To: Neil Conway <neilc@samurai.com>
cc: Peter Eisentraut <peter_e@gmx.net>, Gevik Babakhani <pgdev@xs4all.nl>,
	pgsql-hackers@postgresql.org
Subject: Re: [DOCS] uuid type not documented 
In-reply-to: <1176233452.6558.52.camel@localhost.localdomain> 
References: <200704101724.27285.peter_e@gmx.net>
	<1176219030.6558.47.camel@localhost.localdomain>
	<200704101828.58118.peter_e@gmx.net>
	<1176233452.6558.52.camel@localhost.localdomain>
Comments: In-reply-to Neil Conway <neilc@samurai.com>
	message dated "Tue, 10 Apr 2007 15:30:52 -0400"
Date: Tue, 10 Apr 2007 17:25:05 -0400
Message-ID: <3597.1176240305@sss.pgh.pa.us>
From: Tom Lane <tgl@sss.pgh.pa.us>

Neil Conway <neilc@samurai.com> writes:
> On Tue, 2007-04-10 at 18:28 +0200, Peter Eisentraut wrote:
>> The problem is that most of the standard methods are platform dependent, as 
>> they require MAC addresses or a "good" random source, for instance.

> http://archives.postgresql.org/pgsql-patches/2007-01/msg00392.php

> ISTM random() or similar sources is a sufficient PSRNG for the purposes
> of UUID generation -- I can't see anything in the RFC that would
> contradict that.

Doesn't seem like quite enough bits of uniqueness.

We could improve matters by incorporating the database's
pg_control.system_identifier into the UUID, substituting for the MAC
address we don't have a good way to get.  system_identifier is currently
determined by the system clock at initdb time (to gettimeofday
precision), so it would add at least some additional bits of
randomness...

			regards, tom lane