AI, Data Science & Cloud Certificates from Google, IBM & Meta
Finance Certifications Goldman Sachs & Amazon Teams Trust
Overview
Google, IBM & Meta Certificates — All 10,000+ Courses at 40% Off
One annual plan covers every course and certificate on Coursera. 40% off for a limited time.
Get Full Access
This presentation by Maxim Kuschnerov, Director of R&D at Huawei Technologies, explores optical network interconnectivity solutions for GPU clusters in AI applications. Learn about half-retimed pluggable transceivers that support 224Gbps and comply with optical interface standards, reducing manufacturing and maintenance costs. Discover how Linear Receiver Optics (LRO) can further decrease costs and power consumption using VCSELs for distances under 100m. The talk demonstrates 112G LRO pluggable solutions for both VCSELs and Silicon Photonics, covering various AI datacenter interconnectivity scenarios from multimode fiber up to 100m to single mode fiber up to 500m. Understand the advantages of these solutions, including superior performance, channel-adaptive functionality, interoperability with retimed pluggable solutions, low power consumption, low latency, and high reliability.
Syllabus
LRO-based Pluggable VCSEL and SiP Transceivers for AI clusters
Taught by
Open Compute Project