1. GADS: A Web Service for accessing large environmental data sets Jon Blower, Keith Haines, Adit Santokhee Reading e-Science Centre University of Reading http://www.resc.rdg.ac.uk resc@rdg.ac.uk

2. Background  At Reading we hold copies of various datasets (~2TB) –Mainly from models of oceans and atmosphere –Also some observational data (e.g. satellite data) –From Met Office, SOC, ECMWF, more  We serve these datasets to many end users –Scientists (1000s of hits per year) –Industry (e.g. British Maritime Technology)  Datasets are in a variety of formats –netCDF, GRIB, HDF, HDF5 …  Data do not conform to naming conventions –E.g. “temp” instead of “sea_water_potential_temperature”

3. Background (2)  There is a clear need to make access to these datasets easier –Users shouldn’t have to know details of how data are stored  Hence development of GADS (Grid Access Data Service)  Developed as part of GODIVA project –Grid for Ocean Diagnostics, Interactive Visualisation and Analysis –NERC e-Science pilot project  Originally developed by Woolf et al (2003)  Allows richer queries and more flexibility than DODS standard –Although we plan to implement a DODS translation layer

4. GODIVA Web Portal Allows users to interactively select data for download using a GUI Users can create movies on the fly cf. Live Access Server

5. Advantages of GADS  User’s don’t need to know anything about storage details  Can expose data with conventional names without changing data files  Users can choose their preferred data format, irrespective of how data are stored  Behaves as aggregation server –Delivers single file, even if original data spanned several files  Deployed as a Web Service –Can be called from any platform/language –Can be called programmatically (easily incorporated into larger systems), workflows –Java / Apache Axis / Tomcat

6. Architecture META- DATA FILES Metadata Manager Utility Metadata Interface dataQuery dataRequest GADS Web Service Client

7. Metadata structure

8. GADS Methods  dataQuery() is used for querying the data holdings –“What datasets are there?” –“What variables are there in the dataset X?”  dataRequest() is used for downloading data –User can choose the data format –Can easily download subsets of data –Uses start-stride-count semantics (familiar in community)  dataRequestNatural() –Same as dataRequest() but in natural units (degrees, metres …)

9. dataQuery – examples of use  dataQuery(dataset, variable, axis) – general form  dataQuery(“”, “”, “”) – gets all dataset names in the catalogue  dataQuery(“FOAM_NINTH”, “”, “”) – gets all the variable names in the FOAM_NINTH dataset  dataQuery(“FOAM_NINTH”, “temperature”, “”) – gets the details of the grid for the temperature variable  dataQuery(“FOAM_NINTH”, “temperature”, “z”) – gets all values that the z coordinate can take  dataQuery(“”, “temperature”, “”) – gets all datasets that contain the “temperature” variable

10. dataRequest – example of use  dataRequest(“FOAM_NINTH”, “temperature”, “CDF”, “t”, 0, 1, 20, “z”, 0, 1, -1, “y”, 100, 4, 400, “x”, 300, 4, 600)  dataRequestNatural(“FOAM_NINTH”, “temp”, “CDF”, “t”, “2004-06-01 00:00:00”, “2004-06-22 00:00:00”, “z”, “0”, “10”, “y”, “42”, “64”, “x”, “-26”, “9”)  Returns URL to extracted dataset

11. Metadata manager (in progress) e.g. Adding a dataset – can “harvest” metadata from netCDF file headers

12. Limitations  Assumes one timestep per file –Hence doesn’t handle timeseries well  Long queries can cause problems (synchronous) –Needs a queuing system  Rotated grids a problem (esp. for dataRequestNatural())  Could have richer metadata queries

13. Application: Search and Rescue  Search And Rescue Information System (SARIS) –British Maritime Technology (BMT)  Used by Coastguard to locate people who have fallen overboard  Runs a model using wind and surface current data –Forecasts where person will be by the time rescue arrives  By incorporating GADS, SARIS can consume up-to-date Met Office forecasts on demand. –Should improve quality of prediction

14. Spatial Databases  Database systems now including capability for storing geospatial data –IBM Informix, Oracle 10g, PostgreSQL, mySQL …  ReSC is evaluating some of these –Informix with Grid DataBlade looks promising (www.barrodale.com)  We need capability to store raster data (i.e. gridded data) –Many only store vector data –Gotcha – some vendors use “raster” to mean “photograph”, not “model data”  We also need to store 3-D data –Some only have native understanding of 2-D data

15. Future plans  Interact more with GIS community –There are already some relevant initiatives out there (e.g. MarineGIS) –Use of databases may help (some are OGC compliant) –But have problem that GIS tends to talk in 2-D  Develop DODS (=OpenDAP) layer  Encourage others to install GADS –We don’t want to hold lots of data in Reading! –POL, Met Office, ECMWF all expressed interest –Software needs “hardening” first…  Find more applications!