Fundamental rsync design question.

[Biju Perumal]

Hi folks,

      Was looking at a commercial product which does something similar
to rsync.
It seems to optimise incremental change determination by logging the
changes that
have occured since the first full transfer happened, using techniques
common in snapshot techniques.

    Is there any reason for rsync not using this approach ? Is it the
space constraints
imposed by a logging approach or something ? Cos, this logging/snapshot
approach
definitely seems less cpu intensive, faster and more reliable (using a
checkpoint
algo).

    Any thoughts welcome,

Regards,
Biju Perumal