ONE Summit 2025 in London Summary

ONE Summit 2025 in London Summary

ONE Summit: seeOpen Networking & Edge Summit 2025 Schedule & Directory

  • AI-RAN by Alex Jinsung Choi, AI-RAN Alliance, SoftBank Corp. : https://static.sched.com/hosted_files/onesummit2025/87/Linux Foundation Open Networking and Edge Summit 2025_Alex Jinsung Choi.pdf?_gl=1*15q9wq1*_gcl_au*MzI4OTA3MTU1LjE3NDM5NzY0MDk.*FPAU*MTQ4NTA4MTMxMy4xNzM2NTE1MDA0

    • AI-RAN and Open Source: Building Scalable Dataset Pipelines for AI Training

    • Accelerating AI development for RAN optimization and automation

    • The evolution of Telecom networks into AI platform

  • AI - Lessons learnt from CMCC by Dr. Junlan Feng

    • Challenges understanding and hallucination collaboration vs. AI ton enable large scale complex system

    • Evolve to level 4 advanced autonomous network by 2025

    • Build AI platforms to address network complexity, shorten innovation cycles, and lower R&D barriers

    • Unified portal; automated closed loop; from invisible to visible ; world model aided RAN parameter adjustment and evaluation

    • Inference aided handover parameter, adaptive multi-receptive field spatial temporal graph network for traffic forecasting;

    • Injecting global information into independent channels for long time series forecasting; graph-based decomposition approach for cellular traffic prediction; human trajectories generation based on variational point processes

    • New model of open collaboration for AI+Network

  • OMDIA: an industry analyst’s perspective on Telco Cloud

    • Network functions becoming cloud native

    • Moving from ETSI NFV to GitOps-based frameworks

    • Cloud-native as enabler for autonomous networks

    • Moving to a horizontal cloud for IT and network functions

  • EdgeAI by STL Partner

    • $100 billion opportunity; the edge AI leverages edge computing for certain AI workloads

  • Network Infrastructure to APIs - how Open Source enables speed of invocation and monetization by DT

  • InstantX: Vodafone’s edge innovation for the future of Telco

    • Safer transport for Europe Platform (STEP) - Vodafones enterprise V2X (vehicle to everything) service

    • InstantX - Vodafones projects with LF Edge - ensuring the secure and reliable exchange and distribution of data in real-time between users in a certain geography

  • IOWN Global Forum and Open Source - Key enablers for next decade by IOWN Technology Director

    • IOWN global forum’s areas of activity

    • an industry consortium established in 2020 by NTT, Intel and Sony

    • Open all-photonic network (APN); Digital Twin computing; Cognitive Foundation, integrating AI and ML for network

  • AI Framework for Networking and Responsible AI by Infosys

    • Why AI for Networking and a Framework? ; roadmap towards digital brain (plan & design slice, testing slice, assurance slice) thru agents;

    • Why responsible AI? ; Enhanced security, data privacy, explainable AI, Fairness/Bias mitigation, Versatility)

    • E2E framework approach for AI in networks

  • Open Networking Operation System for all, by Peter Plak from Dell Technologies

    • Dell networking use cases including ISG, DC to Edge, merging AI fabrics

    • SONiC use cases in Microsoft; Kubernetes based SONiC Device Management; AI Traffic characteristics and solutions

  • AI 101 for Networking and Edge by Fatih Nar - RedHat, Ranny Haiby - LF

    • AI adoption for networking and edge computing is accelerating

    • Enterprise AI use; Real world AI consumption Example, DeepSeek Impact using MoE (neural network design where only a subset of model components (called experts) are activated for a given input

    • Agentic life; deployment architectures (Centralized vs. Distributed, Hybrid with workload assessment & placement)

    • Security is not an external component; Telco-AIX; Ent-ObX (Observability experiments); Domain Specific AI

  • Panel discussions: Spotlight on innovation: the latest developments shaping open source networking (Lingli Deng - CMCC, Olaf Renner - Nokia, Ciaran Johnston - Ericsson, Ranny Haiby - LF)

  • A journey towards automated Telco Cloud by Odline Labs, Turkcell

    • their Telco Cloud Automation architecture, including ONAP

    • Why do they choose ONAP? ; Alignment with ETSI MANO; Open-Source, vendor-neutral benefits, Active community and roadmap, new modular architecture - ONAP streamlining, ONAP’s capabilities on orchestration, policy, closed-loop automation, synergy with other LF projects and AI roadmap

    • Architecture:

    •  

    • Expectations from the ONAP project and community

      • better CNF lifecycle support (helm, k8s-native orchestration, simplified onboarding workflows and model tooling, enhanced modularity and testability

      • More hands-on tutorials and quick-start guides, alignment with ETSI MANO evolution, Foster feedback channels between users and developers, promote interoperability across ecosystem tools

  • LF - Horizontal and vertical foundations, panel discussion Dr Junlan Feng from CMCC, Johnne Sominent from Nokia, christian Olrog from Ericsson, Arpit Joshipura from LF -

    • relationships (manufacturing, energy, gas, commerce & retail, automotive, Fleet & transportation, building automation, cities & government

    • LFN, collaboration hub for opensource networking, edge and access

    • Pushing local innovation with global collaboration - all built off open standards and open source

    • APAC & China: Networking contribution / leadership

    • India strategic drivers - LF India launched in Dec 2024 to enable India’s growing open source ecosystem across telecom, blockchain, edge, security, cloud

    • EU Open Source Networking Drivers - DT deploys ONAP in O-RAN TOWN, CAMARA, 6G SNS, Sylva

    • Technology 2025 - reality check; cloud-native is in over 2/3 networks/product deployment, domain specific AI focus on AI frameworks and data sharing (2 new project in LFN  Essedum, LF Salus)

    • Edge - Cloud continuum & AI+data at Edge are two main focus areas

    • Network monetization with CAMARA APIs reaching peak awareness and early adoption