Learn about Dense Wavelength Division Multiplexing (DWDM) Fabric platform technology in this 17-minute conference talk from Xscape Photonics presented at the Open Compute Project. Explore how DWDM fabric platforms enable energy-efficient bandwidth scaling specifically designed for AI clusters, addressing the growing demands of high-performance computing environments. Discover the technical innovations and architectural approaches that make this solution effective for modern data center interconnects and AI infrastructure requirements.
DWDM Fabric Platform for Energy Efficient Bandwidth Scaling for AI Clusters
Open Compute Project via YouTube
Overview
Syllabus
DWDM Fabric platform for energy efficient bandwidth scaling for AI clusters
Taught by
Open Compute Project
Related Courses
-
Energy-Efficient High Bandwidth Optical Interconnect Solutions for Generative AI
-
Linear Optics for AI Fabric Scale-Up and Scale-Out
-
Next Generation Fabric for AI/ML Use Cases - Disaggregated Scheduled Fabric Using SONiC
-
Photonic Fabric for AI Interconnect
-
The Art of AI Networking Scaling to Support 100K GPU Nodes
