A Layered Approach to Data Access - How to Enable Enhanced VO Services

We describe a way to extend the 'small box VO' for time-series datasets in order to enable higher order data access services such as data browsing/visualization, sub-setting below the individual file level, format translation, and coordinate system conversions. As described and implemented by the Solar physics community in the Virtual Solar Observatory (VSO), the small box approach by itself focuses on file discovery - finding a list of files matching specific query criteria. This works well for the Solar community because most of the discovered files are in the well documented FITS format. In the realm of time-series space physics data however, there are numerous formats, and many data layouts even within each format. In order to truly provide seamless data access, a VO will need to be able to open up discovered files and extract data values from these files. We describe a layered approach to data access in which each type of science data is characterized by a specialized interface, and all extraction of data is performed through calls to the appropriate interface. This layered approach offers tremendous benefits long term, because the interface layer mimics the presence of a uniform file format and enables one set of tools (i.e. plotting, subsetting, translation, etc.) to work for all datasets. One such system implementing this approach will be described (http://sd-www.jhuapl.edu/MIDL), as will our future intentions for incorporating the data interface layer into emerging space physics VOs.