HR: 1340h
AN: IN33B-1177    [Abstracts]
TI: Storing, Browsing, Querying, and Sharing Data: the THREDDS Data Repository (TDR)
AU: * Wilson, A
EM: anne@unidata.ucar.edu
AF: UCAR/Unidata, P.O. Box 3000, Boulder, CO 80307-3000 United States
AU: Lindholm, D
EM: lind@unidata.ucar.edu
AF: UCAR/Unidata, P.O. Box 3000, Boulder, CO 80307-3000 United States
AU: Baltzer, T
EM: tbaltzer@unidata.ucar.edu
AF: UCAR/Unidata, P.O. Box 3000, Boulder, CO 80307-3000 United States
AB: The Unidata Internet Data Distribution (IDD) network delivers gigabytes of data per day in near real time to sites across the U.S. and beyond. The THREDDS Data Server (TDS) supports public browsing of metadata and data access via OPeNDAP enabled URLs for datasets such as these. With such large quantities of data, sites generally employ a simple data management policy, keeping the data for a relatively short term on the order of hours to perhaps a week or two. In order to save interesting data in longer term storage and make it available for sharing, a user must move the data herself. In this case the user is responsible for determining where space is available, executing the data movement, generating any desired metadata, and setting access control to enable sharing. This task sequence is generally based on execution of a sequence of low level operating system specific commands with significant user involvement. The LEAD (Linked Environments for Atmospheric Discovery) project is building a cyberinfrastructure to support research and education in mesoscale meteorology. LEAD orchestrations require large, robust, and reliable storage with speedy access to stage data and store both intermediate and final results. These requirements suggest storage solutions that involve distributed storage, replication, and interfacing to archival storage systems such as mass storage systems and tape or removable disks. LEAD requirements also include metadata generation and access in order to support querying. In support of both THREDDS and LEAD requirements, Unidata is designing and prototyping the THREDDS Data Repository (TDR), a framework for a modular data repository to support distributed data storage and retrieval using a variety of back end storage media and interchangeable software components. The TDR interface will provide high level abstractions for long term storage, controlled, fast and reliable access, and data movement capabilities via a variety of technologies such as OPeNDAP and gridftp. The modular structure will allow substitution of software components so that both simple and complex storage media can be integrated into the repository. It will also allow integration of different varieties of supporting software. For example, if replication is desired, replica management could be handled via a simple hash table or a complex solution such as Replica Locater Service (RLS). In order to ensure that metadata is available for all the data in the repository, the TDR will also generate THREDDS metadata when necessary. Users will be able to establish levels of access control to their metadata and data. Coupled with a THREDDS Data Server, both browsing via THREDDS catalogs and querying capabilities will be supported. This presentation will describe the motivating factors, current status, and future plans of the TDR. References: IDD: http://www.unidata.ucar.edu/content/software/idd/index.html THREDDS: http://www.unidata.ucar.edu/content/projects/THREDDS/tech/server/ServerStatus.html LEAD: http://lead.ou.edu/ RLS: http://www.isi.edu/~annc/papers/chervenakRLSjournal05.pdf
DE: 0434 Data sets
DE: 0525 Data management
SC: Earth and Space Science Informatics [IN]
MN: Fall Meeting 2005