Continuation of Earth System Knowledge Environment reports from FY2006
This project is now classified as cyberinfrastructure: CISL Portal.
Grid-BGC is a collaborative project to develop an end-to-end solution that seamlessly couples models, data, and resources to allow geoscientists to simulate terrestrial biogeochemistry over large domains at high spatial resolution. The Grid-BGC project's external funding concluded in October 2006 with the release of a production-grade version of the Grid-BGC science portal to the biogeochemistry community along with user and system documentation. The project received a no-cost extension until 12/31/06, and thus concluded in FY2007.
While development of BGC itself has completed, its workflows remain a primary motivation for continued work in the area of Grid-based service oriented architectures (SOAs). These services form the back end to the Grid-BGC science portal that allow scientists to interactively set up simulations used to model the global carbon cycle. Experiments with the POP, CAM, and WRF models have shown that this SOA can indeed integrate these applications into a grid environment. Grid-BGC is the first working computational grid for carbon cycle simulations, and the SOA designed to support it provides a seamless set of services from which the geoscience community can potentially build other end-to-end solutions. We are generalizing these SOA components into an Extensible Service Provider (ESP) toolkit supporting rapid service deployment for new applications.
The computational components of Grid-BGC workflows are embarrassingly parallel and are traditionally run on Linux clusters, large IBM SMP systems, and other systems that can execute many single-processor tasks efficiently. These types of workflows lead to inefficient usage of massively parallel architectures such as the IBM Blue Gene/L (BG/L) because of allocation constraints forced by its unique system design. Recently, IBM introduced the ability to schedule individual processors on BG/La feature named High Throughput Computing (HTC)creating an opportunity to exploit the system's power efficiency for other classes of computing. In FY2007, we developed a Grid-enabled interface supporting HTC on BG/L and applied it to Grid-BGC workflows. This interface accepts single-processor tasks using Globus GRAM, aggregates HTC tasks into BG/L partitions, and requests partition execution using the underlying system scheduler. By separating HTC task aggregation from scheduling, we provide the ability for workflows constructed using standard Grid middleware to run both parallel and serial jobs on BG/L. This feature allows us to run jobs from Grid-BGC (and similar workflows) on the BG/L.
This effort supports NCAR's strategic priorities of "Developing and providing advanced services and tools" and "Creating an Earth system knowledge environment." NASA has provided funding for the Grid-BGC project through the Advanced Information Systems Technology Office (NASA AIST Grant NAG2-1646) and the Terrestrial Ecology Program.