OpenAI and Tech Giants Launch Open Networking Protocol to Boost AI Cluster Speed and Efficiency
Summary
OpenAI and major tech giants including Microsoft, Nvidia, AMD, and Intel unveil MRC, a groundbreaking open networking protocol using 'packet spraying' to dramatically boost AI GPU cluster speed, reliability, and energy efficiency — already live in major supercomputing sites — as the industry races to combat a looming compute shortage.
Key Points
- OpenAI, alongside researchers from AMD, Broadcom, Intel, Microsoft, and Nvidia, is unveiling a new compute networking protocol called MRC (Multipath Reliable Connection), designed to make GPU clusters faster, more reliable, and more energy-efficient by tackling network congestion and failures.
- MRC uses 'packet spraying' to scatter data across hundreds of network paths simultaneously, while pairing with the SRv6 routing protocol to reduce energy demands on network switches — and is already live in major training clusters including Microsoft's Fairwater supercomputers and an Oracle site in Abilene, Texas.
- OpenAI is releasing the MRC specification as an open standard through the Open Compute Project, aiming to eliminate industry fragmentation and push the entire AI sector toward more efficient compute use as a looming compute shortage intensifies.