@wiredfire @ernie How do you go about moving IA to a more federated or at least less centralised system? Figure the data for all of the archives must be into the high end of the petabyte range at this point.