Long-term archiving of the survey data

How much data?
How would we understand the data?
File-level metadata (FITS) and other project output (TBD).
Why the library?
Because it was recognized that we're in it for the long haul.