Re: Last Call: 'The Plain SASL Mechanism' to Proposed Standard

Simon Josefsson <jas@extundo.com> Tue, 22 February 2005 21:13 UTC

Received: from above.proper.com (localhost.vpnc.org [127.0.0.1]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j1MLD6MJ091728; Tue, 22 Feb 2005 13:13:06 -0800 (PST) (envelope-from owner-ietf-sasl@mail.imc.org)
Received: (from majordom@localhost) by above.proper.com (8.12.11/8.12.9/Submit) id j1MLD6oK091726; Tue, 22 Feb 2005 13:13:06 -0800 (PST)
X-Authentication-Warning: above.proper.com: majordom set sender to owner-ietf-sasl@mail.imc.org using -f
Received: from yxa.extundo.com (root@178.230.13.217.in-addr.dgcsystems.net [217.13.230.178]) by above.proper.com (8.12.11/8.12.9) with ESMTP id j1MLD5mK091697 for <ietf-sasl@imc.org>; Tue, 22 Feb 2005 13:13:05 -0800 (PST) (envelope-from jas@extundo.com)
Received: from latte.josefsson.org (c494102a.s-bi.bostream.se [217.215.27.65]) (authenticated bits=0) by yxa.extundo.com (8.13.3/8.13.3/Debian-6) with ESMTP id j1MLCisf015724 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=OK); Tue, 22 Feb 2005 22:12:45 +0100
From: Simon Josefsson <jas@extundo.com>
To: "Kurt D. Zeilenga" <Kurt@OpenLDAP.org>
Cc: iesg@ietf.org, ietf-sasl@imc.org
Subject: Re: Last Call: 'The Plain SASL Mechanism' to Proposed Standard
References: <E1CziRz-0003w5-Hm@megatron.ietf.org> <ilu7jl0pulv.fsf@latte.josefsson.org> <6.2.0.14.0.20050222093430.05ec9e50@mail.openldap.org>
OpenPGP: id=B565716F; url=http://josefsson.org/key.txt
X-Hashcash: 1:21:050222:iesg@ietf.org::Mt7dhV81g75y3UC5:7FeG
X-Hashcash: 1:21:050222:kurt@openldap.org::oDVOZ83hc+EPXRqm:3na2
X-Hashcash: 1:21:050222:ietf-sasl@imc.org::YhtmhQ756dgjamAZ:g074
X-Hashcash: 1:21:050222:kurt@openldap.org::rDKDzr7PmPXqJvfw:1ZL6
Date: Tue, 22 Feb 2005 22:12:36 +0100
In-Reply-To: <6.2.0.14.0.20050222093430.05ec9e50@mail.openldap.org> (Kurt D. Zeilenga's message of "Tue, 22 Feb 2005 12:36:42 -0800")
Message-ID: <iluoeecmh8b.fsf@latte.josefsson.org>
User-Agent: Gnus/5.110003 (No Gnus v0.3) Emacs/22.0.50 (gnu/linux)
MIME-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
X-Spam-Status: No, score=0.1 required=5.0 tests=FORGED_RCVD_HELO autolearn=failed version=3.0.2
X-Spam-Checker-Version: SpamAssassin 3.0.2 (2004-11-16) on yxa-iv
X-Virus-Scanned: ClamAV version 0.81, clamav-milter version 0.81b on yxa.extundo.com
X-Virus-Status: Clean
Sender: owner-ietf-sasl@mail.imc.org
Precedence: bulk
List-Archive: <http://www.imc.org/ietf-sasl/mail-archive/>
List-ID: <ietf-sasl.imc.org>
List-Unsubscribe: <mailto:ietf-sasl-request@imc.org?body=unsubscribe>

"Kurt D. Zeilenga" <Kurt@OpenLDAP.org> writes:

> The requirement to use the prepared authcid in deriving of
> authorization identity when the client provided none is
> technical sound and quite appropriate.  Without this, a user
> (not doing proxy authorization) using two different clients,
> say one sending NFKC strings and one sending NFC, might end
> up different authorization identities.  This would be a bad
> thing.

Indeed, but that wouldn't occur with my proposal.

I'm saying that it is the application profile that should define the
form of authzid, and in particular, the application profile should say
how to derive (prepare) an authzid from authcid, when the authzid is
absent.

Setting authzid=SASLPrep(authid) inside a SASL mechanism, like PLAIN,
assume too much about the structure of authorization identities.

One point is that deriving a default authzid is not something that has
to be described by each and every SASL mechanism, especially if it
must always be the same algorithm, i.e., authzid=SASLprep(authid).
The application profiles should describe it, or the core document
could do it.

> The WG concluded that requiring the prepared form of
> the authcid be used in determining the authorization identity
> when no authzid field is absent is the best way to ensure
> interoperability here.

That is not in conflict with my proposal.  (As an aside, as I
understood it, the WG concluded to strongly suggest, but not require,
the use of SASLprep.)

I want the invocation of 'authzid=SASLprep(authcid)' derivation moved
from the PLAIN mechanism into the application profiles or core
document.

For most applications, the end result will be the same.

With my approach, PLAIN can be used by application profiles that have
special authzid structures; structures that is otherwise permitted by
the core SASL specification.

It seems to me that if the current text stand, PLAIN may invalidate
assumptions for application profiles that follow from the core
document.

>>They could be two different
>>Unicode representations of essentially the same string.  I.e., had
>>Unicode 4.0 NFKC been used on the two strings, the output would have
>>been identical.
>
> Well, if the server is applying Unicode 4.0 NFKC, then the
> server is obviously not using the RECOMMENDED preparation
> algorithm, SASLprep.  In such cases, the server needs to consider
> all the ramifications of not using the RECOMMENDED preparation
> algorithm, including ironing out all such details.

I was not talking about server using Unicode 4.0 NFKC.  I was talking
about two strings which would be identical after U4NFKC().

Thanks,
Simon