ENTERPRISE
AI Infrastructure Networking
High-bandwidth, low-latency fabrics for GPU clusters: RoCE, InfiniBand concepts, spine-leaf at scale, and observability for ML platforms.
36 hours · 3 modules
Course contents
Select any lesson to open slides, video, or lab.
Module 1
AI Cluster Fabrics
Master ai cluster fabrics through guided lessons, labs, and assessments.
240 min
- • Fat-tree
- • Rail-optimized
- • ECN/PFC
4 lessons
Module 2
RDMA & RoCE
Master rdma & roce through guided lessons, labs, and assessments.
200 min
- • Lossless Ethernet
- • PFC/ECN tuning
4 lessons