HR: 0830h
AN: NG11A-0177 [PDF]
TI: SCEC Community Modeling Environment (SCEC/CME) - Data and Metadata Management Issues
AU: * Minster, J
EM: jbminster@ucsd.edu
AF: Scripps Institution of Oceanography, University of California San Diego
9500 Gilman Drive, La Jolla, CA 92093-0225 United States
AU: Faerman, M
AF: San Diego SuperComputer Center, University of California San Diego
9500 Gilman Drive, La Jolla, CA 92093 United States
AU: Ely, G
AF: Scripps Institution of Oceanography, University of California San Diego
9500 Gilman Drive, La Jolla, CA 92093-0225 United States
AU: Maechling, P
AF: Southern California Earthquake Center, University of Southern California, Los Angeles, CA 90089-0740
United States
AU: Gupta, A
AF: Dept. of Geological Sciences, San Diego State University, San Diego, CA 92182 United States
AU: Xin, Q
AF: San Diego SuperComputer Center, University of California San Diego
9500 Gilman Drive, La Jolla, CA 92093 United States
AU: Kremenek, G
AF: San Diego SuperComputer Center, University of California San Diego
9500 Gilman Drive, La Jolla, CA 92093 United States
AU: Shkoller, B
AF: Scripps Institution of Oceanography, University of California San Diego
9500 Gilman Drive, La Jolla, CA 92093-0225 United States
AU: Olsen, K
AF: Institute of Crustal Studies, University of California Santa Barbara, Santa Barbara, CA 93106 United States
AU: Day, S
AF: Dept. of Geological Sciences, San Diego State University, San Diego, CA 92182 United States
AU: Moore, R
AF: San Diego SuperComputer Center, University of California San Diego
9500 Gilman Drive, La Jolla, CA 92093 United States
AB:
One of the goals of the SCEC Community Modeling Environment is to facilitate the execution of substantial collections of
large numerical simulations. Since such simulations are resource-intensive, and can generate extremely large outputs,
implementing this concept raises a host of data and metadata management challenges.
Due to the high computational cost involved in running these simulations, one must balance the cost of repeating such
simulations against the burden of archiving the produced datasets making them accessible for future use such as post
processing or visualization, without the need of re-computation. Further, a carefully selected collection of such data sets
might be used as benchmarks for assessing accuracy and performance of future simulations, developing post-processing software
such as visualization tools, and testing data and metadata management strategies. The problem is rapidly compounded if one
contemplates the possibility of computing ensemble averages for simulations of complex nonlinear systems. The definition and
organization of a complete set of metadata to describe fully any given simulation is a surprisingly complex task, which we
approach from the point of view of developing a community digital library, which provides the means to organize the material,
as well as standard metadata attributes. Web-based discovery mechanisms are then used to support browsing and retrieval of
data. A key component is the selection of appropriate descriptive metadata. We compare existing metadata standards from the
digital library community, federal standards, and discipline specific metadata attributes.
The digital library community has developed a standard for organizing metadata, called the Metadata Encoding and Transmission
Standard (METS). This schema supports descriptive (provenance), administrative (location), structural (component
relationships), and behavioral (display and manipulation applications). The organization can be augmented with discipline
specific extension schemata. Candidates include the FGDC spatial data standard, the ISO 19115 schema for geographic data,
and the Storage Resource Broker authenticity metadata. Other candidates include various metadata schemata used in
observational seismology. We are also considering metadata attributes that are being developed within the SCEC community and
are specific to the requirements of that community. A comparison of the metadata attributes will be presented, along with
their use in the organization of simulation output from a large-scale anelastic wave prediction simulation, The SDSC Storage
Resource Broker (SRB) provides the data handling capabilities to manage the Terabyte scale simulation output, providing
support for ingestion, organization, description, preservation and access of datasets. The metadata attributes include, in
particular, descriptive information about the simulation run, simulation input parameters, the computational infrastructure,
the physical geometry of the problem, and output structure.
DE: 3230 Numerical solutions
DE: 7209 Earthquake dynamics and mechanics
DE: 7212 Earthquake ground motions and engineering
DE: 7260 Theory and modeling
SC: Nonlinear Geophysics [NG]
MN: 2003 Fall Meeting