GNA-G Community Meeting at TNC25

The Global Network Advancement Group (GNA-G) side meeting at TNC25 will feature updates to several working groups, including GREN security, resiliency, data-intensive science, and more, aiming to foster collaboration, discuss emerging trends, strengthen global intercontinental connectivity and services for scientific research and education.

Participants wishing to attend side meetings on Friday 13 June will be charged a daily fee of €30 ex. VAT. This pass must be purchased ahead of attending the meeting through the TNC website at tnc25.geant.org

If you are unable to attend this meeting in person, you can still join online via Zoom through the following link:
https://aarnet.zoom.us/j/85349924897?pwd=efigAwlw8mOobGFV7hd0j6UZTsgJlH.1

Agenda

• Welcome, agenda – Ivana Golub (PSNC)
• Leadership Team and Working Group updates – David Wilde (AARNet)
  • NomCom – Alexander van den Hil (SURF)
  • Network eAcademy – Maria Isabel Gandia Carriedo (CSUC)
  • GREN Resilience – Steve Maddocks (AArnet)
  • GREN Map – Gustavo Araujo (RNP)
  • Securing GRENs – Ana Alves (GÉANT)
  • Automation WG
• Research Engagement – David Wilde (AARNet)
  • Global Research Platform – Joe Mambretti (Northwestern University/Starlight)
  • SKA Regional Centre Networking update – Ian Collier (STFC UK)
• Break
• Network Advancement using AI – Ivana Golub (PSNC)
  • GÉANT’s SIG AI: Advancing AI for Network Operations and Services –  Leonie Schäfer (DFN), Daniela Brauner (GÉANT)
  • Network Troubleshooting with Generative AI – Karl Newell (Internet2)
  • ESnet Data and AI Workshop Readout – Chin Gouk (ESnet)
  • Implementing a secure and trusted AI Factory in an NREN environment – Bartosz Belter (PSNC)
  • AI-Driven Traffic Characterisation for Scientific Applications in Research and Education Networks – Daniela Brauner (GÉANT)
• Discussion & Close

Session Detail

Global Research Platform – Joe Mambretti (Northwestern University/Starlight)

The Global Research Platform (GRP) is a worldwide collaboration designing, implementing, and operating the largest, globally distributed, high-performance computing and communications environments for data-intensive science, including high-energy physics, astrophysics, bioinformatics, computational science, fusion energy, atmospheric and oceanographic science. GRP communities create advanced, innovative communication services embedded in high performance networks, e.g., 100/400/800 Gbps and Tbps. Participants collaborate to customize services, fabrics, and distributed cyberinfrastructure in support of optimal data-intensive scientific workflows and to facilitate high-performance data gathering, analytics, transport, computing, and storage among scientific research instruments and computing resources worldwide. The annual GRP Workshop (co-located with the IEEE International Conference on eScience) is a unique forum that enables participants to share architectural concepts, strategies, innovations and technologies, and to discuss progress, outcomes, and best practices.

SKA Regional Centre Networking update – Ian Collier (STFC UK)

In March 2025 the SKA Regional Centre Network completed the first round of global data transfer tests, with further functional tests due through 2025. We will report on these tests as well as the latest estimates and roadmap for experiment data rates and discussion of the impact for the global R&E networks.

GÉANT’s SIG AI: Advancing AI for Network Operations and Services –  Leonie Schäfer (DFN), Daniela Brauner (GÉANT)

Artificial Intelligence is poised to revolutionize network operations, security, and user services within the Research and Education Networking domain. This presentation will provide an overview of GÉANT’s new Special Interest Group on AI for NRENs, showcasing our work to identify key AI applications, share best practices, and drive innovation. We will discuss potential areas of collaboration with the GNA-G community, focusing on how AI can address shared challenges and advance next-generation network infrastructure globally.

Network Troubleshooting with Generative AI – Karl Newell (Internet2)

Internet2 will present our progress in leveraging generative AI technologies for network troubleshooting. Through active collaboration with the R&E community, we’re establishing a framework for the development and implementation of a tool-calling chatbot designed specifically for network analysis and diagnostic purposes.

ESnet Data and AI Workshop Readout – Chin Gouk (ESnet)

In February 2025, ESnet hosted a workshop focused on AI and data. The workshop aimed to pinpoint specific challenges within ESnet that could be effectively tackled using data-driven approaches. The goal of the workshop is to inform ESnet’s data analysis needs and influence its AI strategy, ultimately guiding the organization’s data management practices and the trajectory of its AI research and AIOps initiatives for ESnet7. This presentation will summarize the key takeaways and recommendations that emerged from the workshop.

Implementing a secure and trusted AI Factory in an NREN environment – Bartosz Belter (PSNC)

PSNC has been awarded a project to build and host an AI Factory as part of the EuroHPC JU initiative. To date, 13 AI Factories have been selected in Europe, building together an innovative EuroHPC infrastructure ecosystem accessible to startups and the wider innovation community. The talk will summarise the main assumptions and architectural choices for the Piast AI Factory and discuss the potential impact of implementing the AI Factories across Europe on the NREN community.

AI-Driven Traffic Characterisation for Scientific Applications in Research and Education Networks – Daniela Brauner (GÉANT)

Research and Education Networks (RENs) are supporting scientific applications with large-scale data transfers from several research communities, requiring consistent high-throughput and low-latency performance. Identify those scientific flows can help national and regional networks on the support for efficient network use, traffic steering, future provisioning and capacity planning. The use of machine learning for traffic characterisation can be an excellent experiment to test the use of AI in the field of networking. This presentation will explore the application of self-supervised graph neural networks (GNNs) and networking datasets on identifying anomalies and the efforts for science traffic characterisation.