Design, Deploy, and Monitor Networks for AI with Aviz

This 27-minute conference talk by Thomas Scheibe, Chief Product Officer at Aviz Networks, explores solutions for designing, deploying, and monitoring networks specifically optimized for AI workloads. Learn about the specialized networking requirements for AI infrastructure, including the need for multiple networks, differentiated Quality of Service (QoS), and integrating compute into end-to-end network topology. Discover how Aviz's ONCE product supports Sonic and Cumulus network operating systems while providing automation and orchestration for faster deployment, service activation, and infrastructure expansion. Follow the evolution of networking in AI environments, from traditional single data center networks to the modern separation between front-end (user access) and back-end (GPU communication) networks. Understand how Aviz partners with hardware providers and leverages reference architectures like NVIDIA Spectrum-X to automate network configuration, enabling enterprises to define networks and implement network separation. Gain insights into Aviz's comprehensive support for Sonic deployments in enterprise data centers and edge environments, with particular focus on automating deployment workflows that simplify network management for AI infrastructure. Recorded live in Santa Clara, California on April 25, 2025, as part of AI Infrastructure Field Day.