The new AI Computing Continuum initiative will deliver an open, interoperable, and scalable infrastructure extending AI from centralized data center to the user

LONDON: The Innovative Optical and Wireless Network Global Forum (IOWN Global Forum), the international industry association developing next-generation, photonics-driven networks, and the Open Compute Project Foundation (OCP), the global non-profit organization bringing hyperscale innovations to all, today announce a cooperative framework, called the AI Computing Continuum, to deliver a seamless computational infrastructure from centralized to edge deployments.

The organizations' work involves developing a connected ecosystem that goes beyond traditional large-scale, centralized data centers to geographically distributed deployment sites. These include regional data centers, colocation data centers, telecom provider-owned facilities, enterprise-operated data centers, industrial locations, and office sites. This, however, presents the challenges of achieving the desired computing performance and efficiency across the entire continuum.

Under the framework, the IOWN Global Forum and OCP will work together to develop a roadmap for a multi-site, high-bandwidth, low-latency compute and network infrastructure. The IOWN Global Forum will define the communications architecture by leveraging photonics-based optical communication and device technologies. OCP will drive the specifications for the open hardware required to deliver the infrastructure.

The IOWN Global Forum will also deliver system designs for industry-specific early adoption use cases, developed with Forum members representing industries including financial services, manufacturing, entertainment, logistics, construction, and urban development. In addition to meeting common goals, the Forum will build on its current work to deliver unique offerings for these industries such as:

• Optimizing the performance and efficiency of multiple layers to avoid detrimental impact by the abstraction layer.
• Democratizing solutions to allow enterprises to be part of the fast-evolving AI computing space.
• Allowing enterprises to maintain their data sovereignty across data lifecycles.
• Creating a commercially viable approach for the early adoption of emerging optical technologies.
• Delivering techno-economic analysis to assess the economic and environmental gains.

OCP will develop approved contributions to meet the requirements for developing an AI Computing Continuum while endeavoring to leverage hyperscale data center innovations including:

• Enabling high-performance scalability through denser standard compute platforms with reuse of components and designs.
• Ensuring data center infrastructure efficiencies through standardized power and cooling solutions, management APIs, and telemetry.
• Developing network technologies that scale up, out, and across the compute continuum.
• Striving for hardware abstractions, allowing choice of software stacks.

“Five years have passed since the establishment of the IOWN Global Forum, and we are finally beginning to see concrete implementations and practical adaptations,” said Dr. Katsuhiko Kawazoe, President and Chairperson of the IOWN Global Forum. “It is truly encouraging that the outcomes of the IOWN Global Forum can now be applied to the rapidly growing power demands of AI data centers, as well as their needs for operating in distributed environments.”

“Partnering with the IOWN Global Forum will allow the OCP Community to address key challenges with extending computational infrastructure for AI outside the centralized data center, including the standardization of high-performance accelerated compute servers that can scale up and down to span the entire continuum," said George Tchaparian, CEO at the Open Compute Project Foundation. “This collaboration with the IOWN Global Forum will allow the OCP Community to leverage hyperscale data center innovations and cascade them to a much larger portion of the market.”

“The AI transformation within the data center is well underway with new inference architectures still emerging to drive performance. To deliver on the range of training and inference use cases across industries, accelerated compute infrastructure should be deployed where data is created and consumed by people and edge devices,” said Roy Chua, Founder and Principal at Avidthink. “OCP and the IOWN Global Forum collaborating on the AI Computing Continuum has the potential to accelerate extending AI from the central data center out to the edge, catching the next wave of AI.”

About the IOWN Global Forum

The IOWN Global Forum was established in 2020 as a private sector organization to develop IOWN Global Forum technologies and use cases, and is comprised of over 180 organizations. The objective of the IOWN Global Forum is to accelerate innovation and adoption of a new communication infrastructure to meet our future data and computing requirements through the development of new technologies, frameworks, specifications, and reference designs in areas such as photonics R&D, distributed computing, use cases, and best practices. For more information, visit IOWN Global Forum.

About OCP

The Open Compute Project (OCP) brings at-scale innovations and hyperscaler best practices to all, spanning technology domains from the data center to the edge, and the technology stack from silicon, to systems, to site facilities and services. The international OCP Community is made up of organizations and people from hyperscale, neocloud and cloud data center operators, communications providers, colocation providers, diverse enterprises, and technology providers. With The OCP Tenets of Openness, Impact, Efficiency, Scale and Sustainability, the OCP Foundation engages with industry ecosystems, our growing membership, and educates thousands of engineers and industry leaders, every year. Across many projects and initiatives, the OCP Foundation and its Community are meeting the AI Data Center market evolution today and shaping the future. Learn more at Open Compute Project.

Fonte: Business Wire