Linear Optics for AI Infrastructure - Power Efficient LPO and LRO Solutions at 112G and 224G
Open Compute Project via YouTube
Learn AI, Data Science & Business — Earn Certificates That Get You Hired
The Most Addictive Python and SQL Courses
Overview
AI, Data Science & Cloud Certificates from Google, IBM & Meta — 40% Off
One plan covers every Professional Certificate on Coursera. 40% off Coursera Plus Annual.
Unlock All Certificates
Explore power-efficient Linear Pluggable Optics (LPO) and Linear Receive Optics (LRO) solutions designed for AI infrastructure in this 17-minute conference talk. Learn how these linear optics technologies serve as efficient alternatives to traditional DSP-based modules for 800G (112G-lane) and 1.6T (224G-lane) connectivity as data centers face increasing demands from AI workloads. Discover how linear optics support key optical types including DR8 and 2xFR4, while examining the system-level benefits in terms of energy savings and scalability. Gain insights into the latest developments in LPO and LRO technologies and understand their critical role in building more power-optimized AI infrastructure to meet the growing bandwidth requirements of modern data centers.
Syllabus
Linear Optics for AI Infrastructure Power Efficient LPO and LRO Solutions at 112G and 224G pe
Taught by
Open Compute Project