xGFabric

This initiative is part of a U.S. Department of Energy (DOE) grant to research and develop the next generation of distributed computational infrastructure.  The goal of xGFabric is to enable the seamless coupling of High-performance Computing (HPC) and AI large-scale computational resources with computers and storage "at the edge" and in the field using 5G/6G networking.  As a single, unified, and multiscale distributed computing capability, xGFabric uses cutting edge wireless capabilities to enable HPC and AI applications "in the loop"  for automated decision-making and control.

 

The project emphasizes the development of new abstractions, programming systems, and operating systems that specifically leverage 5G/6G technologies.  xGFabric aims to provide a resilient and performant computing capability that can be applied to a range of scientific applications.  For empirical evaluation and validation, xGFabric focuses on applications from digital agriculture as "science drivers." While the project also includes the development of sensor virtualization capabilities and plans to release open-source prototypes for broader use, it is also forwarding technological innovation in rural communities and in the agricultural sector.  The project draws together, as a partnership, researchers and field testbed facilities from UCSB, University of Nebraska, Lincoln, University of Notre Dame, the Princeton Plasma Physics Laboratory, and Brookhaven National Laboratory.