compare-by-hash (was Re: sharing /etc/passwd)

[Colin Percival]

Giorgos Keramidas wrote:
> Increasing the number of bits the hash key uses will decrease the
> possibility of a collision but never eliminate it entirely, AFAICT.

How small does a chance of error need to be before you're willing to 
ignore it?

> What I pointed out was that when a non-zero possibility of two data
> blocks comparing as equal (even though they are no) exists, the method
> in question should not be used for password data or other sensitive bits
> of information.  A larger hash key will never yield a possibility of
> zero, so it doesn't mean that you can sleep untroubled at night while
> the rsync server overwrites /etc/*pwd.db files periodically.

If an appropriately strong hash is used (eg, SHA1), then the probability 
of obtaining an incorrect /etc/*pwd.db with a correct hash is much 
smaller than the probability of a random incorrect password being 
accepted.  Remember, passwords are stored by their MD5 hashes, so a 
random password has a 2^(-128) chance of working.

If, on the other hand, you're concerned about accidentally locking 
yourself out of the server as a result of an undetected mangling of the 
password database... you should be more worried about the server, and 
all your backups, being simultaneously hit by lightning. :-)

Colin Percival