TeraGrid

TeraGrid
Name	TeraGrid
Established	2004
Dissolved	2011 (transitioned)
Country	United States
Type	Distributed high-performance computing cyberinfrastructure
Operator	National Science Foundation and partner institutions

TeraGrid TeraGrid was a United States national cyberinfrastructure initiative that united large-scale high-performance computing resources, data repositories, and network links to support scientific research across disciplines. It connected supercomputers, data storage, and visualization facilities at multiple institutions to enable projects spanning atmospheric science, astrophysics, bioinformatics, chemistry, and materials science. The collaboration linked academic centers, national laboratories, and federal agencies to provide researchers with computational cycles, data management, and software services.

Overview

TeraGrid integrated distributed resources provided by institutions such as Argonne National Laboratory, Oak Ridge National Laboratory, Lawrence Berkeley National Laboratory, National Center for Supercomputing Applications, and Purdue University into a coordinated environment. It provided access to supercomputers like systems at NCSA and Texas Advanced Computing Center, and it leveraged networking contexts involving Internet2 and ESnet. The platform supported workflows in fields connected to NASA, National Oceanic and Atmospheric Administration, National Institutes of Health, Department of Energy, and collaborative projects with universities including University of Illinois at Urbana–Champaign, University of Texas at Austin, University of California, Berkeley, and Indiana University.

History and Development

The development of TeraGrid grew from initiatives funded by the National Science Foundation to expand computational infrastructure following programs like the Blue Gene collaborations and predecessors at centers such as NCSA and Oak Ridge Leadership Computing Facility. Early milestones involved procurement and deployment milestones, procurement awards to vendors such as IBM, Cray Inc., and Sun Microsystems, and collaborative agreements among institutions like SDSC and Purdue. Notable events include the formal launch in the mid-2000s, resource expansions addressing demands from projects associated with NOAA climate modeling, NASA astrophysics simulations, and computational chemistry studies related to National Institutes of Health funding. The program evolved operationally through governance councils and steering committees tied to agencies such as NSF and transitioned into subsequent cyberinfrastructure programs influenced by initiatives like XSEDE.

Architecture and Components

The TeraGrid architecture combined heterogeneous supercomputing systems, high-performance storage arrays, authentication and authorization services, and high-bandwidth networking. Compute nodes ranged from clusters built by Cray Inc. and systems using processors developed by companies like Intel Corporation and AMD. Storage components used technology from vendors including EMC Corporation and NetApp. Networking relied on backbone providers such as ESnet and Internet2 with peering points in metropolitan areas like Chicago, San Diego, and Houston. Middleware and software stacks included resource managers influenced by projects at Argonne National Laboratory and developer tools from groups at University of Illinois, University of Michigan, and University of California, San Diego. Identity management integrated standards and services associated with federations used by InCommon and authentication protocols researched by teams at Massachusetts Institute of Technology and Carnegie Mellon University.

Services and Applications

TeraGrid offered services including batch scheduling, data transfer, remote visualization, science gateways, and user support. Science gateway projects were developed at centers such as Purdue University and Indiana University to serve communities in bioinformatics linked to work at Broad Institute and projects in astronomy coordinated with Space Telescope Science Institute and National Radio Astronomy Observatory. Applications spanned climate modeling collaborations with NOAA and NCAR, molecular dynamics simulations relevant to research at Rice University and University of California, San Diego, and large-scale data analysis used by teams from Lawrence Livermore National Laboratory and Los Alamos National Laboratory. Domain-specific portals enabled researchers at institutions like Columbia University, Princeton University, Stanford University, Harvard University, and Yale University to run workflows on centralized resources.

Governance and Funding

Governance involved a consortium model with principal investigators, campus champions, and site directors from partner institutions, working with programs at the National Science Foundation and coordination with federal laboratories including Argonne National Laboratory and Oak Ridge National Laboratory. Funding mixes included NSF awards, cost-sharing by universities such as University of Illinois at Urbana–Champaign and University of Texas at Austin, and procurement contracts with vendors such as IBM, Cray Inc., and Dell Technologies. Advisory bodies drew membership from representatives of DOE-affiliated labs, academic research computing leaders at University of Michigan and University of California, Berkeley, and stakeholders from organizations like ESnet and Internet2.

Impact and Legacy

TeraGrid influenced subsequent cyberinfrastructure programs and helped establish practices in resource sharing, science gateways, and distributed computing that informed projects like XSEDE and international efforts coordinated with European Grid Infrastructure and collaborations involving PRACE. Its legacy includes enabling influential research outputs from teams at Johns Hopkins University, MIT, Caltech, University of Washington, and University of Chicago, training a generation of computational scientists and shaping policies adopted by research facilities such as Argonne National Laboratory and Oak Ridge National Laboratory. The operational lessons from TeraGrid contributed to standards for middleware, data management, and user support adopted by later infrastructures aligned with funding from agencies like National Science Foundation and Department of Energy.

Category:Supercomputing