Adump Backup Suite

Issues I see:

There is no specified method of saying 'minumum version for extraction'.
Possibly it can be done by changing the group header tag. The version
doesn't seem to be useful for this as it's a text string.

The compression block size is limited to single files; even with gzip (but
more so with other algorithms) you get better ratios if you compress a
group of files
.

There is no metadata compression.

Can data and zdat blocks be mixed?

There is no support for larger CRCs on a group, however, an SHA1 tag will
probably fix this.

A NULL tag should be defined to allow for aliging of large data blocks
between input and output. This could improve archiving performance.

You haven't said what the restore should do with unknown tags and what
should happen with tags after the unknown one that appear to refer to tags
before it; but possibly do not.

Thats all off the top of my head.