ai-networks#
168 picked talks tagged ai-networks, newest conference first.
NANOG 98#
- A Blueprint for Securing Networks for AI : From Architecture to Operations
- AI Training Parallelism, From a Network Engineer's Point of View
- Multi-Path Traffic Engineering Workshop
- SRv6 and MRC for AI Data Centers
PyTorch Conference 2026#
- Beyond Collectives: Building Fused Compute-Communication Kernels with PyTorch Symmetric Memory
- Demo Theater: OpenCLIC: Open Communication for Distributed Inference
- EP-Overlap: Hiding Expert-Parallel Communication in MoE Training
- Five Bottlenecks, Five Fixes: Systematic Performance Diagnosis for PyTorch Distributed Training
- Future of Distributed Communication in PyTorch: New APIs for Fault Tolerance, RDMA and Extensibility
- Introducing NCCL Extensions: Communication Patterns for Modern AI
- MCCL: Fault-Tolerant Collective Communication for Large-Scale GPU Training
- Monarch - Single-Machine DevX at Distributed Scale
- Native Tiered KV Cache Offloading in vLLM: From Storage Offloading to Disaggregated Serving
- Precompile for Training
- rocSHMEM Symmetric Memory in PyTorch for AMD GPUs
- Scaling Foundation Models: From Broken to Near-Linear Multi-Node Training
- Self-Healing PyTorch Training Through Live GPU Migration
- State-of-the-Art KV Transfer for Disaggregated LLM Serving in vLLM
- XCCL: Scaling PyTorch Collectives to Exascale on Intel GPUs with TorchComms
SwiNOG #42#
Capacity Europe 2026#
- GrowthX Dragons Den Pitch 5
- Keynote Panel: Wiring the AI Boom: Building Fibre Corridors for Training, Inference and Scale
- Panel: From FLAP to Frontier: Rethinking Europe's footprint for AI
- Panel: Reclaiming the voice opportunity with voice AI
- Presentation: Data Center Interconnect for the AI Era: Enabling Distributed AI Infrastructure
- Presentation: New control points, new telco value in an AI world
OCP Global Summit 2026#
- Achieving AI Cluster Reliability at Scale: Lessons in Reducing Workload Interruptions
- Beyond Traditional BGP: Decoupling Route Distribution for Large-Scale AI Clusters at Alibaba
- Clock-Aligned Telemetry for AI Clusters: From BMC to GPU to Fleet
- Clustermapper: Deterministic path diagnosis for AI backend networks using SRv6
- Co-Designing AI Fabrics: Inside Google's Virgo Network and the Campus-as-a-Computer
- Enabling AI Networking @ Scale with Multi-path Reliable Connections (MRC)
- Large-Scale Deployment of 800Gb/s Active Electrical Cables (AEC) in Meta Data Centers for AI Applications
- Lessons learned from GW-scale AI deployments
- MetaRoCE: Rethinking RDMA for the AI presented by Meta
- Mitigating AI Training Power Oscillations: Utility Requirements and In-Rack Power Filtering
- Scale-across Networking: Panel discussion on routing and transport equipment implications
- Traffic Engineering in AI DCI/WAN for AI Traffic
OPI Summit on DPUs/IPUs 2026#
- Feeding the AI Factory: The Role of DPU/IPUs in AI Infrastructure - Tina Tsou, Board Chair, LF Edge; Rajalaxmi Angadi, Senior Engineering Manager, Intel & Moderated by Paul Pindell, F5 Networks
- KV Cache in the Agentic Era: The Five Forces that Decide Where the Cache Lives, and What a DPU Can Do About It - Jigar Kuverji Savla, Tensordyne
- Reimagining DPUs for Storage Control and End-to-End Confidential AI - Jitendra Singh & Anshuman Ghosh, Microsoft
Netnod Tech Meeting 2026#
AusNOG 2026#
@Scale: Networking 2026#
- Anatomy of @scale Machine Learning Network Operations
- Co-Designing Communication for AI Accelerators
- Fiber Networks Innovations for AI at Scale
- Live Panel: Scale-Out Networking for AI Supercomputers
- MetaRoCE: From Spec to NIC to Open Source
- MetaRoCE: Meta's RDMA Transport
- Resilient AI Supercomputer Networking using MRC and SRv6
- Virgo: scale-out data center network for AI
IETF 126 (Vienna)#
- Enhanced ECMP for AI Cluster
- Fully Adaptive Routing Ethernet in Multi-Plane Scale-Out Networks
- Fully Adaptive Routing Ethernet using BGP
- GSE architecture for AIDC
- PFC PAUSE Frame Forwarded Transparently in Wide Area Networks
- SRv6 behavior extention for Flow Control in WAN
- SRv6 for Deterministic Path Placement in AI Backends
Netdev 0x1A (Rome, 13-16 July 2026)#
- Networking Headless CXL Devices for AI Memory Services
- Performance Comparison of Transport Mechanisms for LLM Inference KVCache Transfers
- Resilient AI Supercomputer Networking using MRC and SRv6
- The CXL Fabric End-Game: Bandwidth Realities and Networked Memory for AI Scale
NetUK3 (The Brewery)#
AI Engineer World's Fair 2026#
@Scale: Systems & Reliability 2026#
AutoCon 5 (Munich)#
TNC26 (GÉANT research networking conference)#
Cisco Live US 2026 (Las Vegas)#
- Building AI/ML Fabrics at Hyperscale: Architecture Evolutions, Challenges, Open Standards, Performance Characterization and Validation on Cisco SiliconOne-based Platforms. - BRKMSI-2014
- Designing RoCEv2 Networks for Distributed AI Training with Cisco Nexus 9000 - BRKDCN-2688
- Real World Deployments: Architectural lessons learnt from a 100+ Node AI Factory Deployment - BRKXAR-2039
NANOG 97#
CHNUG #3#
LACNIC 45#
MLSys 2026#
- Beyond the Buzz: A Pragmatic Take on Inference Disaggregation
- Communication-Efficient Distributed Inference for Transformer Models via Vector Quantized Context
- Designing Communication-Efficient AI Systems: An Interconnect-Aware HPC Perspective
- DreamDDP: Accelerating Low-Bandwidth Geo-Distributed LLM Training with Layer-wise Partial Synchronization
- fabric-lib: RDMA Point-to-Point Communication for LLM Systems
- GUARD: SCALABLE STRAGGLER DETECTION AND NODE HEALTH MANAGEMENT FOR LARGE-SCALE TRAINING
- MLCommons Chakra: Advancing Performance Benchmarking and Co-design using Standardized Execution Traces
- NEST: Network- and Memory-Aware Device Placement for Distributed Deep Learning
- Optimizing Deployment Configurations for LLM Inference
- SHIP: SRAM-Based Huge Inference Pipelines for Fast LLM Serving
- Sparing Strategies to Minimize Reliability Impact On Large Training Jobs
- TokenWeave: Efficient Compute-Communication Overlap for Distributed LLM Inference
- Unleashing Scalable Context Parallelism for Foundation Models Pre-Training via FCP
RIPE 92 (Edinburgh)#
NSDI '26#
- BURST: Seeking High-performance, Interoperability and Scalability in Soft-RDMA
- From Bits to Tokens: Knowledge-Driven Generative Communication of Multimodal Data
- Matryoshka: Realizing Hyperscale Data Center Network Design for the AI Era
ITNOG10#
NZNOG 2026#
KubeCon + CloudNativeCon Europe 2026#
- KubeVirt on GB200: Virtualizing a Rack-Scale Supercomputer
- Making Topology-Aware Scheduling Practical for AI Workloads: From Discovery to Simulation at Scale
- SIG Network: The State of Networking for AI on Kubernetes
- Virtualizing Large Scale GPU Cluster for Sovereign AI: Petasus AI Cloud Journey with Kubernetes
IETF 125 (Shenzhen)#
- BGP Deterministic Path Forwarding (DPF)
- Fully Adaptive Routing Ethernet in Scale-Up Networks
- Fully Adaptive Routing Ethernet using BGP
- IP Fast Reroute for AI/ML Fabrics
- Requirements and Gap Analysis of Multicast in AI Data Centers
- Use cases and Requirement for Flow Control Collaboration Across DCNs and WAN
GTC 2026#
- Build a High-Performance Research Cluster
- Building Sovereign AI: Scaling 100B+ Model Training on NVIDIA Blackwell Infrastructure (Presented by Lablup, Inc.)
- How AI Factories Can Shift-left to Accelerate Infrastructure and Ecosystems Readiness
- Redefining Scalable AI Performance: OCI Supercomputing in the Cloud (Presented by Oracle)
- Robust LLM Training Infrastructure at ByteDance
- Scaling Out and Across: Networking Innovations for Giga-Scale AI Systems
- Supercharging AI With Multi-Gigawatt AI Factories
- Transforming Bottlenecks into Breakthroughs: Scaling Secure AI Networks for AI Factories (Presented by Cisco Systems)
Cisco Live Europe 2026 (Amsterdam)#
- Benchmarking AIML Infrastructure at Scale: AIML Architectures, AI Workloads and Reproducibility for High Performance Data Centers - BRKDCN-2651
- Demystifying AI Infrastructure for a Network Engineer - BRKDCN-2971
- How the World's Largest Backbone Provider is Leveraging the AI Opportunity - CSSAI-1596
- How to Network LLMs. Splitting Large Language Models Running on GPUs Across a Network - BRKOPS-2931
NANOG 96#
- AI Backend: Deploying SRv6 uSID and SONiC for Deterministic Load Balancing
- From Datacenter to AI Center, building the networks that build AI
- Keynote: AI Data Center Networking: Lessons from Meta's Evolution
- Keynote: Beyond the Chip: The Unprecedented Infrastructure Demands of AI and HPC
- Lightning Talk: Seeing Through the RDMA Fog: Monitoring RoCEv2 with sFlow
- Multipath Traffic Engineering for WAN & AIDC
FOSDEM 2026#
- Building Cloud Infrastructure for AI
- So you want to do RDMA programming? RTRS: An easy to use, reliable high speed transport library over RDMA
OVS + OVN 2025 Fall Conference (Prague)#
KubeCon + CloudNativeCon North America 2025#
- AI Inference Without Boundaries: Dynamic Routing With Multi-Cluster Inference Gateway
- Istio Project Updates: AI Inference, Ambient Multicluster & Default Deny
- Lessons Applied Building a Next-generation AI Proxy
- Navigating the AI/ML Networking Maze in Kubernetes: Lessons From the Trenches
- Routing Stateful AI Workloads in Kubernetes
DENOG 17 (Essen)#
- Ultra Ethernet: Addressing AI/HPC Interconnect Challenges Beyond the InfiniBand Monopoly
- Ultra Ethernet: Addressing AI/HPC Interconnect Challenges Beyond the InfiniBand Monopoly
IETF 124 (Montreal)#
- BGP Next-next Hop Nodes
- Credit-based Flow Control Based on RSVP for RDMA transmission in WAN
- Fast Network Notifications Problem Statement
- Fast Notification for tunnel-based lossless RDMA transmission in WAN
- Scheduling Network Resources for Machine Learning Clusters
OCP Global Summit 2025#
- Agile AI Architectures: The Fungible Data Center for the AI Era
- Capacity Aware Adaptive Routing
- CSIG: Congestion Signaling in the AI Era
- Enabling Technologies for Next Generation Large Scale AI Backend Networking
- Evaluation of CPO Performance and Pluggable Optics Health for Reliable AI Infrastructure
- Evolution of Ethernet based switch platforms and fabrics to meet Meta's AI training clusters' scale-out AI fabric demands
- INSPECT: Data analysis tool for Proactive Link Failure Detection on Meta's AI Infrastructure
- Methodology and Observation of Congestion Control Impact on MoE Training Job Completion Time
- New Strategic Initiative: Open Cluster Designs for AI
- SRv6 for AI Backend
- Technical Aspects of Ultra Ethernet 1.0
- The Path to Zero Flap: Reinventing Optical Reliability for Scalable AI Clusters presented by Credo
Netnod Tech Meeting 2025#
- Building Modern and Scalable DC Fabrics
- Meta – Big Trends and how it will change the Network
- Optics was instrumental in building the Internet—and will be crucial in an AI-driven world
- Ultra Ethernet
SIGCOMM 2025#
- Alibaba Stellar: A New Generation RDMA Network for Cloud AI
- Falcon: A Reliable, Low Latency Hardware Transport
- From ATOP to ZCube: Automated Topology Optimization Pipeline and a Cost-Effective Topology for Large Model Training
- InfiniteHBD: Datacenter-Scale High-Bandwidth Domain for LLM with Optical Circuit Switching Transceivers
- Orderlock: A New Type of Deadlock and its Implications on HPN Protocol Design
ITNOG9#
MLSys 2025#
- COMET: Fine-grained Computation-communication Overlapping for Mixture-of-Experts
- Context Parallelism for Scalable Million-Token Inference
- Extreme PyTorch: Inside the Most Demanding ML Workloads—and the Open Challenges in Building AI Agents to Democratize Them
- Photon: Federated LLM Pre-Training
- Radius: Range-based Gradient Sparsity for Large Foundation Model Pre-training
- ThunderServe: High-performance and Cost-efficient LLM Serving in Cloud Environments
- TileLink: Generating Efficient Compute-Communication Overlapping Kernels using Tile-Centric Primitives