The R/GTS MMR system
In order to successfully exchange and manipulate data in a real
time data environment, RDAC and GDAC must know at any instant what data are available, where they are and what level
of processing they have completed. This is a particularly demanding task given the global distribution of
RDAC and GDAC that are working together in the GHRSST-PP Regional/Global Task sharing framework. In order to
address these issues, a global GHRSST-PP Master Metadata Repository (MMR) service has been developed.
Click the interface to image to go straight to the
GDAC MMR live system and start searching for GHRSST-PP data.
The MMR ensures that all GDS data products are "visible" to the
RDAC and GDAC centres at any given time. From a user perspective, the MMR provides a simple searchable
index to all GHRSST-PP data products, irrespective of where they are physically located.
The MMR is therefore the foundation of the GHRSST-PP data management framework. It provides a searchable
catalogue of the distributed GHRSST-PP data holdings providing information on its physical location, c
ontents and any constraints on its use. Using the results of a search submitted to the MMR system,
a user may access data sets directly via ftp/OPeNDAP/LAS. Without a MMR master catalogue, it would be
extremely difficult (if not impossible) to locate a single data resource within the GHRSST-PP without
physically connecting to the computers storing the data and searching each one individually.
The GHRSST-PP data files themselves have been chosen to follow the
Climate and Forecast netCDF convention
because these conventions provide a practical
standard for storing oceanographic data, and have already been adopted for the Data Sharing Pilot project within GODAE.
The global attributes of netCDF files are also metadata, since they follow the CF conventions, although they differ
slightly from the GHRSST-PP MMR standard. To a certain extent, this is reasonable, since the MMR is intended to
help users and processing centres search for data sets and to keep track of processing tasks,
while the netCDF attributes are to help users apply the data. The MMR is implemented as a physical relational database (MySQL) that contains metadata descriptions of all static
(i.e., data products) data holdings within the GHRSST-PP. Metadata describing each data product including are
stored within MMR database system in real time.