Plenary Sessions Auditorium Bordeaux
Chairman
Roy Chua, Founder & Principal, AvidThink
In overture of the Congress, The Global IP Assembly gathers some of the most renowned experts from Telecom Service Providers
They discuss aspects of what they see for IP going forward
- API-fication, disaggregation, programmability
- Increasing value and decreasing costs: what strategies are needed, what technologies can help (AIOps, virtualization, white box). Does SRv6 help?
- Big, bigger, biggest data: impact of AI training, AI inference, large multimodal dataset movement, hybrid clouds
- Managing complexity: multi domains, multi layers, multi technologies — how to manage, troubleshoot?
- Quantum and classical networks: interplay and considerations - is it too early or is the time now?
Zhenbin Li, Chief IP Standard Representative, Huawei Technologies
Zhenbin Li is in charge of the research and standard promotion work of IP protocols. Before that he had been the Chief IP/MPLS Service Architect and the Chief Architect of SDN controller to be responsible for the service and architecture design of IP software platform and SDN controller. Zhenbin Li and his team are taking an active part in the IP/MPLS/SDN standardization work in IETF and promote the innovation of SR, BGP/VPN, PCE, SBI/NBI, Telemetry, etc.
Jürgen Hatheier, International CTO, Ciena
Kireeti Kompella, Juniper Networks
Michael Beesley, CTO SP Networking, Cisco
Michael is Vice President and CTO of the Cisco Networking Group leading technical direction, architectural transition, solution architecture and roadmap planning for Cisco’s SP networking portfolio. He drives Cisco’s leadership position in core, edge, aggregation and access routing, mobility, cable, optics and optical networking and automation software solutions in collaboration with partners, customers, internal cross-functional leaders and the industry in reinventing the mass-scale networks of the future. Prior to rejoining Cisco in February 2018, Michael was co-founder and CTO of Skyport Systems, developing cloud managed, on premise, virtualized compute platforms with security, ease of use, and always-on analytics built in. Before Skyport, Michael served as engineering CTO at Juniper Networks, leading new architectures, embedded software evolution and silicon roadmap development across the router, switch and Wi-Fi portfolios. Michael has held many senior technical leadership roles at both Cisco and Juniper and was a key contributor to many successful product lines and technology including C7200, C12000, M40, T640, JUNOS, ASR1K and IOS-XE. Michael holds CS and Math degrees from Trinity College, Dublin, is a named inventor on many networking related patents, is a frequent speaker at industry events, and often shares his thoughts through Twitter, LinkedIn and blogs.
Wim Henderickx, Head of Technology and IP Architecture, Nokia
Reviewing the role of networking for AI infrastructures and showing case industry examples related to this critical area of evolution.
Michael Ploeg, PLM Manager Data Center, Nokia
Networking enterprises are not only leveraging AI technologies in their products, for the benefit of their customers, but are also transforming their own internal processes. Describing a first-hand experience with adopting AI for software engineering – including its potential and pitfalls.
John McKinnon, Vice President, Software Engineering R&D, Ciena
The rapid growth of Artificial Intelligence (AI) and Machine Learning (ML) workloads is driving the demand for low latency, low loss, and resilient network fabrics to connect the GPU servers. Global Load Balancing (GLB) for AI/ML fabrics can help to avoid congestions end-to-end which cannot be achieved by traditional load balancing techniques such as random hashing, dynamic load balancing, etc.
Kevin Wang, Distinguished Engineer, Juniper Networks
Providing an overview of the different types of ethernet-based networks for AI and their respective challenges. Evaluating how existing ethernet/IP principals and evolution are used. Covering how SONiC plays important role in AL backend and frontend network.
Jiri Chaloupka, Principal Engineer, Cisco
Sharing experiences of building a Gen AI based assistant for VoLTE and VoNR operations. Discussing the closed loop automation models and their role with predictive and Gen AI to build autonomous operations of mobile voice.
Azhar Sayeed, CTO, ng‑voice
GenAI powered autonomous agents are the key building block. Explaining what they are, how they can be used, real use cases, and how they truly enable a solid path towards the Autonomous Network implementation.
Javier Antich, Product Manager, Cisco
Describing a high-throughput and lossless high-elastic intelligent computing network for consumer services and enterprise AI applications.
Wu Qin, Network Architect, Huawei Data Comm
Carsten Rossenhoevel, Managing Director, EANTC
Exploring an innovative, real-world IP/MPLS network digital twin use case that empowers service providers to improve network planning speed and efficiency by enabling “what-if” scenarios to determine how various changes would impact network performance.
Bill Kaufmann, Director of Product Management, Assurance and Analytics, Blue Planet
As a leading Canadian telecom provider, Telus is transforming the network and control software to be self-healing, self-configuring, self-organizing, and self-optimizing, ultimately achieving autonomy.
Ali Tizghadam, TELUS Fellow
Examining how the combination of network digital twins and large-scale AI models can enhance network management. By making it more adaptive and intelligent, this approach facilitates higher-level decision-making compared to traditional network modeling methods.
Speaker from Huawei Technologies
Moderator
Azhar Sayeed, CTO, ng‑voice
Morning Chairman
Carsten Rossenhoevel, Managing Director, EANTC
Limitations of IPv4 addressing and how IPv6 lays a strong foundation for a more simpler network operation. Practical aspects of SRv6 deployment, examining the different paths of greenfield deployments and complex brownfield environments.
Juan Rodriguez, Senior Sales Engineer, DriveNets
IPv6 has a lower average delay than IPv4, and the difference is growing from the end user's point of view. This IPv6 modification is historic. An increasingly significant benefit for businesses that prioritize user experience is IPv6 deployment.
Li Cheng, Senior IP Standards Representative, Huawei Technologies
IX.br is making significant progress in evolving its network infrastructure by incorporating and deploying advanced technologies such as EVPN (Ethernet VPN) and SRv6 (Segment Routing over IPv6).
Diego Achaval, Product Line Manager, Nokia
Fabio Pessoa Nunes, Network Architect, IX.br
MPLS to SRv6 migration process can be very challenging. For example, equipment vendors typically provide SRv6 support for IS-IS only, which implies additional IGP migration, when existing IGP is OSPF. Another challenge is migration of Layer 2 services, like E-Line or E-LAN, implemented using legacy Layer 2 technologies, like LDP based Layer 2 circuits or VPLS (Virtual Private LAN Service). Again, equipment vendors typically provide SRv6 support for Layer 2 services based on EVPN (Ethernet Virtual Private Network).
Krzysztof Grzegorz Szarkowicz, AWAN PLM, Solutions Architect, Juniper Networks
As Service Provider network infrastructure evolution towards unified Segment Routing transport continues worldwide, is there a common deployment model? Exploring three models: centralized with SR-TE, distributed with Flex-Algo and a hybrid approach. Practical SR-MPLS / SRv6 design tradeoffs will also be discussed, with an emphasis on learnings from worldwide deployments.
Jan Straznicky, Senior Director, Product Line Management, Ciena
Describing the overall framework defined in the IETF to cover all aspects related to the data plane, control plane, and data analysis, and providing guidance for implementation and deployment. It will work with Swisscom to introduce the integration of IFIT technology with network analysis tools.
Jin Minwei, Chief Expert of Metro Router Solution, Huawei Technologies
- AI training backends with SRv6 deterministic path placement
- Enhancing AI inter-clusqter connectivity
- End-to-end SRv6 from cloud front-end to egress internet peering
- IPM hardware connectivity
- From hardware to NaaS
- TITAN Swisscom network design
- Circuit-style services for deterministic networking
- SRv6 MUP progress and update
Clarence Filsfils, Cisco Fellow
Akash Agrawal, Technical Director, Rakuten
Michael Valentine, Technology Fellow, Network Architecture, Goldman Sachs
Satoru Matsushima, Technical Meister, SoftBank
Bart Janssens, Senior Specialist Packet Architecture, Colt Technology
Gyan Mishra, Associate Fellow, Verizon
Afternoon Chairman
Roland Thienpont, Director IP Division Product Marketing, Nokia
Discussing simplified network operations/scalability/enhanced network Efficiency/cost efficiency/service innovation/mobility and 5G benefits (specific to MNOs)/high availability and resilience.
Latif Ladid, President, IPv6 Forum
Covering the most recent advancements in IPv6 enhancement technologies and their impact on various AI network application scenarios. Introducing the related new innovations.
Xiao Xipeng, Head of Datacom Standard & Industry Development (SID), Huawei Europe
Focusing on the implementation of IPv6 using Mapping of Address and Port with Translation (MAP-T) solutions, highlighting real-world deployment experiences from tier-1 CSPs that have successfully navigated this transition. MAP-T enables CSPs to extend their IPv4 services while paving the way for IPv6 integration, offering a hybrid approach that balances immediate operational needs with future requirements.
Juan P Rodriguez, Sr Director, IP Consulting Engineering, IP Networks, Nokia
Exploring real-world experience and use cases for seamless integration of IP routing with coherent DWDM, delivering higher performance with operational simplicity — including lessons learned and how to prepare for 800G coherent optics and beyond.
Rafael Francis, Senior Director, Product Line Management, Ciena
Covering the standards work related to managing IP+Optical networks with ZR/ZR+ digital coherent optics and delivering real world examples illustrating provisioning and assurance of end-to-end multi-layer services utilizing DCO using open and standards-based methods.
Phil Bedard, Distinguished TME, Cisco
Starting with an overview of today’s technology and its evolution path. Then discussing the practicalities and some relevant deployments, and the lessons learned.
Bruno de Troch, Director of EMEA PLM IP Routing, Nokia
Exploring solution innovations, including novel system design to enable full density of coherent pluggables. Presenting field experience of Service Providers and Enterprises deploying IPoDWDM with 400G-ZR/ZR+ and for the first time 800G-ZR/ZR+ pluggables.
Moran Roth, Director, Product Management, Juniper Networks
Detainling our experience in overcoming varied OpenConfig YANG router implementations to achieve successful end-to-end wavelength management at the optical layer pluggable-to-pluggable, simplifying converged network operations.
Robert Friskney, Director, Product Line Management, Ciena
Moderator
Roy Chua, Founder & Principal, AvidThink
Chairman
Amir Zmora, CEO & Co-founder, flexiWAN
Introducing a new router on the network is a challenge in a sustainable network. Exploring the complexity of the network design when taking into account the sustainability criterias in the choice of the router.
Etienne Roux, IP Network Engineer, Orange
Examining the state-of-art in terms of reducing power consumption of individual components within a router, and powering down unused components automatically. We then take a network-wide viewpoint and describe advanced traffic-engineering techniques to route traffic in a power-efficient manner across the network.
Julian Lucek, Senior Distinguished Systems Engineer, Juniper Networks
What is meant by Hyperscale routing? What attributes make a router a good fit for this space? Is there a role for optical switching? This presentation explores the use routing and optical switching for AI traffic patterns.
Gary Swinkels, Distinguished Architect, Ciena
FIB in the data plane needs to meet performance, scale and work seamlessly during changes. Performance is not easy to achieve with simple Table lookups due to memory bandwidth limits. Solving this by using Bloom Filter, achieveing FIB scale and resolving race conditions in the data plane during route changes.
Chandrasekaran Venkatraman, Distinguished Engineer, Juniper Networks
Demonstrating how the new architecture for the KPN infrastructure, in which functionality are distributed as much as possible to the edge (but not yet over), and use SRv6 as the new underlay, allow to scale and distribute functions further out. Even over the edge?
Eduard Metz, Network Architect, KPN
- The impact of AI edge applications on IP network evolution and the emerging market opportunities
- The role and requirements of IP networks in delivering the value of AI applications to end users
- The IP technologies and system requirements essential for enabling AI edge inferencing deployments
Arnold Jansen, Senior Marketing Manager, Nokia
Introducing the standards and industry advances for Compute-Aware Traffic Steering (CATS). SRv6 can be used as a typical tunneling technology CATS, providing service providers with better load balancing across multiple edge clouds.
Giuseppe Fioccola, Senior Standardization Manager, Huawei Technologies
Edge cloud empowers a Communication Services Provider to expand their service portfolio, unveiling innovative approaches that enable value added services for subscribers and improve their quality of experience.
Kashif Islam, Principal Telco Architect, Red Hat
Syed Hassan, Principal Telco Architect, Red Hat
Discussing challenges such as multi-vendor networks and current solutions that use newer technologies, router implementations and a Software layer-based approach that enables Software defined networking. Advances in network router programmability including OpenConfig based interfaces, simplified path control using segment routing and the SDN Controller layer to abstract the network are covered.
Krishnan Thirukonda, Principal Engineer, Cisco
Complex multi-layer networks can benefit from the application of Generative AI (GenAI) and multi-agent Large Language Models (LLMs) to automate the detection, analysis, and resolution of issues, for improved reliability and operational efficiency. Multi-agent systems facilitate collaborative problem-solving, whereby autonomous agents interact and share insights to pinpoint and address faults across both IP and optical layers.
Reza Rokui, Director, Product Line Management, Ciena
A new breed of services (e.g. VR, AR) have emerged demanding more careful consideration of latency and jitter, as relevant parameters to ensure correct service delivery. This requires to define, measure and enforce latency as crucial network KPIs.
Luis Miguel Contreras, Technology Expert at Global CTIO Unit, Telefónica
The Network Performance Measurement (NPM) toolkit for accurate latency and loss measurements, connectivity verification as well as recording packet path, enable network operators to provide strict SLAs. The IETF standards-based toolkit using Simple Two-Way Active Performance Measurement (STAMP) and its extensions for IPv6 and MPLS data planes are essential for seamless performance monitoring.
Rakesh Gandhi, Principal Engineer, Cisco
Showing how a border node can use a context table to automatically filter all labels that have not been previously advertised to a specific neighbor via BGP-LU, thus creating an effective MPLS label anti-spoofing mechanism. Describing how context tables for MPLS labels help improve convergence times, and address scalability in larger networks.
Anton Elita, Technical Solutions Architect in Routing and Automation, EMEA, Juniper Networks
Offering a deep technical dive into the nature of modern DDoS attacks and discussing recent cyberattacks on commercial and government targets, especially from groups like Russia-affiliated NoName, highlighting the pressing need for adaptive DDoS defence strategies.
Jerome Meyer, Deepfield Security Researcher, Nokia
Some major mobile operators have been discussing a new user plane architecture that optimizes both the data plane and signaling. The architecture is also unified for both wireless and wireline networks, allowing for the seamless integration of wireline/wireless services. Discussing the user plane evolution and the new architecture.
Jeffrey Zhang, Distinguished Engineer, Juniper Networks
Demonstrating how network intelligence captures end-to-end user experience data per cell, application, and user level. Discussing various use cases like network performance monitoring, customer care, and CAPEX planning.
Waris Sagheer, Director, Product Management, Cisco
With the continuous development of 2B, 2C, and 2H services and the emergence of cloud and AI, new services pose new requirements on bearer networks. Describing the requirements, technical capabilities, and solutions related to the Net5.5G network.
Speaker from Huawei