At Open Compute Project Summit (OCP) 2024, we’re sharing details about our next-generation network fabric for our AI training clusters.
We’ve expanded our network hardware portfolio and are contributing two new disaggregated network fabrics and a new NIC to OCP.
Meta believes that open hardware drives innovation, and open hardware takes on an important role in assisting with disaggregation.
DSF fabric allows us to build large, non-blocking fabrics to support high-bandwidth AI clusters, and supports an open and standard Ethernet-based RoCE interface.
Meta will deploy two next-generation 400G fabric switches, the Minipack3 and the Cisco 8501, both of which are backward compatible.
Meta's data center fabrics have evolved from 200 Gbps/400 Gbps to 400 Gbps/800 Gbps, and 2x400G optics have already been deployed in our data centers.
Developers and engineers from all over the world can work with this open hardware and contribute their own software that they, in turn, can use themselves and share with the wider industry.
FBNIC, a true multi-host foundational NIC designed by Meta, contains the first of our Meta-designed network ASICs for our server fleet and MTIA solutions.
FBNIC's key features include network interfaces for up to 4x100/4x50/4x25 GE with SerDes support, HW offloads, header-data split to assist zero-copy, and compliance with OCP NIC 3.0 design specification.
Meta envisions a future of AI hardware systems that are not only scalable but also open and collaborative, and encourages anyone who wants to help advance the future of networking hardware for AI to engage with OCP and Meta.