INDUSTRY

CHAMPAIGN, IL -- A new technology developed at the National Center for Supercomputing Applications (NCSA) is helping to simplify grid computing by integrating software applications and tools so that they will work together easily and seamlessly over the Internet. Grid computing, which facilitates sharing of online resources and applications across multiple sites, requires software tools that provide standards for security, resource and data management, communication, job scheduling, and other functions. NCSA’s Grid Packaging Toolkit (GPT) helps combine these tools in an integrated package. It handles compatibility, conversion, and version dependencies, works across multiple platforms, and makes it easier to upgrade and deploy grid-enabled applications. “GPT is one of the unsung heroes of grid computing,” said NCSA Director Dan Reed. “GPT is invisible to the user, which is how it should be. It is an essential tool that will allow grid software to be rapidly and simply installed and deployed. The GPT user has less to worry about and can concentrate on the compute job at hand.” “If you had to install a dozen separate software tools and make them all work together, you would need to be an expert on each one,” said Randy Butler, head of NCSA’s Networking, Security and Middleware Directorate. “We’ve developed a cross platform packaging tool that simplifies the bundling and integration of related software for grid computing. GPT makes it easier to install and upgrade component packages, supports complicated multi-package build processes through version and dependency management, and provides cross platform support. These features allow software developers and distributors to bundle multiple packages into a single customized distribution that can be cleanly and uniformly installed across the grid.” GPT is an integral component of the first release of the NSF Middleware Initiative software (NMI-R1), a suite of applications developed by the National Science Foundation’s Middleware Initiative (NMI). NMI-R1 aims to provide an easy-to-use, full-featured, front end for grid computing. The tools in NMI-R1 comprise a standardized package of middleware that will allow scientists to collaborate by sharing data, applications and resources over grids made up of heterogeneous resources GPT is also a key element used to package and distribute the Globus Toolkit version 2.0, the defacto standard grid services software package and a key component of NMI-R1. "GPT has allowed the Globus Project to modularize the Globus Toolkit into distinct components while maintaining integration of the entire toolkit," said Ian Foster, head of the Globus Project at Argonne National Laboratory and a professor of computer science at the University of Chicago. "This gives Globus Toolkit users additional configuration flexibility and allows the Globus Project to upgrade individual components without having to issue a new release of the entire toolkit. With GPT, our users are able to customize a project-specific distribution and to more effectively manage the upgrade path of that distribution through individual component upgrades." The GPT development project began in 1999 with funding from the Department of Energy and was expanded during the next two years with additional funding through NASA's Information Power Grid project. Continued development is supported by the NMI project. Through NMI, the GPT will become a standard packaging tool used by some of the largest projects to create grid communities for scientific computing and research. These projects include: *The NSF TeraGrid project, a multi-year effort to build and deploy the world's largest, fastest, most comprehensive, distributed computing infrastructure for open scientific research. *The Network for Earthquake Engineering and Simulation grid (NEESgrid), an NSF-funded effort led by NCSA to build a national virtual engineering laboratory for designing earthquake-safe structures. *The Grid Physics Network (GriPhyN), an effort to implement the first petabyte-scale computational environments for data intensive science.