Hi Jose,
I am working on an interface to produce IDV catalogs in an application
that has an embedded web server (based on tcl). Everything works except
when IDV goes to download a specific data file. At that point it complains
about the server not supporting the Bytes Ranges header. Indeed the web
server does not support it.
Before I decide what route to take, I would like to ask if
this will be a firm requirement of IDV, or are there plans to make
it optional and use that feature only if the web server supports it?
The Java/NetCDF layer that the IDV uses can read a remote netcdf file
via http if the server is capable of doing byte offsets. It allows the
IDV to essentially treat the remote file as a random access file. It
needs to do this because the files can be quite large and it wouldn't be
feasible to suck over the whole file.
Normally, for remote access to files, the IDV relies on the OpenDAP
protocol. Unidata provides 2 server applications that can serve up data
via opendap: TDS and RAMADDA:
http://www.unidata.ucar.edu/projects/THREDDS/tech/TDS.html
http://www.unidata.ucar.edu/software/ramadda/
The TDS has capabilites to aggregate separate files (e.g., multiple time
steps) as a single virtual dataset. RAMADDA has a broader range of
capabilities and runs in a variety of ways.
-Jeff
