Deliverable D5.3: Report on Communication, dissemination, and exploitation results and updated CoDEP for Y2

ORIGAMI Optimized resource integration and global architecture for mobile infrastructure for 6G Deliverable D5.3: Report on Communication, dissemination, and exploitation results and updated CoDEP for Y2 Date: 28/11/2025 Version: V1.0 Grant Agreement 101139270 — ORIGAMI — HORIZON-JU-SNS-2023 Deliverable D5.3 DISCLAIMER This document contains information, which is proprietary to the ORIGAMI ("Optimized resource integration and global architecture for mobile infrastructure for 6G”) Consortium that is subject to the rights and obligations and to the terms and conditions applicable to the Grant Agreement number: 101139270. The action of the ORIGAMI Consortium is funded by the European Commission. Neither this document nor the information contained herein shall be used, copied, duplicated, reproduced, modified, or communicated by any means to any third party, in whole or in parts, except with prior written consent of the ORIGAMI Consortium. In such case, an acknowledgement of the authors of the document and all applicable portions of the copyright notice must be clearly referenced. In the event of infringement, the consortium reserves the right to take any legal action it deems appropriate. This document reflects only the authors’ view and does not necessarily reflect the view of the European Commission. Neither the ORIGAMI Consortium as a whole, nor a certain party of the ORIGAMI Consortium warrant that the information contained in this document is suitable for use, nor that the use of the information is accurate or free from risk and accepts no liability for loss or damage suffered by any person using this information. The information in this document is provided as is and no guarantee or warranty is given that the information is fit for any particular purpose. The user thereof uses the information at its sole risk and liability. Grant Agreement 101139270 — ORIGAMI — HORIZON-JU-SNS-2023 Deliverable D5.3 Grant Agreement 101139270 Document number D5.3 Document title Report on Communication, dissemination, and exploitation results and updated CoDEP for Y2 Lead Beneficiary NETAI Editor(s) Paul Patras (NETAI) Author(s) Paul Patras (NETAI) Nikos Passas (FOGUS) Dimitris Tsolkas (FOGUS) Marco Fiore (IMDEA) Mika Skarp (CMC) Andres Saavedra-Garcia (NEC) Jose Ayala Romero (NEC) Steffen Gebert (EMN) Marco Gramaglia (UC3M) Javier Garcia Rodrigo (TID) Dena Markudova (TID) Andra Lutu (TID) Maria Pia Galante (TIM) Md Arifur Rahman (ISRD) Simone Bizzarri (FBC) Maurizio Fodrini (FBC) Marco Gramaglia (UC3M) Esteban Municio (i2CAT) Georgios Iosifidis (TUD) Dissemination level Public Contractual date of delivery 30.11.2025 Status Final File name ORIGAMI_D5.3_V1.0.pdf Revision History Version V0.1 Initial Table of Contents V0.2 First draft V0.3 Pre-final version V1.0 Final version Grant Agreement 101139270 — ORIGAMI — HORIZON-JU-SNS-2023 Deliverable D5.3 GLOSSARY Abbreviations/ Acronym Description 3GPP 3rd Generation Partnership Project 6GArch 6G Architecture Workshop Athena Adaptive Frameworks and Global Architectural Evolution AI Artificial Intelligence CoDEP Communication, Dissemination, and Exploitation CCL Compute Continuum Layer DTW Digital Transformation World DU Distributed Unit ETSI European Telecommunications Standards Institute GMNO Global Mobile Network Operator GSBA Global Service-based Architecture GB Governing Board HA Hardware Accelerator ISAC Integrated Sensing and Communication ITU International Telecommunication Union KER Key Exploitable Result KPI Key Performance Indicator KVI Key Value Indicators ML Machine Learning MNO Mobile Network Operator MOCN Multi-Operator Core Network MWC Mobile World Congress NI Network Intelligence NR New Radio O-RAN Open Radio Access Network RAN Radio Access Network RIC Radio Intelligent Controller SA Service and System Aspects SBA Service-Based Architecture SDO Standards Developing Organization SME Small and Medium-Sized Enterprise SNS JU Smart Networks and Services Undertaking SNS CA SNS Collaboration Agreement SB Steering Board TF Task Force TPC Technical Committee Member UDF Unified Data Framework VNF Virtual Network Function vRAN virtualized Radio Access Networks WP Work Package WG Working Group XR eXtended Reality ZTL Zero-Trust Exposure Layer ZSM Zero-touch Network and Service Management Grant Agreement 101139270 — ORIGAMI — HORIZON-JU-SNS-2023 Deliverable D5.3 EXECUTIVE SUMMARY This deliverable summarises the communication, dissemination, exploitation, and standardization (CoDEP) activities conducted during the second year (Y2) of the ORIGAMI project and presents the updated plan for the final year (Y3). During Y2, ORIGAMI significantly strengthened its influence on 3GPP standardization, particularly through a coordinated strategy targeting RAN3 to pave the way for a study item on 6G RAN architectural evolution. These efforts aim to position ORIGAMI’s key architectural concepts (GSBA, ZTL, and CCL) as relevant enablers for future 6G systems. ORIGAMI also deepened its involvement within the SNS-JU and 6G-IA community, contributing to strategic white papers thus helping shape the shared European vision for 6G. Dissemination activities continued to excel, with a large number of high-impact publications, several best paper awards, and strong visibility across major scientific and industrial events. Communication activities expanded the project’s reach through sustained online presence, press releases, and professional video content. The deliverable concludes with an enhanced CoDEP plan for Y3, focused on strengthening standardization impact, maximizing visibility of final project achievements, and supporting the exploitation of ORIGAMI’s technical results. KEYWORDS 6G, Communication, dissemination, exploitation, standardization, open source, strategy Grant Agreement 101139270 — ORIGAMI — HORIZON-JU-SNS-2023 Deliverable D5.3 TABLE OF CONTENTS 1 Introduction ............................................................................................................... 1 1.1 Timeline of the ORIGAMI CoDEP ..................................................................................... 2 2 Communication activities ........................................................................................... 3 2.1 Communication activities during Y2 ................................................................................ 3 2.2 Communication activities planned for Y3 ...................................................................... 12 3 Dissemination activities ............................................................................................ 13 3.1 Dissemination activities during Y2 ................................................................................. 13 3.1.1 Specific dissemination activities ....................................................................................................... 13 3.1.2 Liaison activities with 6G-IA and external fora ................................................................................. 34 3.2 Updated dissemination plan for Y3 ............................................................................... 42 4 Exploitation activities ............................................................................................... 44 4.1 PROJECT RESULTS AND EXPECTED OUTCOMES BY Y2 ..................................................... 44 4.1.1 Barrier #1: Unsustainable RAN virtualization ................................................................................... 45 4.1.2 Barrier #2: Poor inter-operability of RAN components .................................................................... 47 4.1.3 Barrier #3: High latency to process complex 6G network problems ................................................ 53 4.1.4 Barrier #4: Under-utilized modern programmable transport ........................................................... 54 4.1.5 Barrier #5: Lack of global service APIs .............................................................................................. 55 4.1.6 Barrier #7: Inadequate networking data representation ................................................................. 60 4.1.7 Barrier #8: High control-plane signaling overhead ........................................................................... 64 4.2 INDIVIDUAL EXPLOITATION ACTIVITIES AND PLANS ....................................................... 67 4.2.1 Industrial Partners ............................................................................................................................ 67 4.2.2 Academic Partners ............................................................................................................................ 68 4.3 Standardization activities and Open-source contributions ............................................. 70 4.3.1 Standardization Activities during Y2 ................................................................................................. 70 4.3.2 Standardization activities planned for Y3 ......................................................................................... 76 4.3.3 Open-source contributions planned for Y3 (per open-source fora) ................................................. 78 5 Target Indicators achievement in Y2 ......................................................................... 79 6 Conclusion and next steps......................................................................................... 82 7 References ............................................................................................................... 83 8 Appendix I – Individual Exploitation Plans (Industrial Partners) ................................. 84 9 Appendix II - Individual Exploitation Plans (Academic Partners) .............................. 105 Grant Agreement 101139270 — ORIGAMI — HORIZON-JU-SNS-2023 Deliverable D5.3 List of FIGURES Figure 1: Illustration of ORIGAMI’s webpage structure, highlighting changes compared to D5.2. ........................ 4 Figure 2: Website statistics: (i) active users per country, (ii) active users per language, (iii) number of active users over time. ............................................................................................................................................................... 5 Figure 3: Website statistics: (i) statistics per page, (ii) device category, (iii) number of views. ............................. 5 Figure 4: Social media presence on LinkedIn. ......................................................................................................... 7 Figure 5: LinkedIn statistics: impressions, engagements, and followers. ............................................................... 7 Figure 6: LinkedIn statistics: demographics. ........................................................................................................... 7 Figure 7: LinkedIn statistics: top performing posts. ............................................................................................... 7 Figure 8: Social media presence on X. .................................................................................................................... 8 Figure 9: Social media presence on YouTube. ....................................................................................................... 8 Figure 10: YouTube statistics. ................................................................................................................................. 9 Figure 11: Social media presence on Instagram. .................................................................................................... 9 Figure 12: Instagram statistics: number of posts, followers, engagement, and user activity. .............................. 9 Figure 13: Instagram statistics: top most liked posts. ......................................................................................... 10 Figure 14: ORIGAMI press release (EMN). ............................................................................................................ 10 Figure 15: Videos of the project vision, solutions, and outcomes. ....................................................................... 11 Figure 16: ORIGAMI team supporting the booth at EUCNC 2025. ...................................................................... 19 Figure 17: IS-Wireless presenting the demo at EUCNC 2025. .............................................................................. 20 Figure 18: Md Arifur Rahman delivering the exhibition pitch at EUCNC 2025. .................................................... 21 Figure 19: LinkedIn updates on EUCNC 2025. ...................................................................................................... 22 Figure 20: X updates on EUCNC 2025. .................................................................................................................. 23 Figure 21: Instagram updates on EUCNC 2025. .................................................................................................... 23 Figure 22: Social media updates on Business Models and Trustworthiness KVI talk. .......................................... 24 Figure 23: Social media updates on Paul Patras’s talk at Open Future Networks Showcase. .............................. 24 Figure 24: Social media updates on BOWW talk. ................................................................................................. 25 Figure 25: Social media updates on Paul Patras’s talk at IEEE HPCC. ................................................................... 26 Figure 26: LinkedIn updates on panel at IEEE PIMRC. .......................................................................................... 27 Figure 27: Social media posts on co-sponsoring WueWoWAS’25. ....................................................................... 28 Figure 28: Promotion of WueWoWAS’25 on 6G SNS webpage. ........................................................................... 29 Figure 29: Screenshots from SNS JU Journal 2025, featuring the ORIGAMI project. ........................................... 30 Figure 30: Interactive map of SNS projects, featuring the ORIGAMI project. ...................................................... 30 Figure 31: Promotion of Elsevier Computer Communications edited special issue on 6G SNS webpage. ........... 31 Figure 32: Social media updates on Devoxx France 2025. ................................................................................... 32 Figure 33: Social media updates on Connected Britain 2025 activities. ............................................................... 33 Figure 34: Social media updates on Barcelona DeepTech Summit 2025 activities. ............................................. 34 Figure 35 ORIGAMI Policy feedback ..................................................................................................................... 37 Figure 36 Meta Registry System (MRS) ................................................................................................................ 41 Figure 37 3GPP Release 20 6G Timeline ............................................................................................................... 71 Grant Agreement 101139270 — ORIGAMI — HORIZON-JU-SNS-2023 Deliverable D5.3 List of TABLES Table 1: ORIGAMI KPIs for communication activities during the third year of the project .................................. 12 Table 2: List of scientific publications produced by M23...................................................................................... 18 Table 3: SNS/6G IA WG representatives ............................................................................................................... 36 Table 4: ORIGAMI KPIs for dissemination activities during the second year of the project ................................. 43 Table 5: Achieved ORIGAMI KPIs for communication, dissemination, and exploitation during the second year of the project ............................................................................................................................................................ 81 1 Grant Agreement 101139270 — ORIGAMI — HORIZON-JU-SNS-2023 Deliverable D5.3 1 INTRODUCTION This deliverable reports the communication, dissemination, exploitation, and standardization achievements of the ORIGAMI project during its second year (Y2), while outlining the strategic priorities and planned activities for the final year (Y3). As ORIGAMI advances its mission to define an intelligent, sustainable, and interoperable architectural framework for 6G networks, the project has strengthened its impact across scientific, industrial, and standardization communities, consolidating its position as one of the leading contributors within the European 6G research landscape. A major focus of Y2 has been the intensification of standardization activities within 3GPP, particularly in preparation for contributions to the RAN3 Working Group. ORIGAMI partners have engaged in a coordinated strategy aiming to open a study item on architectural evolution for 6G RAN, including the investigation of service-based principles, compute-connectivity convergence, and cross-layer exposure mechanisms. By proactively shaping the agenda toward Release 20 discussions, ORIGAMI positions its core architectural pillars—the Global Service-Based Architecture (GSBA), the Zero-Trust Exposure Layer (ZTL), and the Compute Continuum Layer (CCL)—as candidate enablers for the nextgeneration RAN. These strategic efforts reflect the project’s commitment to ensuring that its technical innovations contribute directly to global standards with long-term impact. In parallel, ORIGAMI has strengthened its role in European collaborative frameworks, most notably within the SNS-JU and the 6G-IA Working Groups. Project partners have taken editorial and leadership roles in key community documents, including the SNS Architecture Working Group White Paper and the SNS Sustainability Task Force White Paper, both of which influence the strategic research direction for 6G in Europe. ORIGAMI’s contributions to these cross-project initiatives ensure strong alignment with the European 6G vision and amplify the project’s visibility across the entire ecosystem of research, industry, and policy stakeholders. The excellence of ORIGAMI’s dissemination activities is evidenced by its outstanding scientific output, with numerous publications in top-tier venues (CORE A*, JCR Q1) and several best paper awards earned during Y2. Demonstrations at flagship international conferences, invited talks, keynotes, industrial showcases, and active participation in panels and workshops have enabled ORIGAMI to communicate its results to both academic communities and operational stakeholders in the telecommunications industry. This broad dissemination approach has positioned ORIGAMI as a reference project for next-generation architecture design, AI-enabled network intelligence, and sustainability-driven innovation. The remainder of this deliverable is structured as follows: • Section 2 details the communication activities conducted during Y2 and the updated plan for Y3. • Section 3 provides an in-depth overview of dissemination activities, including scientific publications, events, and engagement with the 6G-IA and SNS-JU community. • Section 4 summarises the project’s exploitation strategy and describes progress toward the Key Exploitable Results (KERs). • Section 5 evaluates the achievement of CoDEP-related KPIs • Section 6 presents conclusions and outlines the next steps as ORIGAMI enters its final year. 8 Grant Agreement 101139270 — ORIGAMI — HORIZON-JU-SNS-2023 Deliverable D5.3 • X: https://x.com/sns_origami Figure 8: Social media presence on X. To date, X has gained 111 followers and published 243 posts. X restricts access to advanced analytics features for non-Premium accounts. Only Premium members can view comprehensive engagement metrics, such as impressions, profile visits, and audience breakdowns. • YouTube: https://www.youtube.com/@sns-origami The YouTube channel has gained 32 subscribers by publishing six videos of the project, as shown in Figure 9. Figure 9: Social media presence on YouTube. Figure 10 presents statistical data from the past year concerning all six videos published on the YouTube platform. The cumulative number of impressions, defined as the frequency with which video 9 Grant Agreement 101139270 — ORIGAMI — HORIZON-JU-SNS-2023 Deliverable D5.3 thumbnails were displayed to viewers, is 6,650. In contrast, the total number of video views amounts to 419. The average view duration is recorded at 1 minute and 14 seconds. Figure 10: YouTube statistics. • Instagram: https://www.instagram.com/sns_origami/ Instagram has gained 52 followers, and the total number of posts amounts to 37 (see Figure 11). In Figure 12, more detailed statistics on the number of posts, followers, engagement, and user activity are presented, while Figure 13 provides the top-most-liked posts. Figure 11: Social media presence on Instagram. Figure 12: Instagram statistics: number of posts, followers, engagement, and user activity. 10 Grant Agreement 101139270 — ORIGAMI — HORIZON-JU-SNS-2023 Deliverable D5.3 Figure 13: Instagram statistics: top most liked posts. Project’s press releases EMN issued a press release in November 2025 following the midterm review to announce the company’s involvement in the ORIGAMI project and outline its role, contributions, and innovation activities within the consortium [3]. The press release presented EMN’s vision for advancing cloudnative IoT connectivity in the transition toward 6G networks, highlighting its work on real-world operator use cases and its leadership in driving architectural research for next-generation communication systems. The press release is depicted in Figure 14 below. Figure 14: ORIGAMI press release (EMN). 11 Grant Agreement 101139270 — ORIGAMI — HORIZON-JU-SNS-2023 Deliverable D5.3 Videos of the project vision, solutions, and outcomes In addition to the three videos reported in D5.2, ORIGAMI has produced three more public videos, as can be seen in Figure 15. Figure 15: Videos of the project vision, solutions, and outcomes. In the professional video “ORIGAMI Project overview: Breaking Barriers to Next-Generation Mobile Networks”, project leaders introduce ORIGAMI’s mission, the collaborative effort of its consortium, and its innovative role in shaping the future of 6G, diving into the vision, goals, and transformative potential of 6G networks. The viewers can discover the ORIGAMI project, a groundbreaking European initiative funded by the European Commission under the 6G Smart Networks and Services Joint Undertaking (SNS JU). Bringing together leading players from industry and academia, ORIGAMI focuses on advancing next-generation mobile network architectures and breaking down barriers to 6G technologies. The demo “Demo: Demonstrating Distributed Inference in the User Plane with DUNE” was presented at IMFOCOM 2025 in London, United Kingdom, on 19-22 May 2025, and at EUCNC 2025 in Poznan, Poland, on 3-6 June 2025. Deploying Machine Learning (ML) models in the user plane enables lowlatency and scalable in-network inference but integrating them into programmable devices faces stringent constraints in terms of memory resources and computing capabilities. In this demo, we show how the newly proposed DUNE, a novel framework for distributed user-plane inference across multiple programmable network devices by automating the decomposition of large ML models into smaller sub-models, mitigates the limitations of traditional monolithic ML designs. We run experiments on a testbed with Intel Tofino switches using measurement data and show how DUNE not only improves the accuracy that the traditional single-device monolithic approach gets but also maintains a comparable per-switch latency. The last video, “ORIGAMI at EUCNC 2025”, provides a brief glimpse behind the scenes at EUCNC 2025 in Poznan, Poland, one of the flagship events for 6G research in Europe, where the ORIGAMI project 12 Grant Agreement 101139270 — ORIGAMI — HORIZON-JU-SNS-2023 Deliverable D5.3 had a significant presence. With 8 key contributions, ORIGAMI showcased its technical innovation, standardization potential, and societal impact. More details on the ORIGAMI’s contributions at EUCNC 2025 are provided in Section 3.1. 2.2 COMMUNICATION ACTIVITIES PLANNED FOR Y3 This section outlines the planned communication activities aimed at effectively disseminating the project’s objectives, milestones, and outcomes during Y3. The communication strategy continues to focus on ensuring consistent engagement with key stakeholders and enhancing the project’s visibility through all official channels, including the website, social media platforms, and other outreach tools. As detailed in Section 5, ORIGAMI has successfully met or exceeded its communication and dissemination Key Performance Indicators (KPIs) for Y1 and Y2, covering the website, social media channels, videos, and press materials. Building on these achievements, Y3 will focus on consolidating project visibility and communicating the final results and impacts. The following paragraphs outline the approach for the main communication activities planned for Y3, highlighting the specific channels, tools, and indicators that will guide implementation and evaluation. In terms of press releases, three press releases were published across the first two years of the project. During Y3, up to seven additional press releases are planned, primarily focusing on major outcomes, demonstrations, and final achievements. These will bring the total to ten press releases by the end of the project, ensuring broad dissemination of ORIGAMI’s results to relevant audiences and media outlets. We note that JMU is already preparing the fourth press release. Regarding video production, several videos have already been released, including thematic videos on technical solutions and a professional overview of the project produced in Y2. For Y3, the plan is to develop at least one additional video—potentially more—summarizing the project’s key innovations, results, and impact stories. This video content will serve as a comprehensive communication tool to promote ORIGAMI’s achievements beyond the project’s duration. Thus, Table 1 lists the specific communication activities to be attained by M36 of the project execution. Nature Community KPI Target at M24/M36 Expected deviations Communication General Press releases 4/10 3 press releases were issued in Y2. We expect this to be rectified by the end of the project. Videos of project vision, solutions, and outcomes 4/5 None Other Table 1: ORIGAMI KPIs for communication activities during the third year of the project 13 Grant Agreement 101139270 — ORIGAMI — HORIZON-JU-SNS-2023 Deliverable D5.3 3 DISSEMINATION ACTIVITIES Dissemination activities encompass initiatives related to enhancing awareness of project outcomes within the technical community engaged in the broad field of computer science, with a specific focus on communication systems. This will primarily be achieved through scientific publications, the organization and/or participation in scientific conferences and industrial events. 3.1 DISSEMINATION ACTIVITIES DURING Y2 3.1.1 SPECIFIC DISSEMINATION ACTIVITIES The planned dissemination activities are carried out continuously when there is the appropriate combination of availability of project results and opportunity, as presented next. 3.1.1.1 SCIENTIFIC PUBLICATIONS To date, the project has generated 52 scientific publications. The list of scientific publications produced by M23 is included below in Table 2. For each publication (either conference or journal), the title of the paper, the name of the venue, the authors, their affiliations to project partners, the ranking of the venue according to CORE, and the relevant project tasks are presented. No. Title/Conference Authors/Partners Ranking/Tasks 1 Risk-Aware Continuous Control with Neural Contextual Bandits J. A. Ayala-Romero, A. GarciaSaavedra, X. Costa-Perez A* AAAI 2024 NEC T3.1, T3.2, T.3.3 2 Mean-Field Multi-Agent Contextual Bandit for EnergyEfficient Resource Allocation in vRANs J. A. Ayala-Romero, L. Lo Schiavo, A. Garcia-Saavedra, X. CostaPerez A* IEEE INFOCOM 2024 NEC, IMDEA, UC3M T3.2 3 Encrypted Traffic Classification at Line Rate in Programmable Switches with Machine Learning A.T.-J. Akem, G. Fraysse, M. Fiore B IEEE NOMS 2024 IMDEA T3.3 4 CloudRIC: Open Radio Access Network (O-RAN) Virtualization with Shared Heterogeneous Computing L. Schiavo, G. Garcia-Aviles, A. Saavedra, M. Gramaglia, M. Fiore, A. Banchs, X. Costa-Perez A* ACM MobiCom 2024 NEC, IMDEA, UC3M T3.2 5 A Shortcut Through the IPX: Measuring Latencies in Global Mobile Roaming with Regional Breakouts V. Vomhoff, M. Sichermann, S. Geissler, M. Giess, A. Lutu, Tobias Hoßfeld Other IEEE/IFIP TMA JMU, EMN, TID T3.1, T3.3, T4.1 6 Fair Resource Allocation in Virtualized O-RAN Platforms F. Aslan, G. Iosifidis, J. R. Ayala, A. G. Saavedra, X. Costa-Perez A* ACM Sigmetrics 2024 TUD, NEC T3.2 7 Adaptive Online Non-Stochastic Control N. Mhaisen, G. Iosifidis Other PMLP L4DC TUD T3.2 14 Grant Agreement 101139270 — ORIGAMI — HORIZON-JU-SNS-2023 Deliverable D5.3 8 ATELIER: service tailored and limited-trust network analytics using cooperative learning M. Milani, D. Bega, M. Gramaglia, P. Serrano, C. Mannweiler Q1 IEEE Open Journal of the Communications Society UC3M T3.1 9 Design and validation of scalable reconfigurable intelligent surfaces M. Rossanese, P. Mursia, A. Garcia-Saavedra, V. Sciancalepore, A. Asadi, X. CostaPerez Q1 Elsevier Computer Networks NEC T4.1 10 YinYangRAN: Resource Multiplexing in GPU-Accelerated Virtualized RANs L. Lo Schiavo, J. A. Ayala-Romero, A. Garcia-Saavedra, M. Fiore, X. Costa-Perez A* IEEE INFOCOM 2024 NEC, UC3M, IMDEA, i2CAT T3.2 11 Optimistic Online Non-stochastic Control via FTRL N. Mhaisen, G. Iosifidis Other IEEE CDC TUD T3.2 12 Attacking O-RAN Interfaces: Threat Modeling, Analysis and Practical Experimentation P. Baguer, G. Yilma, E. Municio, G. Garcia-Aviles, A. GarciaSaavedra, M. Liebsch, X. CostaPerez Q1 IEEE Open Journal of the Communications Society I2C, NEC T3.2, T4.1 13 Enabling Beyond-Visual-Line-ofSight Drones Operation over Open RAN 5G Networks with Slicing P. Baguer, E. Municio, G. GarciaAviles, X. Costa-Pérez Q1 IEEE Network I2C, NEC T3.2, T4.1 14 CloudRIC demo: Open Radio Access Network (O-RAN) Virtualization with Shared Heterogeneous Computing L.L. Schiavo, G. Garcia-Aviles, A. Garcia-Saavedra, M. Gramaglia, M. Fiore, A. Banchs; X. Costa Pérez A* ACM Mobicom 2024 NEC, IMDEA, UC3M, i2CAT T3.2 15 Parameterizing 5G New Radio: A Comparative Measurement Study on Throughput and Delay S. Raffeck, S. Grøsvik, S. Lange, T. Hoßfeld, T. Zinner, S. Geißler B IEEE CNSM 2024 JMU T4.1 16 Through the Telco Lens: A Countrywide Empirical Study of Cellular Handovers M. Kalntis, J. Suárez-Varela, J. Omaña Iglesias, A. Kiran Bhattacharjee, G. Iosifidis, F. A. Kuipers, A. Lutu A* ACM IMC 2024 TUD, TID T2.2 17 Adaptive Resource Allocation for Virtualized Base Stations in ORAN with Online Learning M. Kalntis, G. Iosifidis, F. A. Kuipers Q1 IEEE Transactions on Communications TUD T3.2 18 A Service Mesh Platform for the Mobile Network Core I. Fotis, G. Gkionis, A. Charismiadis, D. Tsolkas - 15 Grant Agreement 101139270 — ORIGAMI — HORIZON-JU-SNS-2023 Deliverable D5.3 6G SNS Software and Standards for Smart Networks and Services Conference & Hackfests FOGUS T4.1 19 Smooth Handovers via Smoothed Online Learning M. Kalntis, A. Lutu, J. Omaña Iglesias, F. A. Kuipers, G. Iosifidis A* IEEE INFOCOM 2025 TUD, TID Τ3.1, Τ3.2 20 Poster: A First Look at IPX Hub Breakout with Airalo H. D. Jang, M. Varvello, A. Lutu, Y. Zaki A ACM IMC 2024 TID T3.1, T3.3, T4.1 21 Towards 6G: Architectural Innovations and Challenges in the ORIGAMI Framework L. E. Chatzieleftheriou, M. Gramaglia, A. Garcia-Saavedra, S. Gebert, G. Garcia-Aviles, S. Geissler, M. Fiore, P. Patras, A. Lutu, D. Tsolkas, Md. A. Rahman Other EuCNC & 6G Summit 2024 UC3M, IMDEA, NEC, i2CAT, EMN, NetAI, TID, Fogus, IS-Wireless T2.1, T2.3 22 Towards Data-Driven Management of Mobile Networks through User Plane Inference A.T.-J. Akem, M. Fiore B IEEE Network Operations and Management Symposium IMDEA T3.3, T4.2 23 Ultra-Low Latency User-Plane Cyberattack Detection in SDNbased Smart Grids A.T.-J. Akem, M. Gucciardo, M. Fiore Other ACM International Conference on Future and Sustainable Energy Systems IMDEA T3.3, T4.2 24 FairRIC: Real-time Fair Allocation in O-RAN with Shared Computing F. Aslan, J. A. Ayala-Romero, A. Garcia-Saavedra, X. Costa-Perez, G. Iosifidis A* IEEE INFOCOM 2025 TUD, NEC, i2CAT T3.2 25 Towards Real-Time Intrusion Detection in P4-Programmable 5G User Plane Functions A.T.-J. Akem, M. Fiore B International Conference on Network Protocols IMDEA T3.3, T4.2 26 Learning to Learn How to Manage Network Resources with Loss Function Meta-Learning A. Collet, A. Bazco-Nogueras, A. Banchs, M. Fiore Q1 IEEE Communications Magazine IMDEA T3.1 27 Real-Time Encrypted Traffic Classification in Programmable Networks with P4 and Machine Learning A.T.-J. Akem, G. Fraysse, M. Fiore Q2 International Journal of Network Management IMDEA T3.3, T4.2 28 An Urban Geography of Mobile Application Usage: Connecting Demand Dynamics and Urban Fabrics S. Mishra, D. Madariaga, C. Ziemlicki, D. Naboulsi, M. Fiore A* IEEE INFOCOM 2025 IMDEA, UC3M T3.2 16 Grant Agreement 101139270 — ORIGAMI — HORIZON-JU-SNS-2023 Deliverable D5.3 29 An Evaluation of RAN Sustainability Strategies in Production Networks O.E. Martínez-Durive, J. SuárezVarela, J. Omaña Iglesias, A. Lutu, M. Fiore A* IEEE INFOCOM 2025 IMDEA, TID, UC3M T3.2 30 DUNE: Distributed Inference in the User Plane B. Bütün, D. de Andrés Hernández, M. Gucciardo, M. Fiore A* IEEE INFOCOM 2025 IMDEA, NEC, UC3M T3.3, T4.2 31 O-RAN Intelligence Orchestration Framework for Quality-Driven Xapp Deployment and Sharing F. Mungari, C. Puligheddu, A. Garcia-Saavedra, C. F. Chiasserini Q1 IEEE Transactions on Mobile Computing NEC T3.3, T4.2 32 AI/ML as a key enabler of 6G Networks - Methodology, approach and AI mechanisms in SNS JU A. Garcia-Saavedra, M. Gramaglia, M. Fiore et al Other 6GSNS White Paper NEC, ALL T2.1, T2.3 33 Practical and General-Purpose Flow-Level Inference with Random Forests in Programmable Switches A.T-J. Akem, B. Bütün, M. Gucciardo, M. Fiore Q1 IEEE/ACM Transactions on Networking IMDEA, NEC T3.3, T4.2 34 RISENSE: Long-Range In-Band Wireless Control of Passive Reconfigurable Intelligent Surfaces S.P. Deram, M. Rossanese, A. Garcia-Saavedra, S. Waqas Haider Shah, V. Sciancalepore, J. Widmer, X. Costa-Perez A ACM MobiSys 2025 IMDEA, NEC, i2CAT, UC3M T3.2, T4.1 35 Demonstrating Distributed Inference in the User Plane with DUNE B. Bütün, D. de Andrés Hernández, J. Aguilar, M. Gucciardo, M. Fiore A* IEEE INFOCOM 2025 IMDEA T3.3, T4.2 36 Demonstrating Deep Learningbased Spatial Diffusion O.E. Martínez-Durive, S. Sotirios Bakirtzis, C. Ziemlicki, M. Fiore A* IEEE INFOCOM 2025 IMDEA T3.2, T4.1 37 6G Standardization Potential of the ORIGAMI Novel Architectures and Use Cases L.E. Chatzieleftheriou, D. de Andrés Hernández, S, Bizzarri, M. Fiore, M. Fodrini, A. GarciaSaavedra, M. Gramaglia, E. Municio, D. Tsolkas Other EuCNC & 6G Summit UC3M, IMDEA, NEC, i2CAT, EMN, University of Wurzburg, NetAI, TID, National and Kapodistrian University of Athens, Fogus, ISRD T2.1, T2.3 38 AegisRAN: A Fair and EnergyEfficient Computing Resource Allocation Framework for vRANs E. Sanchez Hidalgo, J. A. AyalaRomero, J.X. Salvat Lozano, A. Garcia-Saavedra, X. Costa Perez Q1 IEEE Transactions on Mobile Computing I2CAT, NEC T3.1 17 Grant Agreement 101139270 — ORIGAMI — HORIZON-JU-SNS-2023 Deliverable D5.3 39 Kairos: Energy-Efficient Radio Unit Control for O-RAN via Advanced Sleep Modes J.X. Salvat Lozano, J.A. AyalaRomero, A. Garcia-Saavedra and X. Costa-Perez A* IEEE INFOCOM 2025 NEC T3.2 40 Quantum Computing in the RAN with Qu4Fec: Closing Gaps Towards Quantum-based FEC processors N. Apostolakis, M. Sierra-Obea, M. Gramaglia, J.A. Ayala-Romera, A. Garcia Saavedra, M. Fiore, A. Banchs, X. Costa-Pérez A* ACM SIGMETRICS 2025 UC3M, IMDEA, NEC, I2CAT T3.2 41 AZTEC+: Long and Short Term Resource Provisioning for ZeroTouch Network Management S. Alcalá-Marín, D. Bega, M. Gramaglia, A. Banchs, X. CostaPérez, M. Fiore Q1 IEEE Transactions on Network and Service Management UC3M, IMDEA, I2CAT, NEC T3.2 42 A Discrete-Time Model of the 5G New Radio Uplink Channel S. Raffeck, S. Grøsvik, L.A. Becker, S. Lange ,S. Geißler, T. Zinner, W. Kellerer, T. Hoßfeld Other IEEE International Teletraffic Congress ITC 36 JMU T3.2, T4.1 43 Fact-Checking 5G Security: Bridging the Gap Between Expectations and Reality O.Lasierra, N. Ludant, G. GarciaAviles, E. Municio, G. Noubir, A. Skarmeta, X. Costa-Pérez Q1 IEEE Open Journal of the Communications Society I2CAT, NEC T3.1 44 On The Dynamic Regret of FTRL: Optimism with History Pruning N. Mhaisen, G. Iosifidis A* PMLR ICML TUD T3.2 45 CHOMET: Conditional Handovers via Meta-Learning M. Kalntis, F. A. Kuipers, G. Iosifidis B IEEE WiOPT TUD T3.1, T3.2, T4.1 46 Minimization of the Training Makespan in Hybrid Federated Split Learning J. Tirana, D. Tsigkari, G. Iosifidis, D. Chatzopoulos Q1 IEEE Transactions on Mobile Computing TUD T3.2 47 Cooperative Streaming Inferences in IoT Networks M. Li, G. Iosifidis, R. R. V. Prasad B IEEE GLOBECOM TUD T3.2 48 Faro: a scalable and reliable outage detection algorithm for IoT Mobile Virtual Network Aggregators M. Milani, V. Vomhoff, D. Bega, M. Gramaglia, S. Geissler, C. Karsai, P. Serrano A Proceedings of the ACM on Networking UC3M, EMN, JMU T3.3 49 The TES framework: Joint Statistical Modeling and Machine Learning for Network KPI Forecasting L. Lo Schiavo, G. Garcia, M. Gramaglia, M. Fiore, A. Banchs Roca; X. Costa Pérez Q1 IEEE Transactions on Network and Service Management UC3M, IMDEA, i2CAT T3.1 24 Grant Agreement 101139270 — ORIGAMI — HORIZON-JU-SNS-2023 Deliverable D5.3 Invited talks On February 6, 2025, Marco Gramaglia (UC3M) delivered a talk on “Business Models and Trustworthiness KVI” during the online 6G4Society TrialsNet Workshop [4]. In his presentation, he explained the ORIGAMI strategy for measuring and integrating the Trustworthiness Key Value Indicator (KVI) within ORIGAMI’s business models. The talk focused on how trustworthiness—covering aspects such as transparency, privacy, security, and user confidence—can be assessed and linked to business value in future 6G ecosystems. Marco emphasized that embedding trustworthiness as a measurable KVI allows 6G systems to move beyond traditional performance metrics toward valuedriven, socially responsible innovation, supporting the broader 6G4Society vision of a trustworthy and human-centered network. Figure 22: Social media updates on Business Models and Trustworthiness KVI talk. On 30 April 2025, Paul Patras (NETAI) delivered a keynote entitled "How Open Networks and AI Define The Future of Communication Networks" at the Open Future Networks Showcase hosted by Digital Catapult in London, UK. In his presentation, he discussed how on how AI and open architectures are reshaping the telecom industry, the importance of explainability and how AI can be harnessed for energy efficiency in Open RAN environments — a real-world step toward more sustainable and intelligent networks. Figure 23: Social media updates on Paul Patras’s talk at Open Future Networks Showcase. 25 Grant Agreement 101139270 — ORIGAMI — HORIZON-JU-SNS-2023 Deliverable D5.3 On 9 September 2025, Paul Patras (NETAI) delivered a talk entitled "How to build network automation? Intelligence is (almost) all you need" at the Berlin Open RAN Working Week hosted by Deutsche Telekom. In his presentation he covered barriers to AI adoption in mobile networks and spoke about the energy efficiency solution developed in the ORIGAMI project, how we are tackling integration challenges, and how the Compute Continuum Layer (CCL) concept developed in ORIGAMI will help manage available resources heterogeneous resources (in terms of technology and execution environments), allowing the highest re-utilization factor across edge and cloud. Figure 24: Social media updates on BOWW talk. Panels On 13 August 2025, the panel on "Hybrid AI and the Edge-Cloud Continuum" at the IEEE International Conferences on High Performance Computing and Communications (HPCC) held in Exeter UK explored the convergence of advanced AI models and next-generation connectivity and how to design, deploy, and govern the emerging edge-cloud continuum - balancing latency, energy, cost, and trust. Panellists (including Paul Patras, NETAI) discussed pushing the boundaries of data processing, how the power of AI can be harnessed across edge and cloud to autoscale resources in virtualised networks, and the need to re-think network function virtualisation for efficient sharing of heterogeneous compute infrastructures while maintaining appropriate abstraction. 26 Grant Agreement 101139270 — ORIGAMI — HORIZON-JU-SNS-2023 Deliverable D5.3 Figure 25: Social media updates on Paul Patras’s talk at IEEE HPCC. On September 3, 2025, in Istanbul, Turkey, the “Architectural Landscape towards 6G: Standardization Trends, New Enablers, and New Challenges” panel explored the evolving framework, technologies, and standardization efforts shaping the next generation of mobile networks. Bringing together experts from academia and industry, it focused on Europe’s coordinated approach to defining a 6G architecture through initiatives such as the Smart Networks and Services Joint Undertaking (SNS JU). Panellists (including Marco Gramaglia) discussed the integration of AI and machine learning as native components of 6G systems, the role of edge computing, security and sustainability innovations, and the transition of telecom operators toward open, API-driven service models. The session emphasized how collaborative research, standardization activities (including 3GPP’s ongoing work), and the outcomes of European 6G projects will jointly shape a unified, intelligent, and sustainable network architecture for the 2030 era. 27 Grant Agreement 101139270 — ORIGAMI — HORIZON-JU-SNS-2023 Deliverable D5.3 Figure 26: LinkedIn updates on panel at IEEE PIMRC. Co-sponsoring ORIGAMI was an officially supporting project of the second edition of the Würzburg Workshop on Next-Generation Communication Networks [5], held from October 6-8, 2025, in Würzburg, Germany. The KuVS Fachgespräch - Würzburg Workshop on Next-Generation Communication Networks (WueWoWAS) focused on preliminary and ongoing research on next-generation communication networks and the different ways of analyzing systems. The discussions encompassed a wide range of subjects, including next-generation network paradigms such as 5G, 6G, IoT, and industrial networks. The conceptual frameworks vital for future networks, including softwarization, edge computing, and the integration of AI and ML, fostering automation, were covered. Additionally, the assessment of network performance, reliability, and resilience through various means, such as measurements, modeling, simulation, and analysis, with a focus on enhancing aspects like QoS, QoE, and energy efficiency, among others, was explored. ORIGAMI also showcased ongoing 6G research activities in the form of posters, and in addition, was represented in the mentoring panel during the workshop. 28 Grant Agreement 101139270 — ORIGAMI — HORIZON-JU-SNS-2023 Deliverable D5.3 Figure 27: Social media posts on co-sponsoring WueWoWAS’25. 29 Grant Agreement 101139270 — ORIGAMI — HORIZON-JU-SNS-2023 Deliverable D5.3 Figure 28: Promotion of WueWoWAS’25 on 6G SNS webpage. Other support activities The ORIGAMI project is among the participating initiatives featured in the SNS JU Journal 2025, contributing to the advancement of next-generation 6G technologies. The SNS JU Journal 2025 provides a comprehensive overview of the 79 research, innovation, and trial projects (out of 80) that form the core of the SNS JU’s portfolio. Backed by approximately €500 million in EU funding, these projects are pivotal to Europe’s goal of becoming a global leader in 6G technology, while also advancing the rollout of 5G. This portfolio reflects the strong commitment of European policymakers, industry stakeholders, and researchers to establishing a robust, competitive, and sustainable digital infrastructure. 30 Grant Agreement 101139270 — ORIGAMI — HORIZON-JU-SNS-2023 Deliverable D5.3 Figure 29: Screenshots from SNS JU Journal 2025, featuring the ORIGAMI project. On 1 June 2025, the 3rd Edition of the SNS Reference Figure was released online [6]. This interactive map provides an accessible overview and concise summaries of all ongoing SNS projects, highlighting initiatives working on related topics. It serves as a valuable external resource for identifying synergies and collaboration opportunities across the 6G research landscape. This tool is highly relevant for understanding the broader context of projects aligned with ORIGAMI’s objectives. Figure 30: Interactive map of SNS projects, featuring the ORIGAMI project. 31 Grant Agreement 101139270 — ORIGAMI — HORIZON-JU-SNS-2023 Deliverable D5.3 Edited special issues Elsevier Computer Communications The ORIGAMI Project is currently co-editing the Special Issue “From Cloud to Edge: Digital Twin Orchestration and Development in the 6G Era”, published by Elsevier Computer Communications (Impact Factor: 4.5). The Special Issue focuses on the design, orchestration, and deployment of Digital Twins in the context of 6G networks, addressing challenges in system-level integration, edge–cloud coordination, AI-driven modelling, and scalable service interaction [7]. The special issue started accepting submissions on June 1, 2025, with an extended submission deadline set for November 17, 2025. Figure 31: Promotion of Elsevier Computer Communications edited special issue on 6G SNS webpage. PACMNET V3 June 2025 Issue and PACMNET V3 September 2025 Issue The ORIGAMI Project, through TID (Andra Lutu), contributes to the management and editorial coordination of the PACMNET Volume 3, June 2025 Issue [8] and PACMNET Volume 3, September 2025 Issue [9] published by Sheridan. The Proceedings of the ACM on Networking (PACMNET) series showcases top-tier research in emerging computer networks and their applications. The topics of submissions include introducing new technologies, innovative experiments, creative applications of networking technologies, and fresh insights gained through analysis and research. 3.1.1.3 PARTICIPATION IN INDUSTRY FORA Mobile World Congress 2025 activities On March 6, 2025, at Mobile World Congress (MWC) 2025 in Barcelona, Spain, the flagship global event for the telecommunications industry, Net AI showcased our AI-driven, QoS-aware RAN energy efficiency solution in collaboration with Red Hat. Alexis Duque (Net AI) represented the team at MWC 32 Grant Agreement 101139270 — ORIGAMI — HORIZON-JU-SNS-2023 Deliverable D5.3 2025, reinforcing Net AI’s commitment to advancing intelligent, energy-efficient network technologies that pave the way toward next-generation mobile connectivity. Devoxx France 2025 activities On April 18, 2025, at Devoxx France 2025 in Paris, Alexis Duque from Net AI presented the ORIGAMI Project framework architecture and objectives, showcasing an AI-driven, QoS-aware RAN energy efficiency solution. In his Lightning Talk, Alexis explored how artificial intelligence can optimize energy consumption in radio access networks (RANs)—which account for roughly 80% of mobile network energy use—without compromising service quality. This innovative solution can reduce infrastructure energy consumption by up to 60% and is fully compatible with OpenRAN deployments. He also highlighted Net AI’s contribution to the ORIGAMI Project, advancing the development of nextgeneration mobile network architectures and helping to overcome key barriers on the path toward 6G. Figure 32: Social media updates on Devoxx France 2025. Connected Britain 2025 activities On 24-25 September 2025, at Connected Britain 2025 in London, UK, the UK’s premier digital economy event, Net AI showcased how the company is harnessing AI-driven traffic forecast to improve energy efficiency in RANs without impacting service quality. The team engaged with multiple stakeholders in the mobile network operator ecosystem and highlighted how the efforts in ORIGAMI is advancing the development of intelligent and sustainable 6G networks. 33 Grant Agreement 101139270 — ORIGAMI — HORIZON-JU-SNS-2023 Deliverable D5.3 Figure 33: Social media updates on Connected Britain 2025 activities. Barcelona DeepTech Summit activities On 4-6 November 2026, at the Barcelona DeepTech Summit 2025, an international congress focused on scientific and technological entrepreneurship, bringing together specialists, experts, corporations, and investment groups to address global and economic challenges through innovation, Alexis Duque (Net AI) showcased Net AI's cutting-edge AI solution for network efficiency and performance, receiving incredibly valuable feedback on the innovation developed. 40 Grant Agreement 101139270 — ORIGAMI — HORIZON-JU-SNS-2023 Deliverable D5.3 the high level of engagement and responsibility that ORIGAMI holds within this working group. This role emphasizes the project's integral involvement in shaping the future of smart networks. Below are key highlights of ORIGAMI contributions in the period covered by the present deliverable: • Contribution to SNWG White Paper on AI/ML Frameworks: ORIGAMI has contributed significantly to the preparation of an SNWG White Paper scheduled for publication by the end of the year. The tentative title of the White Paper is "AI/ML Frameworks for Smart Networks and Services" 5 , and ORIGAMI's specific contribution focuses on AI/ML-enabled user and data plane optimizations. This contribution is central to how artificial intelligence and machine learning can enhance the efficiency and performance of smart networks, aligning with ORIGAMI’s overarching goals of optimizing 6G architecture. • Inputs to Cross-Project Research Topics: In addition to the White Paper, ORIGAMI has actively contributed to various cross-project research topics selected by the SNWG. For example, ORIGAMI provided detailed inputs on the impact of open-source software components within the smart network ecosystem, completing the related and requires SNWG form. This work helps define the role of open-source solutions in the broader context of network evolution, aligning with ORIGAMI’s focus on innovation and collaborative development. Test, Measurement and KPIs During this period, ORIGAMI is deeply involved in the TMV KVI sub-group within the TMV WG, which was established to develop common practices for measuring and evaluating KVIs across SNS projects. Within this framework, a KVI estimation template was designed and circulated among participating projects, with ORIGAMI being one of the ten contributors providing detailed input on its key values, KVI analyses, technology enablers, and related KPIs. During this period, ORIGAMI also contributed with its methodology for KVI estimation, highlighting how its architectural innovations (such as AI-driven optimization, dynamic resource orchestration, and energy-aware design) deliver measurable impact across critical dimensions like energy efficiency, reliability, and adaptability. The White Paper on “6G KVIS – SNS Projects Initial Survey Results 2025” 6 , published on May 12, 2025 by the 6G Smart Networks and Services (6G SNS) initiative, represents an important step in assessing the collective progress of fifteen SNS projects, in defining and quantifying Key Values (KVs) and Key Value Indicators (KVIs). For the time being, this white paper only includes inputs from Call 1 projects; however, by the time of submission of this deliverable, the sub-working group has started collecting inputs from Call 2 projects, including ORIGAMI. Another important activity carried out within the context of this WG has been the development of the Meta Registry System (MRS) to centralize and provide structured access to metadata from all active SNS JU projects, supporting reproducibility and alignment of results. ORIGAMI has not only worked with the WG to help define the structure of the MRS, but we have also contributed six datasets generated within the project: 5 https://smart-networks.europa.eu/wp-content/uploads/2025/02/ai_ml_white-paper-sns_tb_v1.0.pdf 6 https://smart-networks.europa.eu/wp-content/uploads/2025/05/sns-ju-white-paper-6g-kvis-survey2025_final-1.pdf 41 Grant Agreement 101139270 — ORIGAMI — HORIZON-JU-SNS-2023 Deliverable D5.3 Figure 36 Meta Registry System (MRS) Sustainability WG The Working Group held its kickoff meeting on 1 October 2025. Upon submission of the present deliverable, the WG is finalising and approving its work plan, including scope, priorities, milestones, and task ownership, as well as confirming responsibilities and establishing an appropriate reporting cadence. All planned activities remain subject to the WG’s formal agreement on the work plan. Initial actions include work on terminology alignment across projects and an assessment methodology building on existing efforts. From an organisational perspective, Christoph Schmelz (Nokia, SUSTAIN6G) serves as WG Chair, and two vice-chairs are Mir Ghoraishi (GIGASYS) and Stefan Wendt (Orange). So far, no scientific activities have been carried out within the WG to which the SNS projects can contribute. Pre-Standardization The WG is currently progressing on several active topics related to standards dissemination and impact facilitation. Ongoing activities include regular debriefings from 3GPP plenaries, updates on 3GPP RAN and SA Release 20 Study Items from rapporteurs, status reports from ETSI ISGs such as RIS and ENI, and briefings on ITU-T Focus Group AINN developments. Looking ahead, new coordination groups are planned, including an ISAC technologies group (proposed by Interdigital), a KVIs and architecture impact group (proposed by 6G4Society CSA project), and an INSTAR coordination group (proposed by INSTAR). In addition, the WG continues to publish 3GPP plenary debriefs in the online tracker. Within the ORIGAMI project, all discussions are being closely monitored in order to identify potential opportunities related to standardisation. In that sense, one of main topics ORIGAMI plans to bring is related to the possibility to apply Service Based Architecture approach also to RAN. 42 Grant Agreement 101139270 — ORIGAMI — HORIZON-JU-SNS-2023 Deliverable D5.3 WiTaR The “Women in telecommunications and research (WiTaR)” Working Group (WG) focuses on promoting gender equality, inclusion and empowerment in the 6G Research & Innovation (R&I) community. This is a WG that seeks to provide a snapshot of efforts to promote the participation and advancement of women in telecommunications. Building on the contribution reported in D5.2, ORIGAMI continues to support the WG from a perspective standpoint, contributing to its social media outreach by preparing and providing content for LinkedIn posts that promote the visibility of women in telecommunications Communications During the monthly SNS Communication Task Force meetings, a roundtable discussion takes place, providing updates from all SNS projects. Following the discussion during the monthly calls, the ORIGAMI project takes the actions, e.g., • ORIGAMI completed the requested Excel file for the SNS Projects: Social Media Channels Presence initiative, providing information on our presence across social media platforms to support future communication planning of the WG; • ORIGAMI shared the interest in jointly participating in the MWC 2026; • ORIGAMI shared insights on the project objectives and outcomes via Mural to better identify collaboration opportunities across the SNS JU portfolio; • ORIGAMI shared relevant news about the project (workshops, webinars, open call, or public deliverables) on the SNS JU project (see, Figure 28, Figure 31, [11],[12]) • ORIGAMI provided up-to-date information about the project [13] Vision WG (Societal and vision sub-group) This WG has been working throughout most of 2025 to redefine its scope and the breadth of its activities. At the time of submitting this deliverable, the Sub-WG has a clear mandate focused on developing a new methodological framework to support projects in integrating processes for the definition and/or validation of KVIs. Although the development of a white paper had initially been considered for 2025, this has now been postponed to 2026. The main reason is that the WG is currently preparing a questionnaire to be circulated among the SNS-JU projects to collect projects’ experiences about key value and KVI work and to understand the main challenges they face, and only then design a framework that effectively addresses these needs. The WG has continued to invite projects from Stream D (those with higher TRLs within the SNS programme and that deal with KPI and KVI validation in trial phases). From ORIGAMI’s perspective, emphasis is being placed on ensuring that the framework should distinguish between Stream B and Stream C/D projects. The feedback provided by ORIGAMI highlights that the framework must account for the TRL level of each project, since the validation approach (particularly in terms of experimentation) can differ significantly depending on technological maturity. Indeed, in the case of ORIGAMI, as Stream B project, the approach is at the PoC level, whereas for Stream C/D projects the focus is on trials or large-scale deployments. 3.2 UPDATED DISSEMINATION PLAN FOR Y3 As demonstrated in Section 5, ORIGAMI exceeded all activities related to dissemination for Y2. Moreover, ORIGAMI prioritizes the quality of publication venues and, consequently, aims to 43 Grant Agreement 101139270 — ORIGAMI — HORIZON-JU-SNS-2023 Deliverable D5.3 disseminate research in the top conferences and journals within the field. During the third year of project implementation, it is planned to publish 28 papers (reaching 80 papers in total). For example, a poster on an updated service-mesh platform for the mobile network core will be presented at Software & Standards for Smart Networks & Services 2026. Regarding participation in industry forums and events, participation in EUCNC 2026 is planned with at least a project booth and demonstrations based on the project’s results. Moreover, at least 1 paper submission and 1 workshop co-organization (with Arch WG) is planned. The 2nd Winter School on AI for 6G Communications, organized within the context of the ORIGAMI project, will be held next year (January 21-23, 2026, in Las Palmas de Gran Canaria, Spain), bringing together PhD students and early-stage researchers to share and discuss their work. This 2.5-day Winter School will feature tutorial-style talks from leading experts in academia and industry, as well as a student poster session. It will offer a unique opportunity for learning, networking, and showcasing your work in an engaging and collaborative environment. In Y3, the project will participate in the 6GArch Workshop at ICC 2026, which has been accepted and scheduled as part of the conference program. This workshop will serve as our MS5 (M23): “Workshop on 6G architecture and NI solutions disseminating results.” The targets for dissemination activities to be achieved by M36 of the project execution are listed in Table 4 below, while the project will continue pursuing the dissemination plan defined in deliverables 5.1 and 5.2 with no expected deviations. Nature Community KPI Target at M24/M36 Expected deviations Dissemination Academia Scientific publications 50/80 None Edited special issues of international peer-reviewed scientific journals 1/2 None Collaborations with other scientific projects More collaborations are planned for Y2 None; the target for Y2 has been fully achieved by M12 ORIGAMI scientific workshops 1/2 None Industry In-presence participation in industry fora and events 8/10 None ORIGAMI industry workshops 1/2 None Academia Industry Demonstrations 4/8 None Table 4: ORIGAMI KPIs for dissemination activities during the second year of the project 44 Grant Agreement 101139270 — ORIGAMI — HORIZON-JU-SNS-2023 Deliverable D5.3 4 EXPLOITATION ACTIVITIES As already expressed in the project description, ORIGAMI innovative solutions, which answer directly to market needs and future roadmaps, and the strong commitment of the consortium’s industrial members are the key elements of ORIGAMI to maximize the project impact. ORIGAMI members are performing a continuous research and market study, targeting to promptly identify relevant new works on related scientific fields, new market/research trends in 6G, and the roadmaps of large industrials and international organizations. For that purpose, a complete exploitation strategy has been described in [1] including both a plan for project exploitable results, as well as individual exploitation initiatives. Following these targets, this section includes the project exploitation activities for the second year of the project, plus the plans for the third year (both at project level and individual per partner. An extra subsection includes the related standardization activities and open-source contributions for the second project year. 4.1 PROJECT RESULTS AND EXPECTED OUTCOMES BY Y2 As described in [1], we differentiate between project results (what is generated during the project implementation) and project outcomes (the expected effects of results, over the medium term). At the end of Y2, thirteen project key exploitable results have been identified, with potential to lead to project outcomes. A Key Exploitable Result (KER) is a prioritized result with high potential for exploitation, meaning its value can be harnessed across the downstream value chain. This could involve a product, process, or solution, or contribute as a key input for policy, further research, or education. ORIGAMI project has followed a twofold approach to Identifying KERs: Strategic KER Development and Assessment: Following the methodology outlined in Section 4.1 of D5.1 [1], the following steps have been completed: • Collaborative Discussion: Project members identified and agreed on a list of KERs during regular work package meetings, aligning them with expected outcomes. • Target TRL Assignment: Each KER was matched with a target Technology Readiness Level (TRL) to ensure clarity on development stages. • SWOT Analysis: A Strengths, Weaknesses, Opportunities, and Threats (SWOT) analysis was conducted for each result to evaluate its strategic potential. • GAP analysis: A process of measuring the difference between a project result (current state) and project outcome (desired state). KER Distribution Across Project Barriers: Below is the updated breakdown of KERs that have identified and mapped across the project's barriers: • Barrier #1: 2 KERs • Barrier #2: 3 KERs • Barrier #3: 1 KER • Barrier #4: 1 KER • Barrier #5: 2 KERs • Barrier #7: 2 KERs • Barrier #8: 2 KERs 45 Grant Agreement 101139270 — ORIGAMI — HORIZON-JU-SNS-2023 Deliverable D5.3 The list of KERs is provisional, as new or modified results may emerge later in the project. The results are organized below according to the primary barriers they address 4.1.1 BARRIER #1: UNSUSTAINABLE RAN VIRTUALIZATION Key Exploitable Result 1.1: CloudRIC++ Partners NEC, UC3M, IMDEA, i2CAT Result name CloudRIC++ Short description Network Intelligent (NI) solution exploiting ORIGAMI's CCL to improve energy and cost-efficiency of virtualized Radio Access Networks Result type Demonstrator Target TRL 4 Expected impact - Increased energy efficiency in O-RAN virtualized RANs Increased cost efficiency in O-RAN virtualized RANs Key innovations - Joint radio and compute control - Heterogeneous processors (Hardware Accelerators and CPUs) - DU pooling Target Markets O-RAN vendors, O-RAN operators Strengths Weaknesses - Significant cost and energy savings compared to traditional vRAN implementations. By sharing hardware accelerators (HAs) among multiple Distributed Units (DUs) and opportunistically offloading tasks to CPUs, it reduces the need for dedicated, power-hungry hardware for each DU. - High reliability (five-nines or 99.999%) even under heavy network loads. This is achieved through a combination of centralized resource allocation and compute-aware radio scheduling policies. - Seamlessly integrates into the O-RAN architecture, making it potentially easier to adopt for operators already invested in O-RAN. - The system incorporates real-time control mechanisms to adapt to the dynamic nature of network traffic and resource demands. This allows it to efficiently handle fluctuations in user load and maintain performance. - The architecture and algorithms involved are relatively complex. This could increase the difficulty of implementation and deployment, particularly for operators with less technical expertise. - While we aim at low overhead for CloudRIC's operations, any additional layer of control and coordination in a realtime system introduces some level of overhead. This needs to be carefully managed to avoid impacting overall network performance. - The effectiveness of this technology relies heavily on the accuracy and efficiency of its data-driven models. These models need to be trained and updated regularly to maintain performance as network conditions change. - Centralizing resource allocation and control could create potential bottlenecks in the system, especially under extremely high loads. Opportunities Threats - The market for vRANs is rapidly expanding, driven by the desire for greater flexibility and cost savings in 5G networks. The ability of this technology to address the cost and energy - The vRAN market is becoming increasingly competitive, with various vendors and solutions vying for market share. This technology will need to 46 Grant Agreement 101139270 — ORIGAMI — HORIZON-JU-SNS-2023 Deliverable D5.3 challenges of vRANs positions it well to capitalize on this growth. - Continued advancements in hardware acceleration technologies, such as GPUs and FPGAs, could further enhance the efficiency and performance of this technology. - The use of data-driven models opens up opportunities for further integration with AI and machine learning techniques to optimize network operations and enable new services. continue to innovate and demonstrate its value proposition to remain competitive. - Some operators may be hesitant to adopt new and complex solutions like this one, preferring to stick with more traditional, albeit less efficient, approaches. - Centralizing control and data in the cloud could raise security concerns, particularly regarding the potential for attacks on the control plane or data breaches. GAP analysis Outcome: A demonstrator of this technology will be presented as an outcome. The demonstrator will show the advantages of this technology, including the increased energy efficiency and cost efficiency in O-RAN virtualized RANs. Gap(s) from result to outcome: The presented outcome 1 involves a minor practical gap, as from the conceptual design and initial prototyping to the demonstrator, it requires several development steps, including the implementation of the algorithms to run timely in real platforms, GUI design, and video recording of the demonstrator. Steps needed: Addressing the gap involves that the partners in charge of this technology keep working on the demonstrator to: first, test that the implementation works as expected in real hardware showing the expected cost efficiency measure in the prototype; second, create the GUI to control and configure the algorithms so that an external user can configure the different parameters, activate benchmark with the final goal of having a comprehensive view of the technology. Key Exploitable Result 1.2: GPU-assisted vRAN Partners NEC, TUD Result name GPU-assisted vRAN Short description Network Intelligent (NI) solution exploiting GPUs to multiplex 5G and ML processing tasks through ORIGAMI's CCL Result type Research Result Target TRL 4 Expected impact Increase cost-efficiency in GPU-accelerated vRANs Key innovations Reliable multiplexing of GPU resources between vRAN workloads and machine learning workloads Target Markets vRAN vendors, vRAN operators Strengths Weaknesses - The technology enables the sharing of expensive GPUs between 5G PHY processing and machine learning (ML) workloads, leading to more efficient - The dynamic resource allocation mechanism and the neural network-based prediction model introduce complexity to the system, potentially making 47 Grant Agreement 101139270 — ORIGAMI — HORIZON-JU-SNS-2023 Deliverable D5.3 utilization of hardware and reduced costs. - The dynamic resource allocation mechanism allows the system to adapt to varying 5G traffic demands and ML workload requirements, ensuring efficient resource utilization in different scenarios. - The proposed technology prioritizes meeting the stringent latency requirements of 5G PHY processing, ensuring high reliability in delivering 5G services even when sharing resources with ML tasks. - The technology is designed to be integrated into the O-RAN architecture, making it compatible with the industry's move towards open and disaggregated RANs. implementation and management more challenging. - While the technology aims at low overhead, the feasibility of resource sharing, there is still some overhead associated with GPU reconfiguration and context switching between 5G and ML tasks, which needs to be carefully managed. - The effectiveness of the system relies on the accuracy of the neural network's predictions of future 5G traffic demands. Inaccurate predictions could lead to suboptimal resource allocation and potential service disruptions. - The use of Multi-Process Sharing (MPS) for resource partitioning provides only partial isolation between 5G and ML tasks, potentially leading to performance interference under heavy workloads. Opportunities Threats - The increasing adoption of virtualized RANs and the rising demand for edge ML applications create a significant market opportunity for technologies that enable efficient resource sharing in these domains. - Continued advancements in GPU architectures and software frameworks could further enhance the performance and efficiency of the proposed technology. - The use of machine learning for resource allocation opens up opportunities for further integration with AI-driven network optimization techniques, leading to even more intelligent and efficient resource management. - The market for vRAN and edge computing solutions is becoming increasingly competitive, with various vendors and technologies vying for market share. - Sharing GPU resources between 5G and ML tasks could raise security concerns, particularly regarding the potential for data leakage or unauthorized access to sensitive information. - The effectiveness of the proposed technology is dependent on the capabilities of the underlying GPU hardware. Limitations in GPU memory or compute capacity could restrict its applicability in certain scenarios. GAP analysis Outcome: We will deliver as a research result a novel methodology to implement this technology. This methodology includes the formulation and description of the algorithms involved and how to implement them into the current GPU partition technology. This result will be delivered in papers published in top venues and the corresponding deliverables. Gap(s) from result to outcome: Outcome 1 involves a minor practical gap as the delivered methodology should be clear, rigorous, and detailed in order to enable reproducibility. Steps needed: To realize Outcome 1, the involved partners should work first on the development of the solution and then on the documentation of the methodology. The involved partners need to ensure that the delivered methodology is comprehensive, clear, and allows reproducibility. 4.1.2 BARRIER #2: POOR INTER-OPERABILITY OF RAN COMPONENTS 48 Grant Agreement 101139270 — ORIGAMI — HORIZON-JU-SNS-2023 Deliverable D5.3 Key Exploitable Result 2.1: RAN Energy Efficiency Optimization Partners NETAI Result name RAN Energy Efficiency Optimization Short description NI solution that leverages AI-driven traffic demand forecasting to control the amount of radio resources needed at any point in time/location to meet the anticipated demand Result type Prototype Target TRL 5 Expected impact Over 30% average RAN energy consumption reduction while maintaining over 99.5% service availability. Key innovations 1) Performance: We obtain highly accurate, real-time insights into network performance, resource usage, and future demands. This is extremely challenging to do and has been the barrier to enabling mobile network operators to switch off network resources confidently. We achieve over 95% accuracy in traffic forecasting. Our data handling pipeline employs preprocessing that improves the sensitivity and generalization ability of trained models. 2) Scalability: We train a domain-specific neural network architecture with hundreds fewer parameters that state-of-the art, which is more scalable than existing solutions and inherently low in power consumption. 3) Reliability: Our proprietary training methodology ensures that AI models avoid underestimation errors, crucial for maintaining service quality. Target Markets Global Market for AI in Networking. Strengths Weaknesses - Unique expertise at the intersection of AI and telecoms. - The designed solution can be deployed on prem or in the cloud and can flexibility scale up/down depending on utilisation; - it is easy to plug into any type of mobile network architecture (2G, 4G, 5G, etc.), being ready to interface with proprietary Operations Support Systems, Open RAN architectures (packaged as xApp or rApp) and the latest 3GPP systems (as a Network Data Analytics Function - NWDAF); - Once trained, it can run effectively in low-cost compute-constrained environments, including on widely deployed CPU servers, overcoming GPU infrastructure and cost barriers. - Insufficient network experimentation infrastructure to enable validation at scale. - Risk of development time being longer than anticipated given available staff resources. - O-RAN adoption slower than expected. - RIC implementation not mature enough to support all features required by x/rApps. Opportunities Threats - Global O-RAN market size is projected to reach $455.81 billion by 2036, registering a compound - Mobile operators continuing to prefer to deal with large established vendors, even at the 49 Grant Agreement 101139270 — ORIGAMI — HORIZON-JU-SNS-2023 Deliverable D5.3 annual growth rate (CAGR) of 53.8% during the forecast period (20242036). - Large solution vendors opening up their solutions to third parties to accelerate integration and support supply chain diversification cost of performance and wait for these to catch up. - Hyperscalers building telco AI capability that outpaces Net AI. GAP analysis Outcome: Net AI expects to have developed a prototype x/rApp that uses AI-driven forecasts to reduce energy consumption in RANs. This will be trained with a domain-specific methodology that minimises the number of underestimation errors and therefore avoids service degradation when radio frequency resources are switched off to save energy. The x/rApp will be integrated with the ISRD RIC as part of the joint efforts in ORIGAMI, and beyond this with the Ericsson EIAP. Gap(s) from result to outcome: - Theoretical (major): AI-model training methodology that prevents demand underestimation - Technical (major): Scalable AI model that can serve multiple antennas and is inherently lowpower Practical (major): Integration with commercial-grade service management and orchestration (SMO) platforms, overcoming the interoperability barrier identified in ORIGAMI Steps needed: With theoretical work completed and a set of ML models evaluated, Net AI needs access to the ISRD RIC API to begin the integration efforts, which is expected in the second half of the project. This will require development resources from both ISRD and Net AI. A proof-ofconcept is expected by the end of Q1/26. In parallel, Net AI has entered an Alliance Agreement with Ericsson, aiming to integrate the developed solution with their EIAP platform. The EIAP offers a rich set of KPIs that can be consumed and an extensive range of configuration parameters that can be adjusted. The Net AI team requires engineering resources to become fully familiar with this ecosystem, so as to be able to complete an integration. A proof-ofconcept and potentially pilot projects with MNOs are expected by Q2/26. Key Exploitable Result 2.2: Conflict mitigation of xApps in the Near-RT RIC Partners ISRD Result name Conflict mitigation of xApps in the Near-RT RIC Short description NI solution that leverages direct conflict mitigation by Enabling seamless coordination between xApps decisions, avoid conflicting decisions, maintaining stable throughput performance. Result type Demonstrator Target TRL 5 Expected impact Improve the interoperability of O-RAN components that optimize different aspects of network performance my having conflicting goals towards underlying RAN. The direct conflict mitigation of the slice-level PRB quota optimization xApp may conflict in-terms of setting up priorities of assigning PRB quota to the users connected to dedicated slices. By rejecting conflicting parameters by the xApps provides stable throughput performance. Key innovations 1) Resolving conflicting objectives: Optimize different aspects of network performance may have conflicting goals. In ORIGAMI, we consider an xApp that provide slice-level PRB quota optimization may conflict in-terms of setting up priorities of assigning PRB quota to the dedicated slices 56 Grant Agreement 101139270 — ORIGAMI — HORIZON-JU-SNS-2023 Deliverable D5.3 Key innovations - Real-time network optimization, automated service management, crossnetwork roaming Target Markets - Tech-savvy consumers, enterprises, global travelers Strengths Weaknesses - Enhanced User Control: Customerfacing APIs give users direct control over network services, allowing them to customize settings such as data usage, call blocking, or roaming preferences. This increased control could significantly enhance customer satisfaction by offering greater flexibility. - Automation of Routine Tasks: APIs could allow users to automate tasks such as bill payments, plan upgrades, data usage monitoring, or switching between different network modes (e.g., from 5G to Wi-Fi), saving time and improving the overall experience. - Real-time Service Optimization: APIs could enable real-time actions, like dynamically adjusting data speeds based on the user’s needs or network conditions. This could prevent service disruptions and ensure an optimized connection, even in congested or rural areas. - Self-service Efficiency: Customerfacing APIs would reduce the need for human customer service interaction by providing automated, programmable solutions for common user issues like troubleshooting connectivity problems or changing service plans. - Differentiation in Service Offerings: MNOs that offer advanced, easy-to-use APIs for their customers could gain a competitive edge. For tech-savvy users, customizable plans, real-time controls, and personalized automation options could become key differentiators in choosing a network provider. - Complexity for Non-Technical Users: While APIs provide significant benefits for tech-savvy users, non-technical customers may find them overwhelming. This could lead to underutilization of these features, resulting in a lack of perceived value or a frustrating user experience. - High Maintenance and Support Requirements: APIs will need constant maintenance, updates, and security patches. Customer-facing APIs would require the MNO to invest in extensive documentation, support systems, and customer education to help users understand and utilize these APIs effectively. - User Data Misuse: By giving customers direct control over many functions, there is a risk that some users might inadvertently make changes that could lead to negative consequences (e.g., disabling safety features). Additionally, if users mishandle their personal data in conjunction with the API, it could lead to complaints or service disruptions. - Security Concerns: Exposing more network functionality to the customer layer through APIs introduces greater risk of cyberattacks. Users may be more vulnerable to phishing or malware attacks if the APIs are not secured adequately, potentially damaging trust in the operator. 57 Grant Agreement 101139270 — ORIGAMI — HORIZON-JU-SNS-2023 Deliverable D5.3 Opportunities Threats - Personalized User Experiences: APIs could enable customers to build highly personalized network experiences, from tailored billing plans to custom usage limits or automated data-saving modes. This could be particularly beneficial for enterprise customers or power users who need granular control over their connectivity. - Integration with Third-Party Services: Customer-facing APIs can allow integration with other thirdparty apps or platforms, such as digital assistants, home automation systems, or IoT devices. For instance, users could use a voice assistant to check data usage or automate tasks like turning off mobile data when connected to WiFi. - Revenue from Premium API Access: MNOs could offer advanced APIs as part of a premium service offering. For example, developers or enterprises needing specialized functionalities (e.g., API-based network provisioning or data allocation) could pay for enhanced API access. - Global Roaming and Seamless User Experience: APIs could enable users to manage their services across borders easily. For instance, a user could automate network switches while traveling internationally to ensure they remain on the best network, control data costs, and optimize usage without needing to manually configure settings. - Facilitating IoT and Smart Device Ecosystems: As IoT expands, customer-facing APIs will become crucial for managing multiple devices connected to the network. For example, APIs could allow a user to monitor data usage across smart devices, automatically adjust their data plan, or allocate network resources between devices. - Data Privacy and Security Concerns: Giving customers more access to their data and control over network functionalities raises concerns around privacy. If not managed properly, these APIs could be exploited by bad actors, leading to data breaches, unauthorized access, or manipulation of user settings. - Increased Regulatory Scrutiny: As APIs give customers more control over their services, they also create potential legal and regulatory challenges, particularly in regions with stringent data protection laws (e.g., GDPR in Europe). Regulators may impose strict requirements on how user data is accessed and managed through these APIs. - Dependency on Reliable Connectivity: APIs are only as good as the network they depend on. If network connectivity is unstable or inconsistent, automation features will not work as intended, leading to user frustration. This could lead to backlash or negative customer sentiment if promises of automation and optimization cannot be met. - Cannibalization of Traditional Services: By automating many functions and offering selfservice options, MNOs risk reducing the need for traditional, high-margin customer services like premium support or hands-on account management. This could lead to revenue cannibalization in some areas. - Interoperability Challenges: If APIs are not standardized across operators, users may face challenges when switching networks or using their services across different regions. Inconsistent implementations could lead to fragmented user experiences, especially for international customers or those frequently switching networks. 58 Grant Agreement 101139270 — ORIGAMI — HORIZON-JU-SNS-2023 Deliverable D5.3 GAP analysis Outcome: An API platform that allows mobile subscribers to automate network tasks, customize connectivity settings, manage billing, and integrate mobile services with IoT and third-party applications. Gap(s) from result to outcome: - Technical Gap: Lack of standardized API frameworks across MNOs and differing network architectures. - Security Gap: Need for robust authentication, authorization, and data protection mechanisms. Steps needed: Develop and test standardized API protocols and security frameworks Key Exploitable Result 5.2: API to order services Partners CMC, TID Result name API to order services Short description We will develop an Application Programming Interface (API) that enables external systems and service providers to request and activate mobile network services without the need for traditional roaming arrangements. This API will allow dynamic allocation of network resources, authentication, and data connectivity for users or devices regardless of their home network, ensuring seamless service delivery across different administrative domains. Result type Demonstrator Target TRL 6 Expected impact - Capability to generate dataflows for single end-user within the slice with specific QoS setting using API/NEF Key innovations - Dynamic network slice control - Network slice operations - QoS managements inside the network slice - Application with connectivity bundle Target Markets Consumer application providers that use either high bitrates and/or needs low jitter connectivity Strengths Weaknesses Enables Network Slice Monetization: The developed API allows network operators and service providers to offer network slices as on-demand, billable services. By exposing standardized interfaces, it becomes possible to dynamically allocate and manage network resources for different customers or applications. This capability transforms network slicing from a technical feature into a concrete business opportunity, enabling flexible pricing models and partnerships across domains. New Business Model: Introducing this API and the associated service paradigm requires a shift from traditional mobile network operations. Operators and partners may face a learning curve to understand, adopt, and trust the new model. Processes for service provisioning, billing, and support must be adapted, which can slow adoption and require additional internal training and coordination. Change in Charging Paradigm: The API enables ondemand access to network slices and services, which disrupts traditional subscriptionor roamingbased revenue models. Implementing dynamic charging mechanisms for these services is complex 59 Grant Agreement 101139270 — ORIGAMI — HORIZON-JU-SNS-2023 Deliverable D5.3 Improves Network Usability When Roaming: By removing the dependency on traditional roaming agreements, the solution simplifies access to mobile network services across different regions and operators. Users and connected devices can seamlessly connect to local networks while maintaining consistent service quality and security. This significantly enhances user experience, reduces administrative overhead, and supports global-scale operations for industries like logistics, media, and emergency services. Enables Use Cases Where Constant Bitrate Is Required: The API supports the creation and management of dedicated network slices with guaranteed Quality of Service (QoS) parameters, including constant bitrate. This is essential for time-sensitive or high-reliability applications such as live video streaming, industrial automation, and remote control of autonomous systems. The solution ensures stable performance even under variable network load, unlocking new use cases that depend on predictable connectivity. and may require significant changes to existing billing systems. Misalignment between technical capabilities and revenue recognition could pose short-term financial and operational challenges for the network operator. Opportunities Threats New Use Cases Requiring High-Quality Connectivity: The API enables the creation of network slices with guaranteed performance parameters, opening the door to innovative applications that demand high reliability and low latency. Examples include industrial automation, live video streaming, remote medical services, autonomous vehicle coordination, and other missioncritical services. This creates opportunities for operators to target industries and verticals that previously High Security Requirements for the API: As the API enables access to mobile network services without traditional roaming arrangements, it becomes a critical point of exposure. Any security vulnerability could be exploited to gain unauthorized access to network slices, user data, or connectivity services, potentially leading to service disruptions, data breaches, or financial loss. Ensuring robust authentication, encryption, and monitoring mechanisms is essential, but failure to maintain high security standards could undermine trust in the service, damage the operator’s reputation, and create regulatory compliance risks. 60 Grant Agreement 101139270 — ORIGAMI — HORIZON-JU-SNS-2023 Deliverable D5.3 could not rely on mobile networks for such stringent requirements. Increased Network Usage When Roaming: By removing traditional roaming charges, the solution encourages users and devices to connect more frequently to partner networks, even when outside their home network. This drives higher traffic volumes, better network utilization, and opens new revenue streams for operators through slicebased monetization or premium service offerings, while improving the overall user experience for global connectivity. Zero-Charge for Certain Applications: The flexibility of the API allows operators to designate certain services or applications as free of charge, supporting promotional campaigns, public services, or strategic partnerships. This can be leveraged to drive adoption, improve customer satisfaction, and stimulate usage in targeted segments without eroding overall revenue, effectively creating marketing and ecosystem-building opportunities. GAP analysis Outcome: API definition that can used by MNOs to generate new services that requires predictable capacity and/or cost when roaming. This is foreseen to be important especially for B2B customer segment. Gap(s) from result to outcome: In the current architecture, it is required to configure RAN and Core separately when creating a slice. This is complex and prevents fully dynamic slice generation. To be fully dynamic, slice definition must happen in the Core that can expose the API. Steps needed: There needs to be an architecture that allows dynamic slice generation, after that API needs to be defined so that same mechanism works in those networks that are providing this dynamic slice service. 4.1.6 BARRIER #7: INADEQUATE NETWORKING DATA REPRESENTATION Key Exploitable Result 7.1: Prediction of Roaming QoS Partners Emnify, TID, JMU Result name Prediction of Roaming QoS Short description Leveraging predictive analytics to forecast and improve the quality of service for mobile users while roaming across networks. Result type Process or tool Target TRL 7-8 61 Grant Agreement 101139270 — ORIGAMI — HORIZON-JU-SNS-2023 Deliverable D5.3 Expected impact Seamless roaming, tailored plans, enhanced resource management Key innovations - Real-time predictive models, cross-network data integration, proactive roaming management Target Markets Telecom operators, international travelers, MVNAs, global enterprises Strengths Weaknesses - Enhanced User Experience: Improved data representation enables MNOs, MVNOs, and network aggregators to predict the quality of service (QoS) for roaming users. Predictive algorithms can forecast potential connectivity issues (e.g., signal degradation, high latency) and ensure a more seamless experience for customers traveling across different networks. - Proactive Network Optimization: By predicting QoS across various roaming scenarios, operators can optimize resource allocation and network configurations in advance, improving overall performance for roaming users. This could lead to fewer dropped calls, faster data speeds, and more reliable coverage. - Dynamic Service Adjustments: Predictive data allows operators to make real-time adjustments to service offerings for roaming users. For example, they can offer location-specific plans or notify users when they are entering areas with lower QoS and suggest alternative solutions. - Enhanced Network Planning: Access to predictive insights regarding roaming QoS can help operators plan for infrastructure upgrades in regions with high roaming traffic. This ensures that resources are allocated where they are needed most, improving network scalability. - Inaccuracy in Prediction Models: While improved data representation can enhance predictions, there is always the risk of inaccuracies in QoS predictions. Inconsistent data from different operators or regions might lead to incorrect forecasts, causing customer dissatisfaction when the predicted QoS is not met. - Limited Data Availability for Specific Regions: Roaming QoS predictions are only as good as the data available. In some regions, particularly rural or underdeveloped areas, there may be limited data on network performance, resulting in less reliable predictions. - High Computational Costs: Predictive analytics for roaming QoS requires substantial computational power and continuous data collection, which could be costly for smaller operators. The resources required to process and analyze this data might create an economic burden. Opportunities Threats - Customized Roaming Plans: By accurately predicting QoS for roaming users, operators can offer tailored plans that cater to specific - Data Sharing Regulations: Cross-border data sharing required for accurate roaming QoS predictions may run into regulatory issues, especially concerning privacy 62 Grant Agreement 101139270 — ORIGAMI — HORIZON-JU-SNS-2023 Deliverable D5.3 user needs, such as data-heavy or latency-sensitive services, creating new revenue streams and enhancing customer satisfaction. - Collaboration Between Operators: MNOs and MVNOs can collaborate to share data on roaming QoS, improving service quality across borders. This collaboration would not only benefit users but also create opportunities for shared infrastructure and cost savings. - Enhanced Customer Loyalty: Delivering superior QoS for roaming users can significantly boost customer loyalty. Users who experience seamless transitions between networks are more likely to stick with their service provider, driving long-term growth. laws and data sovereignty in different regions. This could hinder the ability to collect necessary data for accurate predictions. - Roaming Data Congestion: Over-reliance on predictive algorithms could lead to network congestion in popular roaming destinations, as operators attempt to allocate resources based on predicted demand. This could strain existing infrastructure and degrade QoS in high-traffic areas. - Increased User Expectations: As predictive QoS models become more accurate, users may develop higher expectations for consistent performance while roaming. Failure to meet these heightened expectations could lead to user dissatisfaction and churn if predictions fall short. GAP analysis Outcome: A visual analytics dashboard providing real-time QoS predictions, alerts, and optimization potential to network operators, integrated into their network management systems. Gap(s) from result to outcome: - Technical Gap: Lack of integration APIs and visualization tools capable of handling largescale, real-time predictive data. - Operational Gap: Need for operator training and workflow adaptation to incorporate predictive insights into daily network operations. Steps needed: Design real-time APIs and data visualization modules. Key Exploitable Result 7.2: Anomaly Detection Partners Emnify, TID, JMU Result name Anomaly Detection Short description Using advanced anomaly detection systems to identify and resolve irregular network behaviors in real-time. Result type Prototype Target TRL 6-7 Expected impact Improved security, proactive issue resolution, reduced downtime Key innovations - AI anomaly detection, cross-network threat detection, automated incident response Target Markets MNOs, MVNOs, enterprises, government, defense Strengths Weaknesses - Proactive Security Threat Mitigation: Improved data representation enhances the ability of MNOs and MVNOs to detect anomalies in network traffic that may indicate security breaches or - False Positives: Even with improved data representation, anomaly detection systems may produce false positives, flagging benign events as threats. This can lead to unnecessary interventions that disrupt service or waste 63 Grant Agreement 101139270 — ORIGAMI — HORIZON-JU-SNS-2023 Deliverable D5.3 attacks (e.g., DDoS attacks, SIM swapping, fraud). Early detection allows operators to mitigate threats before they escalate, reducing the risk of data breaches or service interruptions. - Faster Issue Resolution: With enhanced anomaly detection, network operators can identify performance degradation or abnormal traffic patterns in realtime. This results in faster issue resolution, minimizing downtime and improving overall network reliability. - Network Stability: Anomaly detection ensures that any unusual behaviours—such as sudden spikes in traffic or unusual user activity— are promptly addressed. This enhances network stability and reduces the likelihood of cascading failures caused by undetected issues. resources, especially if automated responses are in place. - Data Overload: As networks grow in complexity and scale, the sheer volume of data to be monitored for anomalies can overwhelm detection systems. This might limit the effectiveness of anomaly detection if the systems struggle to process and analyze large datasets in real-time. - High Technical and Financial Costs: Developing and maintaining advanced anomaly detection systems requires significant investment in both infrastructure and technical expertise. Smaller operators may find it challenging to allocate sufficient resources to deploy effective systems. Opportunities Threats - Advanced AI/ML Implementation: Improved data representation provides a foundation for integrating AI/ML algorithms into anomaly detection systems. These technologies can enhance detection accuracy by learning from historical data and improving over time, allowing for more precise identification of potential threats. - Real-time Threat Intelligence Sharing: MNOs and MVNOs can collaborate to share threat intelligence in realtime, identifying and addressing emerging security threats as they occur across networks. This collaborative approach strengthens the overall security posture of the global cellular ecosystem. - IoT and 5G Network Protection: As 5G and IoT continue to proliferate, anomaly detection systems can help secure these emerging technologies by identifying vulnerabilities and threats unique to IoT devices, such - Evolving Cybersecurity Threats: Cybercriminals are constantly evolving their tactics, and there is a risk that anomaly detection systems may fail to keep up with new, sophisticated attack methods. This arms race between attackers and defenders poses an ongoing challenge for maintaining effective anomaly detection. - Over-reliance on Automation: While automation in anomaly detection is a strength, over-reliance on automated systems can be a threat if human oversight is reduced. Automated systems might miss complex or subtle threats that require human interpretation, or they may act on false positives without proper verification. - Regulatory and Compliance Risks: As anomaly detection systems collect and process large amounts of user data, operators may face regulatory scrutiny around how this data is used and stored. Data privacy laws, such as GDPR, could impose significant penalties if these systems infringe on user privacy. 64 Grant Agreement 101139270 — ORIGAMI — HORIZON-JU-SNS-2023 Deliverable D5.3 as unauthorized access or botnet activity. GAP analysis Outcome: A fast-paced anomaly detection algorithm capable of identifying and classifying network irregularities. It is used in MVNO networks to detect anomalies related to performance degradation, and to trigger automated or operator-assisted recovery actions. Gap(s) from result to outcome: - Technical Gap: Lack of large-scale, labeled datasets representing normal vs. abnormal network behaviors. - Algorithm Gap: Need for robust ML algorithms that can distinguish between benign fluctuations and true anomalies. - Operational Gap: Difficulty integrating anomaly detection outputs into existing network monitoring tools. Steps needed: - Collect and anonymize historical network logs for training and validation. - Develop models for time-series anomaly detection. 4.1.7 BARRIER #8: HIGH CONTROL-PLANE SIGNALING OVERHEAD Key Exploitable Result 8.1: Cloud-native structure for network core functionality Partners FOGUS, CMC Result name Cloud-native structure for network core functionality Short description Showcasing control and telemetry services in 5G and beyond networks. Result type Demonstrator Target TRL 6 Expected impact - Improved Automation and Management - Network Resilience and Availability - Flexibility, Scalability and Resource Optimization Key innovations - Service Mesh model for the core Network - 5G and Network Slicing - Edge Computing Integration - Microservices Architecture - Containerization and Orchestration Target Markets - Telecommunication Service Providers - Experimentation platform owners Strengths Weaknesses - Scalability: Cloud-native networks offer dynamic scaling that adjusts according to real-time network demand. - Cost Efficiency: Cloud-native approach can reduce both capital expenditures (CapEx) and operational expenditures (OpEx). - Automation Capabilities: Automation capabilities through orchestration tools such as - Security Concerns: Microservices and containerized architectures, are more complex and have more entry points, increasing the risk of security attacks. - Dependency on Cloud Providers: Concerns regarding service continuity and data confidentiality. - Expertise: Cloud-native architectures may require specialized skills. 65 Grant Agreement 101139270 — ORIGAMI — HORIZON-JU-SNS-2023 Deliverable D5.3 Kubernetes, reduce the need for manual checks, providing load balancing, service deployment and faster error detection. Opportunities Threats - Growing for Network-as-a-Service (NaaS): Cloud-native core networks can serve the demand for NaaS. - Edge Computing: Cloud-native networks can be extended to the edge, ensuring lower latency and better service delivery. - 5G: Cloud-native networks enable greater flexibility and scalability, simplifying processes like edge computing, network slicing, and more. - AI-Driven Network Optimization: Cloud-native networks can leverage AI and ML to optimize traffic flow and automate network maintenance. - Cybersecurity Threats: Microservices and containerized architectures are more complex and have more entry points, increasing the risk of security attacks. - Cloud Providers' Reliability: Network services will depend on cloud service providers, carrying the risk of downtime or service interruptions. GAP analysis Outcome: An open network core experimentation platform that offers multi-level monitoring and observability capable to embed and validate various optimisation mechanisms including core traffic management, load balancing and fault prevention. - Gap(s) from result to outcome: From the technical point of view, it is challenging to achieve stability of the performance as well as maintenance/continuous upgrade for the several software pieces that define the platform. Integration cycles might be painful since the platform integrates open-source software pieces each one developed for a specific scope and individual use. Steps needed: - Frame the scope in a subset of the capabilities that the service model offers. - Conduct stress tests (end-to-end verification) focusing on only 2-3 concepts/scenarios. - Validate the platform by including end user tests and get feedback from third parties. Key Exploitable Result 8.2: Analysis and optimization of network core traffic Partners FOGUS Result name Analysis and optimization of network core traffic Short description Scientific publication on network core traffic analysis and optimisation Result type Research Result Target TRL 4 Expected impact - Improved network efficiency - Better understanding on network traffic and signaling - Increased Network Reliability and Stability - Scalability 72 Grant Agreement 101139270 — ORIGAMI — HORIZON-JU-SNS-2023 Deliverable D5.3 3GPP RAN WG3FiberCop submitted the following three contributions for discussion at the 3GPP RAN3 meeting on 13-17 October: • R3-256544 “Discussion on RAN Service Based Solution” – It provides a general overview and motivation for studying the introduction of SBA in the RAN, discussing its potential benefits, constraints, and associated requirements • R3-256545 “Discussion on possible RAN internal architecture evolution towards SBA” – It outlines high-level SBA architectural schemes within the RAN that should be analyzed to understand the possible applicability and granularity of the SBA approach in the RAN • R3-256719 “Discussion on possible RAN-CN architecture evolution towards SBA” – It outlines high-level SBA architectural schemes for RAN-CN interface These contributions have been co-signed by Jio Platforms, KT Corp., Qualcomm Inc., and Telstra. Although these were not discussed, as the chair prioritized documents addressing high-level and principle-based requirements, the work made have significant impact, namely: • it brought, for the first time in a structured way, the need for RAN architectural evolution to the group’s attention. The documents received strong appreciation during offline discussion from the operator ecosystem, including co-signing partners (Jio, KT Corp., Qualcomm, Telstra) and other global operators such as Verizon, Deutsche Telekom, TIM, China Mobile, and TMobile US. We are also in ongoing discussions with Charter, Rakuten, and NTT Docomo. Vodafone, while cautious about the proposed evolution, expressed interest in continuing the dialogue to seek convergence; • a constructive exchange with the RAN3 chairman (Nokia) took place, who acknowledged the value of our contribution, even in this early phase focused on requirement gathering. This discussion highlighted how the service-based approach could apply not only to the RAN-CN interface (already foreseen in the Study Item approved by RAN plenary and also on SA2’s agenda) but also within the RAN itself; • this concept of RAN internal SBA is gaining traction in proposals from ZTE and Samsung, particularly for enabling new services (e.g., AI/ML, sensing), while maintaining a protocolbased approach for legacy services. Such an alternative positioning could represent a compromise solution, with the possibility of progressively extending SBA to legacy functions; • the main concern comes from traditional vendors like Ericsson and Nokia, who are proposing monolithic 6G architectures lacking clear functional separation (e.g., CU/DU, CP/UP). This approach undermines modular interaction and limits the effectiveness of a flexible, servicebased communication model, reducing its applicability to RAN-CN and inter-node links only. Operators have strongly opposed this, requesting to maintain architectural splits in the 6G study item; • in parallel, a broader cross-domain architectural vision is emerging (CN, OAM, RAN), introducing a dedicated data plane alongside traditional UP and CP. This would enable a unified data framework for cross-domain data collection, which only makes sense under a service-based approach. Such an architecture would be particularly suited to support AI/ML capabilities, now central to 6G. If introduced, this Unified Data Framework (UDF) will significantly impact innovative activities for 6G (including ORIGAMI overall architecture and NIs); 73 Grant Agreement 101139270 — ORIGAMI — HORIZON-JU-SNS-2023 Deliverable D5.3 • Finally, Qualcomm confirmed its interest in SBA-oriented architectural work and expressed willingness to explore ORIGAMI activities further, not only for GSBA but also for CCL and ZTL components, in line with their cloud-native vision. Following this, a call with them has been scheduled to present the activities and align on a common view. Based on these outcomes, FiberCop have finalized and submitted two contributions for the 3GPP RAN3 meeting on 17-21 November: • R3-258232 – “Requirements for 6G RAN architecture design” – It addresses a set of requirements for the 6G RAN architecture design, focusing on flexibility, scalability, and dynamicity (including proposed definitions). These requirements are derived from service use cases in SA1 TR 22.870 and include descriptions, rationale, and limitations of the current 6G system (considering also the RAN barriers identified in Origami), as well as functional and architectural impacts. The proposal is to capture these requirements within TR 38.760-3 related to the 6G SID under RAN3, and was co-signed by FiberCop, CEWiT, Fujitsu, Jio Platform, KT Corp., Qualcomm Inc., Rakuten, Teja Networks, TIM-Telecom Italia, T-Mobile USA, Verizon Wireless. • R3-258234 – “Preliminary analysis on RAN-CN interface evolution towards SBI” – It provides the rationale for introducing a service-based approach for the RAN-CN interface, as foreseen in the study item approved by RAN plenary. After a brief recap of the benefits (building on R3256544 from the previous meeting), it focuses on aspects to consider when adopting a servicebased model to identify, expose, and manage interactions among network functions. The proposal is to capture these aspects within TR 38.760-3 related to the 6G SID under RAN3 and it was co-signed by FiberCop, CEWiT, Deutsche Telekom, Qualcomm Inc., KT Corp. Based on the discussions held during the meeting, several steps forward were achieved, and the main outcomes related to the areas of interest are as follows: • Requirements – FiberCop document served as the basis for the discussion, with a particular focus on the sections addressing RAN service awareness and RAN resilience, in accordance with the guidance provided by the RAN plenary. The discussion also highlighted the need to introduce corresponding definitions, and given the proposal scope, FiberCop was tasked with coordinating an offline discussion to seek preliminary agreement on these aspects. The work carried out during the week, summarized in document R3-258805, enabled the approval of text to be included in the RAN3 6G technical report (TR 38.760-3), adding definitions and requirements related to these two dimensions in the RAN, i.e. awareness and resilience. Both requirements call for drivers such as flexibility, scalability, and dynamicity, which align well with the service-based architecture proposed for 6G. The final approved document is R3258856, a revision of the initial proposal in R3-258747, co-signed by the majority of companies participating in RAN3 (FiberCop, T-Mobile USA, TIM-Telecom Italia, Qualcomm, Verizon Wireless, Nokia, Ericsson, LG Electronics, Huawei, Vodafone, NEC, NTT Docomo, Rakuten, CMCC, Ofinno, Lenovo, BT, Google, Teja Networks, China Telecom, Xiaomi, Deutsche Telekom, ZTE, Jio Platforms, CATT, OPPO, Samsung). • Architecture – The FiberCop document on SBA for RAN-CN interface was not discussed, as the chair instead focus the discussion on incorporating into the RAN3 TR an initial high-level description of the two protocol-layer approaches, protocol-based and service-based, regarding the interface between the RAN and the CN. Consequently, the documents 74 Grant Agreement 101139270 — ORIGAMI — HORIZON-JU-SNS-2023 Deliverable D5.3 addressing functional, procedural, and evaluation aspects will be handled in subsequent meetings, following the direction set and agreed upon in this session. The document submitted by FiberCop can therefore be resubmitted, appropriately updated to reflect the decisions taken at this meeting, at future RAN3 meetings. It was also confirmed that, for internal RAN interfaces, no solution is excluded, including the service-based option for internal RAN communication interfaces. In fact, we succeeded in having a high-level description, along with a corresponding figure of a very general 6G RAN architecture, approved; this will be further detailed in line with the approaches to be discussed in the upcoming meetings. • The data collection framework was also discussed, and the first general descriptions were defined and added to the TR to support the need for a framework in which data can be generated and used by different entities. As FiberCop, we emphasized that such entities may reside not only within the RAN but also in external domains, reinforcing the view that a unified reference framework is required, consistent with the ongoing SA5 discussions on the Unified Data Framework (UDF). However, the detailed discussion, including the corresponding architectural approach, was postponed to the next meeting. In this regard, FiberCop intends to contribute at the upcoming RAN3 meetings, aligned with SA5, for proposing a unified crossdomain approach featuring a data repository (either single or per domain) and a service-based communication model. • Some high-level agreements on the AI/ML framework were also reached, primarily concerning the input and output data required for training and inference models, and therefore closely linked to the data collection framework. 3GPP SA WG5 The following contributions presented at the 3GPP SA5 Working Group provide important insights into the ongoing discussion on 6G management, orchestration, and data-related architectural evolution. Each document highlights aspects that are closely aligned with the ORIGAMI project’s objectives, particularly regarding unified data frameworks, service-based architectures, automation, and crossdomain management processes. • S5-253282 – “FiberCop view on SA5 Rel-20 6G Priorities” - This contribution outlines the strategic priorities identified by FiberCop for SA5 activities in the 6G timeframe. It places strong emphasis on architectural evolution toward a Service-Based Architecture (SBA), security enhancements, and network automation enabled through AI/ML techniques and Closed Control Loops (CCL). The document also highlights cloud-native design principles and energy-efficiency requirements as foundational elements of future network deployments. A central message is the need for a unified and coherent approach to data management, which is essential for supporting the envisioned automation and intelligence capabilities across all network domains. This perspective aligns well with current discussions on the Unified Data Framework (UDF) and reinforces the importance of consistent data models and cross-domain data exposure mechanisms. • S5-253321 – “Potential Impacts of RAN Evolution Towards a Service-Based Architecture (SBA) on Management Processes and Infrastructures in 6G” - This contribution analyses the implications of transitioning the RAN to a fully service-based architecture, with a specific focus on the resulting impacts on management processes and supporting infrastructures. It stresses that adopting an SBA-based RAN will significantly influence configuration, performance, and fault-management models, requiring new interaction patterns, refined data-collection 75 Grant Agreement 101139270 — ORIGAMI — HORIZON-JU-SNS-2023 Deliverable D5.3 architectures, and enhanced coordination with system-wide automation functions. The contribution highlights the need for SA5 to anticipate these changes and to ensure that management specifications evolve in a way that remains coherent with architectural and procedural developments in RAN and Core Network domains. • S5-255272 – “Discussion on the Scope and Role of SA5 in the Definition of the 6G Unified Data Framework” - This contribution provides an initial analysis of the scope and responsibilities of SA5 in defining the 6G Unified Data Framework. It discusses how management-plane functions and processes may contribute to, interact with, and benefit from the introduction of the UDF as a new architectural and functional enabler for 6G. The document highlights the need for SA5 involvement in shaping data-collection mechanisms, data-exposure models, assurance workflows, and automated management procedures that will rely on the UDF as a foundational element. It also recognises the importance of aligning SA5 work with the broader architectural studies on 6G data frameworks, ensuring consistency across domains such as RAN, CN, and analytics/AI functions. 4.3.1.2 O-RAN ALLIANCE During Y2, ORIGAMI has continued contributed to WG2, primarily in the context of the definition of APIs and services for AI/ML model management and deployment within the RAN. More specifically, we made 11 contributions on AI-ML topics: • WG2-CR-0084-A1GAP - Add cancel AIML model training V1-drafting • WG2-CR-0083-A1AP - Add service operations for cancel AIML model training V2-drafting • WG2-CR-0083-A1AP - Add service operations for cancel AIML model training V1-drafting • WG2-CR-0082-A1UCR - Cancel model training use case-v3-drafting • WG2-CR-0082-A1UCR - Cancel model training use case-v2-drafting • WG2-CR-0082-A1UCR - Cancel model training use case-v1-drafting • WG2-CR-0081-A1UCR - Add requirement on delete model training -v1-drafting • WG2-CR-0080-A1AP - Add the introduction of service operations of A1-ML model training API V2-drafting • WG2-CR-0078-A1AP - Add service operations for create AIML model training V2-drafting • WG2-CR-0072-R1AP - AIML model retrieve API-resources v1-drafting • WG2-CR-0071-R1AP - AIML model retrieve API-service operations v1-drafting 7 contributions on the Non-RT RIC architecture: • WG2-CR-0078 - Non-RT RICFederated learning v3 • WG2-CR-0078 - Non-RT RICFederated learning v2 • WG2-CR-0067 - Non-RT RICAI ML model lifecycle v8 • WG2-CR-0067 - Non-RT RICAI ML model lifecycle v7 • WG2-CR-0067 - Non-RT RICAI ML model lifecycle v6 • WG2-CR-0067 - Non-RT RICAI ML model lifecycle v5 • WG2-CR-0067 - Non-RT RICAI ML model lifecycle v4 10 contributions to R1AP and R1GAP:WG2-CR-0075-R1AP - AIML model retrieve v3 • WG2-CR-0075-R1AP - AIML model retrieve v2 • WG2-CR-0074-R1AP - AIML model retrieve API-data model v2 76 Grant Agreement 101139270 — ORIGAMI — HORIZON-JU-SNS-2023 Deliverable D5.3 • WG2-CR-0074-R1AP - AIML model retrieve API-data model v1 • WG2-CR-0073-R1AP - AIML model retrieve API-URI structure v2 • WG2-CR-0073-R1AP - AIML model retrieve API-URI structure v1 • WG2-CR-0072-R1AP - AIML model retrieve API-resources v3 • WG2-CR-0072-R1AP - AIML model retrieve API-resources v2 • WG2-CR-0071-R1AP - AIML model retrieve API-service operations v3 • WG2-CR-0071-R1AP - AIML model retrieve API-service operations v2 And 20 contributions to A1AP: • WG2-CR-0088-A1AP - Add OpenAPI for AIML model training V6 • WG2-CR-0088-A1AP - Add OpenAPI for AIML model training V5 • WG2-CR-0088-A1AP - Add OpenAPI for AIML model training V4 • WG2-CR-0088-A1AP - Add OpenAPI for AIML model training V2 • WG2-CR-0088-A1AP - Add OpenAPI for AIML model training V1 • WG2-CR-0086-A1APAdd data model for AIML model training V3 • WG2-CR-0086-A1AP - Add data model for AIML model training V2 • WG2-CR-0086-A1AP - Add data model for AIML model training V1 • WG2-CR-0083-A1AP - Add service operations for cancel AIML model training V4 • WG2-CR-0083-A1AP - Add service operations for cancel AIML model training V3 • WG2-CR-0080-A1AP - Add the introduction of service operations of A1-ML model training API V3 • WG2-CR-0083-A1AP - Add service operations for cancel AIML model training V5 • WG2-CR-0080-A1AP - Add the introduction of service operations of A1-ML model training API V3 • WG2-CR-0078-A1AP - Add service operations for create AIML model training V4 • WG2-CR-0078-A1AP - Add service operations for create AIML model training V4 • WG2-CR-0078-A1AP - Add service operations for create AIML model training V3 • WG2-CR-0077-A1AP - Add A1-ML model training API v4 • WG2-CR-0077-A1AP - Add A1-ML model training API v3 • WG2-CR-0077-A1AP - Add A1-ML model training API v2 • WG2-CR-0077-A1AP - Add A1-ML model training API v1 4.3.2 STANDARDIZATION ACTIVITIES PLANNED FOR Y3 4.3.2.1 3GPP Listed below are the most important topics related to ORIGAMI’s goals 4.3.2.1.1 KEY TOPICS FOR 6G IN 3GPP BEYOND 2025 As 3GPP transitions from its focus on 5G Advanced (5G-A) with Release 19 to the first phases of 6G development starting from Release 20 Study Item phase, several key areas of innovation are emerging. These topics reflect the need for a leap in technological capabilities to meet the demands of the 2030s and beyond. The key areas are outlined below: In Year 3, the ORIGAMI project will intensify its standardization activities in 3GPP, with a particular focus on the architectural evolution of the Radio Access Network (RAN) and management functionalities in the context of 6G. The planned contributions will address the following key areas, 77 Grant Agreement 101139270 — ORIGAMI — HORIZON-JU-SNS-2023 Deliverable D5.3 reflecting both the latest research directions and the priorities identified within the project and the broader SNS/6G-IA community: • 6G RAN Requirements and Network Intelligence: the project will propose architectural requirements for the 6G RAN, emphasizing flexibility, scalability, and dynamicity, based on advanced use cases and aligned with the ongoing discussions Within 3GPP. Particular attention will be devoted to the integration of pervasive, native AI/ML frameworks across all in RAN domains, recognizing AI/ML both as a key enabler and as a primary use case that fully exploits the network’s data assets. In this context, The identification and specification of network intelligence functions will be guided by the project’s technical outcomes, ensuring that the resulting requirements address real-world needs for automation, closed-loop control, and datadriven optimization. • Service-Based Architecture (SBA) Evolution: ORIGAMI will continue to advance the study of service-based architectural principles, focusing on the RAN–CN interface as well as also on internal RAN architecture interfaces. This effort aims to promote modularity, interoperability, and the exposure of RAN capabilities as services, thereby enabling new business models and operational paradigms as outlined by the project. The contribution will also address the definition of high-level service-based architecture schemes, functional splits, and interface models, in line with the vision defined within the project itself. • Unified Data Framework (UDF) for a Data Plane definition: a key area of work will be the definition of a Unified Data Framework (UDF) for 6G, conceived to provide a cross-domain data plane that complements the traditional control and user planes in the RAN as wells as the management plane of OAM domain. This framework will be designed to support AI/ML workflows, federated learning, context-aware inference, and advanced management functions, ensuring consistent data representation and exposure across the RAN, CN, and OAM domains. A harmonized approach to data collection, processing, and sharing will be promoted, working in close coordination in both RAN3 and SA5 working groups. • Native AI/ML Frameworks and Network Sensing: the project will promote the adoption of native and pervasive AI/ML frameworks, closely integrated with the UDF to enable intelligent automation and support new classes of services. It will also contribute to the integration of network sensing capabilities, leveraging the RAN as a platform for environmental awareness and enabling innovative applications, e.g., like joint communication and sensing (ISAC). • Energy Efficiency and Sustainability: energy efficiency remains a central priority, with contribution to the definition of requirements and architectural enablers aimed at reducing power consumption and fostering sustainable network operation. The activity will continue to advance solutions for energy-efficient RAN management, making use of AI/ML and dynamic resource allocation to optimize overall energy performance. • 5G-A (Rel-20) Evolution and Bridging to 6G: A strong link with the ongoing evolution of 5G will be maintained, particularly 5G-A (Rel-20), focusing on the introduction of AI/ML for new use cases, concepts related to ISAC and the reinforcement of sustainability and energy efficiency as guiding principles. 78 Grant Agreement 101139270 — ORIGAMI — HORIZON-JU-SNS-2023 Deliverable D5.3 These activities are designed to maximize the project’s influence on 6G research and development, fostering collaboration and innovation within the standardization ecosystem. All planned contributions will be closely coordinated between RAN3 (for architectural and interface evolution) and SA5 (for management, orchestration, and data frameworks), ensuring a unified and impactful approach to 6G standardization. 4.3.2.2 O-RAN ALLIANCE ORIGAMI substantially contributed to O-RAN WG 3 (Year 1) and WG 2 (Years 1 and 2) with a total of 62 contributions to date. All of these contributions targeted the integration of AI/ML workloads that are relevant to ORIGAMI use cases defined in D2.1 and D2.2. Year 3 will focus on final implementation and evaluation of all of the pending use cases that have not been reported in D4.2. Consequently, we plan to substantially slow down our contributions to O-RAN during Year 3, with only essential contributions that we missed during the design phase of the related ORIGAMI use cases. 4.3.3 OPEN-SOURCE CONTRIBUTIONS PLANNED FOR Y3 (PER OPEN-SOURCE FORA) As part of their ongoing commitment to open-source collaboration, TID have made initial contributions to the CAMARA project. Specifically, we submitted a commit (ID submitted by TID) to the CAMARA repository. Additionally, we have developed concrete plans to integrate several solutions, particularly those created within the context of the CAMARA initiative. These integration efforts will be further advanced and completed in D5.4. 79 Grant Agreement 101139270 — ORIGAMI — HORIZON-JU-SNS-2023 Deliverable D5.3 5 TARGET INDICATORS ACHIEVEMENT IN Y2 Table 5 below summarizes the KPIs, targets set to be achieved by M24 of the project execution, the current status (M23), and the overall associated targets (by M36) defined by the ORIGAMI CoDEP. Nature Community KPI Target at M24 Current status Target at M36 Notes Communication General Official project website 1 1 1 Online by M1 Official project social media accounts 4 (channels) 4 4 Twitter/X, LinkedIn, YouTube, Instagram Official project leaflet 1 1 1 Produced by M3 Press releases 4 3 10 TID+IMDEA, TID+UC3M, EMN Videos of the project vision, solutions, and outcomes 4 6 5 5G Techritory video, SNS webinar, professional video, CloudRIC demo, DUNE demo, EUCNC video Dissemination Academia Scientific publications 50 52 80 29 at top venues (CORE A*) and journals (JCR Q1) Edited special issues of international peer-reviewed scientific journals 1 3 2 Elsevier ComCom, PACMNET V3 June 2025 and September Collaborations with other scientific projects 2 2 2 Joint Stream B and C workshop, joint Stream B and D workshop on KVIs ORIGAMI scientific workshops 2 6 2 GLOBECOM 2024, 2 workshops and 3 special sessions at EUCNC 80 Grant Agreement 101139270 — ORIGAMI — HORIZON-JU-SNS-2023 Deliverable D5.3 Industry In-presence participation in industry fora and events 8 11 10 2024: MWC, 5G Forum, Digital Transformation World Ignite, London Tech Week, AI for Good Summit 2025: MWC, Devoxx France, EUCNC, Connected Britain, Barcelona DeepTech Summit ORIGAMI industry workshops 1 1 2 SA5 joint workshop with SA5 plenary meeting Academia Industry Demonstrations 4 5 8 CloudRIC, DUNE at INFOCOM25, Demonstrating Deep Learningbased Spatial Diffusion at INFOCOM25, 2 Demos at EUCNC Exploitation Industry Contributions to standards 2 73 10 3 contributions to 3GPP WG SA5, 1 contribution to ITU-T FG-AINN, 62 contributions to O-RAN, 3 contributions to 3GPP WG RAN3, 3 contributions to 3GPP 6G WS, 1 contribution to IRTF SUSTAIN 81 Grant Agreement 101139270 — ORIGAMI — HORIZON-JU-SNS-2023 Deliverable D5.3 Patent applications 0 4 9 1 patent on energy efficiency features for ORAN base stations [14], 1 patent on RAN virtualization techniques [15], 1 patent on anomaly detection [16], 1 patent on distributed network traffic decomposition [17] Large opensource project contributions 0 1 2 1 contribution to CAMARA Table 5: Achieved ORIGAMI KPIs for communication, dissemination, and exploitation during the second year of the project As outlined in Sections 2 and 3 and summarized in Table 5, ORIGAMI aligns with the strategic framework established in D5.1 and D5.2 concerning the outcomes for Y2 and the ultimate objectives. In terms of scientific dissemination, ORIGAMI emphasizes the quality of publication venues, thereby striving to disseminate research in the most prestigious conferences and journals within the discipline. Currently, 29 out of 49 papers have been accepted in top-tier conferences (CORE A*) and journals (JCR Q1). The only KPI that is underachieved by M23 is the number of press releases (3 out of 4). We plan to address this gap and issue the remaining seven press releases in Y3 to reach the final target. 88 Grant Agreement 101139270 — ORIGAMI — HORIZON-JU-SNS-2023 Deliverable D5.3 through thought leadership articles, podcasts, etc. to increase awareness of value among the prospective customer base. I.3 NEC Key Exploitable Results (KERs) Relevant to the Partner List the KERs (identified by the project) that your organization intends to exploit. KER # Title Partner's Role/Contribution 1 CloudRIC++ NEC contributed to the design and implementation of this solution. 2 GPU-assisted vRAN NEC built (designed and evaluated) a novel OpenRAN-compliant system to dynamically multiplex the resources of GPUs in GPUpowered base stations between 5G workloads and third-party services. Section A: Exploitation Activities so far 1. Internal Use / Strategic Adoption - Integration into internal R&D or roadmaps: The two KERs, CloudRIC++ and GPU-assisted vRAN, are foundational research outcomes currently integrated into NEC's internal R&D efforts. They represent key advancements in energy and cost-efficiency for virtualized and Open RANs, forming a crucial basis for future product development. These technologies are being explored within NEC's research departments, influencing the strategic direction for next-generation network solutions. - Products/processes influenced: The core innovations from both KERs, such as joint radio and compute control, heterogeneous processor utilization, and GPU multiplexing, directly influence NEC's Open RAN and RAN virtualization product lines. These advancements aim to enhance the performance, energy efficiency, and cost-effectiveness of NEC's offerings in these critical market segments, driving internal process improvements for network optimization. - Departments involved: Primarily, NEC's R&D divisions and engineering teams focused on Radio Access Networks (RAN) and virtualization are deeply involved. Product management and strategy departments are also engaged in understanding how these research outcomes can be translated into competitive features and solutions for NEC's commercial portfolio, ensuring alignment with market demands. 2. Commercial / Pre-commercial Efforts - Pilots, PoCs, or early business plans: As both KERs are currently at TRL 4 (Demonstrator/Research Result), external commercial pilots or Proofs of Concept (PoCs) are not yet explicitly underway. However, internal validation and feasibility studies are ongoing to prepare these technologies for future pre-commercial engagements. Early business considerations are focused on quantifying the energy and cost savings potential for operators. 89 Grant Agreement 101139270 — ORIGAMI — HORIZON-JU-SNS-2023 Deliverable D5.3 - Target market / segments: The primary target markets for both KERs are O-RAN vendors and operators, as well as the broader vRAN vendor and operator ecosystem. These technologies address the growing demand for flexible, cost-efficient, and energy-saving solutions in 5G networks, particularly within the evolving Open RAN landscape. - Client discussions or demo engagements: Direct client discussions or external demo engagements are limited at this TRL. However, internal demonstrations and technical deep-dives are being conducted to showcase the capabilities and potential benefits to internal stakeholders, paving the way for future customer interactions as the technology matures. - IPR filed (patents, trademarks): NEC has filed Intellectual Property Right (IPR) applications directly stemming from these KERs. 3. Standardization Contributions (if applicable) - SDOs targeted (3GPP, O-RAN, etc.): Both KERs are designed with strong compatibility and integration into the O-RAN architecture, making the O-RAN Alliance the primary Standardization Development Organization (SDO) targeted for future contributions. The focus on Open RAN and RAN virtualization aligns directly with O-RAN's objectives. - Working groups joined: NEC participates in O-RAN WG1, WG2, WG3 and WG6, among others, which are directly related to these KERs. - Documents submitted: Please, refer to D5.2 (NEC contributions to O-RAN Alliance). Section B: Exploitation Plans 1. Planned Internal Adoption / Integration - Targeted internal milestones (e.g., trials, internal reviews, lab tests): A key milestone is the dedicated workshop in Q4 of Year 2, where NEC will present the project's findings and applications to internal business units. This aims to facilitate the integration of ORIGAMI technologies into the roadmap of NEC's Open RAN and RAN virtualization products, ensuring strategic alignment and fostering internal adoption. - KPIs or metrics to be tracked (e.g., energy savings, latency improvements): For CloudRIC++, the primary KPIs include increased energy efficiency and cost efficiency in O-RAN virtualized RANs. For GPU-assisted vRAN, the focus is on increasing cost-efficiency in GPU-accelerated vRANs, alongside reliable multiplexing of GPU resources. These metrics will guide internal trials and reviews, demonstrating tangible benefits. 2. Commercial Strategy - Business use cases or services to be developed: The KERs enable advanced resource sharing and optimization, leading to business use cases such as highly energy-efficient 5G deployments, flexible network slicing with guaranteed performance, and cost-optimized edge AI/ML integration within RAN. These capabilities support new services requiring low latency and high throughput. - Potential customers or verticals: The primary potential customers remain O-RAN vendors and operators, along with enterprises adopting private 5G networks and edge computing solutions. 90 Grant Agreement 101139270 — ORIGAMI — HORIZON-JU-SNS-2023 Deliverable D5.3 Verticals include telecommunications, manufacturing (for industrial IoT with edge ML), and smart cities, all benefiting from optimized, high-performance, and cost-effective network infrastructure. - Exploitation readiness level (TRL): While the KERs are currently at TRL 3, the exploitation plans aim to advance this. The patent filing and internal workshops are steps towards increasing the TRL, moving towards TRL 4. 3. Standardization Objectives (if applicable) - SDOs/WGs where contributions will be made NEC will continue to target the O-RAN Alliance, particularly working groups focused on architecture, interfaces, and energy efficiency. Contributions will aim to influence the evolution of Open RAN specifications to incorporate the principles of AIpowered resource optimization and heterogeneous compute utilization - Contribution topics: Key contribution topics will include AI-driven resource allocation mechanisms for energy and cost efficiency, methods for multiplexing 5G and ML workloads on shared hardware accelerators (like GPUs), and advanced compute-aware radio scheduling policies within the O-RAN framework. - Expected timeline: The submission of scientific publications to top-tier conferences is planned for Q1 2026. While not direct SDO contributions, these publications will lay the groundwork for future standardization discussions, with formal SDO contributions likely following the internal workshop in Q4 Y2 and additional patent filing in Q1 Y3. 4. Risks & Mitigations - Key risks for exploitation (e.g., technical readiness, market alignment, IPR issues): Risks include the inherent complexity of the architecture and algorithms, potential performance interference from NVIDIA MPS technologies, the need for continuous training and updating of data-driven models, and the highly competitive vRAN market. - Mitigation actions planned: To mitigate complexity, NEC will focus on robust internal technology transfer through workshops. Market competitiveness will be addressed through IPR filing (patent in Q1 Y3) for differentiation and showcasing achievements via scientific publications. Ongoing R&D will aim to refine models, reduce overhead, and enhance isolation. I.4 FOGUS Key Exploitable Results (KERs) Relevant to the Partner List the KERs (identified by the project) that your organization intends to exploit. KER # Title Partner's Role/Contribution 8.1 Cloud-native structure for network core functionality FOGUS / Developer and integrator 8.2 Analysis and optimization for network core traffic FOGUS / Conducts the study 91 Grant Agreement 101139270 — ORIGAMI — HORIZON-JU-SNS-2023 Deliverable D5.3 Section A: Exploitation Activities so far 1. Internal Use / Strategic Adoption - Integration into internal R&D or roadmaps: Fully aligned with the R&D activities. Action started within ORIGAMI and is expected to evolve after the project lifetime. - Products/processes influenced: Independent service line (no influence) - Departments involved: R&D 2. Commercial / Pre-commercial Efforts - Pilots, PoCs, or early business plans: An initial PoC presented in SNS4SNS event Nov 2024 - Target market / segments: Not specified yet. - Client discussions or demo engagements: NO - IPR filed (patents, trademarks): NO 3. Standardization Contributions (if applicable) – N/A - SDOs targeted (3GPP, O-RAN, etc.): - Working groups joined: - Documents submitted: Section B: Exploitation Plans 1. Planned Internal Adoption / Integration FOGUS R&D activities are tightly related to network architecture enhancements and to network performance evaluations. Thus, FOGUS will take advantage of the project results to upgrade its software tools and testbeds with new features. Given the specific project scope, FOGUS aims to materialize and monetize, when possible, solutions emerging from the project that reside at two primary sectors: mobile network openness and mobile network virtualization. These sectors are interconnected, influencing each other's development, and the advancement of one will strengthen the application of the other. The potential impact of these sectors is significant and will be pivotal in shaping the presence of the company in the 6G research ecosystem. Overall, the involvement of the company in the project is expected to strengthen the company’s position against the competition in the fields of experimentation and benchmarking. Also, since FOGUS invests in training and consulting services, the know-how acquired by the project will be exploited by the training and consulting sector in FOGUS to devise new courses and training material. - Targeted internal milestones (e.g., trials, internal reviews, lab tests): • Oct.15, 2025: Full set of components available, results from the second study available. - KPIs or metrics to be tracked (e.g., energy savings, latency improvements): • Communication heat maps 92 Grant Agreement 101139270 — ORIGAMI — HORIZON-JU-SNS-2023 Deliverable D5.3 • Data rates • Topology • Energy consumption 2. Commercial Strategy - Business use cases or services to be developed: Action not started - Potential customers or verticals: Vertical agnostic solution - Exploitation readiness level (TRL): 6 3. Standardization Objectives (if applicable) – N/A - SDOs/WGs where contributions will be made: - Contribution topics: - Expected timeline: 4. Risks & Mitigations - Key risks for exploitation (e.g., technical readiness, market alignment, IPR issues): - Mitigation actions planned: I.5 EMN Key Exploitable Results (KERs) Relevant to the Partner List the KERs (identified by the project) that your organization intends to exploit. KER # Title Partner's Role/Contribution 1 Global operator monitoring emnify designed, implemented and operates the KER 2 Operator Anomaly Detection emnify worked together with UC3M Section A: Exploitation Activities so far 1. Internal Use / Strategic Adoption Integration into internal R&D or roadmaps: emnify’s development of a global operator monitoring is tightly aligned with the operational experiences of running a global IoT operator. The KER developed within ORIGAMI has been partially adopted within the operational teams for monitoring of latency and availability of roaming partners during incidents. Part of emnify’s internal roadmap is to extend features and use cases of this monitoring solution to also generate alerts when networks go down or latency between core networks of roaming partners and emnify increases. 93 Grant Agreement 101139270 — ORIGAMI — HORIZON-JU-SNS-2023 Deliverable D5.3 Products/processes influenced: The main product of emnify is its connectivity platform. Customers of this platform benefit from an increased awareness of network outages, as in many cases emnify must block faulty networks to force devices over to alternative operators. The KRE influenced the internal process of investigating (potential) network outages and facilitates the root cause analysis. Departments involved: Several teams of emnify are included into the exploitation activities: The Engineering team implements and extends the operator monitoring solution. The Support team validates customer complaints based on monitoring data The Incident Management team uses the monitoring solution to understand scope, impact and root causes of operator outages The Sales and Marketing team demonstrate the capabilities that emnify has for providing reliable, global connectivity. 2. Commercial / Pre-commercial Efforts Pilots, PoCs, or early business plans: Internal validation and early use to integrate it into the standard operational procedures (SOPs) Target market / segments: Internal users Client discussions or demo engagements: Network-to-network latency has been supporting ~10 client conversations, where knowledge about end-to-end latency has been crucial. The ability to demonstrate such deep understanding helps emnify to demonstrate competence and control over its global operator architecture. IPR filed (patents, trademarks): none 3. Standardization Contributions (if applicable) - SDOs targeted (3GPP, O-RAN, etc.): N/A - Working groups joined: N/A - Documents submitted: N/A Section B: Exploitation Plans 1. Planned Internal Adoption / Integration Targeted internal milestones (e.g., trials, internal reviews, lab tests): After adoption for ad-hoc monitoring purposes, automated alerting as well as routing to breakout regions based on latencies. KPIs or metrics to be tracked (e.g., energy savings, latency improvements): Latency reduced by selecting better breakout region; time reduced to gain understanding about outages of roaming partners. 94 Grant Agreement 101139270 — ORIGAMI — HORIZON-JU-SNS-2023 Deliverable D5.3 2. Commercial Strategy Business use cases or services to be developed: Besides the internal use, displaying the health of emnify’s global roaming connectivity could be displayed publicly on status.emnify.com. Potential customers or verticals: All customers regarding improvement of availability; latencysensitive customers for tuning of breakout selection. Exploitation readiness level (TRL): 5 3. Standardization Objectives (if applicable) - SDOs/WGs where contributions will be made: N/A - Contribution topics: N/A - Expected timeline: N/A 4. Risks & Mitigations Key risks for exploitation (e.g., technical readiness, market alignment, IPR issues): Technical risks of monitoring latency of S-GWs / UPFs of roaming partners at high frequency (few minutes) are mainly driven by potential complaints of these roaming partners. While we expect no practical issues of mobile core equipment to handle GTP echo requests, operations teams at network operators often complain about not following (legacy) practices. Regarding publication of global roaming health, we see risk of roaming partners being – in doubt falsely – put on the spot for having issues. Further, emnify-own marketing needs to align with the fact that operational issues are publicly exposed. Mitigation actions planned: Currently, measurements are executed once per hour per S-GW only. The fact that issues are shown publicly could be remediated by increasing the thresholds (severity, duration) until when an issue is publicly shown, as well as a manual review through the NOC team. I.6 CMC Key Exploitable Results (KERs) Relevant to the Partner List the KERs (identified by the project) that your organization intends to exploit. KER # Title Partner's Role/Contribution 1 Global operator model from private mobile network perspective Development of network APIs 95 Grant Agreement 101139270 — ORIGAMI — HORIZON-JU-SNS-2023 Deliverable D5.3 Section A: Exploitation Activities so far 1. Internal Use / Strategic Adoption Integration into Internal R&D and Roadmaps: Cumucore’s development of network APIs is tightly aligned with real-world market feedback and customer requirements. Feedback from system integrators, vertical industry partners, and end users is systematically collected and analyzed to shape the company’s research and development priorities. This agile approach ensures that the evolving needs of industrial, defense, and mission-critical communications markets are directly reflected in the product roadmap. The continuous loop between field insights and engineering implementation enables rapid innovation and adaptation. Products and Processes Influenced: The primary product influenced by this feedback-driven development approach is Cumucore’s 5G mobile core, purpose-built for private mobile network deployments. This includes core components such as AMF, SMF, UPF, and NEF, all designed to support edge-native deployment, low-latency use cases, and integration with industrial systems. API development enables advanced features like dynamic network slicing, traffic shaping, device management, and integration with OT environments via open interfaces. Additionally, internal processes such as quality assurance, software testing, and documentation have been refined to accommodate modular and API-driven development workflows. Departments Involved: The integration of market feedback into product development is a companywide effort. The engineering team translates requirements into features and maintains the modular, containerized software architecture. The sales and business development teams collect feedback from deployments and maintain close communication with partners and customers. The product management team prioritizes feature requests and updates the roadmap accordingly. The support and operations teams ensure that customer-driven updates are effectively delivered, tested, and documented. This holistic involvement ensures that Cumucore’s entire organization remains responsive, customer-focused, and aligned with the real-world demands of private mobile network deployments. 2. Commercial / Pre-commercial Efforts Pilots, Proof-of-Concepts (PoCs), and Early Commercialization: Cumucore’s technology has been successfully deployed in multiple pilot projects and proof-of-concept (PoC) trials, validating its applicability across diverse use cases. These initiatives demonstrate the readiness of Cumucore’s 5G mobile core to support industrial and media applications with stringent requirements for latency, reliability, and integration with existing systems. Commercialization efforts are currently underway through a growing network of system integrator partners and regional resellers. These partnerships enable scalable go-to-market strategies, allowing Cumucore to focus on core technology development while leveraging local expertise and customer relationships for deployment and support. 96 Grant Agreement 101139270 — ORIGAMI — HORIZON-JU-SNS-2023 Deliverable D5.3 Target Markets and Segments: Cumucore’s solution is tailored for two high-impact verticals: Industry 4.0, where private 5G networks support automation, robotics, real-time process control, and predictive maintenance in manufacturing, logistics, energy, and mining sectors. Media and Entertainment, where reliable, high-bandwidth wireless communication is required for remote broadcasting, temporary event networks, and mobile production units. In both segments, Cumucore’s advanced features such as Time-Sensitive Networking (TSN), 5GLAN, and support for nonpublic network slicing enable new business models and operational efficiencies. Client Discussions and Demo Engagements: Cumucore is actively engaged in ongoing projects with clients and research institutions around the world. These include live network deployments, joint demonstrations, and collaborative research initiatives. Notable projects span Europe, North America, and Asia, including deployments in smart factories, mines, ports, TV production environments, and emergency response scenarios. These engagements validate the adaptability of Cumucore’s technology to diverse operational conditions and regulatory environments. Intellectual Property (IPR) N/A 3. Standardization Contributions (if applicable) - SDOs targeted (3GPP, O-RAN, etc.): N/A - Working groups joined: N/A - Documents submitted: N/A Section B: Exploitation Plans 1. Planned Internal Adoption / Integration Targeted Internal Milestones: Cumucore has set clear internal milestones to guide the evolution of its product portfolio. The next major objective is the release of a new version of the 5G core platform, optimized for market trials by the end of 2025. This release will incorporate enhancements based on lessons learned from current deployments and will include extended API functionality, improved orchestration features, and deeper integration capabilities with industrial control systems. Prior to external trials, the product will undergo rigorous lab testing, internal validation, and compatibility assessments with selected radio vendors and network function components. Key Performance Indicators (KPIs) and Metrics: To ensure the solution meets the performance and reliability demands of its target markets, Cumucore tracks several critical KPIs: Bit Error Rate (BER): Continuous improvement of BER is a priority, especially for use cases requiring ultra-reliable communication such as autonomous vehicles, industrial automation, and media transmission. Capacity Control: Monitoring and managing network load in real time is essential for enabling dynamic slicing, optimizing spectrum use, and ensuring service continuity in mixed-use environments. Additional internal metrics include system latency, boot-up time for edge nodes, QoS enforcement efficiency, and compatibility with third-party radio and device ecosystems. These indicators support 97 Grant Agreement 101139270 — ORIGAMI — HORIZON-JU-SNS-2023 Deliverable D5.3 ongoing product refinement and help Cumucore meet the operational expectations of demanding private network scenarios. 2. Commercial Strategy Business Use Cases and Services Under Development: Cumucore continues to actively develop and refine business use cases that leverage its advanced private 5G core capabilities. Ongoing development efforts focus on enabling new services such as real-time monitoring and control for autonomous systems, mobile video production platforms, and resilient communication for emergency response. These use cases are shaped through close collaboration with industry stakeholders and are supported by the company’s modular architecture, which allows for rapid customization and integration with existing workflows and infrastructure. Potential Customers and Industry Verticals: The primary target customers include global industrial automation companies and media equipment manufacturers seeking to embed connectivity into their solutions or deploy full-stack communication systems in their operational environments. Cumucore’s technology addresses the specific needs of these sectors by enabling features such as deterministic networking, low-latency communication, and seamless device onboarding through eSIMs and open APIs. Additional verticals with strong potential include defense, logistics, transportation, and energy. Exploitation Readiness Level (TRL) Cumucore’s private 5G core platform is currently at TRL 7 - demonstrated in operational environments under real-world conditions. Multiple pilots and customer deployments have validated both the technology and the operational model, and the company is now scaling toward broader commercial rollout through strategic partnerships and integrator channels. 3. Standardization Objectives (if applicable) - SDOs/WGs where contributions will be made: N/A - Contribution topics: N/A - Expected timeline: N/A 4. Risks & Mitigations Key Risks for Exploitation The primary risk to widespread exploitation lies in the market acceptance of cellular technology, particularly in industrial environments where legacy systems and entrenched wireless solutions (e.g., Wi-Fi or proprietary radio) are still widely used. Some stakeholders may perceive cellular technologies as complex, costly, or over-engineered for their specific use cases, especially in small and medium-sized enterprises. Mitigation Actions Planned Cumucore addresses these concerns by working closely with customers to demonstrate the practical benefits of private cellular networks, such as reliability, scalability, deterministic performance, and secure remote access. Tailored pilot deployments, hands-on demonstrations, and close integration with existing OT environments help build trust and reduce perceived risk. In addition, Cumucore’s simplified user interface, modular architecture, and flexible pricing model are designed to lower the barrier to entry and ease adoption. Continued engagement with system integrators and industry-specific partners also ensures that the solution is aligned with real operational needs. 104 Grant Agreement 101139270 — ORIGAMI — HORIZON-JU-SNS-2023 Deliverable D5.3 - Target market / segments: edge cloud based solution (e.g. industrial customers) and evolved network management system - Client discussions or demo engagements: not applicable - IPR filed (patents, trademarks): not applicable 3. Standardization Contributions (if applicable) - SDOs targeted (3GPP, O-RAN, etc.): mainly 3GPP for 6G system. - Working groups joined: 3GPP RAN3 for radio Access Network architecture and interfaces aspects and 3GPP SA5 for OAM aspects - Documents submitted: for SA5 mainly regarding EE aspects and AI/ML; for RAN3: mainly regarding RAN requirement to effectively support 6G services and RAN Service Based Architecture and toward Core Network. In both groups FiberCop is committed on AI/ML framework through Network Intelligence to enhance RAN functionalities and network management. Section B: Exploitation Plans 1. Planned Internal Adoption / Integration - Targeted internal milestones (e.g., trials, internal reviews, lab tests): not applicable. - KPIs or metrics to be tracked (e.g., energy savings, latency improvements): number of 3GPP standard contributions. 2. Commercial Strategy - Business use cases or services to be developed: not applicable - Potential customers or verticals: not applicable - Exploitation readiness level (TRL): not applicable 3. Standardization Objectives (if applicable) - SDOs/WGs where contributions will be made: 3GPP RAN and SA (especially RAN WG3 and SA WG5) - Contribution topics: see above - Expected timeline: Aligned with 3GPP 6G timeline 4. Risks & Mitigations - Key risks for exploitation (e.g., technical readiness, market alignment, IPR issues): no specific risk or issue detected - Mitigation actions planned: none 105 Grant Agreement 101139270 — ORIGAMI — HORIZON-JU-SNS-2023 Deliverable D5.3 9 APPENDIX II - INDIVIDUAL EXPLOITATION PLANS (ACADEMIC PARTNERS) II.1 UC3M Key Exploitable Results (KERs) Relevant to the Partner List the KERs that your academic institution contributes to or intends to exploit. KER # Title Partner's Role/Contribution 1 CloudRIC++ UC3M contributed to the design of the demonstration, including the machine learning solutions used by the algorithm. 2 CPU-optimized vRAN UC3M contributed to the design and the evaluation framework of the solution. Section A: Exploitation Activities so far 1. Research Advancements - Contributions to knowledge and methodology: The Key Exploitable Results (KERs) identified by UC3M within the framework of the ORIGAMI project represent significant advancements in the optimized operation of virtualized Radio Access Networks (vRAN). These contributions address Barrier #1 of the project and stand out for their methodological innovation and technical relevance. The work conducted in this area positions UC3M for high visibility through potential publications in leading venues within the mobile networking domain, thereby enhancing the university’s reputation for research excellence at national, European, and international levels. - Integration of KERs into research agenda: The development of KERs 1 and 2 will enable UC3M to deepen its expertise in several cutting-edge areas that go beyond traditional mobile networking, including deep learning and advanced software tools such as FlexRAN, DPDK, and NVIDIA Aerial. Acquiring proficiency in these technologies is expected to have a lasting impact on UC3M’s research agenda, as they are going to become important elements in the evolution of 6G networks and in adjacent fields such as cybersecurity. 2. Education & Training - Courses and Lectures Updated with ORIGAMI Content: The research activities carried out within the ORIGAMI project have contributed to the enhancement of several bachelor’s and master’s level courses at UC3M, particularly those focused on mobile and wireless networking. In addition, the most advanced findings and methodologies developed in the project are being integrated into the training of PhD students through dedicated seminars, thereby strengthening the academic curriculum and fostering knowledge transfer at all levels of education. - MSc/PhD Theses Aligned with KERs: A new MSc student will begin working on topics related to the ORIGAMI project starting on July 1st, contributing to the project’s scientific and technological objectives. 106 Grant Agreement 101139270 — ORIGAMI — HORIZON-JU-SNS-2023 Deliverable D5.3 3. Publications and Academic Dissemination - Peer-reviewed publications and conference presentations: The research activities conducted under the ORIGAMI project have substantially enriched UC3M’s portfolio of scientific publications, with most results being disseminated through leading international conferences and high-impact journals. These contributions are expected to enhance the visibility and reputation of the UC3M researchers involved, both within the mobile networking community and across related scientific domains. - Collaborative academic workshops or panels: Participation in the ORIGAMI project and related SNSJU initiatives has provided UC3M researchers with valuable opportunities to engage in panels and workshops organized by e.g. the 6G Architecture Working. - Public repositories and licenses: N/A Section B: Exploitation Plans 1. Planned Research Activities - Continuation or expansion of current research efforts: The research activities associated with the KERs, as well as the ongoing architectural design work, are expected to continue and intensify during the second half of the ORIGAMI project. These efforts will be key to finalizing the technical developments and achieving the overarching objectives of the project. - New interdisciplinary or collaborative proposals: The work carried out within the ORIGAMI project strategically positions UC3M to participate in new interdisciplinary and collaborative research proposals. These initiatives will help extend the impact of ORIGAMI beyond the project’s duration, fostering long-term research synergies and broadening UC3M’s engagement in future European and international research programs. 2. Planned Education & Outreach - New or updated curricula: Insights and technological advancements from the ORIGAMI project are being leveraged to update existing curricula and develop new academic content at UC3M, particularly in areas such as mobile networks and artificial intelligence. Moreover, the project's outcomes have contributed to the design of new academic programs, including emerging degrees in Quantum Technologies and Artificial Intelligence. - Planned student engagement (projects, internships): UC3M will proactively seek to involve undergraduate and master’s students in the activities of the ORIGAMI project, offering them handson research experience and the opportunity to contribute to research in mobile networking and related fields. 3. Planned Scientific Dissemination - Target journals and conferences: UC3M will continue to pursue high-impact dissemination of ORIGAMI-related research by targeting top-tier venues such as ACM MobiCom, IEEE INFOCOM, and the IEEE Journal on Selected Areas in Communications (JSAC). These venues will ensure visibility and recognition of the project's scientific contributions within the global research community. - Joint publications with project partners: UC3M will continue to regularly publish with ORIGAMI partners such as NEC, IMDEA, TID and pursue new collaborations such as the one with EMN and JMU. 107 Grant Agreement 101139270 — ORIGAMI — HORIZON-JU-SNS-2023 Deliverable D5.3 4. Risks & Mitigations - Risks specific to academic exploitation (e.g., funding continuity, publication delays): While publication delays are an inherent risk in academic dissemination, UC3M’s prior experience with publishing results from the ORIGAMI project indicates a strong likelihood of acceptance in top-tier venues. This track record mitigates concerns related to visibility and reinforces confidence in the academic impact of ongoing and future findings. - Mitigation actions planned: UC3M will continue to prioritize early identification of high-impact results, ensuring timely preparation and submission to relevant conferences and journals. II.2 TUD Key Exploitable Results (KERs) Relevant to the Partner List the KERs that your academic institution contributes to or intends to exploit. KER # Title Partner's Role/Contribution 1 GPU-assisted vRAN TUD collaborates with NEC on developing a Network Intelligent (NI) solution that uses GPUs to multiplex 5G PHY and machine learning workloads, improving cost-efficiency and hardware utilization through ORIGAMI’s Compute Continuum Layer (CCL). Section A: Exploitation Activities so far 1. Research Advancements - Contributions to knowledge and methodology: TUD has contributed to advancing the state of the art in resource allocation and fairness mechanisms for virtualized RAN systems. In collaboration with NEC, TUD developed and evaluated algorithms supporting sustainable and energy-efficient RAN virtualization. The joint publication Fair Resource Allocation in Virtualized O-RAN Platforms (ACM Sigmetrics, with NEC) demonstrates the applicability of dynamic fairness mechanisms in managing base station workloads. Further contributions include Through the Telco Lens: A Countrywide Empirical Study of Cellular Handovers (ACM IMC, with TID), which provides valuable empirical insights into handover dynamics across mobile networks. Overall, TUD published eleven (11) conference and journal papers (including four A* publications), with two papers directly related to ORIGAMI use cases (FLB and CFA) and their evaluation. - Integration of KERs into research agenda: The GPU-assisted vRAN KER has been incorporated into TUD’s research activities on network intelligence and compute resource optimization. The team’s ongoing work contributes to the development of sustainable and flexible architectures for Open RAN systems, aligning with ORIGAMI’s goals of efficiency and interoperability. 108 Grant Agreement 101139270 — ORIGAMI — HORIZON-JU-SNS-2023 Deliverable D5.3 2. Education & Training - Courses or lectures updated with ORIGAMI content: ORIGAMI-related material has been incorporated into PhD-level seminars at TUD, particularly in topics related to communication networks. - MSc/PhD theses aligned with KERs: Doctoral students at TUD have engaged in thesis work that builds upon ORIGAMI-related research outcomes, especially in areas such as network optimization and resource allocation. 3. Publications and Academic Dissemination - Peer-reviewed publications and conference presentations: Fair Resource Allocation in Virtualized O-RAN Platforms, ACM Sigmetrics (Best Student Paper, with NEC, i2CAT) Through the Telco Lens: A Countrywide Empirical Study of Cellular Handovers, ACM IMC (with TID) Adaptive Online Non-stochastic Control, L4DC Optimistic Online Non-Stochastic Control via FTRL, IEEE CDC Adaptive Resource Allocation for Virtualized Base Stations in O-RAN with Online Learning, IEEE Trans. On Communications On The Dynamic Regret of FTRL: Optimism with History Pruning, ICML Smooth Handovers via Smoothed Online Learning, IEEE INFOCOM (with TID) FairRIC: Real-time Fair Allocation in O-RAN with Shared Computing, IEEE INFOCOM (with NEC, i2CAT) CHOMET: Conditional Handovers via Meta-Learning, IEEE WiOPT Minimization of the Training Makespan in Hybrid Federated Split Learning, IEEE Trans. On Mobile Computing Cooperative Streaming Inferences in IoT Networks, IEEE GLOBECOM - Collaborative academic workshops or panels: TUD participated in joint dissemination efforts through workshops and presentations, including contributions to PhD schools and technical seminars featuring material derived from ORIGAMI research. - Public repositories and licenses: N/A Section B: Exploitation Plans 1. Planned Research Activities - Continuation or expansion of current research efforts: TUD will continue developing and validating the GPU-assisted vRAN solution, focusing on optimizing fairness and reliability in compute resource management. Planned work includes extending the evaluation of GPU multiplexing mechanisms 109 Grant Agreement 101139270 — ORIGAMI — HORIZON-JU-SNS-2023 Deliverable D5.3 between 5G and ML workloads and integrating them within the ORIGAMI Compute Continuum Layer (CCL). - New interdisciplinary or collaborative proposals: TUD plans to expand collaborations with ORIGAMI partners, including NEC and TID, to pursue new research directions at the intersection of AI, network management, and energy-efficient communication systems. 2. Planned Education & Outreach - New or updated curricula: Results from ORIGAMI will continue to be incorporated into postgraduate education, including seminars on wireless networking and intelligent systems. - Planned student engagement (projects, internships): MSc and PhD students will be encouraged to contribute to ongoing research efforts related to RAN optimization, GPU scheduling, and network intelligence through projects and internships within ORIGAMI and related initiatives. 3. Planned Scientific Dissemination - Target journals and conferences: TUD will target high-impact venues such as IEEE INFOCOM, IEEE JSAC, and ACM MobiCom for dissemination of ORIGAMI-related results, including new findings on GPU resource optimization and sustainable RAN systems. - Joint publications with project partners: Future publications will include continued collaboration with ORIGAMI partners, following the successful joint work with NEC and TID. 4. Risks & Mitigations - Risks specific to academic exploitation (e.g., funding continuity, publication delays): Potential challenges include limited access to high-performance GPU infrastructure for experimentation and delays in integrating research outcomes into real-world validation environments. - Mitigation actions planned: TUD will leverage partner infrastructure, such as NEC’s experimental testbeds, and pursue early-stage validation using simulation environments. Publication planning will prioritize early preparation and submission to ensure timely dissemination of results. II.3 IMDEA Key Exploitable Results (KERs) Relevant to the Partner List the KERs that your academic institution contributes to or intends to exploit. KER # Title Partner's Role/Contribution 1 Distributed inference in programmable user planes IMDEA designed, implemented and demonstrated the KER 2 Streamlined deployment of user-plane intelligence IMDEA designed and is presently implementing the KER 110 Grant Agreement 101139270 — ORIGAMI — HORIZON-JU-SNS-2023 Deliverable D5.3 Section A: Exploitation Activities so far 1. Research Advancements - Contributions to knowledge and methodology: In the context of KER #1, IMDEA has developed a framework to decompose complex/large ML models into a set of simpler/smaller ones, which can then be deployed and chained across a programmable user-plane composed of multiple switch ASICs, FPGAs, SmartNICs, NPUs, etc. In the context of KER #2, IMDEA has developed a toolset to partially automate the training and test phases of ML models designed for user-plane deployment. Jointly, KER #1 and #2 make it possible to train and deploy the very first distributed implementations of ML model in programmable user planes for line-rate inference on data traffic. The solution presently focuses on programmable switch ASICs only and it is a first step towards a more complete framework that encompasses diverse programmable network equipment. - Integration of KERs into research agenda: IMDEA is committed to support the transfer of technology to the industrial sector to improve its capacity for innovation and competitiveness. The partner is also fostering the creation of spin-off companies to promote the release of new products and services to the global market. In this context, the developed KERs offers new opportunities to IMDEA researchers to develop novel Network Intelligence solutions to be transferred into products through the wide network of collaborations that IMDEA has established with the telecommunication industry. 2. Education & Training - Courses or lectures updated with ORIGAMI content: Personnel if IMDEA involved in ORIGAMI has presented material from KER #1 in various invited lectures and seminars, including at the First European Mobile Systems Winter School (Como, Italy, Feb 2025), the IEEE SPS 1st Winter School on AI for 6G Communications (Las Palmas de Gran Canaria, Spain, Feb 2025), the Comitê Técnico de Monitoramento de Redes (Rede brasileira para educação e pesquisa, Sep 2024), the University of Porto (Porto, Portugal, Jun 2024), or the MedComNet conference (Nice, France, Jun 2024). - MSc/PhD theses aligned with KERs: During the first half of the project, IMDEA recruited and started the training of one PhD student who actively worked on the design and development of the innovative NI solution above, hence creating new expertise at the interface of machine learning and mobile networking that will be critical to the European industry. Network programmability is indeed poised to become a primary trend in the telco space in the coming years, as soon as the softwarization and cloudification of the network infrastructures is completed. By training the PhD student, IMDEA contributed to the creation of domain experts with a profound knowledge of P4 coding on programmable switch platforms. 3. Publications and Academic Dissemination - Peer-reviewed publications and conference presentations: Descriptions of the design, implementation and evaluation of the solution for KER #1 have been published at prestigious and highly selective venues such as IEEE INFOCOM and IEEE/ACM Transactions on Networking. It is to be remarked that the solution also won the Best Paper Award at IEEE INFOCOM 2025. - Collaborative academic workshops or panels: A complete demo of the solution for KER #1 has been developed that allows observing the performance of the ORIGAMI-proposed solution with a real- 111 Grant Agreement 101139270 — ORIGAMI — HORIZON-JU-SNS-2023 Deliverable D5.3 world use case and comparing them against state-of-the-art approaches. The demo has been showcased at IEEE INFOCOM 2025 and EuCNC 2025. - Public repositories and licenses: The current version of the solution for KER #1 is available at the public repository https://github.com/nds-group/DUNE. Section B: Exploitation Plans 1. Planned Research Activities - Continuation or expansion of current research efforts: While it has exploitation potential, the current version of the NI solution above is still far from a state where it could transition to technology transfer. During the second half of the project, IMDEA plans to enhance substantially the NI solution for distributed line-rate user-plane inference. The main directions for improvement include (i) extensions to diverse programmable network hardware that feature heterogeneous compute and memory capabilities, (ii) extension to very large scenarios (e.g., datacenter topologies) via emulation, and (iii) improvements in the user-to-control-plane communication. - New interdisciplinary or collaborative proposals: Once the solution is mature enough, IMDEA will take advantage of its capacity of member of the Intel Connectivity Research Program and of the Intel Connectivity Academy to ensure that it is communicated and highly visible at Intel, which is a leading company in the field. IMDEA will also exploit its preferential links to the Intel Connectivity Academy to disseminate the project vision and solutions across the community around network programmability that such an Intel initiative gathers. This includes interacting with the many researchers and industry partners via the forums provided by the initiative and making sure that all relevant ORIGAMI results are present on and advertised by the official Intel Connectivity Academy website and P4 social media. During the remainder of the ORIGAMI lifetime, IMDEA will also explore opportunities to leverage the existing 5TONIC platform, jointly created with Telefonica, which involves the leading European industry players in 5G, for demonstration purposes. This includes investigating the possibility to integrate any ORIGAMI-developed technology in the 5TONIC platform, within one or more of the many experimental setups deployed in the laboratory. 2. Planned Education & Outreach - New or updated curricula: The updated content from the development of KER #1 and KER #2 will keep feeding material presented in seminars and lectures during the second half of the project. - Planned student engagement (projects, internships): The PhD student will keep working on KER #1 and KER #2 and towards his PhD thesis. IMDEA will consider hosting interns and graduate students to support the research activities linked with these KERs: in fact, one visit from a Master student from École Normale Supérieure de Lyon in France is already planned for the summer 2025. 3. Planned Scientific Dissemination - Target journals and conferences: IMDEA will submit scientific papers describing the enhancements to the solution to major international peer-reviewed venues, such as IEEE INFOCOM, ACM CoNEXT, and IEEE/ACM Transactions journals to ensure that the principles behind the NI solutions nurture further research in the academic and industrial research community. 112 Grant Agreement 101139270 — ORIGAMI — HORIZON-JU-SNS-2023 Deliverable D5.3 - Joint publications with project partners: No publication with project partners is planned on activities related to KER #1 or KER #2. However, IMDEA will gladly explore such opportunities in the second half of the project execution. Risks & Mitigations - Risks specific to academic exploitation (e.g., funding continuity, publication delays): Risks like publication delays from peer-review are unavoidable in academic research and exploitation. No further risks are identified as of now. - Mitigation actions planned: IMDEA will mitigate the risks above by primarily adopting a strategy of high-quality research that produces strong and technically sound papers with higher chances of begin accepted for publication at top venues. ΙΙ.4 I2CAT Key Exploitable Results (KERs) Relevant to the Partner List the KERs that your academic institution contributes to or intends to exploit. KER # Title Partner's Role/Contribution 1 CloudRIC++ I2CAT contributed to the design and implementation of the demonstrator, focusing on the adaptation abstraction layer provided by DPDK-BBDev to the novel mechanisms. 2 CPU-Optimized vRAN Exploration of different approaches for energy-efficient data processing within virtualized RANs. I2CAT contributed to the design of the evaluation framework as well as to the exploration of the different alternatives. Section A: Exploitation Activities so far 1. Research Advancements - Contributions to knowledge and methodology: i2CAT has made significant methodological and technical contributions to progress in enhancing the efficiency of virtualized RANs through the design of an abstraction layer. The efforts directly address both Barriers #1 and #2 of the project, laying the groundwork for more scalable, flexible and efficient vRAN architectures. - Integration of KERs into research agenda: The identified KERs will strengthen i2CAT’s expertise in critical architectural components and technologies directly related to RAN virtualisation (e.g. FlexRAN, Sionna, DPDK/BBDev or Nvidia AERIAL), supporting the advancement of its research agenda. These developments are expected to attract researchers, foster engagement with global industry stakeholders, and enhance i2Cat’s visibility and positioning within the international research community. 113 Grant Agreement 101139270 — ORIGAMI — HORIZON-JU-SNS-2023 Deliverable D5.3 2. Education & Training - Courses or lectures updated with ORIGAMI content: Leveraging i2CAT’s close collaboration with academic institutions, the research conducted within ORIGAMI project is being actively transferred to educational courses. For example, some of the outcomes have been integrated into university lectures on mobile wireless communications, enriching both Bachelor’s and Master’s content toward improving the background of the students related to RAN virtualization architectures and techniques. Additionally, the research findings are also used to enrich specialized training sessions dedicated to PhD students, fostering early-stage researcher engagement and ensuring alignment between cuttingedge research and higher education. - MSc/PhD theses aligned with KERs: N/A 3. Publications and Academic Dissemination - Peer-reviewed publications and conference presentations: As part of its exploitation strategy within ORIGAMI project, i2CAT is committed to publishing key research outcomes in top-tier, peer-reviewed journals and presenting findings at leading international conferences. These efforts aim to maximize the scientific visibility of the project results in the areas of beyond 5G, virtualized RANs and open innovation ecosystems. By targeting high-impact dissemination, i2CAT strengthens its role as a reference center for advanced research in mobile networking, while fostering future collaborations with academic and industrial stakeholders. - Collaborative academic workshops or panels: i2CAT actively participate in and co-organize academic workshops and expert panels focused on the technological domains addressed by ORIGAMI. These events serve as a platform for engaging with researchers, students and industry representatives for knowledge exchange and discussions on emerging topics. - Public repositories and licenses: N/A Section B: Exploitation Plans 1. Planned Research Activities - Continuation or expansion of current research efforts: During the second half of the project, i2CAT plans to finalize the technical developments performed toward KER #1 and continue working on KER #2, exploring new techniques and advances that may improve the performance in terms of energy efficiency of CPU-Optimized vRANs. - New interdisciplinary or collaborative proposals: Building on the advancements made in KERs #1 and #2, i2CAT is exploring new collaborative research opportunities intersecting networking, energy efficiency and hardware-aware optimization. These efforts aim to bring together the expertise from multiple disciplines to address the challenges of energy-efficient CPU-optimized vRANs. I2CAT is initiating discussions with academic and industrial partners toward the development of joint project proposals targeting both national and European funding calls. 2. Planned Education & Outreach - New or updated curricula: The research outcomes during the second half of the project will continue to be used at PhD-student oriented seminars and academic material within the cellular networks field.