[582] in linux-security and linux-alert archive
Re: Proposal: s[gu]id standards.
daemon@ATHENA.MIT.EDU (Bruce Murphy)
Mon Jan 29 12:43:10 1996
To: Jeff Uphoff <juphoff@tarsier.cv.nrao.edu>
cc: linux-security@tarsier.cv.nrao.edu
In-reply-to: Your message of "Sun, 28 Jan 1996 14:21:55 EST."
<199601281921.OAA03394@tarsier.cv.nrao.edu>
Reply-To: packrat@tartarus.uwa.edu.au
Date: Mon, 29 Jan 1996 17:56:09 +0800
From: Bruce Murphy <packrat@ratbox.rattus.uwa.edu.au>
[mod: quoting trimmed. --okir]
In message <199601281921.OAA03394@tarsier.cv.nrao.edu>,
Jeff Uphoff wrote:
> However, IMO this should be more of an advisory list than any attempt at
> a standard....
{snip}
> The FSSTND group has already decided that this issue, in general, would
> not be addressed by them.
{snip}
>
> Now this isn't to say that it won't ever be done--it just says that the
> FSSTND currently considers this issue to be outside of its territory.
> For id specifications to be successful they really would have to be
> incorporated, eventually, into the FSSTND. (I personally think such an
> effort would be a waste of time, and thus agree with the FSSTND as it
> stands on this issue.)
>
The impact of such guidelines would only be on linux boxes that are
installed "out of the box" from distributions. Sites that have enough
on-hand unix experience to have s[ug]id policies wouldn't come into
this category.
A separate advisory list should be maintained independantly from the
FSSTND, as the goal is really to have a 'minimum' level of security
for out-of-box (or off cd) distributions, rather than specify
standards for linux systems which are installed.
Such a list would have two components, being a central repository of
the relative "safety" of various compents of the linux environment as
far as s[gu]id installations, and to allow a central resource to be
shown to distributors of Linux. Perhaps the two functions are mutally
incompatible enough that they should be maintained separately.
(motivation)
I would be interested in participating in any discussion lists that
are started on this topic and or people getting together with the goal
of putting one more piece of work towards linux being regarded as a
stable secure platform from which commercial things can be run without
having to write too much C code.
Or at least as much as other commerical Unices are considered safe and
stable. It's more about perception that fact in the commercial world
however.
Cheers
Bruce.
(Unsure whether this should have been posted to the linux-security but
have done so anyway, seeing as how you moderate it anyway ;)
--
Packrat (BSc/BE;COSO;Wombat Admin)
Nihil illegitemi carborvndvm.
