There is a "privacy hole" from using the MAC address. (Read it in the
WIKI article someone else posted).
Probably, it would be better to use a one way hash of the MAC address.
> -----Original Message-----
> From: pgsql-general-owner@postgresql.org [mailto:pgsql-general-
> owner@postgresql.org] On Behalf Of Chris Travers
> Sent: Wednesday, July 27, 2005 2:27 PM
> To: Magnus Hagander; pgsql-general
> Subject: Re: [GENERAL] GUID for postgreSQL
>
> Magnus Hagander wrote:
>
> >>
> >
> >At least on Windows, the GUID is derived in part from the computers
> >primary MAC address. No, it's not a guarantee, but it's pretty
unlikely
> >:-)
> >
> >
> The danger is not that the MAC address will be duplicated, but that
> other factors will lead to an MD5 collision.
>
> Unless you can show me that there is a 1:1 correspondence of all
> possible unique factors going into the GUID generation and the output,
> then I will say it is still no guarantee.
>
> Just because two documents or files have the same MD5 doesn't mean
that
> they are the same files either. I.e. you can't go searching all files
> by MD5 checksums and expecting to find the right one. OTOH, MD5
> provides reasonable assurance that any given file (once you know its
> intended MD5) has not been tampered with. I.e. MD5 is not meant to
> preclude collisions, but rather it is meant to preclude *intentional*
> collisions. Similarly, if we want a guaranteed uniqueness to a GUID
we
> have to have some sort of unique string to the GUID prepended to it
(not
> merely used in a hash).
>
> So you could use the Mac address of the machine, I guess, if you
wanted
> to....
>
> Best Wishes,
> Chris Travers
> Metatron Technology Consulting
>
> ---------------------------(end of
broadcast)---------------------------
> TIP 5: don't forget to increase your free space map settings
