- starting from an flist copy files from one storage location (mycofs server) to other - efficient, tell all files which are needed (based on hashes), do in batches e.g. per 1000 obj (hashes) - starting from flist, check on source serve, do you have the objects, if not error - starting from flist, check on dest server, per 1000, what are you missing, the missing ones are efficiently uploaded (not one per one)