NET3 – Network & Service Orchestration
Thursday, 5 June 2025, 16:00-17:30, room 1.A
Session Chair: Angelos Antonopoulos (Nearby Computing, ES)
Dynamic QoS Adaptation via Exposed Service APIs for Advanced CAM Use Cases
Belma Turkovic, Rintse Van de Vlasakker, Srinath Potnuru, Yonatan Woldeleul Shiferaw, Ramon S. Schwartz and Peter-Paul Schackmann (TNO, The Netherlands); Gergely Attila Kovács (Commsignia Research, Irinyi József u. 4-20, 1117 Budapest, Hungary); Alexandru Forrai (Siemens, The Netherlands); Geerd Kakes and Pieter Span (KPN, The Netherlands)
The exposure of network functions has been addressed in different 3GPP specifications (such as in NEF, CAPIF and SEAL) to support vertical applications and services. Despite significant effort the integration of vertical services with the 3GPP exposure frameworks still requires network-specific knowledge and tight coordination between service providers and telcos, which may hinder the wider adoption of a 5G system for advanced vertical use cases. The ENVELOPE project addresses this limitation by transforming the reference 5G-Advanced architecture into a vertical-oriented one with the necessary interfaces tailored to vertical use cases of Connected and Automated Mobility (CAM). In line with this objective, this paper presents the architecture and 3GPP compliant message flow specifications that can serve as underlying implementations for two CAMARA service APIs: Device Location and Quality-on-Demand (QoD). We demonstrate the feasibility of these APIs by conducting a testbed campaign for a digital twin use case as part of the Dutch trial site activity of the project.
Network-Wide Service Deployment Using Centrally Orchestrated, eBPF-Based Programmable Dataplanes
Filip Holik (University of Glasgow, United Kingdom (Great Britain) & Norwegian University of Science and Technology, Norway); Marco M Cook, Awais Aziz Shah and Dimitrios P Pezaros (University of Glasgow, United Kingdom (Great Britain))
The separation of control and data planes along with controller centralisation have enabled more efficient network orchestration to support current diverse use cases. However, typical southbound protocols, for instance OpenFlow, have technical limitations such as statically defined match fields and no support for line rate per-packet processing for complex actions. Network programmatically implemented through P4 provides a more agile and dynamic solution to achieve line rate performance, however it can only be realistically deployed on specific hardware targets that are prohibitively expensive. Conversely, eBPF has recently gained traction as an alternative offering increased deployment flexibility and stateful programmability compared to conventional OpenFlow or P4 deployments, while also being hardware agnostic, hence addressing a wider range of devices and important use cases including 6G networks. In this paper, we propose the centralised control plane orchestration of eBPF-based programmable data plane devices. Our orchestrator uses eBPF programs instead of OpenFlow rules to perform packet processing and supports more complex code structures than P4. We emulate a three-layer hierarchical network model, and implement a congestion avoidance network service that autonomously manages eBPF functions to achieve sustainable and zero-touch network operations.
CAMINO: Cloud-Native Autonomous Management and Intent-Based Orchestrator
Konstantinos Antonakoglou, Ioannis Mavromatis, Saptarshi Ghosh, Mark Rouse and Konstantinos Katsaros (Digital Catapult, United Kingdom (Great Britain))
This paper introduces CAMINO, a Cloud-native Autonomous Management and Intent-based Orchestrator designed to address the challenges of scalable, declarative, and cloud-native service management and orchestration. CAMINO leverages a modular architecture, the Configuration-as-Data (CaD) paradigm, and real-time resource monitoring to facilitate zero-touch provisioning across multi-edge infrastructure. By incorporating intent-driven orchestration and observability capabilities, CAMINO enables automated lifecycle management of network functions, ensuring optimized resource utilization. The proposed solution abstracts complex configurations into high-level intents, offering a scalable approach to orchestrating services in distributed cloud-native infrastructures. This paper details CAMINO’s system architecture, implementation, and key benefits, highlighting its effectiveness in cloud-native telecommunications environments.
Evolution of Policy Verification and Control for 6G Secure Service Orchestration in Telco Clouds
Souvik Paul, Sheeba Backia Mary Baskaran and Andreas Kunz (Lenovo, Germany); Apostolis K. Salkintzis (Lenovo, Greece)
The telecommunications industry has undergone a significant transformation in the post-virtualization era of cloud computing. The deployment of 5G core network functions as stateless Virtualized Network Functions (VNFs) in cloud-native environments-such as Docker containers or Kubernetes pods-has enhanced operational flexibility and cost efficiency. These Network Functions (NFs) are managed based on Service Level Agreements (SLAs) and Quality of Service (QoS) requirements, with coordination between the Element Manager (EM) and the NFV Management and Orchestration (NFV MANO) framework. However, this coordination relies on policy enforcement by the NFV Orchestrator (NFVO) and VNF Manager (VNFM), which are vulnerable to policy manipulation or misinterpretation in the presence of cyber threats. Attacks on NFV MANO or the operator network can lead to orchestration compromises, policy violations, and service disruptions. While access control mechanisms provide some security, current NFVO implementations lack the capability to detect policy anomalies, potentially resulting in resource depletion and outages. To address this challenge, we propose a novel policy verification function within NFV MANO, following the zero-trust principle. Our approach employs an ML-based anomaly detection technique using One-Class Support Vector Machine (OCSVM) to identify anomalous policy parameters. We evaluated this method through integration with an operator’s core network using the open-source ETSI OSM MANO. Experimental results demonstrated an anomaly detection precision of 0.65, recall of 0.96, and an F1-score of 0.77, highlighting the effectiveness of our approach in enhancing orchestration security and reliability.
Exploiting 6G RAN and Core Network Information for Intelligent Edge-Cloud Service Orchestration
Michail Dalgitsis, Godfrey Mirondo Kibalya and Maria A. Serrano (Nearby Computing, Spain); Jordi Serra (CTTC, Spain); Angelos Antonopoulos (Nearby Computing, Spain)
The advent of 5G networks and the imminent evolution to 6G have driven significant changes in mobile network architecture, emphasizing edge-cloud integration, network softwarization, and Application Programmable Interface (API)-driven service delivery. To optimize resource allocation, service scaling, and service migration, mobile network operators must leverage both Radio Access Network (RAN) and Core Network (CN) information in orchestration frameworks. This paper introduces a novel cloud-native orchestration framework designed for 5G/6G networks, incorporating RAN and CN data to enable dynamic and efficient management of edge services. The framework includes three key orchestration strategies: (i) Horizontal Service Instance Autoscaler (HSIA), which adjusts service availability based on the number of active users derived from core network session data; (ii) Horizontal Service Resource Autoscaler (HSRA), which scales resources based on aggregated radio traffic from base stations; and (iii) Horizontal Service Edge-Cloud Migration (HSECM), which dynamically migrates services to the cloud when edge resources are insufficient. Experimental results demonstrate the effectiveness of these strategies in reducing energy consumption and minimizing users affected by service migration (UASM) while maintaining service quality. The findings highlight the transformative potential of mobile network-aware orchestration for enabling more sustainable and adaptive edge service deployments in next-generation networks.
