[clug] Unique Id's and CD's

Robert Edwards bob at cs.anu.edu.au
Fri May 8 02:14:54 GMT 2009

Jim Croft wrote:
> To hit an unintentional duplicate, you would have to be unlucky.
> "... only after generating 1 billion UUIDs every second for the next
> 100 years, the probability of creating just one duplicate would be
> about 50%. The probability of one duplicate would be about 50% if
> every person on earth owns 600 million UUIDs" (source: (cough)
> Wikip....)
> On Fri, May 8, 2009 at 9:08 AM, Robert Edwards <bob at cs.anu.edu.au> wrote:
>> If you only want IDs for data CDs that you produce yourself, and you
>> can live with someone else accidentally/deliberately copying the same
>> ID, then UUIDs work. You can also place an encrypted string with the
>> UUID, creation date, author etc. on the CD as well, to increase your
>> level of confidence.
> jim

By "accidentally", I meant accidentally copying the chunk of disk where
the "unique ID" (however it is generated) is being stored and putting
that onto a new CD with other data on it.

Eg. if the "unique ID" is stored in a hidden file on the CD somewhere
and someone does (eg.) "cp -a" to get a copy of the disk, then adds a
new file/changes some existing file/whatever and burns the result back
out onto a new CD, it could well end up with the same "unique ID" as
the original...

Adding a hash of the result of "ls -{whatever}" (as others have
suggested) would counteract such an "accidental" copying of the
"unique ID".

Of course, use of "ls -R..." won't protect you from someone
deliberately trying to create the same hash and hence the same "unique
ID", as "ls" only checks the file's metadata (inode data), which can
all be manipulated...


Bob Edwards.

More information about the linux mailing list