Are the hex strings that make up the bulk of filenames random? Hopefully, with a flat distribution?
So, if I am handling a large backup, and it’s going in filename order, when I get to the first filename (out of, let’s say, 7,500) that begins with “duplicati-b20” I can be confident I have gone through about one-eighth of it? Liekwise, the first one that starts “duplicati-b80” flags the halfway point?
Yes, the part after “duplicati-b” is a random sequence of hexadecimal numbers. So I guess your assumption is correct.
Note that - when downloading files in filename order - the DBLOCK files will be downloaded first (character b in duplicati-b indicates this). Every DBLOCK file has an assiciated DINDEX file (starting with duplicati-i). These index files usually have a much smaller file size. So after downloading 2000 DBLOCK files, there are 2000 DINDEX files left to download.
Each backup version has one small DLIST file, so 120 DLIST files have to be downloaded if you have 120 backup versions.