Grid Support for Collaborative Control Room in Fusion Science

  • Authors:

  • K. Keahey;M. E. Papka;Q. Peng;D. Schissel;G. Abla;T. Araki;J. Burruss;E. Feibush;P. Lane;S. Klasky;T. Leggett;D. Mccune;L. Randerson

  • Affiliations:

  • Argonne National Laboratory, Argonne 60439-4844;Argonne National Laboratory, Argonne 60439-4844;General Atomics, San Diego 92186-5608;General Atomics, San Diego 92186-5608;General Atomics, San Diego 92186-5608;Aff1 Aff3;General Atomics, San Diego 92186-5608;Princeton Plasma Physics Laboratory, Princeton 08543;Argonne National Laboratory, Argonne 60439-4844;Princeton Plasma Physics Laboratory, Princeton 08543;Argonne National Laboratory, Argonne 60439-4844;Princeton Plasma Physics Laboratory, Princeton 08543;Princeton Plasma Physics Laboratory, Princeton 08543

  • Venue:
  • Cluster Computing
  • Year:
  • 2005

Quantified Score

Hi-index 0.00

Visualization

Abstract

The National Fusion Collaboratory project seeks to enable fusion scientists to exploit Grid capabilities in support of experimental science. To this end we are exploring the concept of a collaborative control room that harnesses Grid and collaborative technologies to provide an environment in which remote experimental devices, codes, and expertise can interact in real time during an experiment. This concept has the potential to make fusion experiments more efficient by enabling researchers to perform more analysis and by engaging more expertise from a geographically distributed team of scientists and resources. As the realities of software development, talent distribution, and budgets increasingly encourage pooling resources and specialization, we see such environments as a necessary tool for future science.In this paper, we describe an experimental mock-up of a remote interaction with the DIII-D control room. The collaborative control room was demonstrated at SC03 and later reviewed at an international ITER Grid Workshop. We describe how the combined effect of various technologies--collaborative, visualization, and Grid--can be used effectively in experimental science. Specifically, we describe the Access Grid, experimental data presentation tools, and agreement-based resource management and workflow systems enabling time-bounded end-to-end application execution. We also report on FusionGrid services whose use during the fusion experimental cycle became possible for the first time thanks to this technology, and we discuss its potential use in future fusion experiments.