Congestion Management in the AI Data Center

Ethernet fabrics are optimized for AI/ML clusters leveraging widely adopted standards-based technology. They deliver proven scalability with lower cost-per-bit and advanced congestion management, making them an optimal choice for GPU clusters. Juniper’s Arun Gandhi, Mahesh Subramaniam, and Michal Styszynski discuss the cutting-edge congestion management technologies needed for the lossless ROCEv2 transport networks when connecting GPU clusters.

AI Data Center Networking:
https://www.juniper.net/us/en/solutions/data-center/ai-infrastructure.html

AI/ML Data Center Networking on Ethernet:
https://blogs.juniper.net/en-us/industry-solutions-and-trends/ai-ml-data-center-networking-on-ethernet