DCD Zero Downtime: The Bi-Weekly Data Center Show Episode 52 - Why Oxide rebuilt the rack from scratch
Apr 4, 2024
Bryan Cantrill, CTO of Oxide Computer, explains why the industry has been using the wrong components in data centers. The cloud's success is due to its easily consumed virtual servers, but it comes with costs. Cantrill discusses the need for an elastic infrastructure on-premise and why his team had to rebuild almost everything to deliver it.
AI Snips
Chapters
Transcript
Episode notes
Rebuilding Rack from Scratch
- Oxide Computer rebuilds entire rack infrastructure from hardware to software for on-premise elastic computing.
- Co-designing hardware and software is key to delivering systems rivaling public cloud flexibility and control.
One U Servers Are Flawed
- One U form factor servers are inefficient due to power and cooling constraints.
- Oxide's design uses larger fans and a different sled geometry to reduce noise and improve airflow efficiently.
Power Supplies Cause Inefficiency
- AC power supplies with internal fans add significant noise, heat, and energy inefficiency in data centers.
- Power supply design complexity and lack of monitoring make power management difficult at rack scale.