Cloud service providers, OEMs, chip firms form Ultra Ethernet Consortium

Consortium formed to deliver on Ethernet-based open, interoperable, high-performance, full-communications stack architecture to meet the growing network demands of AI and high-performance computing at scale

AMD, Arista, Broadcom, Cisco, Eviden, HPE, Intel, Meta and Microsoft have revealed that they will be among the founding members of the Ultra Ethernet Consortium (UEC), an industry-wide cooperation to build a complete Ethernet-based communication stack architecture for high-performance networking.

Explaining the reasons for why it was setting up, the UEC, a Joint Development Foundation project hosted by The Linux Foundation, observed that artificial intelligence (AI) and high-performance computing (HPC) workloads were rapidly evolving and require best-in-class functionality, performance, interoperability and total cost of ownership, without sacrificing developer and end-user friendliness.

The Ultra Ethernet solution stack will aim to capitalise on Ethernet’s ubiquity and flexibility for handling a wide variety of workloads while being scalable and cost-effective. 

The consortium is founded by companies that see themselves as having a long-standing history and experience in high-performance solutions and who collectively have decades of networking, AI, cloud, and high-performance computing-at-scale deployments.

Each member will endeavour to contribute to the broader ecosystem of high-performance in an egalitarian manner and the e consortium will work on minimising communication stack changes while maintaining and promoting Ethernet interoperability. UEC will begin accepting applications for new members in Q4 2023.

Among the specific technical goals for the consortium is a plan to develop AI and HPC at-scale Ethernet specifications, APIs and source code to define protocols, electrical and optical signalling characteristics, application programming interfaces, and/or data structures for Ethernet communications.

UEC said it will follow a systematic approach with modular, compatible, interoperable layers with tight integration to provide a holistic improvement for demanding workloads. The founding companies are seeding the consortium with highly valuable contributions in four working groups: physical layer, link layer, transport layer and software layer. 

In addition, it will look to develop link-level and end-to-end network transport protocols to extend or replace existing link and transport protocols and assess link-level and end-to-end congestion, telemetry and signalling mechanisms. Each of these are seen as suitable for artificial intelligence, machine learning, and high-performance computing environments. The consortium will also investigate software, storage, management, and security constructs to facilitate a variety of workloads and operating environments.

Industry Analysts have offered positive feedback to the setting up of the UEC. Hyperion Research said it was exciting to see what it called an impressive group of leading companies work together to create a new common higher-performance interconnect solution.

“Many HPC and AI users are finding it difficult to obtain the full performance from their systems due to weaknesses in the system interconnect capabilities,” it added.

“It’s also difficult for users to integrate and learn multiple new or different solutions. Buyers in the HPC and AI areas have very demanding workloads, which the Ultra Ethernet Consortium approach could greatly help improve interoperability, performance and capabilities. We look forward to seeing a new set of products enter the market in the near future.”

“There has been an ongoing discussion, dare I say battle, over the best networking to use for infrastructure supporting the training and inference of large language models for generative AI,” said Karl Freund, founder and principal analyst at Cambrian-AI Research.

“Some companies have been shifting to Ethernet-based networking, preferring its ease of installation and use. The UEC initiative will be a welcome addition to the AI community.”

Read more about HPC infrastructure

Read more on Internet infrastructure

Data Center
Data Management