First I will begin with the completed text for comments to Organizers. I have not been able to change my original incomplete text:

This talk will focus on grid capabilities and information technologies in the context of the planned teragrid. This planned very high speed network with advanced computing and storage resources is under review. I have not been able to talk about details but hope to do so by the time the preprint is due. The availability of infrastructure like that envisioned in the Teragrid will enhance our ability to process and interactively interrogate data from multiple sources including laboratory, modeling, and observational platforms.

Now for the Abstract:

The Teragrid is a plan to advance the cyberinfrastructure for 21st century science and engineering. It is a response to the pressing need for greater computational power to enable experimentation, modeling, data analysis, and visualization activities that often involve large volumes of data and use of an expanding national computational grid. It is a vision to develop and deploy a comprehensive computational, data management, and networking infrastructure of unprecedented scale and capability that couples distributed scientific instruments, terascale and petascale computing facilities, multiple petabyte data archives, and gigabit and beyond networks, all widely accessible by researchers including those in the atmospheric sciences, oceanography, and hydrology.

The Teragrid is an outgrowth of the NSF PACI program which has two leading edge sites, NPACI in San Diego and NCSA in Champaign/Urbana. Facilities and resources from these and affiliate sites have been used by a growing number of geoscientists during the past decade. Over the past several years there has been increased emphasis within PACI on use of the national grid in support of distributed computation, distributed data access and processing, distributed datamining, and distributed collaboration. Infrastructure developments in support of this emphasis have included Globus middleware for authentication and security and the Access Grid for collaborative group meetings and distributed presentations. The Teragrid is a natural extension of these and other developments.

The Teragrid plans call for provision to the national research community of multiple teraflop computing capabilities in the form of Linux clusters that support multi-teraflop model simulation, petabyte data analysis capabilities, high-performance visualization on remote platforms, and high resolution wall displays. Further, the software infrastructure necessary to support research productivity in this environment will be enhanced. The presentation will focus on the implementation of the plan and how it can impact geoscience research with an emphasis on atmospheric science, oceanography, and hydrology.