About
Abstract: This talk covers the basics of Networking for AI, focusing on open Ethernet-based solutions. We'll describe some common types of AI workloads and how they differ from traditional datacentre workloads. We'll then discuss three types of dedicated network fabrics that are built to support large-scale AI, presenting some of the design principles and implementation options that arise. We will close with a discussion of key open issues and industry trends.
Bio: Alan Judge is a Distinguished Engineer and Head of Cloud & AI Architecture at Arista. Prior to joining Arista, he spent two years with a gigawatt-scale AI infrastructure startup and eighteen years with Amazon building network infrastructure for the cloud. Even earlier, he was an ISP guy back when people still used modems. He holds B.A. (Mod) and PhD degrees in Computer Science from Trinity College, Dublin.
This seminar is part of the CERN STEAM Academy Seminar Series.
Networking cocktail will follow the seminar.
With the support of CERN's Next Generation Triggers Project.
00:00:00 Slide 1
00:01:24 Slide 2
00:02:01 Slide 3
00:02:34 Slide 4
00:07:14 Slide 5
00:11:10 Slide 6
00:16:54 Slide 7
The talk
Credits and licence
- Institution
- CERN
- Speaker
- Alan Dr Judge
- Year
- 2026
- Language
- English
- Subjects
- Computer science
- Source
- CERN
- Licence
- CERN, as CERN states it. CERN's terms
- Attribution
- © CERN. CERN provides its audiovisual media free of charge for educational and informational use, credited to CERN; this lecture is shown in CERN's own player.
- Who can watch
- CERN gives no rating. Invisible Network shows it to Tweens profiles and older, judged from what the source says about it. How courses are placed