All Archives / humbug / 1996-12
<<< Date >>>
Permanent Link
Date: Mon, 30 Dec 1996 09:36:48 +1000
From: Martin Pool
To: humbug-l
Message-Id: <199612292336.JAA00647@buffalo.pharos.com.au>
Subject: [humbug] fwd: Certified systems evaluated product list
No followups recorded.

------- Start of forwarded message -------
Of interest, perhaps

--/\/\----------
  Martin Pool                                     m . p o o l @ p h a r o s . c o m . a u
  Designer, Pharos Business Solutions                  +61-7-3222-2155
  "What part of mu don't you understand?"


---- forwarded message -----------

Return-Path: f i r e w a l l s - o w n e r @ G r e a t C i r c l e . C O M
Comments: Authenticated sender is <m j r @ m a i l . c l a r k . n e t
From: "Marcus J. Ranum" <m j r @ m a i l . c l a r k . n e t
Organization: V-ONE Corp Baltimore office
To: f i r e w a l l s @ G r e a t C i r c l e . C O M
Date: Sun, 29 Dec 1996 10:03:53 +0000
MIME-Version: 1.0
Content-type: text/plain; charset=US-ASCII
Content-transfer-encoding: 7BIT
Subject: Re:  EPL (Evaluated Products List)
X-mailer: Pegasus Mail for Win32 (v2.42a)
Sender: f i r e w a l l s - o w n e r @ G r e a t C i r c l e . C O M
Precedence: bulk

Gene Lee <g e n e l @ i n f o r a m p . n e t writes:
>From what I understand of the evaluation process, any change whatsoever to
> the original certified system, be it hardware, software changes or system
> configuration, totally invalidates the certification. This includes things
> like OS versions, the type of hardware running the code, the addition or
> removal of devices like floppy drives and network interfaces, and more
> importantly bug fixes or patches to the code (yeah, yeah, I know - If it's
> certified, why should there be any bugs?

OS' are not "certified" they are "evaluated" -- this is an important
distinction. It means "we looked at it" not "we guarantee it."
Even the orange book guys haven't got the chutzpah to certify
an operating system's security!

Yes, you're right, getting a product through evaluation is
insanely time-consuming and expensive. By the time it is
complete, the product is woefully obsolete. Take, for example,
Trusted Information Systems' Trusted Xenix (tm) -- remember
Xenix?? That's Version 7 UNIX. It's evaluated to run on an
80286. Remember those? We're talking museum piece
software running on museum piece hardware at art gallery
prices. Want X-windows? Forget it: X would break the B-level
trust model by having a device (the bitmap display) containing
multiple levels of data at once and that's a nono. I think that
there was a trusted X-windows effort that probably cost a
million bucks by the time it was done -- maybe by now they
have X11R2 completed.

Imagine doing software development by committee.
Now imagine that there are really 2 committees.
Next imagine the 2 committees are in an adversarial relationship.
Then imagine that one is a *GOVERNMENT* committee.
Remember that one of the committees likes $16,000 toilet seats.
Lastly assume that lots of the process (and programmers in
     the committees) need to be cleared for classified material.

Scared yet? You'd better be! Evaluation makes ISO standards
working groups look nimble by comparison. Cost effectiveness
isn't in the picture.

There's a process called RAMP (I forget what the acronym
stands for) in which you can formally tweak an evaluation
to keep it relevant across versions of the OS and platform.
I think, basically, that all the changes get reviewed. As if
someone actually knows what they are, when we're talking
modern hardware.

The evaluated system biz is the emperor's new clothes. The
government has spent tens (more) of millions of taxypayers'
dollars to design systems that are inherently obsolete and
practically unusable. The end result is that most computing
in the classified world is done on ordinary joe-bob systems
that are on carefully isolated (we hope!) networks. Hidden
away, you'll find SunOS, Windows, BSD, and so forth. But
since so much money has been spent on trusted systems,
nobody can admit that it's gone nowhere because that would
entail admitting that a huge amount of effort has been wasted.

mjr.
- -----
Marcus J. Ranum, Chief Scientist, V-ONE Corporation
Work:       http://www.v-one.com
Personal:   http://www.clark.net/pub/mjr
"I'll have time to be laid back when I'm laid out on a slab"
------- End of forwarded message -------

<<< Date >>>
This page was automatically generated, based on a complete record of postings made to the nominated list. Copyright issues, blame or gratitude belongs to the entity that wrote the content.