Tuesday Night Live
summary
32,000 attendees and concrete operational detail: AWS shows tens of thousands of virtual servers spun up for Prime Day, 14 regions (growing to 18) with multi‑AZ buildings, a private 100Gb backbone and the Hawaiki 14,000 km cable using ±10,000V DC repeater trick. Infrastructure includes custom routers built around a 3.2 Tb Tomahawk ASIC and a 25Gb optics strategy, P2 GPU instances for deep learning with MXNet, JPL runs Mars/Europa workflows on AWS, and storage racks reach 1,110 drives (~11 PB); 50% renewable target and ~970 MW of projects.
AI-written summary · about
chapters (11)
- 0:00Welcome and introduction
- 1:46AWS infrastructure overview
- 7:33Global AWS network and regions
- 15:24Deep dive into AWS region
- 23:12Custom-built networking hardware
- 35:14Fault management and reliability
- 40:54Custom storage and compute servers
- 44:50Mainframe migration to AWS
- 46:58NASA JPL and cloud computing
- 53:19Machine learning and deep learning
- 1:12:15Sustainability and renewable energy
