Globus Case Studies

The following list contains descriptions of and pointers to various studies or efforts illustrating how scientists from different disciplines are using the Globus Toolkit to take advantage of grid computing. For more information on any individual study, click on its title.

CMS Data Analysis Using Alliance Grid Resources
Using the Globus Toolkit and Condor, a prototype grid environment was developed to allow the researchers at the California Institute of Technology working on the Compact Muon Solenoid (CMS) of CERN's Large Hadron Collider (LHC) to more effectively and efficiently exploit Alliance computational resources. The CMS experiment will begin producing data in 2005 at the rate of 100 MB/s or approximately 5000 TB per year.

GridGaussian
Gaussian 98 is a general-purpose ab initio electronic structure package used by more than 50 scientific groups in the Alliance. The package is available on a number of compute hosts around the Alliance, but because of heterogeneity of architectures, batch systems, and local policies, users are burdened with mastering the often cumbersome details of how to run Gaussian-98 on each system. GridGaussian will eliminate this burden by providing a single, unified interface built on top of the Globus Toolkit and Condor infrastructures.

Image Server for High-Resolution Movie Playback on Tiled Displays
The goal of this effort is to store high-resolution frames of visualization of scientific data in a standardized format at a centralized location and to use the stored frames for playback on tiled displays. The intent is to use the striped/parallel gsiftp server packaged with the Grid-in-a-Box software to transfer the appropriately configured portion of the full-resolution image to each of the machines responsible for displaying the individual tiles.

Extending Cactus Capabilities to Address Network Performance Problems
This study documents production runs in which over 1500 processors were pooled from the Origin2000 cluster at NCSA and the Blue Horizon SP system at SDSC for a single execution sustaining 190Glop/s with over 70 percent scaling efficiency. To do this, Cactus MPI code was modified so that researchers could distinguish between inter- and intra- machine communication. Intra-machine communication remained as it was, but to enhance communication over the OC-12 WAN, a Globus Toolkit RSL script was written that coalesced small messages into a single large message, which was then compressed and sent over the WAN to the peer machine, where it was decompressed.

Gnome file manager
This case study documents work that was done at CERN to extend the Gnome file manager using specially patched FTP servers using GSI authentication.

Compiling Cactus with the Globus Toolkit
This study describes how the Astrophysics Simulation Collaboratory (ACS), a specialized Cactus Portal, uses various Globus Toolkit components to enable configuration and submission of Cactus components from Web browsers. The ASC also uses MPICH-G2 to enable distributed execution across multiple computers.

The Grid initiatives linked below involve "real end users working with computer scientists to build a new generation of tools and infrastructure for their day-to-day work in disciplines such as physics, biology, and engineering."

• Grid Physics Network project: www.griphyn.org



• European Data Grid: www.eu-datagrid.org



• Particle Physics Data Grid: www.ppdg.net



• Network for Earthquake Engineering Simulation Grid: www.neesgrid.org



• NASA IPG: www.ipg.nasa.gov/aboutipg/aboutipg_index.htm
  The Information Power Grid (IPG) is NASA's high-performance computational grid. The IPG is a collaborative effort between NASA Ames, NASA Glenn, and NASA Langley Research Centers, and the NSF PACI programs at SDSC and NCSA, and is funded by the IT/ACNS program at NASA Ames Research Center.



• Alliance VMR: http://www.ncsa.uiuc.edu/UserInfo/Alliance/VMR.html
  NCSA and its Alliance partners are currently deploying the infrastructure necessary to establish the Alliance Virtual Machine Room (VMR). The goal is to make the many resources available at Alliance sites appear to be managed in a single coherent fashion as if they were co-located and under the direct management of a single center. The VMR consists of six sites with 24 x 7 operations and resource and usage monitoring. Distributed research environments like the VMR will likely emerge as the collaborative spaces in which science and interdisciplinary research will be conducted in the future.



• Alliance DTF:
  (To come)