OIF Launches Energy-Efficient Interfaces Project for Data Centers and AI
The Optical Internetworking Forum (OIF) has initiated the Energy Efficient Interfaces (EEI) Framework project to study new energy-efficient electrical and optical interfaces. These interfaces, including "Linear" or "Direct Drive" optics, aim to meet the demands of data center networking, artificial intelligence (AI) training, machine learning (ML), and disaggregation.
The project's objective is to identify critical applications and requirements for next-generation electrical and optical links, such as die-to-die, co-packaged, near-packaged, and pluggable solutions. The findings will be outlined in a Framework Document, serving as a technical white paper.
Rob Stone from Meta Platforms emphasized the importance of energy efficiency, particularly for data center operators. The project aims to enable groundbreaking AI-driven applications and immersive experiences by exploring energy-efficient interfaces for next-gen data centers.
The project also seeks to identify opportunities for interoperability standards, paving the way for potential future collaborations at OIF or other standards organizations.
Yi Tang from Cisco Systems was appointed as PLL Working Group Electrical Vice Chair. OIF, with over 140 network operators, vendors, and test equipment vendors, fosters open network connectivity by developing interoperable electrical, optical, and control platforms.