Calnex Solutions Achieves PFC Transparent 400G Network Emulation for Validation Testing of Lossless Ethernet AI Clusters
Network behavior is a critical determinant of efficiency in large scale AI clusters. Microbursts, congestion and PFC interactions introduce bottlenecks that increase tail latency, directly impacting GPU utilization and job completion time. These conditions are difficult to reproduce consistently in lab environments, making it challenging for engineering teams to validate how lossless Ethernet fabrics will perform under real production workloads. Calnex addresses this by enabling controlled, repeatable validation of these behaviors, bringing real-world network conditions into the lab.
"AI infrastructure investment is accelerating rapidly, but return on that investment still depends heavily on the network," said
This milestone reinforces Calnex's growing role in AI validation, supporting customers across inference, AI-RAN and data center networking. As AI deployments expand across different environments, Calnex is focused on enabling testing and assurance solutions that help teams deliver predictable performance.
For more information, please contact
Calnex designs, produces and markets test instrumentation and solutions for network synchronization and network emulation, enabling its customers to validate the performance of the critical infrastructure associated with telecoms networks.
Logo - https://mma.prnewswire.com/media/2918120/5818349/Calnex_Solutions_Logo.jpg
View original content:https://www.prnewswire.com/news-releases/calnex-solutions-achieves-pfc-transparent-400g-network-emulation-for-validation-testing-of-lossless-ethernet-ai-clusters-302743373.html
SOURCE