Delivering Energy & Exergy Efficiency in the Converged 5G RAN/Edge Compute Network

Default ARPA-E Project Image


Program:
OPEN 2021
Award:
$1,990,282
Location:
Murray Hill, New Jersey
Status:
ACTIVE
Project Term:
05/02/2022 - 05/01/2025
Website:

Critical Need:

Current global data center (electricity consumption is estimated to be as high as 400–500 terawatt hours, corresponding to ~2% of total global electricity consumption. The rise of computationally intensive artificial intelligence (AI) workloads is also a contributing factor to increasing rack (and data center) power densities. The power usage associated with AI workloads is so significant that web-scale operators are actively exploring how to optimally distribute and schedule AI training activities across data center facilities to maximize energy efficiency and minimize carbon intensity.

Project Innovation + Advantages:

Nokia will reuse the heat energy AI workloads produce while delivering digital services to supply high quality thermal energy that can be used directly for building heating, cooling, and/or thermal energy storage. The proposed technology will pursue a low-cost, passive, ultra-reliable, high-performance, two-phase cooling philosophy from chip to room scale. The team will rearchitect the computing infrastructure for secondary use as a valuable heat source in heat reuse applications with minimal supporting infrastructure. Available work efficiency will be maximized via heat reuse enabled by the increased flow of high-quality thermal power out of a densified server environment with corresponding energy reuse effectiveness.

Potential Impact:

Nokia will develop a highly efficient, resource-conserving, thermal energy architecture that will simultaneously improve AI server cooling energy efficiency and deliver high quality thermal energy that can be used directly for heating and cooling buildings.

Security:

This effort will contribute to ensuring that the U.S maintains a technological lead in developing and deploying advanced energy and information and communications technology.

Environment:

: Deploying AI in networks for building heating, cooling, and/or thermal energy storage without using additional energy would play a positive role in tackling climate change.

Economy:

The new technology will enable significant cost and space savings associated with minimizing air conditioning and distribution requirements.

Contact

ARPA-E Program Director:
Dr. Peter de Bock
Project Contact:
Todd Salamon
Press and General Inquiries Email:
ARPA-E-Comms@hq.doe.gov
Project Contact Email:
todd.salamon@nokia-bell-labs.com

Partners

University of Illinois, Urbana Champaign

Related Projects


Release Date:
02/11/2021