LEAD Status Report
April 2005
Anne Wilson
Brian Kelly
Doug Lindholm
Mohan Ramamurthy
Tom Baltzer
Overview
The Linked Environments for Atmospheric Discovery (LEAD) is a National Science Foundation
(NSF) Large Information Technology Research (ITR) project that is creating
an integrated, scalable cyberinfrastructure for mesoscale meteorology
research and education. It is a five year effort that started in the
Fall of 2003.
The UPC is working in collaboration with several other University partners to set
up a collection of test bed computer systems to support the research
and educational goals of the LEAD project. Each test bed, including
one at the UPC, will integrate existing and new tools (e.g. for data
gathering, serving, mining, assimilation, and forecast modeling) and
enable them to interoperate under the dynamic control of a Grid
environment.
Unidata's Role in LEAD
The primary roles that Unidata will play in the near term in the LEAD effort include:
- Support data delivery to the LEAD test beds via the IDD, LDM, and decoders.
- Provide data discovery and access mechanisms via THREDDS, OPeNDAP, ADDE, and NetCDF.
- Provide visualization and analysis capabilities via the IDV.
- Enhance Unidata tools to work in a Grid and Web Services environment.
- Help design and implement the software architecture of the LEAD Grid test beds.
- Build out the Unidata LEAD testbed with underlying tools and services that will fulfill the LEAD vision.
- Help define the meteorology research and education use case scenarios that drive the
development.
- Help coordinate the technical aspects of the LEAD project.
Benefits to Unidata Community
In addition to meeting the goals of the LEAD project, Unidata's
efforts toward these ends will enhance our ability to serve the
community. In fact, we are concentrating our efforts initially on the
aspects of LEAD in which we think our community is most interested.
Some of the community benefits will include:
- Improved procedures for tool installation and configuration.
- Improved documentation of existing Unidata tools.
- Enhanced tool features and capabilities.
- Improved interoperability between and better integration of Unidata tools and systems.
- UPC expertise in the installation and operation of WRF, data assimilation, data mining,
and Grid and Web Service technologies.
- Access to atmospheric discovery capabilities unachievable today.
Recent Accomplishments
- Unidata LEAD Testbed
- Installed Rocks on two testbed machines. Had difficulty with what seemed to be a service toolkit incompatibility that was subsequently resolved.
- Built test bed cluster/grid from scratch due to apparent incompatibility. Thus we currently have two different Globus installations. We will likely remove the Rocks installation in the future.
- Worked with Indiana University's (IU) Service Wrapper Toolkit and sample Decoder Service
- Got decoder service working on lead testbed under both Rocks and "from scratch" Globus installations.
- Reported decoder service install challenges to other lead tb sites and IU to allow for upgrades to the toolkit.
- Provided installation info to install gribtonc to other lead tb sites for use by decoder service.
- Worked with Weather Research and Forecasting Model (WRF)
- Installed WRF on the Unidata Test Bed system (note: only worked with PGI compiler, Intel is no good for our platform)
- Configured WRF to run twice daily over Boulder and place results into the LEAD THREDDS catalog.
- Updated LEAD THREDDS top level catalog (found at http://lead.unidata.ucar.edu:8080/thredds/topcatalog.xml) to include WRF data from DTC, NCAR, UIUC and UPC
- General
- Worked with Unidata staff to present a self-education session for other LEAD participants on the topics of THREDDS and OPeNDAP.
- Presented LEAD related papers at the Fall 2004 meeting of the American Geophysical Union (AGU) and the 2005 meeting of the American Meteorological Society (AMS).
- Demonstrated at the AMS an end to end system that involved running the LDM, decoders, THREDDS, and OPeNDAP on the UPC and other LEAD testbeds and visualizing the results in the IDV.
- Actively participated in LEAD meetings via the Access Grid. Also participated in revamping of these meetings to reduce their quantity while maintaining quality.
- Worked with LEAD PI to provide technical project management across institutions. This effort identified critical paths and critical pieces within those paths towards meeting our July site review deadline.
- Provided feedback on Portal Requirements document, Project Implementation Plan and fleshing out of Canonical Problem #3 in the LEAD architectural context.
- Provided agendas for and conducted significant percentage of LEAD meetings
Ongoing Efforts
- Working with LEAD team and THREDDS team in contributing to the definition of an interchange language for LEAD
metadata.
- Working with Indiana University to include adequate metadata in THREDDS catalogs
to support rich queries.
- Working with the University of Oklahoma to define NetCDF conventions
for their ARPS Data Analysis System (ADAS).
- Completing support for WRF NetCDF data in the IDV.
- Supported University of North Carolina in bringing up LDM, Decoders, THREDDS and OPeNDAP on their LEAD test bed.
- Supporting and participating in upcoming meetings: LEAD All-Hands in May, 2005 and the NSF Site Visit in July, 2005.
- Working with LEAD PI to provide technical project coordination across institutions.
- Working to configure WRF to run in multi-processor mode on the Unidata Test Bed cluster.
- Working to make the WRF runs "steerable" by creating a script that takes lat/lon information.
