OpenCAPIF) and a 3rd Generation Partnership Project (3GPP)-compliant Network Exposure Function (NEF) Event Exposure service. The proposed framework also introduces a Transformation Function (TF) that maps CAMARA-compliant APIs to 3GPP MonitoringEvent operations. By translating high level requests into 5GC payloads, the TF enables developers to interact with complex network functions through simplified interfaces. The implementation is realized using an open-source 5GC environment and repository that provides the complete CAMARA, OpenCAPIF and NEF integration logic, including configuration and execution objects. In this light, the proposed framework is validated through an experimental workflow that exercises UE last-known-location retrieval from an external Artificial Intelligence (AI)-native crowd application assessing crowd mobility in urban environments. The results demonstrate the feasibility of exposing UE location capabilities through standardized network APIs and highlight design considerations that are directly applicable to future 6G network exposure platforms.
