
From klaas@cisco.com  Tue Aug  3 03:15:24 2010
Return-Path: <klaas@cisco.com>
X-Original-To: kitten@core3.amsl.com
Delivered-To: kitten@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id B0B6E3A6915 for <kitten@core3.amsl.com>; Tue,  3 Aug 2010 03:15:20 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -10.2
X-Spam-Level: 
X-Spam-Status: No, score=-10.2 tagged_above=-999 required=5 tests=[AWL=0.399,  BAYES_00=-2.599, RCVD_IN_DNSWL_HI=-8]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id i1ypxb5-Mz2O for <kitten@core3.amsl.com>; Tue,  3 Aug 2010 03:15:14 -0700 (PDT)
Received: from rtp-iport-2.cisco.com (rtp-iport-2.cisco.com [64.102.122.149]) by core3.amsl.com (Postfix) with ESMTP id 2B5763A6951 for <kitten@ietf.org>; Tue,  3 Aug 2010 03:15:13 -0700 (PDT)
Authentication-Results: rtp-iport-2.cisco.com; dkim=neutral (message not signed) header.i=none
X-IronPort-Anti-Spam-Filtered: true
X-IronPort-Anti-Spam-Result: AvsEAIuIV0xAZnwN/2dsb2JhbACgCnGmNJtTgnCCSQSJAg
X-IronPort-AV: E=Sophos;i="4.55,308,1278288000"; d="scan'208";a="142693950"
Received: from rtp-core-2.cisco.com ([64.102.124.13]) by rtp-iport-2.cisco.com with ESMTP; 03 Aug 2010 10:15:40 +0000
Received: from macmini.wierenga.net (ams-kwiereng-8715.cisco.com [10.55.220.246]) by rtp-core-2.cisco.com (8.13.8/8.14.3) with ESMTP id o73AFdtd027265; Tue, 3 Aug 2010 10:15:39 GMT
Message-ID: <4C57EC4A.6040405@cisco.com>
Date: Tue, 03 Aug 2010 12:15:38 +0200
From: Klaas Wierenga <klaas@cisco.com>
User-Agent: Mozilla/5.0 (Macintosh; U; Intel Mac OS X 10.6; en-GB; rv:1.9.1.11) Gecko/20100711 Thunderbird/3.0.6
MIME-Version: 1.0
To: kitten@ietf.org
Subject: difference between sasl-saml and sasl-saml-ec drafts (was: Re: Review of draft-wierenga-ietf-sasl-saml-01.txt:)
References: <4C4C4730.4040000@isode.com> <00f901cb2c33$16df2c70$449d8550$@osu.edu>
In-Reply-To: <00f901cb2c33$16df2c70$449d8550$@osu.edu>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit
X-BeenThere: kitten@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Common Authentication Technologies - Next Generation <kitten.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/kitten>
List-Post: <mailto:kitten@ietf.org>
List-Help: <mailto:kitten-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=subscribe>
X-List-Received-Date: Tue, 03 Aug 2010 10:15:25 -0000

Hi,

At the KITTEN WG meeting I noticed that below answer from Scott was not 
completely clear to everyone, so hear an attempt to give a bit more clarity:

SASL-SAML:

Here the assumption is that the SASL-client receives a SAML 
authentication requests and hands that to a browser that deals with the 
SAML authentication, the SASL client is NOT part of the authentication 
exchange anymore, apart from receiving the final ok from the SASL server 
(hence the empty reply to the challenge). The  result of the 
authentication at the SAML IdP is communicated to the SASL-server 
out-of-band.

+ minimal change to SASL client (just fire up a browser with the right URL)
+ no extra application that is used to enter authentication credentials
+ Doesn't need to touch the IdP

- rely on external program
- 'strange' user experience

SASL-SAML-EC

Here the assumption is that the SASL-client is ALSO a SAML client. So 
the SASL client receives the SAML authentication request and performs 
all the SAML magic itself, the Authentication Statement from the SASL 
server is send as the SASL response to the SASL servers challenge. In 
order to do so the SASL client must implement the SAML Enhanced Client 
or Proxy profile and the SAML IdP must support that.

+ 'cleaner' solution
+ doesn't require external program

- yet another piece of software to trust with user credentials
- IdP needs to support the ECP profile
- SASL client needs to implement the ECP profile

One way of merging these two proposals would be if in the SASL-SAML case 
the control can be handed back to the SASL client. Options like having 
the client listen on a port on localhost (essentially becoming a 
webserver) to which the browser is redirected back after authentication 
or serving out a MIME-type like x-sasl-jabber for which the SASL client 
is the handler come to mind, but they don't give me the warm fuzzy 
feeling yet...
It would however allow the SASL exchange be the same for both 
mechanisms, rather than out-of-band for one and in-band for the other, 
so ideas in that space are welcomed.

Hope this helps!

Klaas


On 7/25/10 9:53 PM, Scott Cantor wrote:
>> Why does IdP need to talk to RP (SASL server)? How would IdP identify
>> the SASL session to the SASL server?
>
> The IdP/RP communication is a result of using a particular SAML binding. I
> don't think this mechanism actually cares which binding is used between the
> web client anf the RP. It could be artifact, which results in a callback, or
> POST, which doesn't.
>
>> Can this be done through the client? The latter would be a natural data
>> flow for SASL.
>
> That's why I submitted an alternative mechanism, because I prefer that to
> using a web browser. The client can't do it here, because the SASL client
> has passed control to the browser, and the SAML response is handled by the
> browser, not the SASL client. That takes place via web, and the SASL server
> has to coordinate that with the web server.
>
>> 4.3.  Server Redirect
>>
>>     The SASL Server transmits a redirect to the IdP that the user
>>     provided, with a SAML authentication request in the form of a SAML
>>     assertion as one of the parameters.
>>
>> I assume that the redirect is an HTTP URI?
>> For interoperability this needs to be spelled out explicitly.
>
> The language is off. It's not a SAML assertion, it's a SAML AuthnRequest.
> The formal specification is the HTTP-Redirect binding.
>
> -- Scott
>
>
> _______________________________________________
> Kitten mailing list
> Kitten@ietf.org
> https://www.ietf.org/mailman/listinfo/kitten


From cantor.2@osu.edu  Tue Aug  3 08:14:20 2010
Return-Path: <cantor.2@osu.edu>
X-Original-To: kitten@core3.amsl.com
Delivered-To: kitten@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 2B3713A687A for <kitten@core3.amsl.com>; Tue,  3 Aug 2010 08:14:20 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -3.099
X-Spam-Level: 
X-Spam-Status: No, score=-3.099 tagged_above=-999 required=5 tests=[AWL=-0.500, BAYES_00=-2.599]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id ga7Dol+JLTQC for <kitten@core3.amsl.com>; Tue,  3 Aug 2010 08:14:17 -0700 (PDT)
Received: from defang1.it.ohio-state.edu (defang1.it.ohio-state.edu [128.146.216.81]) by core3.amsl.com (Postfix) with ESMTP id A7B5B3A68A7 for <kitten@ietf.org>; Tue,  3 Aug 2010 08:14:16 -0700 (PDT)
Received: from defang9.it.ohio-state.edu (defang9.it.ohio-state.edu [128.146.216.78]) by defang1.it.ohio-state.edu (8.13.7/8.13.1) with ESMTP id o73FEgrN021320; Tue, 3 Aug 2010 11:14:42 -0400
Received: from SNOWDOG (dhcp-128-146-28-191.osuwireless.ohio-state.edu [128.146.28.191]) by defang9.it.ohio-state.edu (8.13.7/8.13.1) with ESMTP id o73FEgid024977; Tue, 3 Aug 2010 11:14:42 -0400
From: "Scott Cantor" <cantor.2@osu.edu>
To: "'Klaas Wierenga'" <klaas@cisco.com>, <kitten@ietf.org>
References: <4C4C4730.4040000@isode.com> <00f901cb2c33$16df2c70$449d8550$@osu.edu> <4C57EC4A.6040405@cisco.com>
In-Reply-To: <4C57EC4A.6040405@cisco.com>
Subject: RE: difference between sasl-saml and sasl-saml-ec drafts (was: Re: Review of draft-wierenga-ietf-sasl-saml-01.txt:)
Date: Tue, 3 Aug 2010 11:14:53 -0400
Organization: The Ohio State University
Message-ID: <046c01cb331e$a013aab0$e03b0010$@osu.edu>
MIME-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Content-Transfer-Encoding: 7bit
X-Mailer: Microsoft Outlook 14.0
Thread-index: AQIlLOP+LmPyITL1dr5/rhKOaRD48QKtQPfAAYXEijWR+koSQA==
Content-language: en-us
X-CanIt-Geo: ip=128.146.216.78; country=US; region=OH; city=Columbus; latitude=39.9968; longitude=-82.9882; metrocode=535; areacode=614; http://maps.google.com/maps?q=39.9968,-82.9882&z=6
X-CanItPRO-Stream: outbound
X-Scanned-By: CanIt (www . roaringpenguin . com) on 128.146.216.81
X-BeenThere: kitten@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Common Authentication Technologies - Next Generation <kitten.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/kitten>
List-Post: <mailto:kitten@ietf.org>
List-Help: <mailto:kitten-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=subscribe>
X-List-Received-Date: Tue, 03 Aug 2010 15:14:20 -0000

> At the KITTEN WG meeting I noticed that below answer from Scott was not
> completely clear to everyone, so hear an attempt to give a bit more
clarity:

Couple of small corrections:

> SASL-SAML-EC
> 
> Here the assumption is that the SASL-client is ALSO a SAML client.

Technically "SAML-aware client", meaning it doesn't have to handle SAML
assertions or know how to parse any SAML messages, but plays a role in the
exchange. It's a SOAP intermediary, actually, not a SAML one.

> So the SASL client receives the SAML authentication request and performs
> all the SAML magic itself, the Authentication Statement from the SASL
> server is send as the SASL response to the SASL servers challenge.

s/from the SASL server/from the IdP

-- Scott



From mrex@sap.com  Wed Aug  4 16:21:12 2010
Return-Path: <mrex@sap.com>
X-Original-To: kitten@core3.amsl.com
Delivered-To: kitten@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 9B7D93A6A89 for <kitten@core3.amsl.com>; Wed,  4 Aug 2010 16:21:12 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -9.723
X-Spam-Level: 
X-Spam-Status: No, score=-9.723 tagged_above=-999 required=5 tests=[AWL=0.526,  BAYES_00=-2.599, HELO_EQ_DE=0.35, RCVD_IN_DNSWL_HI=-8]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id M21Fq6TBzzL0 for <kitten@core3.amsl.com>; Wed,  4 Aug 2010 16:21:11 -0700 (PDT)
Received: from smtpde02.sap-ag.de (smtpde02.sap-ag.de [155.56.68.140]) by core3.amsl.com (Postfix) with ESMTP id E0B7B3A6AB8 for <kitten@ietf.org>; Wed,  4 Aug 2010 16:21:09 -0700 (PDT)
Received: from mail.sap.corp by smtpde02.sap-ag.de (26) with ESMTP id o74NLZw5015932 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=OK); Thu, 5 Aug 2010 01:21:35 +0200 (MEST)
From: Martin Rex <mrex@sap.com>
Message-Id: <201008042321.o74NLYEd023603@fs4113.wdf.sap.corp>
Subject: Re: OpenID+GSS-API and naming: GSS_C_NT_OPENID?
To: simon@josefsson.org (Simon Josefsson)
Date: Thu, 5 Aug 2010 01:21:34 +0200 (MEST)
In-Reply-To: <87pqz0y7od.fsf@mocca.josefsson.org> from "Simon Josefsson" at Jul 6, 10 05:02:58 pm
MIME-Version: 1.0
Content-Type: text/plain; charset=ISO-8859-1
Content-Transfer-Encoding: 8bit
X-Scanner: Virus Scanner virwal07
X-SAP: out
Cc: kitten@ietf.org
X-BeenThere: kitten@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
Reply-To: mrex@sap.com
List-Id: Common Authentication Technologies - Next Generation <kitten.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/kitten>
List-Post: <mailto:kitten@ietf.org>
List-Help: <mailto:kitten-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=subscribe>
X-List-Received-Date: Wed, 04 Aug 2010 23:21:13 -0000

Simon Josefsson wrote:
> 
> I currently have the following text, is there any feedback from GSS-API
> people on it?
> 
> It seems useful to specify a GSS-API name attribute for the normalized
> OpenID identifier, if that is an appropriate use of name attributes.
> Thoughts?

I have not been following the details -- what exactly do you mean by
GSS-API name attribute (something that doesn't exist in the base GSS-API
architecture)?

There should be a mechanism-specific nametype defined for an OpenID
GSS-API mechanism along with a canonical representation (and guidance
on how to perform name canonicalization), so that applications can
create ACLs based on this and these ACLs remain independent of a
particular implementation (instead useable for all versions of
that mechanisms and ultimately for independent implementations
of the OpenID gssapi mechanism).

-Martin

From simon@josefsson.org  Wed Aug  4 16:34:52 2010
Return-Path: <simon@josefsson.org>
X-Original-To: kitten@core3.amsl.com
Delivered-To: kitten@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 326043A6AC3 for <kitten@core3.amsl.com>; Wed,  4 Aug 2010 16:34:52 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -102.705
X-Spam-Level: 
X-Spam-Status: No, score=-102.705 tagged_above=-999 required=5 tests=[AWL=-0.106, BAYES_00=-2.599, USER_IN_WHITELIST=-100]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id 9H7duVPpNZkH for <kitten@core3.amsl.com>; Wed,  4 Aug 2010 16:34:49 -0700 (PDT)
Received: from yxa-v.extundo.com (yxa-v.extundo.com [83.241.177.39]) by core3.amsl.com (Postfix) with ESMTP id 5C7E33A69EE for <kitten@ietf.org>; Wed,  4 Aug 2010 16:34:27 -0700 (PDT)
Received: from mocca (c80-216-27-64.bredband.comhem.se [80.216.27.64]) (authenticated bits=0) by yxa-v.extundo.com (8.14.3/8.14.3/Debian-5+lenny1) with ESMTP id o74NYlQ0027187 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES128-SHA bits=128 verify=NOT); Thu, 5 Aug 2010 01:34:48 +0200
From: Simon Josefsson <simon@josefsson.org>
To: mrex@sap.com
Subject: Re: OpenID+GSS-API and naming: GSS_C_NT_OPENID?
References: <201008042321.o74NLYEd023603@fs4113.wdf.sap.corp>
OpenPGP: id=B565716F; url=http://josefsson.org/key.txt
X-Hashcash: 1:22:100804:kitten@ietf.org::n0zKU30w+PUfkW+u:4CTt
X-Hashcash: 1:22:100804:klaas@cisco.com::UpnjcVyg5v8InLED:GIN8
X-Hashcash: 1:22:100804:mrex@sap.com::YbT6m8XVbC3/8d/6:RuCO
Date: Thu, 05 Aug 2010 01:34:45 +0200
In-Reply-To: <201008042321.o74NLYEd023603@fs4113.wdf.sap.corp> (Martin Rex's message of "Thu, 5 Aug 2010 01:21:34 +0200 (MEST)")
Message-ID: <877hk6rlyi.fsf@mocca.josefsson.org>
User-Agent: Gnus/5.110011 (No Gnus v0.11) Emacs/23.1 (gnu/linux)
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
X-Virus-Scanned: clamav-milter 0.96.1 at yxa-v
X-Virus-Status: Clean
Cc: kitten@ietf.org
X-BeenThere: kitten@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Common Authentication Technologies - Next Generation <kitten.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/kitten>
List-Post: <mailto:kitten@ietf.org>
List-Help: <mailto:kitten-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=subscribe>
X-List-Received-Date: Wed, 04 Aug 2010 23:34:52 -0000

Martin Rex <mrex@sap.com> writes:

> Simon Josefsson wrote:
>> 
>> I currently have the following text, is there any feedback from GSS-API
>> people on it?
>> 
>> It seems useful to specify a GSS-API name attribute for the normalized
>> OpenID identifier, if that is an appropriate use of name attributes.
>> Thoughts?
>
> I have not been following the details -- what exactly do you mean by
> GSS-API name attribute (something that doesn't exist in the base GSS-API
> architecture)?

See http://tools.ietf.org/html/draft-ietf-kitten-gssapi-naming-exts-08

I was thinking that one way for the application to learn the normalized
(by the mechanism) OpenID identifier could be through a name attribute.
I'm not certain this is necessary, and given the complexity and lack of
response on this topic, I have been considering to skip this for the
time being.  This aspect could be revisited in the future, if
applications wants a way to learn the normalized OpenID identifier.

I am also not certain this is an appropriate use of name attributes.

> There should be a mechanism-specific nametype defined for an OpenID
> GSS-API mechanism along with a canonical representation (and guidance
> on how to perform name canonicalization), so that applications can
> create ACLs based on this and these ACLs remain independent of a
> particular implementation (instead useable for all versions of
> that mechanisms and ultimately for independent implementations
> of the OpenID gssapi mechanism).

Ok.  Several people appears to believe a nametype for OpenID is the
right thing, so I think we should add it.

The name canonicalization rules are specified in the OpenID
specification, however some parts of it requires interaction with the
network, and is thus not trivial for an OpenID-unaware application to
do.  Hence my idea that it should be the OpenID mechanism that performs
the canonicalization and let the application know what the canonical
name is.  Does this make sense?

/Simon

From shawn.emery@oracle.com  Fri Aug  6 22:56:44 2010
Return-Path: <shawn.emery@oracle.com>
X-Original-To: kitten@core3.amsl.com
Delivered-To: kitten@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 2687F3A69C0; Fri,  6 Aug 2010 22:56:44 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -6.619
X-Spam-Level: 
X-Spam-Status: No, score=-6.619 tagged_above=-999 required=5 tests=[AWL=-0.021, BAYES_00=-2.599, HTML_MESSAGE=0.001, RCVD_IN_DNSWL_MED=-4]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id lqYlwpI0L44A; Fri,  6 Aug 2010 22:56:42 -0700 (PDT)
Received: from rcsinet10.oracle.com (rcsinet10.oracle.com [148.87.113.121]) by core3.amsl.com (Postfix) with ESMTP id E767A3A69F4; Fri,  6 Aug 2010 22:53:13 -0700 (PDT)
Received: from rcsinet15.oracle.com (rcsinet15.oracle.com [148.87.113.117]) by rcsinet10.oracle.com (Switch-3.4.2/Switch-3.4.2) with ESMTP id o775rWfQ017596 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=OK); Sat, 7 Aug 2010 05:53:33 GMT
Received: from acsmt353.oracle.com (acsmt353.oracle.com [141.146.40.153]) by rcsinet15.oracle.com (Switch-3.4.2/Switch-3.4.1) with ESMTP id o773X4iT006800; Sat, 7 Aug 2010 05:53:32 GMT
Received: from abhmt010.oracle.com by acsmt353.oracle.com with ESMTP id 493594131281160321; Fri, 06 Aug 2010 22:52:01 -0700
Received: from [10.7.250.156] (/10.7.250.156) by default (Oracle Beehive Gateway v4.0) with ESMTP ; Fri, 06 Aug 2010 22:52:01 -0700
Message-ID: <4C5CF47F.5040102@oracle.com>
Date: Fri, 06 Aug 2010 23:51:59 -0600
From: Shawn Emery <shawn.emery@oracle.com>
User-Agent: Mozilla/5.0 (X11; U; SunOS i86pc; en-US; rv:1.9.2.4) Gecko/20100610 Lightning/1.0b2 Thunderbird/3.1
MIME-Version: 1.0
To: "kitten@ietf.org" <kitten@ietf.org>, sasl@ietf.org
Subject: New Work Items - Kitten Recharter
Content-Type: multipart/alternative; boundary="------------060008070807000208090700"
X-BeenThere: kitten@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Common Authentication Technologies - Next Generation <kitten.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/kitten>
List-Post: <mailto:kitten@ietf.org>
List-Help: <mailto:kitten-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=subscribe>
X-List-Received-Date: Sat, 07 Aug 2010 05:56:44 -0000

This is a multi-part message in MIME format.
--------------060008070807000208090700
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit


We are looking for consensus on whether the Kitten WG should adopt the 
following drafts as work items:

     draft-cantor-ietf-sasl-saml-ec
     draft-mills-kitten-sasl-oauth

You can respond to lists (kitten or SASL), but please indicate your 
decision regardless if you are for or against the proposal.

If we do decide to shepherd these drafts then we will recharter again to 
include the following patch:

@@ -50,10 +50,14 @@

  * A SASL Mechanism for OpenID
      draft-lear-ietf-sasl-openid-00
  * A SASL Mechanism for SAML
      draft-wierenga-ietf-sasl-saml-00
+* A SASL Mechanism for SAML Enhanced Client
+    draft-cantor-ietf-sasl-saml-ec
+* A SASL Mechanism for OAuth
+    draft-mills-kitten-sasl-oauth

  The transition from SASL to GSS-API mechanisms will allow a greater set of
  applications to utilize said mechanisms with SASL implementations that 
support
  the use of GSS-API mechanisms in SASL (draft-ietf-sasl-gs2).

Shawn.
--

--------------060008070807000208090700
Content-Type: text/html; charset=ISO-8859-1
Content-Transfer-Encoding: 7bit

<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN">
<html>
  <head>

    <meta http-equiv="content-type" content="text/html; charset=ISO-8859-1">
  </head>
  <body text="#000000" bgcolor="#ffffff">
    <font size="+1"><tt><br>
        We are looking for consensus on whether the Kitten WG should
        adopt the following drafts as work items:<br>
        <br>
        &nbsp;&nbsp;&nbsp; draft-cantor-ietf-sasl-saml-ec<br>
        &nbsp;&nbsp;&nbsp; draft-mills-kitten-sasl-oauth<br>
        <br>
        You can respond to lists (kitten or SASL), but please indicate
        your decision regardless if you are for or against the proposal.<br>
        <br>
        If we do decide to shepherd these drafts then we will recharter
        again to include the following patch:<br>
        <br>
        @@ -50,10 +50,14 @@<br>
        &nbsp;<br>
        &nbsp;* A SASL Mechanism for OpenID<br>
        &nbsp;&nbsp;&nbsp;&nbsp; draft-lear-ietf-sasl-openid-00<br>
        &nbsp;* A SASL Mechanism for SAML<br>
        &nbsp;&nbsp;&nbsp;&nbsp; draft-wierenga-ietf-sasl-saml-00<br>
        +* A SASL Mechanism for SAML Enhanced Client<br>
        +&nbsp;&nbsp;&nbsp; draft-cantor-ietf-sasl-saml-ec<br>
        +* A SASL Mechanism for OAuth<br>
        +&nbsp;&nbsp;&nbsp; draft-mills-kitten-sasl-oauth<br>
        &nbsp;<br>
        &nbsp;The transition from SASL to GSS-API mechanisms will allow a
        greater set of<br>
        &nbsp;applications to utilize said mechanisms with SASL
        implementations that support<br>
        &nbsp;the use of GSS-API mechanisms in SASL (draft-ietf-sasl-gs2).<br>
        <br>
        Shawn.<br>
        --<br>
      </tt></font>
  </body>
</html>

--------------060008070807000208090700--

From shawn.emery@oracle.com  Fri Aug  6 23:32:04 2010
Return-Path: <shawn.emery@oracle.com>
X-Original-To: kitten@core3.amsl.com
Delivered-To: kitten@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 8D0603A6937; Fri,  6 Aug 2010 23:32:04 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -6.619
X-Spam-Level: 
X-Spam-Status: No, score=-6.619 tagged_above=-999 required=5 tests=[AWL=-0.020, BAYES_00=-2.599, RCVD_IN_DNSWL_MED=-4]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id Xky-TtPdNULJ; Fri,  6 Aug 2010 23:32:02 -0700 (PDT)
Received: from rcsinet10.oracle.com (rcsinet10.oracle.com [148.87.113.121]) by core3.amsl.com (Postfix) with ESMTP id C0EDD3A684D; Fri,  6 Aug 2010 23:32:02 -0700 (PDT)
Received: from rcsinet15.oracle.com (rcsinet15.oracle.com [148.87.113.117]) by rcsinet10.oracle.com (Switch-3.4.2/Switch-3.4.2) with ESMTP id o776WWo9008700 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=OK); Sat, 7 Aug 2010 06:32:34 GMT
Received: from acsmt353.oracle.com (acsmt353.oracle.com [141.146.40.153]) by rcsinet15.oracle.com (Switch-3.4.2/Switch-3.4.1) with ESMTP id o772Vtef020608; Sat, 7 Aug 2010 06:32:32 GMT
Received: from abhmt008.oracle.com by acsmt355.oracle.com with ESMTP id 493658761281162679; Fri, 06 Aug 2010 23:31:19 -0700
Received: from [10.7.250.156] (/10.7.250.156) by default (Oracle Beehive Gateway v4.0) with ESMTP ; Fri, 06 Aug 2010 23:31:18 -0700
Message-ID: <4C5CFDB5.5010307@oracle.com>
Date: Sat, 07 Aug 2010 00:31:17 -0600
From: Shawn Emery <shawn.emery@oracle.com>
User-Agent: Mozilla/5.0 (X11; U; SunOS i86pc; en-US; rv:1.9.2.4) Gecko/20100610 Lightning/1.0b2 Thunderbird/3.1
MIME-Version: 1.0
To: "kitten@ietf.org" <kitten@ietf.org>, sasl@ietf.org
Subject: WGLC on draft-ietf-kitten-digest-to-historic-00
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit
X-BeenThere: kitten@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Common Authentication Technologies - Next Generation <kitten.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/kitten>
List-Post: <mailto:kitten@ietf.org>
List-Help: <mailto:kitten-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=subscribe>
X-List-Received-Date: Sat, 07 Aug 2010 06:32:04 -0000

This message officially starts the Kitten Working Group Last Call for
the following document:

     Moving DIGEST-MD5 to Historic
     http://tools.ietf.org/html/draft-ietf-kitten-digest-to-historic-00

The Working Group Last Call for this document starts today on Saturday,
August 7th and will end on Saturday, August 21st.

Please send any comments to the Kitten or SASL mailing list or directly to
the chairs. Reviews that found no issues are also welcome, so if you
review the document and find it acceptable, please let the mailing
list or chairs know as well.

Thank you,

Tom Yu and Shawn Emery, Kitten WG chairs
--

From klaas@cisco.com  Sat Aug  7 00:22:04 2010
Return-Path: <klaas@cisco.com>
X-Original-To: kitten@core3.amsl.com
Delivered-To: kitten@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 71C1B3A6819; Sat,  7 Aug 2010 00:22:04 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -2.449
X-Spam-Level: 
X-Spam-Status: No, score=-2.449 tagged_above=-999 required=5 tests=[AWL=0.149,  BAYES_00=-2.599, HTML_MESSAGE=0.001]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id tFcCr8P8fMyh; Sat,  7 Aug 2010 00:22:03 -0700 (PDT)
Received: from teletubbie.het.net.je (teletubbie.het.net.je [IPv6:2001:610:508:110:192:87:110:29]) by core3.amsl.com (Postfix) with ESMTP id 0A60F3A686D; Sat,  7 Aug 2010 00:22:03 -0700 (PDT)
Received: from 64-103-25-233.cisco.com ([64.103.25.233] helo=[10.55.220.245]) by teletubbie.het.net.je with esmtpsa (TLSv1:AES128-SHA:128) (Exim 4.71 (FreeBSD)) (envelope-from <klaas@cisco.com>) id 1OhdjN-000AQx-7q; Sat, 07 Aug 2010 09:22:09 +0200
References: <4C5CF47F.5040102@oracle.com>
Message-Id: <56C516BE-37B4-4E34-B66F-EB92C7CB3FE2@cisco.com>
From: Klaas Wierenga <klaas@cisco.com>
To: Shawn Emery <shawn.emery@oracle.com>
In-Reply-To: <4C5CF47F.5040102@oracle.com>
Content-Type: multipart/alternative; boundary=Apple-Mail-5-63394487
Content-Transfer-Encoding: 7bit
X-Mailer: iPad Mail (7B405)
Mime-Version: 1.0 (iPad Mail 7B405)
Subject: Re: [sasl] New Work Items - Kitten Recharter
Date: Sat, 7 Aug 2010 09:23:08 +0200
X-Antivirus: no malware found
Cc: "kitten@ietf.org" <kitten@ietf.org>, "sasl@ietf.org" <sasl@ietf.org>
X-BeenThere: kitten@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Common Authentication Technologies - Next Generation <kitten.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/kitten>
List-Post: <mailto:kitten@ietf.org>
List-Help: <mailto:kitten-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=subscribe>
X-List-Received-Date: Sat, 07 Aug 2010 07:22:04 -0000

--Apple-Mail-5-63394487
Content-Type: text/plain;
	charset=us-ascii
Content-Transfer-Encoding: quoted-printable

I support adopting both drafts.

Klaas

Sent from my iPad

On Aug 7, 2010, at 7:57 AM, "Shawn Emery" <shawn.emery@oracle.com> =
wrote:

>=20
> We are looking for consensus on whether the Kitten WG should adopt the =
following drafts as work items:
>=20
>     draft-cantor-ietf-sasl-saml-ec
>     draft-mills-kitten-sasl-oauth
>=20
> You can respond to lists (kitten or SASL), but please indicate your =
decision regardless if you are for or against the proposal.
>=20
> If we do decide to shepherd these drafts then we will recharter again =
to include the following patch:
>=20
> @@ -50,10 +50,14 @@
> =20
>  * A SASL Mechanism for OpenID
>      draft-lear-ietf-sasl-openid-00
>  * A SASL Mechanism for SAML
>      draft-wierenga-ietf-sasl-saml-00
> +* A SASL Mechanism for SAML Enhanced Client
> +    draft-cantor-ietf-sasl-saml-ec
> +* A SASL Mechanism for OAuth
> +    draft-mills-kitten-sasl-oauth
> =20
>  The transition from SASL to GSS-API mechanisms will allow a greater =
set of
>  applications to utilize said mechanisms with SASL implementations =
that support
>  the use of GSS-API mechanisms in SASL (draft-ietf-sasl-gs2).
>=20
> Shawn.
> --
> _______________________________________________
> sasl mailing list
> sasl@ietf.org
> https://www.ietf.org/mailman/listinfo/sasl

--Apple-Mail-5-63394487
Content-Type: text/html;
	charset=utf-8
Content-Transfer-Encoding: 7bit

<html><body bgcolor="#FFFFFF"><div>I support adopting both drafts.</div><div><br></div><div>Klaas<br><br>Sent from my iPad</div><div><br>On Aug 7, 2010, at 7:57 AM, "Shawn Emery" &lt;<a href="mailto:shawn.emery@oracle.com">shawn.emery@oracle.com</a>&gt; wrote:<br><br></div><div></div><blockquote type="cite"><div>

    <font size="+1"><tt><br>
        We are looking for consensus on whether the Kitten WG should
        adopt the following drafts as work items:<br>
        <br>
        &nbsp;&nbsp;&nbsp; draft-cantor-ietf-sasl-saml-ec<br>
        &nbsp;&nbsp;&nbsp; draft-mills-kitten-sasl-oauth<br>
        <br>
        You can respond to lists (kitten or SASL), but please indicate
        your decision regardless if you are for or against the proposal.<br>
        <br>
        If we do decide to shepherd these drafts then we will recharter
        again to include the following patch:<br>
        <br>
        @@ -50,10 +50,14 @@<br>
        &nbsp;<br>
        &nbsp;* A SASL Mechanism for OpenID<br>
        &nbsp;&nbsp;&nbsp;&nbsp; draft-lear-ietf-sasl-openid-00<br>
        &nbsp;* A SASL Mechanism for SAML<br>
        &nbsp;&nbsp;&nbsp;&nbsp; draft-wierenga-ietf-sasl-saml-00<br>
        +* A SASL Mechanism for SAML Enhanced Client<br>
        +&nbsp;&nbsp;&nbsp; draft-cantor-ietf-sasl-saml-ec<br>
        +* A SASL Mechanism for OAuth<br>
        +&nbsp;&nbsp;&nbsp; draft-mills-kitten-sasl-oauth<br>
        &nbsp;<br>
        &nbsp;The transition from SASL to GSS-API mechanisms will allow a
        greater set of<br>
        &nbsp;applications to utilize said mechanisms with SASL
        implementations that support<br>
        &nbsp;the use of GSS-API mechanisms in SASL (draft-ietf-sasl-gs2).<br>
        <br>
        Shawn.<br>
        --<br>
      </tt></font>
  

</div></blockquote><blockquote type="cite"><div><span>_______________________________________________</span><br><span>sasl mailing list</span><br><span><a href="mailto:sasl@ietf.org">sasl@ietf.org</a></span><br><span><a href="https://www.ietf.org/mailman/listinfo/sasl">https://www.ietf.org/mailman/listinfo/sasl</a></span><br></div></blockquote></body></html>
--Apple-Mail-5-63394487--

From simon@josefsson.org  Sat Aug  7 04:30:34 2010
Return-Path: <simon@josefsson.org>
X-Original-To: kitten@core3.amsl.com
Delivered-To: kitten@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id D9A4D3A68B1; Sat,  7 Aug 2010 04:30:34 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -102.697
X-Spam-Level: 
X-Spam-Status: No, score=-102.697 tagged_above=-999 required=5 tests=[AWL=-0.098, BAYES_00=-2.599, USER_IN_WHITELIST=-100]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id tAE0W79Dv1zO; Sat,  7 Aug 2010 04:30:33 -0700 (PDT)
Received: from yxa-v.extundo.com (yxa-v.extundo.com [83.241.177.39]) by core3.amsl.com (Postfix) with ESMTP id B0E313A6846; Sat,  7 Aug 2010 04:30:32 -0700 (PDT)
Received: from mocca (c80-216-27-64.bredband.comhem.se [80.216.27.64]) (authenticated bits=0) by yxa-v.extundo.com (8.14.3/8.14.3/Debian-5+lenny1) with ESMTP id o77BUqOR016996 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES128-SHA bits=128 verify=NOT); Sat, 7 Aug 2010 13:30:55 +0200
From: Simon Josefsson <simon@josefsson.org>
To: Shawn Emery <shawn.emery@oracle.com>
Subject: Re: WGLC on draft-ietf-kitten-digest-to-historic-00
References: <4C5CFDB5.5010307@oracle.com>
OpenPGP: id=B565716F; url=http://josefsson.org/key.txt
X-Hashcash: 1:22:100807:sasl@ietf.org::Z0GNzBP1wLrOd2Mf:qIV
X-Hashcash: 1:22:100807:shawn.emery@oracle.com::Drad3FWVjr9JfpZv:2QMQ
X-Hashcash: 1:22:100807:kitten@ietf.org::JetaYbKEARqODLdn:TFsL
Date: Sat, 07 Aug 2010 13:30:48 +0200
In-Reply-To: <4C5CFDB5.5010307@oracle.com> (Shawn Emery's message of "Sat, 07 Aug 2010 00:31:17 -0600")
Message-ID: <877hk2pslz.fsf@mocca.josefsson.org>
User-Agent: Gnus/5.110011 (No Gnus v0.11) Emacs/23.1 (gnu/linux)
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
X-Virus-Scanned: clamav-milter 0.96.1 at yxa-v
X-Virus-Status: Clean
Cc: "kitten@ietf.org" <kitten@ietf.org>, sasl@ietf.org
X-BeenThere: kitten@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Common Authentication Technologies - Next Generation <kitten.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/kitten>
List-Post: <mailto:kitten@ietf.org>
List-Help: <mailto:kitten-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=subscribe>
X-List-Received-Date: Sat, 07 Aug 2010 11:30:35 -0000

I have re-read the document and this version looks good to me.  Two
comments below.

1)

I believe the document should contain a pointer to SCRAM, RFC 5802.
Then readers will understand that they are supposed to be implementing
SCRAM instead of DIGEST-MD5.  I suggest adding at the end of section 1:

  The SCRAM mechanism [RFC 5802] has been developed to provide similar
  features as DIGEST-MD5 but with a better design.

2)

In section 1, I would add the following bullet under 8:

  C.  The DES cipher for the security layer is considered insecure
      due to its small key space.

/Simon

From shawn.emery@oracle.com  Sat Aug  7 20:49:02 2010
Return-Path: <shawn.emery@oracle.com>
X-Original-To: kitten@core3.amsl.com
Delivered-To: kitten@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 4387928C132; Sat,  7 Aug 2010 20:49:02 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -6.618
X-Spam-Level: 
X-Spam-Status: No, score=-6.618 tagged_above=-999 required=5 tests=[AWL=-0.020, BAYES_00=-2.599, HTML_MESSAGE=0.001, RCVD_IN_DNSWL_MED=-4]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id v2JIbwHgZ0iW; Sat,  7 Aug 2010 20:49:01 -0700 (PDT)
Received: from rcsinet10.oracle.com (rcsinet10.oracle.com [148.87.113.121]) by core3.amsl.com (Postfix) with ESMTP id 1D7DB28C12F; Sat,  7 Aug 2010 20:49:01 -0700 (PDT)
Received: from acsinet15.oracle.com (acsinet15.oracle.com [141.146.126.227]) by rcsinet10.oracle.com (Switch-3.4.2/Switch-3.4.2) with ESMTP id o783nWNr003212 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=OK); Sun, 8 Aug 2010 03:49:33 GMT
Received: from acsmt353.oracle.com (acsmt353.oracle.com [141.146.40.153]) by acsinet15.oracle.com (Switch-3.4.2/Switch-3.4.1) with ESMTP id o782cFEl026326; Sun, 8 Aug 2010 03:49:32 GMT
Received: from abhmt002.oracle.com by acsmt353.oracle.com with ESMTP id 476269961281239306; Sat, 07 Aug 2010 20:48:26 -0700
Received: from [10.7.250.156] (/10.7.250.156) by default (Oracle Beehive Gateway v4.0) with ESMTP ; Sat, 07 Aug 2010 20:48:26 -0700
Message-ID: <4C5E2908.10507@oracle.com>
Date: Sat, 07 Aug 2010 21:48:24 -0600
From: Shawn Emery <shawn.emery@oracle.com>
User-Agent: Mozilla/5.0 (X11; U; SunOS i86pc; en-US; rv:1.9.2.4) Gecko/20100610 Lightning/1.0b2 Thunderbird/3.1
MIME-Version: 1.0
To: "kitten@ietf.org" <kitten@ietf.org>, sasl@ietf.org
Subject: IETF 78 - Kitten Minutes
Content-Type: multipart/alternative; boundary="------------010706060103010601000605"
X-BeenThere: kitten@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Common Authentication Technologies - Next Generation <kitten.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/kitten>
List-Post: <mailto:kitten@ietf.org>
List-Help: <mailto:kitten-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=subscribe>
X-List-Received-Date: Sun, 08 Aug 2010 03:49:02 -0000

This is a multi-part message in MIME format.
--------------010706060103010601000605
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit


I've uploaded the session minutes for Kitten:

http://www.ietf.org/proceedings/78/minutes/kitten.txt

Please let us know if you have any updates/corrections.

Shawn.
--

--------------010706060103010601000605
Content-Type: text/html; charset=ISO-8859-1
Content-Transfer-Encoding: 7bit

<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN">
<html>
  <head>

    <meta http-equiv="content-type" content="text/html; charset=ISO-8859-1">
  </head>
  <body text="#000000" bgcolor="#ffffff">
    <font size="+1"><tt><br>
        I've uploaded the session minutes for Kitten:<br>
        <br>
        <a class="moz-txt-link-freetext" href="http://www.ietf.org/proceedings/78/minutes/kitten.txt">http://www.ietf.org/proceedings/78/minutes/kitten.txt</a><br>
        <br>
        Please let us know if you have any updates/corrections.<br>
        <br>
        Shawn.<br>
        --<br>
      </tt></font>
  </body>
</html>

--------------010706060103010601000605--

From alexey.melnikov@isode.com  Mon Aug  9 01:15:51 2010
Return-Path: <alexey.melnikov@isode.com>
X-Original-To: kitten@core3.amsl.com
Delivered-To: kitten@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id B2B803A6891; Mon,  9 Aug 2010 01:15:51 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -101.53
X-Spam-Level: 
X-Spam-Status: No, score=-101.53 tagged_above=-999 required=5 tests=[AWL=-0.790, BAYES_20=-0.74, USER_IN_WHITELIST=-100]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id S4VoVTzHosyP; Mon,  9 Aug 2010 01:15:50 -0700 (PDT)
Received: from rufus.isode.com (rufus.isode.com [62.3.217.251]) by core3.amsl.com (Postfix) with ESMTP id 59BF13A685A; Mon,  9 Aug 2010 01:15:50 -0700 (PDT)
Received: from [10.0.0.111] (204.80-202-212.nextgentel.com [80.202.212.204])  by rufus.isode.com (submission channel) via TCP with ESMTPA  id <TF-5VgBSpcXx@rufus.isode.com>; Mon, 9 Aug 2010 09:16:23 +0100
Message-ID: <4C5FB94F.4080003@isode.com>
Date: Mon, 09 Aug 2010 09:16:15 +0100
From: Alexey Melnikov <alexey.melnikov@isode.com>
User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.7.12) Gecko/20050915
X-Accept-Language: en-us, en
To: Simon Josefsson <simon@josefsson.org>
Subject: Re: [sasl] WGLC on draft-ietf-kitten-digest-to-historic-00
References: <4C5CFDB5.5010307@oracle.com> <877hk2pslz.fsf@mocca.josefsson.org>
In-Reply-To: <877hk2pslz.fsf@mocca.josefsson.org>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii; format=flowed
Content-Transfer-Encoding: 7bit
Cc: "kitten@ietf.org" <kitten@ietf.org>, sasl@ietf.org
X-BeenThere: kitten@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Common Authentication Technologies - Next Generation <kitten.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/kitten>
List-Post: <mailto:kitten@ietf.org>
List-Help: <mailto:kitten-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=subscribe>
X-List-Received-Date: Mon, 09 Aug 2010 08:15:51 -0000

Simon Josefsson wrote:

>I have re-read the document and this version looks good to me.  Two
>comments below.
>
>1)
>
>I believe the document should contain a pointer to SCRAM, RFC 5802.
>Then readers will understand that they are supposed to be implementing
>SCRAM instead of DIGEST-MD5.  I suggest adding at the end of section 1:
>
>  The SCRAM mechanism [RFC 5802] has been developed to provide similar
>  features as DIGEST-MD5 but with a better design.
>
>2)
>
>In section 1, I would add the following bullet under 8:
>
>  C.  The DES cipher for the security layer is considered insecure
>      due to its small key space.
>
I am happy to add these. Tim Polk also suggested 1) to me verbally.


From alexey.melnikov@isode.com  Mon Aug  9 01:35:40 2010
Return-Path: <alexey.melnikov@isode.com>
X-Original-To: kitten@core3.amsl.com
Delivered-To: kitten@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id A7E633A6AA3; Mon,  9 Aug 2010 01:35:40 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -102.196
X-Spam-Level: 
X-Spam-Status: No, score=-102.196 tagged_above=-999 required=5 tests=[AWL=0.403, BAYES_00=-2.599, USER_IN_WHITELIST=-100]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id 8A+mq9mjumND; Mon,  9 Aug 2010 01:35:39 -0700 (PDT)
Received: from rufus.isode.com (rufus.isode.com [62.3.217.251]) by core3.amsl.com (Postfix) with ESMTP id B51E83A6A99; Mon,  9 Aug 2010 01:35:39 -0700 (PDT)
Received: from [10.0.0.111] (204.80-202-212.nextgentel.com [80.202.212.204])  by rufus.isode.com (submission channel) via TCP with ESMTPA  id <TF-9-gBSpUy0@rufus.isode.com>; Mon, 9 Aug 2010 09:36:12 +0100
Message-ID: <4C5FBDF0.9010909@isode.com>
Date: Mon, 09 Aug 2010 09:36:00 +0100
From: Alexey Melnikov <alexey.melnikov@isode.com>
User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.7.12) Gecko/20050915
X-Accept-Language: en-us, en
To: Shawn Emery <shawn.emery@oracle.com>
Subject: Re: [sasl] New Work Items - Kitten Recharter
References: <4C5CF47F.5040102@oracle.com>
In-Reply-To: <4C5CF47F.5040102@oracle.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit
Cc: "kitten@ietf.org" <kitten@ietf.org>, sasl@ietf.org
X-BeenThere: kitten@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Common Authentication Technologies - Next Generation <kitten.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/kitten>
List-Post: <mailto:kitten@ietf.org>
List-Help: <mailto:kitten-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=subscribe>
X-List-Received-Date: Mon, 09 Aug 2010 08:35:40 -0000

Shawn Emery wrote:

> We are looking for consensus on whether the Kitten WG should adopt the 
> following drafts as work items:
>
>     draft-cantor-ietf-sasl-saml-ec
>     draft-mills-kitten-sasl-oauth
>
> You can respond to lists (kitten or SASL), but please indicate your 
> decision regardless if you are for or against the proposal.
>
> If we do decide to shepherd these drafts then we will recharter again 
> to include the following patch:
>
> @@ -50,10 +50,14 @@
>  
>  * A SASL Mechanism for OpenID
>      draft-lear-ietf-sasl-openid-00
>  * A SASL Mechanism for SAML
>      draft-wierenga-ietf-sasl-saml-00
> +* A SASL Mechanism for SAML Enhanced Client
> +    draft-cantor-ietf-sasl-saml-ec

<hat type="AreaDirector">
I think IESG might want to know why there are 2 separate SAML 
mechanisms. While I can be convinced that both of them might be needed, 
the WG should have some good arguments to defend this.
</hat>

<hat type="IndividualContributor">
I think both of the documents should be discussed in the WG.
</hat>

> +* A SASL Mechanism for OAuth
> +    draft-mills-kitten-sasl-oauth

<hat type="IndividualContributor">
I am also happy to take this one as a WG item.
</hat>

>  The transition from SASL to GSS-API mechanisms will allow a greater 
> set of
>  applications to utilize said mechanisms with SASL implementations 
> that support
>  the use of GSS-API mechanisms in SASL (draft-ietf-sasl-gs2).
>
> Shawn.



From lear@cisco.com  Mon Aug  9 02:12:17 2010
Return-Path: <lear@cisco.com>
X-Original-To: kitten@core3.amsl.com
Delivered-To: kitten@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 4305A3A6AB7; Mon,  9 Aug 2010 02:12:17 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -110.398
X-Spam-Level: 
X-Spam-Status: No, score=-110.398 tagged_above=-999 required=5 tests=[AWL=0.200, BAYES_00=-2.599, HTML_MESSAGE=0.001, RCVD_IN_DNSWL_HI=-8, USER_IN_WHITELIST=-100]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id dDpxojSNTvJk; Mon,  9 Aug 2010 02:12:15 -0700 (PDT)
Received: from rtp-iport-1.cisco.com (rtp-iport-1.cisco.com [64.102.122.148]) by core3.amsl.com (Postfix) with ESMTP id 849E63A68C3; Mon,  9 Aug 2010 02:12:15 -0700 (PDT)
Authentication-Results: rtp-iport-1.cisco.com; dkim=neutral (message not signed) header.i=none
X-IronPort-Anti-Spam-Filtered: true
X-IronPort-Anti-Spam-Result: Av0EAD5jX0xAZnwN/2dsb2JhbACDFZ0vcac7iVKRD4RHcwSJOw
X-IronPort-AV: E=Sophos;i="4.55,341,1278288000";  d="scan'208,217";a="145158696"
Received: from rtp-core-2.cisco.com ([64.102.124.13]) by rtp-iport-1.cisco.com with ESMTP; 09 Aug 2010 09:12:49 +0000
Received: from dhcp-10-61-97-133.cisco.com (dhcp-10-61-97-133.cisco.com [10.61.97.133]) by rtp-core-2.cisco.com (8.13.8/8.14.3) with ESMTP id o799CmGa000905; Mon, 9 Aug 2010 09:12:48 GMT
Message-ID: <4C5FC699.8060902@cisco.com>
Date: Mon, 09 Aug 2010 11:12:57 +0200
From: Eliot Lear <lear@cisco.com>
User-Agent: Mozilla/5.0 (Macintosh; U; Intel Mac OS X 10.6; en-US; rv:1.9.2.8) Gecko/20100802 Thunderbird/3.1.2
MIME-Version: 1.0
To: Shawn Emery <shawn.emery@oracle.com>
Subject: Re: [sasl] New Work Items - Kitten Recharter
References: <4C5CF47F.5040102@oracle.com>
In-Reply-To: <4C5CF47F.5040102@oracle.com>
X-Enigmail-Version: 1.1.1
Content-Type: multipart/alternative; boundary="------------070407040704040805060506"
Cc: "kitten@ietf.org" <kitten@ietf.org>, sasl@ietf.org
X-BeenThere: kitten@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Common Authentication Technologies - Next Generation <kitten.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/kitten>
List-Post: <mailto:kitten@ietf.org>
List-Help: <mailto:kitten-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=subscribe>
X-List-Received-Date: Mon, 09 Aug 2010 09:12:17 -0000

This is a multi-part message in MIME format.
--------------070407040704040805060506
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 7bit


>
> We are looking for consensus on whether the Kitten WG should adopt the
> following drafts as work items:
>
>     draft-cantor-ietf-sasl-saml-ec
>     draft-mills-kitten-sasl-oauth
>
> You can respond to lists (kitten or SASL), but please indicate your
> decision regardless if you are for or against the proposal.

I support the inclusion of both drafts in our milestones.  To answer
Alexey's reaonable question, draft-wierenga-ietf-sasl-saml attempts to
rely on existing infrastructure as much as possible, both on the client
and on the IdP, while requiring some substantial changes to the RP.
  draft-cantor-ietf-sasl-saml-ec requires more substantial changes to
the client, but keeps SAML within the application protocol.  The
implication of the design choices relate more to how actual
authentication gets performed between the SASL client and the IdP.
 Existing IdPs make use of HTTP/HTML and unstructured exchanges for that
authentication.

In Scott's draft, that occurs in step (4).  This requires the client to
have substantially more capabilities than it might have today with
either a fully functional web browser either built into the application
or tied to the application via some form of IPC with sufficient semantic
abilities to discern when to move through step 4 to step 5, but at the
same time, provides for an overall simpler protocol flow than the
document that Klaas and I have put forth.  

Discovery is also handled in Scott's draft.  That is something that we
should consider incorporating into the other.

Eliot


--------------070407040704040805060506
Content-Type: text/html; charset=UTF-8
Content-Transfer-Encoding: 8bit

<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN">
<html>
  <head>
    <meta content="text/html; charset=UTF-8" http-equiv="Content-Type">
  </head>
  <body bgcolor="#ffffff" text="#000000">
    <blockquote cite="mid:4C5CF47F.5040102@oracle.com" type="cite">
      <meta http-equiv="content-type" content="text/html; charset=UTF-8">
      <font size="+1"><tt><br>
          We are looking for consensus on whether the Kitten WG should
          adopt the following drafts as work items:<br>
          <br>
              draft-cantor-ietf-sasl-saml-ec<br>
              draft-mills-kitten-sasl-oauth<br>
          <br>
          You can respond to lists (kitten or SASL), but please indicate
          your decision regardless if you are for or against the
          proposal.<br>
        </tt></font></blockquote>
    <br>
    I support the inclusion of both drafts in our milestones.  To answer
    Alexey's reaonable question, draft-wierenga-ietf-sasl-saml attempts
    to rely on existing infrastructure as much as possible, both on the
    client and on the IdP, while requiring some substantial changes to
    the RP.   draft-cantor-ietf-sasl-saml-ec requires more substantial
    changes to the client, but keeps SAML within the application
    protocol.  The implication of the design choices relate more to how
    actual authentication gets performed between the SASL client and the
    IdP.  Existing IdPs make use of HTTP/HTML and unstructured exchanges
    for that authentication.<br>
    <br>
    In Scott's draft, that occurs in step (4).  This requires the client
    to have substantially more capabilities than it might have today
    with either a fully functional web browser either built into the
    application or tied to the application via some form of IPC with
    sufficient semantic abilities to discern when to move through step 4
    to step 5, but at the same time, provides for an overall simpler
    protocol flow than the document that Klaas and I have put forth.  <br>
    <br>
    Discovery is also handled in Scott's draft.  That is something that
    we should consider incorporating into the other.<br>
    <br>
    Eliot<br>
    <br>
  </body>
</html>

--------------070407040704040805060506--

From alexey.melnikov@isode.com  Mon Aug  9 05:58:37 2010
Return-Path: <alexey.melnikov@isode.com>
X-Original-To: kitten@core3.amsl.com
Delivered-To: kitten@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 420363A6ACC; Mon,  9 Aug 2010 05:58:37 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -102.206
X-Spam-Level: 
X-Spam-Status: No, score=-102.206 tagged_above=-999 required=5 tests=[AWL=0.393, BAYES_00=-2.599, USER_IN_WHITELIST=-100]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id hL1HzKTX8gVO; Mon,  9 Aug 2010 05:58:36 -0700 (PDT)
Received: from rufus.isode.com (rufus.isode.com [62.3.217.251]) by core3.amsl.com (Postfix) with ESMTP id D73F33A69B0; Mon,  9 Aug 2010 05:58:35 -0700 (PDT)
Received: from [10.0.0.112] (204.80-202-212.nextgentel.com [80.202.212.204])  by rufus.isode.com (submission channel) via TCP with ESMTPA  id <TF=7mwBSpVaQ@rufus.isode.com>; Mon, 9 Aug 2010 13:59:08 +0100
Message-ID: <4C5FFB8F.6030406@isode.com>
Date: Mon, 09 Aug 2010 14:58:55 +0200
From: Alexey Melnikov <alexey.melnikov@isode.com>
User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.7.12) Gecko/20050915
X-Accept-Language: en-us, en
To: Eliot Lear <lear@cisco.com>
Subject: Re: [sasl] New Work Items - Kitten Recharter
References: <4C5CF47F.5040102@oracle.com> <4C5FC699.8060902@cisco.com>
In-Reply-To: <4C5FC699.8060902@cisco.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8; format=flowed
Content-Transfer-Encoding: 7bit
Cc: "kitten@ietf.org" <kitten@ietf.org>, sasl@ietf.org
X-BeenThere: kitten@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Common Authentication Technologies - Next Generation <kitten.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/kitten>
List-Post: <mailto:kitten@ietf.org>
List-Help: <mailto:kitten-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=subscribe>
X-List-Received-Date: Mon, 09 Aug 2010 12:58:37 -0000

Hi Eliot,

Eliot Lear wrote:

>> We are looking for consensus on whether the Kitten WG should adopt 
>> the following drafts as work items:
>>
>>     draft-cantor-ietf-sasl-saml-ec
>>     draft-mills-kitten-sasl-oauth
>>
>> You can respond to lists (kitten or SASL), but please indicate your 
>> decision regardless if you are for or against the proposal.
>
> I support the inclusion of both drafts in our milestones.  To answer 
> Alexey's reaonable question, draft-wierenga-ietf-sasl-saml attempts to 
> rely on existing infrastructure as much as possible, both on the 
> client and on the IdP, while requiring some substantial changes to the 
> RP.   draft-cantor-ietf-sasl-saml-ec requires more substantial changes 
> to the client, but keeps SAML within the application protocol.  The 
> implication of the design choices relate more to how actual 
> authentication gets performed between the SASL client and the IdP. 
>  Existing IdPs make use of HTTP/HTML and unstructured exchanges for 
> that authentication.
>
> In Scott's draft, that occurs in step (4).  This requires the client 
> to have substantially more capabilities than it might have today with 
> either a fully functional web browser either built into the 
> application or tied to the application via some form of IPC with 
> sufficient semantic abilities to discern when to move through step 4 
> to step 5, but at the same time, provides for an overall simpler 
> protocol flow than the document that Klaas and I have put forth.

Yes, I've seen this argument on the mailing list. Two documents suggest 
reasonable solutions for the assumption made. However I am not yet fully 
convinced that the assumption is something important enough to warrant 
having 2 documents.

> Discovery is also handled in Scott's draft.  That is something that we 
> should consider incorporating into the other.

Right.


From hannes.tschofenig@gmx.net  Mon Aug  9 06:37:18 2010
Return-Path: <hannes.tschofenig@gmx.net>
X-Original-To: kitten@core3.amsl.com
Delivered-To: kitten@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 681353A695A for <kitten@core3.amsl.com>; Mon,  9 Aug 2010 06:37:18 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -102.598
X-Spam-Level: 
X-Spam-Status: No, score=-102.598 tagged_above=-999 required=5 tests=[BAYES_00=-2.599, HTML_MESSAGE=0.001, USER_IN_WHITELIST=-100]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id TCabe+JkIlNL for <kitten@core3.amsl.com>; Mon,  9 Aug 2010 06:37:17 -0700 (PDT)
Received: from mail.gmx.net (mailout-de.gmx.net [213.165.64.22]) by core3.amsl.com (Postfix) with SMTP id 8AE7B3A6944 for <kitten@ietf.org>; Mon,  9 Aug 2010 06:37:16 -0700 (PDT)
Received: (qmail invoked by alias); 09 Aug 2010 13:37:49 -0000
Received: from a88-115-222-204.elisa-laajakaista.fi (EHLO [192.168.1.5]) [88.115.222.204] by mail.gmx.net (mp070) with SMTP; 09 Aug 2010 15:37:49 +0200
X-Authenticated: #29516787
X-Provags-ID: V01U2FsdGVkX1/PIsYRMILP8leNniaLeDmapGsKizjb+LguSx1xyV RleIVhiGeFgHoD
Message-ID: <4C6004AC.2060806@gmx.net>
Date: Mon, 09 Aug 2010 16:37:48 +0300
From: Hannes Tschofenig <hannes.tschofenig@gmx.net>
User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.9.1.11) Gecko/20100711 Thunderbird/3.0.6
MIME-Version: 1.0
To: Shawn Emery <shawn.emery@oracle.com>
Subject: Re: New Work Items - Kitten Recharter
References: <4C5CF47F.5040102@oracle.com>
In-Reply-To: <4C5CF47F.5040102@oracle.com>
Content-Type: multipart/alternative; boundary="------------000505030901030302010702"
X-Y-GMX-Trusted: 0
Cc: "kitten@ietf.org" <kitten@ietf.org>, sasl@ietf.org
X-BeenThere: kitten@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Common Authentication Technologies - Next Generation <kitten.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/kitten>
List-Post: <mailto:kitten@ietf.org>
List-Help: <mailto:kitten-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=subscribe>
X-List-Received-Date: Mon, 09 Aug 2010 13:37:18 -0000

This is a multi-part message in MIME format.
--------------000505030901030302010702
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit

Hey Shawn,

this makes sense to me (as discussed during the WG session at the recent 
IETF meeting).

Ciao
Hannes

On 8/7/2010 8:51 AM, Shawn Emery wrote:
>
> We are looking for consensus on whether the Kitten WG should adopt the 
> following drafts as work items:
>
>     draft-cantor-ietf-sasl-saml-ec
>     draft-mills-kitten-sasl-oauth
>
> You can respond to lists (kitten or SASL), but please indicate your 
> decision regardless if you are for or against the proposal.
>
> If we do decide to shepherd these drafts then we will recharter again 
> to include the following patch:
>
> @@ -50,10 +50,14 @@
>
>  * A SASL Mechanism for OpenID
>      draft-lear-ietf-sasl-openid-00
>  * A SASL Mechanism for SAML
>      draft-wierenga-ietf-sasl-saml-00
> +* A SASL Mechanism for SAML Enhanced Client
> +    draft-cantor-ietf-sasl-saml-ec
> +* A SASL Mechanism for OAuth
> +    draft-mills-kitten-sasl-oauth
>
>  The transition from SASL to GSS-API mechanisms will allow a greater 
> set of
>  applications to utilize said mechanisms with SASL implementations 
> that support
>  the use of GSS-API mechanisms in SASL (draft-ietf-sasl-gs2).
>
> Shawn.
> --
>
>
> _______________________________________________
> Kitten mailing list
> Kitten@ietf.org
> https://www.ietf.org/mailman/listinfo/kitten
>    


--------------000505030901030302010702
Content-Type: text/html; charset=ISO-8859-1
Content-Transfer-Encoding: 7bit

<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN">
<html>
<head>
  <meta content="text/html; charset=ISO-8859-1"
 http-equiv="Content-Type">
  <title></title>
</head>
<body bgcolor="#ffffff" text="#000000">
Hey Shawn, <br>
<br>
this makes sense to me (as discussed during the WG session at the
recent IETF meeting). <br>
<br>
Ciao<br>
Hannes<br>
<br>
On 8/7/2010 8:51 AM, Shawn Emery wrote:
<blockquote cite="mid:4C5CF47F.5040102@oracle.com" type="cite">
  <meta http-equiv="content-type"
 content="text/html; charset=ISO-8859-1">
  <font size="+1"><tt><br>
We are looking for consensus on whether the Kitten WG should adopt the
following drafts as work items:<br>
  <br>
&nbsp;&nbsp;&nbsp; draft-cantor-ietf-sasl-saml-ec<br>
&nbsp;&nbsp;&nbsp; draft-mills-kitten-sasl-oauth<br>
  <br>
You can respond to lists (kitten or SASL), but please indicate your
decision regardless if you are for or against the proposal.<br>
  <br>
If we do decide to shepherd these drafts then we will recharter again
to include the following patch:<br>
  <br>
@@ -50,10 +50,14 @@<br>
&nbsp;<br>
&nbsp;* A SASL Mechanism for OpenID<br>
&nbsp;&nbsp;&nbsp;&nbsp; draft-lear-ietf-sasl-openid-00<br>
&nbsp;* A SASL Mechanism for SAML<br>
&nbsp;&nbsp;&nbsp;&nbsp; draft-wierenga-ietf-sasl-saml-00<br>
+* A SASL Mechanism for SAML Enhanced Client<br>
+&nbsp;&nbsp;&nbsp; draft-cantor-ietf-sasl-saml-ec<br>
+* A SASL Mechanism for OAuth<br>
+&nbsp;&nbsp;&nbsp; draft-mills-kitten-sasl-oauth<br>
&nbsp;<br>
&nbsp;The transition from SASL to GSS-API mechanisms will allow a greater
set of<br>
&nbsp;applications to utilize said mechanisms with SASL implementations that
support<br>
&nbsp;the use of GSS-API mechanisms in SASL (draft-ietf-sasl-gs2).<br>
  <br>
Shawn.<br>
--<br>
  </tt></font>
  <pre wrap="">
<fieldset class="mimeAttachmentHeader"></fieldset>
_______________________________________________
Kitten mailing list
<a class="moz-txt-link-abbreviated" href="mailto:Kitten@ietf.org">Kitten@ietf.org</a>
<a class="moz-txt-link-freetext" href="https://www.ietf.org/mailman/listinfo/kitten">https://www.ietf.org/mailman/listinfo/kitten</a>
  </pre>
</blockquote>
<br>
</body>
</html>

--------------000505030901030302010702--

From cantor.2@osu.edu  Mon Aug  9 06:55:25 2010
Return-Path: <cantor.2@osu.edu>
X-Original-To: kitten@core3.amsl.com
Delivered-To: kitten@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 4EE333A6AD6; Mon,  9 Aug 2010 06:55:25 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -2.599
X-Spam-Level: 
X-Spam-Status: No, score=-2.599 tagged_above=-999 required=5 tests=[BAYES_00=-2.599]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id 0as1Z3Yumxhr; Mon,  9 Aug 2010 06:55:24 -0700 (PDT)
Received: from defang1.it.ohio-state.edu (defang1.it.ohio-state.edu [128.146.216.81]) by core3.amsl.com (Postfix) with ESMTP id 747D63A67FB; Mon,  9 Aug 2010 06:55:24 -0700 (PDT)
Received: from defang10.it.ohio-state.edu (defang10.it.ohio-state.edu [128.146.216.79]) by defang1.it.ohio-state.edu (8.13.7/8.13.1) with ESMTP id o79DtuCT003344; Mon, 9 Aug 2010 09:55:56 -0400
Received: from SNOWDOG (SNOWDOG.dyn.cio.osu.edu [164.107.161.86]) by defang10.it.ohio-state.edu (8.13.7/8.13.1) with ESMTP id o79Dtuu8008647; Mon, 9 Aug 2010 09:55:56 -0400
From: "Scott Cantor" <cantor.2@osu.edu>
To: "'Eliot Lear'" <lear@cisco.com>, "'Shawn Emery'" <shawn.emery@oracle.com>
References: <4C5CF47F.5040102@oracle.com> <4C5FC699.8060902@cisco.com>
In-Reply-To: <4C5FC699.8060902@cisco.com>
Subject: RE: [sasl] New Work Items - Kitten Recharter
Date: Mon, 9 Aug 2010 09:55:58 -0400
Organization: The Ohio State University
Message-ID: <012a01cb37ca$984ce9d0$c8e6bd70$@osu.edu>
MIME-Version: 1.0
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable
X-Mailer: Microsoft Outlook 14.0
Thread-index: AQJ85rZC8rLev/gGs0jS+zIiq+O3tALDC0ypkV+ugtA=
Content-language: en-us
X-CanIt-Geo: ip=128.146.216.79; country=US; region=OH; city=Columbus; latitude=39.9968; longitude=-82.9882; metrocode=535; areacode=614; http://maps.google.com/maps?q=39.9968,-82.9882&z=6
X-CanItPRO-Stream: outbound
X-Scanned-By: CanIt (www . roaringpenguin . com) on 128.146.216.81
Cc: kitten@ietf.org, sasl@ietf.org
X-BeenThere: kitten@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Common Authentication Technologies - Next Generation <kitten.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/kitten>
List-Post: <mailto:kitten@ietf.org>
List-Help: <mailto:kitten-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=subscribe>
X-List-Received-Date: Mon, 09 Aug 2010 13:55:25 -0000

> In Scott's draft, that occurs in step (4).  This requires the client =
to have
> substantially more capabilities than it might have today with either a =
fully
> functional web browser either built into the application or tied to =
the
> application via some form of IPC with sufficient semantic abilities to
> discern when to move through step 4 to step 5, but at the same time,
> provides for an overall simpler protocol flow than the document that =
Klaas
> and I have put forth.

My proposal uses ECP, which means the exchange with the IdP is generally =
(and can be required to be) SOAP over HTTP. There is generally no need =
for a web browser in the client, or anything like that. That's the IPC =
you're referring to.
=20
> Discovery is also handled in Scott's draft.  That is something that we
> should consider incorporating into the other.

I believe the latest draft of Klaas' proposal does this, though in a =
different way from ECP of course.

-- Scott



From cantor.2@osu.edu  Mon Aug  9 07:06:12 2010
Return-Path: <cantor.2@osu.edu>
X-Original-To: kitten@core3.amsl.com
Delivered-To: kitten@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 249673A67FB; Mon,  9 Aug 2010 07:06:12 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -2.599
X-Spam-Level: 
X-Spam-Status: No, score=-2.599 tagged_above=-999 required=5 tests=[BAYES_00=-2.599]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id jsgmEjJBQxMv; Mon,  9 Aug 2010 07:06:11 -0700 (PDT)
Received: from defang1.it.ohio-state.edu (defang1.it.ohio-state.edu [128.146.216.81]) by core3.amsl.com (Postfix) with ESMTP id 2D3BB3A67AE; Mon,  9 Aug 2010 07:06:09 -0700 (PDT)
Received: from defang10.it.ohio-state.edu (defang10.it.ohio-state.edu [128.146.216.79]) by defang1.it.ohio-state.edu (8.13.7/8.13.1) with ESMTP id o79E6ctx008112; Mon, 9 Aug 2010 10:06:38 -0400
Received: from SNOWDOG (SNOWDOG.dyn.cio.osu.edu [164.107.161.86]) by defang10.it.ohio-state.edu (8.13.7/8.13.1) with ESMTP id o79E6bLV015560; Mon, 9 Aug 2010 10:06:37 -0400
From: "Scott Cantor" <cantor.2@osu.edu>
To: "'Eliot Lear'" <lear@cisco.com>, "'Shawn Emery'" <shawn.emery@oracle.com>
References: <4C5CF47F.5040102@oracle.com> <4C5FC699.8060902@cisco.com> <012a01cb37ca$984ce9d0$c8e6bd70$@osu.edu>
In-Reply-To: <012a01cb37ca$984ce9d0$c8e6bd70$@osu.edu>
Subject: RE: [sasl] New Work Items - Kitten Recharter
Date: Mon, 9 Aug 2010 10:06:40 -0400
Organization: The Ohio State University
Message-ID: <012f01cb37cc$16954b10$43bfe130$@osu.edu>
MIME-Version: 1.0
Content-Type: text/plain; charset="US-ASCII"
Content-Transfer-Encoding: 7bit
X-Mailer: Microsoft Outlook 14.0
Thread-index: AQJ85rZC8rLev/gGs0jS+zIiq+O3tALDC0ypAgaOvy2RT31pEA==
Content-language: en-us
X-CanIt-Geo: ip=128.146.216.79; country=US; region=OH; city=Columbus; latitude=39.9968; longitude=-82.9882; metrocode=535; areacode=614; http://maps.google.com/maps?q=39.9968,-82.9882&z=6
X-CanItPRO-Stream: outbound
X-Scanned-By: CanIt (www . roaringpenguin . com) on 128.146.216.81
Cc: kitten@ietf.org, sasl@ietf.org
X-BeenThere: kitten@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Common Authentication Technologies - Next Generation <kitten.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/kitten>
List-Post: <mailto:kitten@ietf.org>
List-Help: <mailto:kitten-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=subscribe>
X-List-Received-Date: Mon, 09 Aug 2010 14:06:12 -0000

> My proposal uses ECP, which means the exchange with the IdP is generally
> (and can be required to be) SOAP over HTTP. There is generally no need for
a
> web browser in the client, or anything like that. That's the IPC you're
> referring to.

Actually rereading you meant IPC between the client and some local agent.
What I'm trying to say is that the client/IdP exchange is structured (SOAP),
not unstructured (HTML). The client requirements of my mechanism proposal
are trivial (until/unless we add holder of key or other kinds of security
options).
 
-- Scott



From jhutz@cmu.edu  Mon Aug  9 11:40:07 2010
Return-Path: <jhutz@cmu.edu>
X-Original-To: kitten@core3.amsl.com
Delivered-To: kitten@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id B936E3A69D1; Mon,  9 Aug 2010 11:40:06 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -106.522
X-Spam-Level: 
X-Spam-Status: No, score=-106.522 tagged_above=-999 required=5 tests=[AWL=0.077, BAYES_00=-2.599, RCVD_IN_DNSWL_MED=-4, USER_IN_WHITELIST=-100]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id q7IUz1gCxJji; Mon,  9 Aug 2010 11:40:05 -0700 (PDT)
Received: from smtp02.srv.cs.cmu.edu (SMTP02.SRV.CS.CMU.EDU [128.2.217.197]) by core3.amsl.com (Postfix) with ESMTP id 03E3B3A683C; Mon,  9 Aug 2010 11:40:04 -0700 (PDT)
Received: from atlantis-home.pc.cs.cmu.edu (SIRIUS.FAC.CS.CMU.EDU [128.2.216.216]) (authenticated bits=0) by smtp02.srv.cs.cmu.edu (8.13.6/8.13.6) with ESMTP id o79IeZkb024867 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=NO); Mon, 9 Aug 2010 14:40:36 -0400 (EDT)
Date: Mon, 09 Aug 2010 14:40:35 -0400
From: Jeffrey Hutzelman <jhutz@cmu.edu>
To: Alexey Melnikov <alexey.melnikov@isode.com>, Eliot Lear <lear@cisco.com>
Subject: Re: [sasl] New Work Items - Kitten Recharter
Message-ID: <4C6210AE12B6E901D2025331@atlantis.pc.cs.cmu.edu>
In-Reply-To: <17613_1281358755_o79CxEZL028303_4C5FFB8F.6030406@isode.com>
References: <4C5CF47F.5040102@oracle.com> <4C5FC699.8060902@cisco.com> <17613_1281358755_o79CxEZL028303_4C5FFB8F.6030406@isode.com>
X-Mailer: Mulberry/4.0.8 (Linux/x86)
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii; format=flowed
Content-Transfer-Encoding: 7bit
Content-Disposition: inline
X-Scanned-By: mimedefang-cmuscs on 128.2.217.197
Cc: kitten@ietf.org, sasl@ietf.org, jhutz@cmu.edu
X-BeenThere: kitten@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Common Authentication Technologies - Next Generation <kitten.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/kitten>
List-Post: <mailto:kitten@ietf.org>
List-Help: <mailto:kitten-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=subscribe>
X-List-Received-Date: Mon, 09 Aug 2010 18:40:07 -0000

--On Monday, August 09, 2010 02:58:55 PM +0200 Alexey Melnikov 
<alexey.melnikov@isode.com> wrote:

> Hi Eliot,
>
> Eliot Lear wrote:
>
>>> We are looking for consensus on whether the Kitten WG should adopt
>>> the following drafts as work items:
>>>
>>>     draft-cantor-ietf-sasl-saml-ec
>>>     draft-mills-kitten-sasl-oauth
>>>
>>> You can respond to lists (kitten or SASL), but please indicate your
>>> decision regardless if you are for or against the proposal.
>>
>> I support the inclusion of both drafts in our milestones.  To answer
>> Alexey's reaonable question, draft-wierenga-ietf-sasl-saml attempts to
>> rely on existing infrastructure as much as possible, both on the
>> client and on the IdP, while requiring some substantial changes to the
>> RP.   draft-cantor-ietf-sasl-saml-ec requires more substantial changes
>> to the client, but keeps SAML within the application protocol.  The
>> implication of the design choices relate more to how actual
>> authentication gets performed between the SASL client and the IdP.
>>  Existing IdPs make use of HTTP/HTML and unstructured exchanges for
>> that authentication.
>>
>> In Scott's draft, that occurs in step (4).  This requires the client
>> to have substantially more capabilities than it might have today with
>> either a fully functional web browser either built into the
>> application or tied to the application via some form of IPC with
>> sufficient semantic abilities to discern when to move through step 4
>> to step 5, but at the same time, provides for an overall simpler
>> protocol flow than the document that Klaas and I have put forth.
>
> Yes, I've seen this argument on the mailing list. Two documents suggest
> reasonable solutions for the assumption made. However I am not yet fully
> convinced that the assumption is something important enough to warrant
> having 2 documents.

The failure here is not in having two documents to discuss, but in 
insisting on having specific documents listed by name in the charter.  A 
charter work item should be something like "produce a SAML-based GSS-API 
mechanism", which could be satisfied by either of these documents by the 
time we're done with them.  Perhaps in the end we'll choose one or the 
other, or perhaps we'll decide that two mechanisms really are necessary, in 
which case we'll make the case to the IESG as to why both should by 
published.

-- Jeff


From alexey.melnikov@isode.com  Mon Aug  9 11:42:13 2010
Return-Path: <alexey.melnikov@isode.com>
X-Original-To: kitten@core3.amsl.com
Delivered-To: kitten@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id D671D3A6940; Mon,  9 Aug 2010 11:42:13 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -102.599
X-Spam-Level: 
X-Spam-Status: No, score=-102.599 tagged_above=-999 required=5 tests=[BAYES_00=-2.599, USER_IN_WHITELIST=-100]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id Rs5vjN1uV6en; Mon,  9 Aug 2010 11:42:12 -0700 (PDT)
Received: from rufus.isode.com (rufus.isode.com [62.3.217.251]) by core3.amsl.com (Postfix) with ESMTP id 922D83A683C; Mon,  9 Aug 2010 11:42:12 -0700 (PDT)
Received: from [172.16.2.104] (shiny.isode.com [62.3.217.250])  by rufus.isode.com (submission channel) via TCP with ESMTPA  id <TGBMGwBIEDN-@rufus.isode.com>; Mon, 9 Aug 2010 19:42:44 +0100
Message-ID: <4C604BEF.8050503@isode.com>
Date: Mon, 09 Aug 2010 20:41:51 +0200
From: Alexey Melnikov <alexey.melnikov@isode.com>
User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.7.12) Gecko/20050915
X-Accept-Language: en-us, en
To: Jeffrey Hutzelman <jhutz@cmu.edu>
Subject: Re: [sasl] New Work Items - Kitten Recharter
References: <4C5CF47F.5040102@oracle.com> <4C5FC699.8060902@cisco.com> <17613_1281358755_o79CxEZL028303_4C5FFB8F.6030406@isode.com> <4C6210AE12B6E901D2025331@atlantis.pc.cs.cmu.edu>
In-Reply-To: <4C6210AE12B6E901D2025331@atlantis.pc.cs.cmu.edu>
MIME-Version: 1.0
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit
Cc: kitten@ietf.org, sasl@ietf.org
X-BeenThere: kitten@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Common Authentication Technologies - Next Generation <kitten.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/kitten>
List-Post: <mailto:kitten@ietf.org>
List-Help: <mailto:kitten-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=subscribe>
X-List-Received-Date: Mon, 09 Aug 2010 18:42:13 -0000

Jeffrey Hutzelman wrote:

> --On Monday, August 09, 2010 02:58:55 PM +0200 Alexey Melnikov 
> <alexey.melnikov@isode.com> wrote:
>
>> Hi Eliot,
>>
>> Eliot Lear wrote:
>>
>>>> We are looking for consensus on whether the Kitten WG should adopt
>>>> the following drafts as work items:
>>>>
>>>>     draft-cantor-ietf-sasl-saml-ec
>>>>     draft-mills-kitten-sasl-oauth
>>>>
>>>> You can respond to lists (kitten or SASL), but please indicate your
>>>> decision regardless if you are for or against the proposal.
>>>
>>>
>>> I support the inclusion of both drafts in our milestones.  To answer
>>> Alexey's reaonable question, draft-wierenga-ietf-sasl-saml attempts to
>>> rely on existing infrastructure as much as possible, both on the
>>> client and on the IdP, while requiring some substantial changes to the
>>> RP.   draft-cantor-ietf-sasl-saml-ec requires more substantial changes
>>> to the client, but keeps SAML within the application protocol.  The
>>> implication of the design choices relate more to how actual
>>> authentication gets performed between the SASL client and the IdP.
>>>  Existing IdPs make use of HTTP/HTML and unstructured exchanges for
>>> that authentication.
>>>
>>> In Scott's draft, that occurs in step (4).  This requires the client
>>> to have substantially more capabilities than it might have today with
>>> either a fully functional web browser either built into the
>>> application or tied to the application via some form of IPC with
>>> sufficient semantic abilities to discern when to move through step 4
>>> to step 5, but at the same time, provides for an overall simpler
>>> protocol flow than the document that Klaas and I have put forth.
>>
>> Yes, I've seen this argument on the mailing list. Two documents suggest
>> reasonable solutions for the assumption made. However I am not yet fully
>> convinced that the assumption is something important enough to warrant
>> having 2 documents.
>
> The failure here is not in having two documents to discuss, but in 
> insisting on having specific documents listed by name in the charter.  
> A charter work item should be something like "produce a SAML-based 
> GSS-API mechanism", which could be satisfied by either of these 
> documents by the time we're done with them.  Perhaps in the end we'll 
> choose one or the other, or perhaps we'll decide that two mechanisms 
> really are necessary, in which case we'll make the case to the IESG as 
> to why both should by published.

Yes, I was about to suggest the same. Not naming specific documents 
would be better in this case.


From hannes.tschofenig@nsn.com  Wed Aug 11 03:41:35 2010
Return-Path: <hannes.tschofenig@nsn.com>
X-Original-To: kitten@core3.amsl.com
Delivered-To: kitten@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 911073A6A6B; Wed, 11 Aug 2010 03:41:35 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -102.541
X-Spam-Level: 
X-Spam-Status: No, score=-102.541 tagged_above=-999 required=5 tests=[AWL=0.057, BAYES_00=-2.599, HTML_MESSAGE=0.001, USER_IN_WHITELIST=-100]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id gC1xDUiecgrI; Wed, 11 Aug 2010 03:41:34 -0700 (PDT)
Received: from demumfd002.nsn-inter.net (demumfd002.nsn-inter.net [93.183.12.31]) by core3.amsl.com (Postfix) with ESMTP id 26E0B3A6807; Wed, 11 Aug 2010 03:41:33 -0700 (PDT)
Received: from demuprx017.emea.nsn-intra.net ([10.150.129.56]) by demumfd002.nsn-inter.net (8.12.11.20060308/8.12.11) with ESMTP id o7BAg9m4012461 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=OK); Wed, 11 Aug 2010 12:42:09 +0200
Received: from demuexc022.nsn-intra.net (demuexc022.nsn-intra.net [10.150.128.35]) by demuprx017.emea.nsn-intra.net (8.12.11.20060308/8.12.11) with ESMTP id o7BAg39Y025777; Wed, 11 Aug 2010 12:42:09 +0200
Received: from FIESEXC015.nsn-intra.net ([10.159.0.23]) by demuexc022.nsn-intra.net with Microsoft SMTPSVC(6.0.3790.4675);  Wed, 11 Aug 2010 12:42:05 +0200
X-MimeOLE: Produced By Microsoft Exchange V6.5
Content-class: urn:content-classes:message
MIME-Version: 1.0
Content-Type: multipart/alternative; boundary="----_=_NextPart_001_01CB3941.D619A9C7"
Subject: Feedback solicited for Privacy and Identity Management Terminology
Date: Wed, 11 Aug 2010 13:38:12 +0300
Message-ID: <3D3C75174CB95F42AD6BCC56E5555B4502E5D295@FIESEXC015.nsn-intra.net>
X-MS-Has-Attach: 
X-MS-TNEF-Correlator: 
Thread-Topic: Feedback solicited for Privacy and Identity Management Terminology
Thread-Index: Acs5PcCKKP+5/l6FR0Cr0uxjuLJ03Q==
From: "Tschofenig, Hannes (NSN - FI/Espoo)" <hannes.tschofenig@nsn.com>
To: <kitten@ietf.org>, <sasl@ietf.org>
X-OriginalArrivalTime: 11 Aug 2010 10:42:05.0378 (UTC) FILETIME=[D70CAA20:01CB3941]
X-Mailman-Approved-At: Wed, 11 Aug 2010 08:51:54 -0700
X-BeenThere: kitten@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Common Authentication Technologies - Next Generation <kitten.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/kitten>
List-Post: <mailto:kitten@ietf.org>
List-Help: <mailto:kitten-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=subscribe>
X-List-Received-Date: Wed, 11 Aug 2010 10:41:35 -0000

This is a multi-part message in MIME format.

------_=_NextPart_001_01CB3941.D619A9C7
Content-Type: text/plain;
	charset="us-ascii"
Content-Transfer-Encoding: quoted-printable

Hi all,=20

we have just submitted a new version of the privacy and identity
management terminology document:
http://www.ietf.org/internet-drafts/draft-hansen-privacy-terminology-01.
txt
=20
The full document titel is "Terminology for Talking about Privacy by
Data Minimization: Anonymity, Unlinkability, Undetectability,
Unobservability, Pseudonymity, and Identity Management" and here is the
abstract:=20

   This document is an attempt to consolidate terminology in the field
   privacy by data minimization.  It motivates and develops definitions
   for anonymity/identifiability, (un)linkability, (un)detectability,
   (un)observability, pseudonymity, identity, partial identity, digital
   identity and identity management.  Starting the definitions from the
   anonymity and unlinkability perspective and not from a definition of
   identity (the latter is the obvious approach to some people) reveals
   some deeper structures in this field.
  =20
Please let us know whether the terminology is useful and complete. Your
input is highly appreciated!

Ciao
Hannes



------_=_NextPart_001_01CB3941.D619A9C7
Content-Type: text/html;
	charset="us-ascii"
Content-Transfer-Encoding: quoted-printable

<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 3.2//EN">
<HTML>
<HEAD>
<META HTTP-EQUIV=3D"Content-Type" CONTENT=3D"text/html; =
charset=3Dus-ascii">
<META NAME=3D"Generator" CONTENT=3D"MS Exchange Server version =
6.5.7654.12">
<TITLE>Feedback solicited for Privacy and Identity Management =
Terminology</TITLE>
</HEAD>
<BODY>
<!-- Converted from text/rtf format -->

<P><FONT SIZE=3D2 FACE=3D"Courier New">Hi all, </FONT>
</P>

<P><FONT SIZE=3D2 FACE=3D"Courier New">we have just submitted a new =
version of the privacy and identity management terminology =
document:</FONT>

<BR><A =
HREF=3D"http://www.ietf.org/internet-drafts/draft-hansen-privacy-terminol=
ogy-01.txt"><U><FONT COLOR=3D"#0000FF" SIZE=3D2 FACE=3D"Courier =
New">http://www.ietf.org/internet-drafts/draft-hansen-privacy-terminology=
-01.txt</FONT></U></A>

<BR><FONT SIZE=3D2 FACE=3D"Courier New">&nbsp;</FONT>

<BR><FONT SIZE=3D2 FACE=3D"Courier New">The full document titel is =
&quot;Terminology for Talking about Privacy by Data Minimization: =
Anonymity, Unlinkability, Undetectability, Unobservability, =
Pseudonymity, and Identity Management&quot; and here is the abstract: =
</FONT></P>

<P><FONT SIZE=3D2 FACE=3D"Courier New">&nbsp;&nbsp; This document is an =
attempt to consolidate terminology in the field</FONT>

<BR><FONT SIZE=3D2 FACE=3D"Courier New">&nbsp;&nbsp; privacy by data =
minimization.&nbsp; It motivates and develops definitions</FONT>

<BR><FONT SIZE=3D2 FACE=3D"Courier New">&nbsp;&nbsp; for =
anonymity/identifiability, (un)linkability, (un)detectability,</FONT>

<BR><FONT SIZE=3D2 FACE=3D"Courier New">&nbsp;&nbsp; (un)observability, =
pseudonymity, identity, partial identity, digital</FONT>

<BR><FONT SIZE=3D2 FACE=3D"Courier New">&nbsp;&nbsp; identity and =
identity management.&nbsp; Starting the definitions from the</FONT>

<BR><FONT SIZE=3D2 FACE=3D"Courier New">&nbsp;&nbsp; anonymity and =
unlinkability perspective and not from a definition of</FONT>

<BR><FONT SIZE=3D2 FACE=3D"Courier New">&nbsp;&nbsp; identity (the =
latter is the obvious approach to some people) reveals</FONT>

<BR><FONT SIZE=3D2 FACE=3D"Courier New">&nbsp;&nbsp; some deeper =
structures in this field.</FONT>

<BR><FONT SIZE=3D2 FACE=3D"Courier New">&nbsp;&nbsp; </FONT>

<BR><FONT SIZE=3D2 FACE=3D"Courier New">Please let us know whether the =
terminology is useful and complete. Your input is highly =
appreciated!</FONT>
</P>

<P><FONT SIZE=3D2 FACE=3D"Courier New">Ciao</FONT>

<BR><FONT SIZE=3D2 FACE=3D"Courier New">Hannes</FONT>
</P>
<BR>

</BODY>
</HTML>
------_=_NextPart_001_01CB3941.D619A9C7--

From hartmans@mit.edu  Wed Aug 11 11:55:24 2010
Return-Path: <hartmans@mit.edu>
X-Original-To: kitten@core3.amsl.com
Delivered-To: kitten@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 247D83A698D for <kitten@core3.amsl.com>; Wed, 11 Aug 2010 11:55:24 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -102.735
X-Spam-Level: 
X-Spam-Status: No, score=-102.735 tagged_above=-999 required=5 tests=[AWL=-0.470, BAYES_00=-2.599, IP_NOT_FRIENDLY=0.334, USER_IN_WHITELIST=-100]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id flfLhmGbGc9a for <kitten@core3.amsl.com>; Wed, 11 Aug 2010 11:55:23 -0700 (PDT)
Received: from mail.suchdamage.org (permutation-city.suchdamage.org [69.25.196.28]) by core3.amsl.com (Postfix) with ESMTP id 45F953A68A2 for <kitten@ietf.org>; Wed, 11 Aug 2010 11:55:23 -0700 (PDT)
Received: from carter-zimmerman.suchdamage.org (carter-zimmerman.suchdamage.org [69.25.196.178]) (using TLSv1 with cipher DHE-RSA-AES256-SHA (256/256 bits)) (Client CN "laptop", Issuer "laptop" (not verified)) by mail.suchdamage.org (Postfix) with ESMTPS id 2BCE42041B for <kitten@ietf.org>; Wed, 11 Aug 2010 14:55:57 -0400 (EDT)
Received: by carter-zimmerman.suchdamage.org (Postfix, from userid 8042) id 9BC62412A; Wed, 11 Aug 2010 14:55:46 -0400 (EDT)
From: Sam Hartman <hartmans-ietf@mit.edu>
To: kitten@ietf.org
Subject: Concerns about oauth sasl mechanism
Date: Wed, 11 Aug 2010 14:55:46 -0400
Message-ID: <tsllj8dyo5p.fsf@mit.edu>
User-Agent: Gnus/5.110009 (No Gnus v0.9) Emacs/22.3 (gnu/linux)
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
X-BeenThere: kitten@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Common Authentication Technologies - Next Generation <kitten.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/kitten>
List-Post: <mailto:kitten@ietf.org>
List-Help: <mailto:kitten-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=subscribe>
X-List-Received-Date: Wed, 11 Aug 2010 18:55:24 -0000

I'll admit that I'm a bit concerned about the never of SASL/GSS
mechanisms we're creating that have no better security than cram-md5 and
may have somewhat less.

I don't think we need any more SASL mechanisms that provide neither
security layers nor channel binding.  I haven't followed the OAuth
standards in a while. If they currently have something that can be used
for channel binding or if they gain such a facility, I think
standardizing such a mechanism here would be great.  Otherwise, I don't
think it's a good use of this WG's time.

Obviously, nothing stops proponents of such a mechanism of registering
it or even getting an non-IETF RFC.  I guess nothing even stops them
from trying to recharter OAuth, although I would bring up the same
concern in that case as well.

I understand that a similar argument can be made about the SAML and
OpenID mechanism. First, I have some hope for the SAML mechanism.  I'll
admit that I'm quite conflicted on the OpenID mechanism and that I can't
entirely justify why I didn't raise this objection there. Roughly, I
think that's very close to the line and this is over the line.

--Sam

From stpeter@stpeter.im  Thu Aug 12 13:59:32 2010
Return-Path: <stpeter@stpeter.im>
X-Original-To: kitten@core3.amsl.com
Delivered-To: kitten@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id EB2293A68CF for <kitten@core3.amsl.com>; Thu, 12 Aug 2010 13:59:32 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -102.637
X-Spam-Level: 
X-Spam-Status: No, score=-102.637 tagged_above=-999 required=5 tests=[AWL=-0.038, BAYES_00=-2.599, USER_IN_WHITELIST=-100]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id 7pxCHpYBwXLq for <kitten@core3.amsl.com>; Thu, 12 Aug 2010 13:59:31 -0700 (PDT)
Received: from stpeter.im (stpeter.im [207.210.219.233]) by core3.amsl.com (Postfix) with ESMTP id C16563A67F2 for <kitten@ietf.org>; Thu, 12 Aug 2010 13:59:31 -0700 (PDT)
Received: from dhcp-64-101-72-121.cisco.com (dhcp-64-101-72-121.cisco.com [64.101.72.121]) (Authenticated sender: stpeter) by stpeter.im (Postfix) with ESMTPSA id 1BD2940074; Thu, 12 Aug 2010 15:01:21 -0600 (MDT)
Message-ID: <4C6460D6.7020800@stpeter.im>
Date: Thu, 12 Aug 2010 15:00:06 -0600
From: Peter Saint-Andre <stpeter@stpeter.im>
User-Agent: Mozilla/5.0 (Macintosh; U; Intel Mac OS X 10.5; en-US; rv:1.9.1.11) Gecko/20100711 Thunderbird/3.0.6
MIME-Version: 1.0
To: Sam Hartman <hartmans-ietf@mit.edu>
Subject: Re: Concerns about oauth sasl mechanism
References: <tsllj8dyo5p.fsf@mit.edu>
In-Reply-To: <tsllj8dyo5p.fsf@mit.edu>
X-Enigmail-Version: 1.0.1
OpenPGP: url=http://www.saint-andre.com/me/stpeter.asc
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 7bit
Cc: kitten@ietf.org
X-BeenThere: kitten@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Common Authentication Technologies - Next Generation <kitten.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/kitten>
List-Post: <mailto:kitten@ietf.org>
List-Help: <mailto:kitten-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=subscribe>
X-List-Received-Date: Thu, 12 Aug 2010 20:59:33 -0000

On 8/11/10 12:55 PM, Sam Hartman wrote:

> I'll admit that I'm a bit concerned about the never of SASL/GSS
> mechanisms we're creating that have no better security than cram-md5 and
> may have somewhat less.
> 
> I don't think we need any more SASL mechanisms that provide neither
> security layers nor channel binding.  I haven't followed the OAuth
> standards in a while. If they currently have something that can be used
> for channel binding or if they gain such a facility, I think
> standardizing such a mechanism here would be great.  Otherwise, I don't
> think it's a good use of this WG's time.

The SAML-OAUTH work that seems to have the most traction in the OAuth WG
is development of methods for using a SAML assertion to get or present
an OAuth token. I agree that using an OAuth token as a way to do SASL
authentication is much less interesting.

> Obviously, nothing stops proponents of such a mechanism of registering
> it or even getting an non-IETF RFC.  I guess nothing even stops them
> from trying to recharter OAuth, although I would bring up the same
> concern in that case as well.

I think that development of SASL mechanisms needs to happen in the
(refactored) KITTEN WG, not in the OAuth WG or anywhere else.

> I understand that a similar argument can be made about the SAML and
> OpenID mechanism. First, I have some hope for the SAML mechanism.  I'll
> admit that I'm quite conflicted on the OpenID mechanism and that I can't
> entirely justify why I didn't raise this objection there. 

Given existing security issues with OpenID as it is usually deployed, I
tend to agree -- although I have heard about deployments under
controlled circumstances (inside enterprises) that have stronger
security properties than the OpenID deployments on the open Internet.

> Roughly, I
> think that's very close to the line and this is over the line.

Could you clarify which line you're talking about? :)

Peter

-- 
Peter Saint-Andre
https://stpeter.im/



From hartmans@mit.edu  Thu Aug 12 14:25:29 2010
Return-Path: <hartmans@mit.edu>
X-Original-To: kitten@core3.amsl.com
Delivered-To: kitten@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 279BE3A69FE for <kitten@core3.amsl.com>; Thu, 12 Aug 2010 14:25:29 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -102.714
X-Spam-Level: 
X-Spam-Status: No, score=-102.714 tagged_above=-999 required=5 tests=[AWL=-0.449, BAYES_00=-2.599, IP_NOT_FRIENDLY=0.334, USER_IN_WHITELIST=-100]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id 1kliNBIixzZB for <kitten@core3.amsl.com>; Thu, 12 Aug 2010 14:25:28 -0700 (PDT)
Received: from mail.suchdamage.org (permutation-city.suchdamage.org [69.25.196.28]) by core3.amsl.com (Postfix) with ESMTP id 5F7D73A69F1 for <kitten@ietf.org>; Thu, 12 Aug 2010 14:25:28 -0700 (PDT)
Received: from carter-zimmerman.suchdamage.org (carter-zimmerman.suchdamage.org [69.25.196.178]) (using TLSv1 with cipher DHE-RSA-AES256-SHA (256/256 bits)) (Client CN "laptop", Issuer "laptop" (not verified)) by mail.suchdamage.org (Postfix) with ESMTPS id 49A372041B; Thu, 12 Aug 2010 17:26:03 -0400 (EDT)
Received: by carter-zimmerman.suchdamage.org (Postfix, from userid 8042) id 57837412A; Thu, 12 Aug 2010 17:25:50 -0400 (EDT)
From: Sam Hartman <hartmans-ietf@mit.edu>
To: Peter Saint-Andre <stpeter@stpeter.im>
Subject: Re: Concerns about oauth sasl mechanism
References: <tsllj8dyo5p.fsf@mit.edu> <4C6460D6.7020800@stpeter.im>
Date: Thu, 12 Aug 2010 17:25:50 -0400
In-Reply-To: <4C6460D6.7020800@stpeter.im> (Peter Saint-Andre's message of "Thu, 12 Aug 2010 15:00:06 -0600")
Message-ID: <tsl1va3v7z5.fsf@mit.edu>
User-Agent: Gnus/5.110009 (No Gnus v0.9) Emacs/22.3 (gnu/linux)
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Cc: kitten@ietf.org, Sam Hartman <hartmans-ietf@mit.edu>
X-BeenThere: kitten@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Common Authentication Technologies - Next Generation <kitten.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/kitten>
List-Post: <mailto:kitten@ietf.org>
List-Help: <mailto:kitten-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=subscribe>
X-List-Received-Date: Thu, 12 Aug 2010 21:25:29 -0000

The line about whether we should spend kitten's time on it.  I think
having multiple insecure SASL mechanisms under development or even
having a string of insecure sasl mechanisms is not an effective use of
our time.


To be secure, I think a mechanism needs to:

* Not provide an attacker with a weak password equivelent (yapp is out)
* Provide mutual authentication
* Provide channel binding or security layers

I understand we've made exceptions for SAML and OpenID.  Perhaps after
Oauth is done we'll look at its deployment and make a similar exception.

A lot of my reasoning here is that registering a SASL mechanism is easy:
it's first-come, first-serve. So, what are you actually getting out of
spending IETF security area review cycles and IESG cycles if at the end
of the day the result is not actually secure.

From Nicolas.Williams@oracle.com  Thu Aug 12 14:30:50 2010
Return-Path: <Nicolas.Williams@oracle.com>
X-Original-To: kitten@core3.amsl.com
Delivered-To: kitten@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 881363A6A46 for <kitten@core3.amsl.com>; Thu, 12 Aug 2010 14:30:50 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -6.517
X-Spam-Level: 
X-Spam-Status: No, score=-6.517 tagged_above=-999 required=5 tests=[AWL=0.081,  BAYES_00=-2.599, RCVD_IN_DNSWL_MED=-4, UNPARSEABLE_RELAY=0.001]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id 7Qs4lrRNIK8O for <kitten@core3.amsl.com>; Thu, 12 Aug 2010 14:30:49 -0700 (PDT)
Received: from rcsinet10.oracle.com (rcsinet10.oracle.com [148.87.113.121]) by core3.amsl.com (Postfix) with ESMTP id 39B0D3A6A43 for <kitten@ietf.org>; Thu, 12 Aug 2010 14:30:49 -0700 (PDT)
Received: from acsinet15.oracle.com (acsinet15.oracle.com [141.146.126.227]) by rcsinet10.oracle.com (Switch-3.4.2/Switch-3.4.2) with ESMTP id o7CLVN2M013438 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=OK); Thu, 12 Aug 2010 21:31:25 GMT
Received: from acsmt353.oracle.com (acsmt353.oracle.com [141.146.40.153]) by acsinet15.oracle.com (Switch-3.4.2/Switch-3.4.1) with ESMTP id o7C68fZ9010650; Thu, 12 Aug 2010 21:31:23 GMT
Received: from abhmt012.oracle.com by acsmt354.oracle.com with ESMTP id 512170851281648631; Thu, 12 Aug 2010 14:30:31 -0700
Received: from oracle.com (/129.153.128.104) by default (Oracle Beehive Gateway v4.0) with ESMTP ; Thu, 12 Aug 2010 14:30:30 -0700
Date: Thu, 12 Aug 2010 16:30:24 -0500
From: Nicolas Williams <Nicolas.Williams@oracle.com>
To: Sam Hartman <hartmans-ietf@mit.edu>
Subject: Re: Concerns about oauth sasl mechanism
Message-ID: <20100812213024.GW5213@oracle.com>
References: <tsllj8dyo5p.fsf@mit.edu> <4C6460D6.7020800@stpeter.im> <tsl1va3v7z5.fsf@mit.edu>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <tsl1va3v7z5.fsf@mit.edu>
User-Agent: Mutt/1.5.20 (2010-03-02)
Cc: kitten@ietf.org
X-BeenThere: kitten@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Common Authentication Technologies - Next Generation <kitten.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/kitten>
List-Post: <mailto:kitten@ietf.org>
List-Help: <mailto:kitten-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=subscribe>
X-List-Received-Date: Thu, 12 Aug 2010 21:30:50 -0000

On Thu, Aug 12, 2010 at 05:25:50PM -0400, Sam Hartman wrote:
> To be secure, I think a mechanism needs to:
> 
> * Not provide an attacker with a weak password equivelent (yapp is out)
> * Provide mutual authentication
> * Provide channel binding or security layers

+1.

(What's YAPP?)

> I understand we've made exceptions for SAML and OpenID.  Perhaps after
> Oauth is done we'll look at its deployment and make a similar exception.
> 
> A lot of my reasoning here is that registering a SASL mechanism is easy:
> it's first-come, first-serve. So, what are you actually getting out of
> spending IETF security area review cycles and IESG cycles if at the end
> of the day the result is not actually secure.

Likely answer: the imprimatur of "RFC", even if it isn't on the
Standards track.

Nico
-- 

From alexey.melnikov@isode.com  Thu Aug 12 14:37:03 2010
Return-Path: <alexey.melnikov@isode.com>
X-Original-To: kitten@core3.amsl.com
Delivered-To: kitten@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id CD6A23A67F0 for <kitten@core3.amsl.com>; Thu, 12 Aug 2010 14:37:03 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -102.67
X-Spam-Level: 
X-Spam-Status: No, score=-102.67 tagged_above=-999 required=5 tests=[AWL=-0.071, BAYES_00=-2.599, USER_IN_WHITELIST=-100]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id q-RvwtXCPZaq for <kitten@core3.amsl.com>; Thu, 12 Aug 2010 14:37:02 -0700 (PDT)
Received: from rufus.isode.com (rufus.isode.com [62.3.217.251]) by core3.amsl.com (Postfix) with ESMTP id C02053A67E3 for <kitten@ietf.org>; Thu, 12 Aug 2010 14:37:02 -0700 (PDT)
Received: from [172.16.2.146] (shiny.isode.com [62.3.217.250])  by rufus.isode.com (submission channel) via TCP with ESMTPA  id <TGRpnwBIEHhX@rufus.isode.com>; Thu, 12 Aug 2010 22:37:38 +0100
Message-ID: <4C64698A.9030405@isode.com>
Date: Thu, 12 Aug 2010 22:37:14 +0100
From: Alexey Melnikov <alexey.melnikov@isode.com>
User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.7.12) Gecko/20050915
X-Accept-Language: en-us, en
To: Sam Hartman <hartmans-ietf@mit.edu>
Subject: Re: Concerns about oauth sasl mechanism
References: <tsllj8dyo5p.fsf@mit.edu> <4C6460D6.7020800@stpeter.im> <tsl1va3v7z5.fsf@mit.edu>
In-Reply-To: <tsl1va3v7z5.fsf@mit.edu>
MIME-Version: 1.0
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit
Cc: kitten@ietf.org
X-BeenThere: kitten@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Common Authentication Technologies - Next Generation <kitten.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/kitten>
List-Post: <mailto:kitten@ietf.org>
List-Help: <mailto:kitten-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=subscribe>
X-List-Received-Date: Thu, 12 Aug 2010 21:37:03 -0000

Hi Sam,

Sam Hartman wrote:
 [...]

>To be secure, I think a mechanism needs to:
>
>* Not provide an attacker with a weak password equivelent (yapp is out)
>* Provide mutual authentication
>* Provide channel binding or security layers
>  
>
Agreed.
Can we change the proposed OAuth mechanism to satisfy these requirements?

>I understand we've made exceptions for SAML and OpenID.  Perhaps after
>Oauth is done we'll look at its deployment and make a similar exception.
>
>A lot of my reasoning here is that registering a SASL mechanism is easy:
>it's first-come, first-serve. So, what are you actually getting out of
>spending IETF security area review cycles and IESG cycles if at the end
>of the day the result is not actually secure.
>  
>
I am a big believer in helping people to avoid damage by designing 
something insecure. So I would rather avoid telling people "go away, we 
are not interested".


From hartmans@mit.edu  Thu Aug 12 14:39:09 2010
Return-Path: <hartmans@mit.edu>
X-Original-To: kitten@core3.amsl.com
Delivered-To: kitten@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id C80883A6922 for <kitten@core3.amsl.com>; Thu, 12 Aug 2010 14:39:09 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -102.694
X-Spam-Level: 
X-Spam-Status: No, score=-102.694 tagged_above=-999 required=5 tests=[AWL=-0.429, BAYES_00=-2.599, IP_NOT_FRIENDLY=0.334, USER_IN_WHITELIST=-100]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id ddoBgJ4JISrS for <kitten@core3.amsl.com>; Thu, 12 Aug 2010 14:39:09 -0700 (PDT)
Received: from mail.suchdamage.org (permutation-city.suchdamage.org [69.25.196.28]) by core3.amsl.com (Postfix) with ESMTP id 0CB083A67E3 for <kitten@ietf.org>; Thu, 12 Aug 2010 14:39:08 -0700 (PDT)
Received: from carter-zimmerman.suchdamage.org (carter-zimmerman.suchdamage.org [69.25.196.178]) (using TLSv1 with cipher DHE-RSA-AES256-SHA (256/256 bits)) (Client CN "laptop", Issuer "laptop" (not verified)) by mail.suchdamage.org (Postfix) with ESMTPS id 2CC002029D; Thu, 12 Aug 2010 17:39:46 -0400 (EDT)
Received: by carter-zimmerman.suchdamage.org (Postfix, from userid 8042) id 4EADB412A; Thu, 12 Aug 2010 17:39:33 -0400 (EDT)
From: Sam Hartman <hartmans-ietf@mit.edu>
To: Nicolas Williams <Nicolas.Williams@oracle.com>
Subject: Re: Concerns about oauth sasl mechanism
References: <tsllj8dyo5p.fsf@mit.edu> <4C6460D6.7020800@stpeter.im> <tsl1va3v7z5.fsf@mit.edu> <20100812213024.GW5213@oracle.com>
Date: Thu, 12 Aug 2010 17:39:33 -0400
In-Reply-To: <20100812213024.GW5213@oracle.com> (Nicolas Williams's message of "Thu, 12 Aug 2010 16:30:24 -0500")
Message-ID: <tslsk2jtsru.fsf@mit.edu>
User-Agent: Gnus/5.110009 (No Gnus v0.9) Emacs/22.3 (gnu/linux)
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Cc: kitten@ietf.org, Sam Hartman <hartmans-ietf@mit.edu>
X-BeenThere: kitten@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Common Authentication Technologies - Next Generation <kitten.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/kitten>
List-Post: <mailto:kitten@ietf.org>
List-Help: <mailto:kitten-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=subscribe>
X-List-Received-Date: Thu, 12 Aug 2010 21:39:09 -0000

>>>>> "Nicolas" == Nicolas Williams <Nicolas.Williams@oracle.com> writes:

    Nicolas> On Thu, Aug 12, 2010 at 05:25:50PM -0400, Sam Hartman wrote:
    >> To be secure, I think a mechanism needs to:
    >> 
    >> * Not provide an attacker with a weak password equivelent (yapp
    >> is out) * Provide mutual authentication * Provide channel binding
    >> or security layers

    Nicolas> +1.

    Nicolas> (What's YAPP?)
What's Kirt's thing called?

    >> I understand we've made exceptions for SAML and OpenID.  Perhaps
    >> after Oauth is done we'll look at its deployment and make a
    >> similar exception.
    >> 
    >> A lot of my reasoning here is that registering a SASL mechanism
    >> is easy: it's first-come, first-serve. So, what are you actually
    >> getting out of spending IETF security area review cycles and IESG
    >> cycles if at the end of the day the result is not actually
    >> secure.

    Nicolas> Likely answer: the imprimatur of "RFC", even if it isn't on
    Nicolas> the Standards track.

There's an rfc-editor independent stream for that.

    Nicolas> Nico --


From hartmans@mit.edu  Thu Aug 12 14:44:01 2010
Return-Path: <hartmans@mit.edu>
X-Original-To: kitten@core3.amsl.com
Delivered-To: kitten@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 38DAA3A6958 for <kitten@core3.amsl.com>; Thu, 12 Aug 2010 14:44:01 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -102.676
X-Spam-Level: 
X-Spam-Status: No, score=-102.676 tagged_above=-999 required=5 tests=[AWL=-0.411, BAYES_00=-2.599, IP_NOT_FRIENDLY=0.334, USER_IN_WHITELIST=-100]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id og4HxsMiBWRf for <kitten@core3.amsl.com>; Thu, 12 Aug 2010 14:44:00 -0700 (PDT)
Received: from mail.suchdamage.org (permutation-city.suchdamage.org [69.25.196.28]) by core3.amsl.com (Postfix) with ESMTP id 602CD3A694E for <kitten@ietf.org>; Thu, 12 Aug 2010 14:44:00 -0700 (PDT)
Received: from carter-zimmerman.suchdamage.org (carter-zimmerman.suchdamage.org [69.25.196.178]) (using TLSv1 with cipher DHE-RSA-AES256-SHA (256/256 bits)) (Client CN "laptop", Issuer "laptop" (not verified)) by mail.suchdamage.org (Postfix) with ESMTPS id 5E046203CB; Thu, 12 Aug 2010 17:44:37 -0400 (EDT)
Received: by carter-zimmerman.suchdamage.org (Postfix, from userid 8042) id 81623412A; Thu, 12 Aug 2010 17:44:24 -0400 (EDT)
From: Sam Hartman <hartmans-ietf@mit.edu>
To: Alexey Melnikov <alexey.melnikov@isode.com>
Subject: Re: Concerns about oauth sasl mechanism
References: <tsllj8dyo5p.fsf@mit.edu> <4C6460D6.7020800@stpeter.im> <tsl1va3v7z5.fsf@mit.edu> <4C64698A.9030405@isode.com>
Date: Thu, 12 Aug 2010 17:44:24 -0400
In-Reply-To: <4C64698A.9030405@isode.com> (Alexey Melnikov's message of "Thu,  12 Aug 2010 22:37:14 +0100")
Message-ID: <tslocd7tsjr.fsf@mit.edu>
User-Agent: Gnus/5.110009 (No Gnus v0.9) Emacs/22.3 (gnu/linux)
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Cc: kitten@ietf.org, Sam Hartman <hartmans-ietf@mit.edu>
X-BeenThere: kitten@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Common Authentication Technologies - Next Generation <kitten.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/kitten>
List-Post: <mailto:kitten@ietf.org>
List-Help: <mailto:kitten-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=subscribe>
X-List-Received-Date: Thu, 12 Aug 2010 21:44:01 -0000

>>>>> "Alexey" == Alexey Melnikov <alexey.melnikov@isode.com> writes:

    Alexey> Hi Sam,
    Alexey> Sam Hartman wrote:
    Alexey> [...]
    >> 
    Alexey> I am a big believer in helping people to avoid damage by
    Alexey> designing something insecure. So I would rather avoid
    Alexey> telling people "go away, we are not interested".


O, if someone like the OpenID folks or Oauth folks come to us and want
to do the work to make something meet those security requirements, I'm
all for it!  That would be great.  However if after discussing what
someone wants with them, they're unable to design something that is
secure and that meets their needs then our goals are not aligned and we
should get out of each others' way.

If the OAuth proposal has channel bindings/mutual auth or the proponents
want to modify it to do so, then we should absolutely focus on it.

From Nicolas.Williams@oracle.com  Thu Aug 12 14:49:59 2010
Return-Path: <Nicolas.Williams@oracle.com>
X-Original-To: kitten@core3.amsl.com
Delivered-To: kitten@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id BC5793A67B1 for <kitten@core3.amsl.com>; Thu, 12 Aug 2010 14:49:59 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -6.519
X-Spam-Level: 
X-Spam-Status: No, score=-6.519 tagged_above=-999 required=5 tests=[AWL=0.079,  BAYES_00=-2.599, RCVD_IN_DNSWL_MED=-4, UNPARSEABLE_RELAY=0.001]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id 8ZIvhDs+0SO3 for <kitten@core3.amsl.com>; Thu, 12 Aug 2010 14:49:59 -0700 (PDT)
Received: from rcsinet10.oracle.com (rcsinet10.oracle.com [148.87.113.121]) by core3.amsl.com (Postfix) with ESMTP id E1D073A69B2 for <kitten@ietf.org>; Thu, 12 Aug 2010 14:49:58 -0700 (PDT)
Received: from acsinet15.oracle.com (acsinet15.oracle.com [141.146.126.227]) by rcsinet10.oracle.com (Switch-3.4.2/Switch-3.4.2) with ESMTP id o7CLoXI0017387 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=OK); Thu, 12 Aug 2010 21:50:35 GMT
Received: from acsmt355.oracle.com (acsmt355.oracle.com [141.146.40.155]) by acsinet15.oracle.com (Switch-3.4.2/Switch-3.4.1) with ESMTP id o7BIx0Fn002958; Thu, 12 Aug 2010 21:50:31 GMT
Received: from abhmt005.oracle.com by acsmt355.oracle.com with ESMTP id 491709231281649767; Thu, 12 Aug 2010 14:49:27 -0700
Received: from oracle.com (/129.153.128.104) by default (Oracle Beehive Gateway v4.0) with ESMTP ; Thu, 12 Aug 2010 14:49:26 -0700
Date: Thu, 12 Aug 2010 16:49:22 -0500
From: Nicolas Williams <Nicolas.Williams@oracle.com>
To: Sam Hartman <hartmans-ietf@mit.edu>
Subject: Re: Concerns about oauth sasl mechanism
Message-ID: <20100812214921.GX5213@oracle.com>
References: <tsllj8dyo5p.fsf@mit.edu> <4C6460D6.7020800@stpeter.im> <tsl1va3v7z5.fsf@mit.edu> <20100812213024.GW5213@oracle.com> <tslsk2jtsru.fsf@mit.edu>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <tslsk2jtsru.fsf@mit.edu>
User-Agent: Mutt/1.5.20 (2010-03-02)
Cc: kitten@ietf.org
X-BeenThere: kitten@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Common Authentication Technologies - Next Generation <kitten.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/kitten>
List-Post: <mailto:kitten@ietf.org>
List-Help: <mailto:kitten-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=subscribe>
X-List-Received-Date: Thu, 12 Aug 2010 21:49:59 -0000

On Thu, Aug 12, 2010 at 05:39:33PM -0400, Sam Hartman wrote:
> >>>>> "Nicolas" == Nicolas Williams <Nicolas.Williams@oracle.com> writes:
> 
>     Nicolas> On Thu, Aug 12, 2010 at 05:25:50PM -0400, Sam Hartman wrote:
>     >> To be secure, I think a mechanism needs to:
>     >> 
>     >> * Not provide an attacker with a weak password equivelent (yapp
>     >> is out) * Provide mutual authentication * Provide channel binding
>     >> or security layers
> 
>     Nicolas> +1.
> 
>     Nicolas> (What's YAPP?)
> What's Kirt's thing called?

Ah, YAP.  It requires channel binding though, and it requires unique
channel bindings.

I like to think of YAP as an optimization of SCRAM that can be used when
you have unique channel bindings.

Nico
-- 

From hartmans@mit.edu  Thu Aug 12 15:19:20 2010
Return-Path: <hartmans@mit.edu>
X-Original-To: kitten@core3.amsl.com
Delivered-To: kitten@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 398EA3A6A13 for <kitten@core3.amsl.com>; Thu, 12 Aug 2010 15:19:20 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -102.711
X-Spam-Level: 
X-Spam-Status: No, score=-102.711 tagged_above=-999 required=5 tests=[AWL=-0.446, BAYES_00=-2.599, IP_NOT_FRIENDLY=0.334, USER_IN_WHITELIST=-100]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id zr3aifReG3Dm for <kitten@core3.amsl.com>; Thu, 12 Aug 2010 15:19:19 -0700 (PDT)
Received: from mail.suchdamage.org (permutation-city.suchdamage.org [69.25.196.28]) by core3.amsl.com (Postfix) with ESMTP id 73E993A67E3 for <kitten@ietf.org>; Thu, 12 Aug 2010 15:19:19 -0700 (PDT)
Received: from carter-zimmerman.suchdamage.org (carter-zimmerman.suchdamage.org [69.25.196.178]) (using TLSv1 with cipher DHE-RSA-AES256-SHA (256/256 bits)) (Client CN "laptop", Issuer "laptop" (not verified)) by mail.suchdamage.org (Postfix) with ESMTPS id 5E4AD200B4; Thu, 12 Aug 2010 18:19:52 -0400 (EDT)
Received: by carter-zimmerman.suchdamage.org (Postfix, from userid 8042) id 8CF0B412A; Thu, 12 Aug 2010 18:19:39 -0400 (EDT)
From: Sam Hartman <hartmans-ietf@mit.edu>
To: Nicolas Williams <Nicolas.Williams@oracle.com>
Subject: Re: Concerns about oauth sasl mechanism
References: <tsllj8dyo5p.fsf@mit.edu> <4C6460D6.7020800@stpeter.im> <tsl1va3v7z5.fsf@mit.edu> <20100812213024.GW5213@oracle.com> <tslsk2jtsru.fsf@mit.edu> <20100812214921.GX5213@oracle.com>
Date: Thu, 12 Aug 2010 18:19:39 -0400
In-Reply-To: <20100812214921.GX5213@oracle.com> (Nicolas Williams's message of "Thu, 12 Aug 2010 16:49:22 -0500")
Message-ID: <tsl4oeztqx0.fsf@mit.edu>
User-Agent: Gnus/5.110009 (No Gnus v0.9) Emacs/22.3 (gnu/linux)
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Cc: kitten@ietf.org, Sam Hartman <hartmans-ietf@mit.edu>
X-BeenThere: kitten@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Common Authentication Technologies - Next Generation <kitten.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/kitten>
List-Post: <mailto:kitten@ietf.org>
List-Help: <mailto:kitten-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=subscribe>
X-List-Received-Date: Thu, 12 Aug 2010 22:19:20 -0000

>>>>> "Nicolas" == Nicolas Williams <Nicolas.Williams@oracle.com> writes:


    Nicolas> I like to think of YAP as an optimization of SCRAM that can
    Nicolas> be used when you have unique channel bindings.

I like to think of it as insecure:-)

From Nicolas.Williams@oracle.com  Thu Aug 12 15:22:17 2010
Return-Path: <Nicolas.Williams@oracle.com>
X-Original-To: kitten@core3.amsl.com
Delivered-To: kitten@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 4BE8C3A69B1 for <kitten@core3.amsl.com>; Thu, 12 Aug 2010 15:22:17 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -6.522
X-Spam-Level: 
X-Spam-Status: No, score=-6.522 tagged_above=-999 required=5 tests=[AWL=0.076,  BAYES_00=-2.599, RCVD_IN_DNSWL_MED=-4, UNPARSEABLE_RELAY=0.001]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id jgWVEJZuEl-v for <kitten@core3.amsl.com>; Thu, 12 Aug 2010 15:22:14 -0700 (PDT)
Received: from rcsinet10.oracle.com (rcsinet10.oracle.com [148.87.113.121]) by core3.amsl.com (Postfix) with ESMTP id 823023A69C2 for <kitten@ietf.org>; Thu, 12 Aug 2010 15:22:14 -0700 (PDT)
Received: from acsinet15.oracle.com (acsinet15.oracle.com [141.146.126.227]) by rcsinet10.oracle.com (Switch-3.4.2/Switch-3.4.2) with ESMTP id o7CMMnI4008061 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=OK); Thu, 12 Aug 2010 22:22:51 GMT
Received: from acsmt355.oracle.com (acsmt355.oracle.com [141.146.40.155]) by acsinet15.oracle.com (Switch-3.4.2/Switch-3.4.1) with ESMTP id o7CMMnKh001995; Thu, 12 Aug 2010 22:22:49 GMT
Received: from abhmt013.oracle.com by acsmt354.oracle.com with ESMTP id 512287431281651719; Thu, 12 Aug 2010 15:21:59 -0700
Received: from oracle.com (/129.153.128.104) by default (Oracle Beehive Gateway v4.0) with ESMTP ; Thu, 12 Aug 2010 15:21:59 -0700
Date: Thu, 12 Aug 2010 17:21:57 -0500
From: Nicolas Williams <Nicolas.Williams@oracle.com>
To: Sam Hartman <hartmans-ietf@mit.edu>
Subject: Re: Concerns about oauth sasl mechanism
Message-ID: <20100812222157.GZ5213@oracle.com>
References: <tsllj8dyo5p.fsf@mit.edu> <4C6460D6.7020800@stpeter.im> <tsl1va3v7z5.fsf@mit.edu> <20100812213024.GW5213@oracle.com> <tslsk2jtsru.fsf@mit.edu> <20100812214921.GX5213@oracle.com> <tsl4oeztqx0.fsf@mit.edu>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <tsl4oeztqx0.fsf@mit.edu>
User-Agent: Mutt/1.5.20 (2010-03-02)
Cc: kitten@ietf.org
X-BeenThere: kitten@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Common Authentication Technologies - Next Generation <kitten.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/kitten>
List-Post: <mailto:kitten@ietf.org>
List-Help: <mailto:kitten-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=subscribe>
X-List-Received-Date: Thu, 12 Aug 2010 22:22:17 -0000

On Thu, Aug 12, 2010 at 06:19:39PM -0400, Sam Hartman wrote:
> >>>>> "Nicolas" == Nicolas Williams <Nicolas.Williams@oracle.com> writes:
> 
> 
>     Nicolas> I like to think of YAP as an optimization of SCRAM that can
>     Nicolas> be used when you have unique channel bindings.
> 
> I like to think of it as insecure:-)

Well, yes :)  I don't find the one-round-trip saving to be sufficiently
compelling to have the application understand the difference between
unique and non-unique CB, much less to have the initiator implementation
inspect the CB to see that it is nominally unique (which would violate
abstractions)

Nico
-- 

From cantor.2@osu.edu  Thu Aug 12 15:41:00 2010
Return-Path: <cantor.2@osu.edu>
X-Original-To: kitten@core3.amsl.com
Delivered-To: kitten@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 70F9A3A69FF for <kitten@core3.amsl.com>; Thu, 12 Aug 2010 15:41:00 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -3.242
X-Spam-Level: 
X-Spam-Status: No, score=-3.242 tagged_above=-999 required=5 tests=[AWL=0.357,  BAYES_00=-2.599, RCVD_IN_DNSWL_LOW=-1]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id HzQvsnnnT94Y for <kitten@core3.amsl.com>; Thu, 12 Aug 2010 15:40:59 -0700 (PDT)
Received: from defang20.it.ohio-state.edu (defang20.it.ohio-state.edu [128.146.216.134]) by core3.amsl.com (Postfix) with ESMTP id 7838D3A6A40 for <kitten@ietf.org>; Thu, 12 Aug 2010 15:40:59 -0700 (PDT)
Received: from defang6.it.ohio-state.edu (defang6.it.ohio-state.edu [128.146.216.86]) by defang20.it.ohio-state.edu (8.13.7/8.13.1) with ESMTP id o7CMfYPt000806; Thu, 12 Aug 2010 18:41:34 -0400
Received: from SNOWDOG (SNOWDOG.dyn.cio.osu.edu [164.107.161.86]) by defang6.it.ohio-state.edu (8.13.7/8.13.1) with ESMTP id o7CMfY9f013772; Thu, 12 Aug 2010 18:41:34 -0400
From: "Scott Cantor" <cantor.2@osu.edu>
To: "'Sam Hartman'" <hartmans-ietf@mit.edu>
References: <tsllj8dyo5p.fsf@mit.edu> <4C6460D6.7020800@stpeter.im> <tsl1va3v7z5.fsf@mit.edu> <4C64698A.9030405@isode.com> <tslocd7tsjr.fsf@mit.edu>
In-Reply-To: <tslocd7tsjr.fsf@mit.edu>
Subject: RE: Concerns about oauth sasl mechanism
Date: Thu, 12 Aug 2010 18:41:37 -0400
Organization: The Ohio State University
Message-ID: <043001cb3a6f$866073c0$93215b40$@osu.edu>
MIME-Version: 1.0
Content-Type: text/plain; charset="US-ASCII"
Content-Transfer-Encoding: 7bit
X-Mailer: Microsoft Outlook 14.0
Thread-index: AQEtpK0+ZMhTGTaNZYkccHtUfLPXDgKpFn1eAjsa3OcCFW1apACo0J5Ek9x/5KA=
Content-language: en-us
X-CanIt-Geo: ip=128.146.216.86; country=US; region=OH; city=Columbus; latitude=39.9968; longitude=-82.9882; metrocode=535; areacode=614; http://maps.google.com/maps?q=39.9968,-82.9882&z=6
X-CanItPRO-Stream: outbound
X-Scanned-By: CanIt (www . roaringpenguin . com) on 128.146.216.134
Cc: kitten@ietf.org
X-BeenThere: kitten@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Common Authentication Technologies - Next Generation <kitten.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/kitten>
List-Post: <mailto:kitten@ietf.org>
List-Help: <mailto:kitten-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=subscribe>
X-List-Received-Date: Thu, 12 Aug 2010 22:41:00 -0000

> If the OAuth proposal has channel bindings/mutual auth or the proponents
> want to modify it to do so, then we should absolutely focus on it.

For the record, possibly repeating myself from an earlier statement, I would
like to modify my SAML proposal, with help, to allow for both.

I don't know how easy that will be or what kinds of modifications of IdPs it
will take, and I think it will be necessary to allow for their absence to
get any progress made in the interim, but I'm not interested in the lowest
bar here, just a migration path. For example, it's likely that part of
introducing such additions will be best accomplished by proposing new SAML
profiles to the OASIS TC so that they are usable outside the context of the
SASL/GSS work.

It probably also bears noting that one of the reasons why I don't care *at
all* about the fact that some IdPs might not support the ECP profile today
is that I think other modifications to IdPs (such as this kind of thing)
will continue to be necessary in the future.

-- Scott



From hartmans@mit.edu  Thu Aug 12 16:15:16 2010
Return-Path: <hartmans@mit.edu>
X-Original-To: kitten@core3.amsl.com
Delivered-To: kitten@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id C1E6A3A680C for <kitten@core3.amsl.com>; Thu, 12 Aug 2010 16:15:16 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -102.739
X-Spam-Level: 
X-Spam-Status: No, score=-102.739 tagged_above=-999 required=5 tests=[AWL=-0.474, BAYES_00=-2.599, IP_NOT_FRIENDLY=0.334, USER_IN_WHITELIST=-100]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id 7SpPPSo0FOBw for <kitten@core3.amsl.com>; Thu, 12 Aug 2010 16:15:15 -0700 (PDT)
Received: from mail.suchdamage.org (permutation-city.suchdamage.org [69.25.196.28]) by core3.amsl.com (Postfix) with ESMTP id A77F63A67B3 for <kitten@ietf.org>; Thu, 12 Aug 2010 16:15:15 -0700 (PDT)
Received: from carter-zimmerman.suchdamage.org (carter-zimmerman.suchdamage.org [69.25.196.178]) (using TLSv1 with cipher DHE-RSA-AES256-SHA (256/256 bits)) (Client CN "laptop", Issuer "laptop" (not verified)) by mail.suchdamage.org (Postfix) with ESMTPS id A2EF7203CB for <kitten@ietf.org>; Thu, 12 Aug 2010 19:15:50 -0400 (EDT)
Received: by carter-zimmerman.suchdamage.org (Postfix, from userid 8042) id 3325E412A; Thu, 12 Aug 2010 19:15:37 -0400 (EDT)
From: Sam Hartman <hartmans-ietf@mit.edu>
To: kitten@ietf.org
Subject: Speaking in support of Scott's SAML proposal
Date: Thu, 12 Aug 2010 19:15:37 -0400
Message-ID: <tslvd7fs9ra.fsf@mit.edu>
User-Agent: Gnus/5.110009 (No Gnus v0.9) Emacs/22.3 (gnu/linux)
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
X-BeenThere: kitten@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Common Authentication Technologies - Next Generation <kitten.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/kitten>
List-Post: <mailto:kitten@ietf.org>
List-Help: <mailto:kitten-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=subscribe>
X-List-Received-Date: Thu, 12 Aug 2010 23:15:16 -0000

Scott's most recent mail caused me to realize I had not explicitly
indicated support for his proposal once Shawn indicated we needed to ask
that question.


I definitely think that especially if we add CB to Scott's proposal, it
will be a good tool.
I think Simon has volunteered to help with this.
If not, I'd be happy to help as well.

From Nicolas.Williams@oracle.com  Thu Aug 12 16:24:20 2010
Return-Path: <Nicolas.Williams@oracle.com>
X-Original-To: kitten@core3.amsl.com
Delivered-To: kitten@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id A63AD3A682F for <kitten@core3.amsl.com>; Thu, 12 Aug 2010 16:24:20 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -6.526
X-Spam-Level: 
X-Spam-Status: No, score=-6.526 tagged_above=-999 required=5 tests=[AWL=0.072,  BAYES_00=-2.599, RCVD_IN_DNSWL_MED=-4, UNPARSEABLE_RELAY=0.001]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id 6ybmBqPm+c2F for <kitten@core3.amsl.com>; Thu, 12 Aug 2010 16:24:20 -0700 (PDT)
Received: from rcsinet10.oracle.com (rcsinet10.oracle.com [148.87.113.121]) by core3.amsl.com (Postfix) with ESMTP id CF9133A680C for <kitten@ietf.org>; Thu, 12 Aug 2010 16:24:19 -0700 (PDT)
Received: from acsinet15.oracle.com (acsinet15.oracle.com [141.146.126.227]) by rcsinet10.oracle.com (Switch-3.4.2/Switch-3.4.2) with ESMTP id o7CNOsi9019751 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=OK); Thu, 12 Aug 2010 23:24:56 GMT
Received: from acsmt353.oracle.com (acsmt353.oracle.com [141.146.40.153]) by acsinet15.oracle.com (Switch-3.4.2/Switch-3.4.1) with ESMTP id o7CHL7HV004221; Thu, 12 Aug 2010 23:24:54 GMT
Received: from abhmt016.oracle.com by acsmt355.oracle.com with ESMTP id 512423311281655494; Thu, 12 Aug 2010 16:24:54 -0700
Received: from oracle.com (/129.153.128.104) by default (Oracle Beehive Gateway v4.0) with ESMTP ; Thu, 12 Aug 2010 16:24:53 -0700
Date: Thu, 12 Aug 2010 18:24:48 -0500
From: Nicolas Williams <Nicolas.Williams@oracle.com>
To: Sam Hartman <hartmans-ietf@mit.edu>
Subject: Re: Speaking in support of Scott's SAML proposal
Message-ID: <20100812232448.GC5213@oracle.com>
References: <tslvd7fs9ra.fsf@mit.edu>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <tslvd7fs9ra.fsf@mit.edu>
User-Agent: Mutt/1.5.20 (2010-03-02)
Cc: kitten@ietf.org
X-BeenThere: kitten@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Common Authentication Technologies - Next Generation <kitten.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/kitten>
List-Post: <mailto:kitten@ietf.org>
List-Help: <mailto:kitten-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=subscribe>
X-List-Received-Date: Thu, 12 Aug 2010 23:24:20 -0000

On Thu, Aug 12, 2010 at 07:15:37PM -0400, Sam Hartman wrote:
> Scott's most recent mail caused me to realize I had not explicitly
> indicated support for his proposal once Shawn indicated we needed to ask
> that question.
> 
> I definitely think that especially if we add CB to Scott's proposal, it
> will be a good tool.

+1.  But, should it be in-charter, or left to be included in a future
re-charter?  I'm assuming the latter, with an excellent chance of
adopting the work if a sufficiently fleshed out proposal materializes.

> I think Simon has volunteered to help with this.
> If not, I'd be happy to help as well.

I'd review it.

From cantor.2@osu.edu  Thu Aug 12 17:42:45 2010
Return-Path: <cantor.2@osu.edu>
X-Original-To: kitten@core3.amsl.com
Delivered-To: kitten@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 631523A687C for <kitten@core3.amsl.com>; Thu, 12 Aug 2010 17:42:45 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -2.449
X-Spam-Level: 
X-Spam-Status: No, score=-2.449 tagged_above=-999 required=5 tests=[AWL=0.150,  BAYES_00=-2.599]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id o4UgUQ420s6J for <kitten@core3.amsl.com>; Thu, 12 Aug 2010 17:42:42 -0700 (PDT)
Received: from hrndva-omtalb.mail.rr.com (hrndva-omtalb.mail.rr.com [71.74.56.123]) by core3.amsl.com (Postfix) with ESMTP id 616E53A6845 for <kitten@ietf.org>; Thu, 12 Aug 2010 17:42:40 -0700 (PDT)
X-Authority-Analysis: v=1.1 cv=nbWoJD01ruQLYOYE44JZbYtzz81AmFDL3LnuPvZDHs8= c=1 sm=0 a=Q2rYhj3HWgEA:10 a=kj9zAlcOel0A:10 a=/Wh0N9815gtYTLoiNwER9g==:17 a=B8Vwz_W7sr4vv2v61vcA:9 a=kOipOE69fPloh-ClxzWd1sq33QMA:4 a=CjuIK1q_8ugA:10 a=/Wh0N9815gtYTLoiNwER9g==:117
X-Cloudmark-Score: 0
X-Originating-IP: 24.210.116.83
Received: from [24.210.116.83] ([24.210.116.83:24964] helo=SNOWDOG) by hrndva-oedge02.mail.rr.com (envelope-from <cantor.2@osu.edu>) (ecelerity 2.2.2.39 r()) with ESMTP id 04/55-01723-425946C4; Fri, 13 Aug 2010 00:43:17 +0000
From: "Scott Cantor" <cantor.2@osu.edu>
To: "'Sam Hartman'" <hartmans-ietf@mit.edu>, <kitten@ietf.org>
References: <tslvd7fs9ra.fsf@mit.edu>
In-Reply-To: <tslvd7fs9ra.fsf@mit.edu>
Subject: RE: Speaking in support of Scott's SAML proposal
Date: Thu, 12 Aug 2010 20:43:17 -0400
Organization: The Ohio State University
Message-ID: <043701cb3a80$866286e0$932794a0$@osu.edu>
MIME-Version: 1.0
Content-Type: text/plain; charset="US-ASCII"
Content-Transfer-Encoding: 7bit
X-Mailer: Microsoft Outlook 14.0
Thread-index: AQHCNyGTwgBvXMFDRVTjG5E2EdVQhJLwkQaQ
Content-language: en-us
X-BeenThere: kitten@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Common Authentication Technologies - Next Generation <kitten.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/kitten>
List-Post: <mailto:kitten@ietf.org>
List-Help: <mailto:kitten-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=subscribe>
X-List-Received-Date: Fri, 13 Aug 2010 00:42:45 -0000

> I definitely think that especially if we add CB to Scott's proposal, it
> will be a good tool.
> I think Simon has volunteered to help with this.

Simon donated GS2 text to the OpenID and SAML browser mechanisms that I
think is the initial step, but as I understood the wording, it does not
allow CB, it just explains you can't have them. I was confused by this given
what I read in the GS2 RFC, which seemed to imply that all such mechanisms
had to support CB. Simon tried to explain that in email, but I don't think I
understood the explanation.

In any case, if a first step is adding the GS2 material without CB, I can do
that fairly soon.

> If not, I'd be happy to help as well.

I think a starting point will probably need to be me spending some time
reading up on CB, and perhaps getting a straw man from somebody with some
rudimentary SAML knowledge who could suggest what the basic outline of it
would have to look like.

-- Scott



From simon@josefsson.org  Fri Aug 13 01:58:24 2010
Return-Path: <simon@josefsson.org>
X-Original-To: kitten@core3.amsl.com
Delivered-To: kitten@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 603AB3A687E for <kitten@core3.amsl.com>; Fri, 13 Aug 2010 01:58:24 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -102.683
X-Spam-Level: 
X-Spam-Status: No, score=-102.683 tagged_above=-999 required=5 tests=[AWL=-0.084, BAYES_00=-2.599, USER_IN_WHITELIST=-100]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id biGOYQTsw8UX for <kitten@core3.amsl.com>; Fri, 13 Aug 2010 01:58:23 -0700 (PDT)
Received: from yxa-v.extundo.com (yxa-v.extundo.com [83.241.177.39]) by core3.amsl.com (Postfix) with ESMTP id E444C3A6830 for <kitten@ietf.org>; Fri, 13 Aug 2010 01:58:22 -0700 (PDT)
Received: from mocca (c80-216-27-64.bredband.comhem.se [80.216.27.64]) (authenticated bits=0) by yxa-v.extundo.com (8.14.3/8.14.3/Debian-5+lenny1) with ESMTP id o7D8wlQC031595 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES128-SHA bits=128 verify=NOT); Fri, 13 Aug 2010 10:58:50 +0200
From: Simon Josefsson <simon@josefsson.org>
To: "Scott Cantor" <cantor.2@osu.edu>
Subject: Re: Speaking in support of Scott's SAML proposal
References: <tslvd7fs9ra.fsf@mit.edu> <043701cb3a80$866286e0$932794a0$@osu.edu>
OpenPGP: id=B565716F; url=http://josefsson.org/key.txt
X-Hashcash: 1:22:100813:cantor.2@osu.edu::Udu63GI/3yK7Ib0y:5vUy
X-Hashcash: 1:22:100813:kitten@ietf.org::cy9IAvqYRZ0kaURo:4pV7
X-Hashcash: 1:22:100813:hartmans-ietf@mit.edu::l6+ruKD1YRH7TThw:6CNc
Date: Fri, 13 Aug 2010 10:58:44 +0200
In-Reply-To: <043701cb3a80$866286e0$932794a0$@osu.edu> (Scott Cantor's message of "Thu, 12 Aug 2010 20:43:17 -0400")
Message-ID: <87hbiysxbv.fsf@mocca.josefsson.org>
User-Agent: Gnus/5.110011 (No Gnus v0.11) Emacs/23.1 (gnu/linux)
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
X-Virus-Scanned: clamav-milter 0.96.1 at yxa-v
X-Virus-Status: Clean
Cc: kitten@ietf.org, 'Sam Hartman' <hartmans-ietf@mit.edu>
X-BeenThere: kitten@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Common Authentication Technologies - Next Generation <kitten.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/kitten>
List-Post: <mailto:kitten@ietf.org>
List-Help: <mailto:kitten-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=subscribe>
X-List-Received-Date: Fri, 13 Aug 2010 08:58:24 -0000

"Scott Cantor" <cantor.2@osu.edu> writes:

>> I definitely think that especially if we add CB to Scott's proposal, it
>> will be a good tool.
>> I think Simon has volunteered to help with this.
>
> Simon donated GS2 text to the OpenID and SAML browser mechanisms that I
> think is the initial step, but as I understood the wording, it does not
> allow CB, it just explains you can't have them.

Right.  This was an initial strawman.  I have not seen a way to support
channel bindings with OpenID, but with SAML there should certainly be a
way.  The text in the SAML spec were just copied from the OpenID text
that I wrote, and eventually I don't think they can be exactly the same
so they will need to be somewhat difference down the line.

> I was confused by this given what I read in the GS2 RFC, which seemed
> to imply that all such mechanisms had to support CB. Simon tried to
> explain that in email, but I don't think I understood the explanation.

Here is the short version:

GS2 requires that the GSS-API mechanism that GS2 wraps supports channel
bindings on the GSS-API level.  This is essential to the GS2 design and
there is no way to make anything work (securely) without it.  This is
not channel binding to a TLS session or anything like that, this is just
the basic support for doing channel binding to application-provided
data.

A GS2 mechanism does not have to support channel binding on the SASL
level, but it is recommended.

Consider real examples: The Kerberos GSS-API mechanism supports channel
bindings (and mutual authentication), so it is possible to wrap it in
GS2 so that the result becomes a SASL mechanism.  This SASL mechanism
can support binding to a TLS channel, or it can be used without binding
to a TLS channel.  This is the SASL level channel binding.  GS2
negotiates whether the SASL level channel binding support through the
GS2-KRB5 and GS2-KRB5-PLUS mechanism names.  Both variants uses GSS-API
level channel binding to make the GS2 construct secure.

> In any case, if a first step is adding the GS2 material without CB, I can do
> that fairly soon.

I think that would be a good start.  I'd be happy to help, although the
basics are what is done in the OpenID document already.

>> If not, I'd be happy to help as well.
>
> I think a starting point will probably need to be me spending some time
> reading up on CB, and perhaps getting a straw man from somebody with some
> rudimentary SAML knowledge who could suggest what the basic outline of it
> would have to look like.

I'll prepare something for you.

/Simon

From simon@josefsson.org  Fri Aug 13 02:13:09 2010
Return-Path: <simon@josefsson.org>
X-Original-To: kitten@core3.amsl.com
Delivered-To: kitten@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id B99783A6993 for <kitten@core3.amsl.com>; Fri, 13 Aug 2010 02:13:09 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -102.677
X-Spam-Level: 
X-Spam-Status: No, score=-102.677 tagged_above=-999 required=5 tests=[AWL=-0.078, BAYES_00=-2.599, USER_IN_WHITELIST=-100]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id qtZBPXiSQkdw for <kitten@core3.amsl.com>; Fri, 13 Aug 2010 02:12:58 -0700 (PDT)
Received: from yxa-v.extundo.com (yxa-v.extundo.com [83.241.177.39]) by core3.amsl.com (Postfix) with ESMTP id C7FB63A694C for <kitten@ietf.org>; Fri, 13 Aug 2010 02:12:48 -0700 (PDT)
Received: from mocca (c80-216-27-64.bredband.comhem.se [80.216.27.64]) (authenticated bits=0) by yxa-v.extundo.com (8.14.3/8.14.3/Debian-5+lenny1) with ESMTP id o7D9DK5F032498 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES128-SHA bits=128 verify=NOT); Fri, 13 Aug 2010 11:13:22 +0200
From: Simon Josefsson <simon@josefsson.org>
To: "Scott Cantor" <cantor.2@osu.edu>
Subject: Re: Speaking in support of Scott's SAML proposal
References: <tslvd7fs9ra.fsf@mit.edu> <043701cb3a80$866286e0$932794a0$@osu.edu>
OpenPGP: id=B565716F; url=http://josefsson.org/key.txt
X-Hashcash: 1:22:100813:cantor.2@osu.edu::GehTr6YoICFUMH1G:7NS/
X-Hashcash: 1:22:100813:kitten@ietf.org::LlJGoD4hccQmCHNl:AKDY
X-Hashcash: 1:22:100813:hartmans-ietf@mit.edu::Z6/VyYuCzItL4guC:Pmih
Date: Fri, 13 Aug 2010 11:13:18 +0200
In-Reply-To: <043701cb3a80$866286e0$932794a0$@osu.edu> (Scott Cantor's message of "Thu, 12 Aug 2010 20:43:17 -0400")
Message-ID: <87aaoqswnl.fsf@mocca.josefsson.org>
User-Agent: Gnus/5.110011 (No Gnus v0.11) Emacs/23.1 (gnu/linux)
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
X-Virus-Scanned: clamav-milter 0.96.1 at yxa-v
X-Virus-Status: Clean
Cc: kitten@ietf.org, 'Sam Hartman' <hartmans-ietf@mit.edu>
X-BeenThere: kitten@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Common Authentication Technologies - Next Generation <kitten.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/kitten>
List-Post: <mailto:kitten@ietf.org>
List-Help: <mailto:kitten-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=subscribe>
X-List-Received-Date: Fri, 13 Aug 2010 09:13:09 -0000

Scott,

Below are the changes I'd like to see in draft-cantor-ietf-sasl-saml-ec
to provide initial support for GS2.  If you'd like to co-author the
document, I'd be happy to edit the XML file to include these changes,
and any more that will come.

There is more needed than what's shown below, but that can come later.
The most important outstanding issues I see are:

* It appears as if we need to define a new GSS-API name type for SAML
  identifiers.  Your document should use the same SAML GSS-API name type
  as Klaas/Eliot's SAML mechanism, so possibly this should be done in a
  separate document that you both can reference?  Or the text could be
  duplicated, of course, but that seems sub-optimal.  Alternatively,
  there could be one name type for SAML EC and one for SAML non-EC, but
  I don't see any reason for that right now, and I see some arguments
  against.

* It should be possible to get (SASL level) channel binding support with
  SAML, and this should be added.  What is needed here is to transport
  the channel binding data (or a hash thereof) integrity protected
  through the SAML assertions.  Is that possible?  My SAML-fu is weak
  here.

* Support for GSS-API per-message tokens and GSS_Pseudo_random, similar
  to what is done for SCRAM.  What is needed here is to generate a
  symmetric key securely out of the SAML negotiation, how can that be
  done securely with SAML?  Is there already specifications on this?
  This seems like a generally useful thing to have, so if there is
  nothing like this today, perhaps we could develop it.

Are there other aspects?

/Simon

* Change title + abstract + intro to mention GSS-API

* Change section 4.2 to say that clients send the GS2 header to initiate
  connection

* Add a new section like this:

X. SAML GSS-API Mechanism Specification

   This section and its sub-sections and all normative references of it
   not referenced elsewhere in this document are INFORMATIONAL for SASL
   implementors, but they are NORMATIVE for GSS-API implementors.

   The SAML SASL Enhanced Clients mechanism is actually also a GSS-API
   mechanism.  The messages are the same, but a) the GS2 header on the
   client's first message and channel binding data is excluded when SAML
   EC is used as a GSS-API mechanism, and b) the RFC2743 section 3.1
   initial context token header is prefixed to the client's first
   authentication message (context token).

   The GSS-API mechanism OID for SAML EC is 1.3.6.1.4.1.11591.4.6.

   SAML EC security contexts always have the mutual_state flag
   (GSS_C_MUTUAL_FLAG) set to TRUE.  SAML EC does not support credential
   delegation, therefore SAML EC security contexts alway have the
   deleg_state flag (GSS_C_DELEG_FLAG) set to FALSE.

   The SAML EC mechanism does not support per-message tokens or
   GSS_Pseudo_random.

X.1. GSS-API Principal Name Types for SAML EC

   SAML EC supports standard generic name syntaxes for acceptors such as
   GSS_C_NT_HOSTBASED_SERVICE (see [RFC2743], Section 4.1).

   SAML EC supports only a single name type for initiators:
   GSS_C_NT_USER_NAME.  GSS_C_NT_USER_NAME is the default name type for
   SAML EC.

   The query, display, and exported name syntaxes for SAML EC principal
   names are all the same.  There are no SAML EC-specific name syntaxes
   -- applications should use generic GSS-API name types such as
   GSS_C_NT_USER_NAME and GSS_C_NT_HOSTBASED_SERVICE (see [RFC2743],
   Section 4).  The exported name token does, of course, conform to
   [RFC2743], Section 3.2, but the "NAME" part of the token should be
   treated as a potential input string to the SAML EC name normalization
   rules.

   GSS-API name attributes may be defined in the future to hold the
   normalized SAML EC Identifier.

From hannes.tschofenig@nsn.com  Fri Aug 13 04:58:48 2010
Return-Path: <hannes.tschofenig@nsn.com>
X-Original-To: kitten@core3.amsl.com
Delivered-To: kitten@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id C216F3A6878 for <kitten@core3.amsl.com>; Fri, 13 Aug 2010 04:58:48 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -102.506
X-Spam-Level: 
X-Spam-Status: No, score=-102.506 tagged_above=-999 required=5 tests=[AWL=0.093, BAYES_00=-2.599, USER_IN_WHITELIST=-100]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id O9O9kgrSmNLN for <kitten@core3.amsl.com>; Fri, 13 Aug 2010 04:58:47 -0700 (PDT)
Received: from demumfd001.nsn-inter.net (demumfd001.nsn-inter.net [93.183.12.32]) by core3.amsl.com (Postfix) with ESMTP id 7AB0D3A6857 for <kitten@ietf.org>; Fri, 13 Aug 2010 04:58:47 -0700 (PDT)
Received: from demuprx017.emea.nsn-intra.net ([10.150.129.56]) by demumfd001.nsn-inter.net (8.12.11.20060308/8.12.11) with ESMTP id o7DBxN5c023011 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=OK); Fri, 13 Aug 2010 13:59:23 +0200
Received: from demuexc025.nsn-intra.net (demuexc025.nsn-intra.net [10.159.32.12]) by demuprx017.emea.nsn-intra.net (8.12.11.20060308/8.12.11) with ESMTP id o7DBxLMw013390; Fri, 13 Aug 2010 13:59:22 +0200
Received: from FIESEXC015.nsn-intra.net ([10.159.0.23]) by demuexc025.nsn-intra.net with Microsoft SMTPSVC(6.0.3790.4675);  Fri, 13 Aug 2010 13:59:04 +0200
X-MimeOLE: Produced By Microsoft Exchange V6.5
Content-class: urn:content-classes:message
MIME-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Content-Transfer-Encoding: quoted-printable
Subject: RE: Concerns about oauth sasl mechanism
Date: Fri, 13 Aug 2010 14:59:02 +0300
Message-ID: <3D3C75174CB95F42AD6BCC56E5555B4502E9BF74@FIESEXC015.nsn-intra.net>
In-Reply-To: <tsllj8dyo5p.fsf@mit.edu>
X-MS-Has-Attach: 
X-MS-TNEF-Correlator: 
Thread-Topic: Concerns about oauth sasl mechanism
Thread-Index: Acs5htsyUgTBZ3oXTNqUAAphCXkfxgBVj+aw
References: <tsllj8dyo5p.fsf@mit.edu>
From: "Tschofenig, Hannes (NSN - FI/Espoo)" <hannes.tschofenig@nsn.com>
To: "ext Sam Hartman" <hartmans-ietf@mit.edu>, <kitten@ietf.org>
X-OriginalArrivalTime: 13 Aug 2010 11:59:04.0025 (UTC) FILETIME=[ECCDC490:01CB3ADE]
X-BeenThere: kitten@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Common Authentication Technologies - Next Generation <kitten.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/kitten>
List-Post: <mailto:kitten@ietf.org>
List-Help: <mailto:kitten-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=subscribe>
X-List-Received-Date: Fri, 13 Aug 2010 11:58:48 -0000

Hi Sam,=20

"I haven't followed xyz but here are my comments" is not a good start.=20
=20
I hear complaints about this and that but then nobody cares about
posting anything to the Oauth mailing list nor to show up at the Oauth
meeting either. Almost nobody at the OAuth face-to-face meeting has read
the specs.=20

Having said that I find it a bit strange to require channel bindings.
The funny thing is that I have heard that comment already at one of the
Oauth meetings. Then, I asked what this actually means. The answer was
*I don't know*. So, what should I think about these comments?

I am happy to have a discussion about channel bindings in the context of
Oauth (and of Oauth - SASL in particular) and even about how to re-write
RFC 5056 in a way that it can be understood by humans. Interestingly
enough I tried to figure out what Channel Bindings would provide us with
SASL OpenID (together with my co-authors) and we weren't convinced about
the value!

Ciao
Hannes

> I'll admit that I'm a bit concerned about the never of SASL/GSS
> mechanisms we're creating that have no better security than=20
> cram-md5 and
> may have somewhat less.
>=20
> I don't think we need any more SASL mechanisms that provide neither
> security layers nor channel binding.  I haven't followed the OAuth
> standards in a while. If they currently have something that=20
> can be used
> for channel binding or if they gain such a facility, I think
> standardizing such a mechanism here would be great. =20
> Otherwise, I don't
> think it's a good use of this WG's time.
>=20
> Obviously, nothing stops proponents of such a mechanism of registering
> it or even getting an non-IETF RFC.  I guess nothing even stops them
> from trying to recharter OAuth, although I would bring up the same
> concern in that case as well.
>=20
> I understand that a similar argument can be made about the SAML and
> OpenID mechanism. First, I have some hope for the SAML=20
> mechanism.  I'll
> admit that I'm quite conflicted on the OpenID mechanism and=20
> that I can't
> entirely justify why I didn't raise this objection there. Roughly, I
> think that's very close to the line and this is over the line.
>=20
> --Sam


From hannes.tschofenig@nsn.com  Fri Aug 13 05:05:54 2010
Return-Path: <hannes.tschofenig@nsn.com>
X-Original-To: kitten@core3.amsl.com
Delivered-To: kitten@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id B5E063A6946 for <kitten@core3.amsl.com>; Fri, 13 Aug 2010 05:05:54 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -102.509
X-Spam-Level: 
X-Spam-Status: No, score=-102.509 tagged_above=-999 required=5 tests=[AWL=0.090, BAYES_00=-2.599, USER_IN_WHITELIST=-100]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id 610obi2F3gUQ for <kitten@core3.amsl.com>; Fri, 13 Aug 2010 05:05:53 -0700 (PDT)
Received: from demumfd001.nsn-inter.net (demumfd001.nsn-inter.net [93.183.12.32]) by core3.amsl.com (Postfix) with ESMTP id 60B453A677C for <kitten@ietf.org>; Fri, 13 Aug 2010 05:05:53 -0700 (PDT)
Received: from demuprx016.emea.nsn-intra.net ([10.150.129.55]) by demumfd001.nsn-inter.net (8.12.11.20060308/8.12.11) with ESMTP id o7DC6TZr006334 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=OK); Fri, 13 Aug 2010 14:06:29 +0200
Received: from demuexc022.nsn-intra.net (demuexc022.nsn-intra.net [10.150.128.35]) by demuprx016.emea.nsn-intra.net (8.12.11.20060308/8.12.11) with ESMTP id o7DC6SqA020531; Fri, 13 Aug 2010 14:06:29 +0200
Received: from FIESEXC015.nsn-intra.net ([10.159.0.23]) by demuexc022.nsn-intra.net with Microsoft SMTPSVC(6.0.3790.4675);  Fri, 13 Aug 2010 14:06:15 +0200
X-MimeOLE: Produced By Microsoft Exchange V6.5
Content-class: urn:content-classes:message
MIME-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Content-Transfer-Encoding: quoted-printable
Subject: RE: Concerns about oauth sasl mechanism
Date: Fri, 13 Aug 2010 15:06:14 +0300
Message-ID: <3D3C75174CB95F42AD6BCC56E5555B4502E9BF85@FIESEXC015.nsn-intra.net>
In-Reply-To: <4C64698A.9030405@isode.com>
X-MS-Has-Attach: 
X-MS-TNEF-Correlator: 
Thread-Topic: Concerns about oauth sasl mechanism
Thread-Index: Acs6ZppDwNR1C36QTDek+6DrGdzMlgAeGwzg
References: <tsllj8dyo5p.fsf@mit.edu> <4C6460D6.7020800@stpeter.im><tsl1va3v7z5.fsf@mit.edu> <4C64698A.9030405@isode.com>
From: "Tschofenig, Hannes (NSN - FI/Espoo)" <hannes.tschofenig@nsn.com>
To: "ext Alexey Melnikov" <alexey.melnikov@isode.com>, "Sam Hartman" <hartmans-ietf@mit.edu>
X-OriginalArrivalTime: 13 Aug 2010 12:06:15.0684 (UTC) FILETIME=[EE17B440:01CB3ADF]
Cc: kitten@ietf.org
X-BeenThere: kitten@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Common Authentication Technologies - Next Generation <kitten.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/kitten>
List-Post: <mailto:kitten@ietf.org>
List-Help: <mailto:kitten-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=subscribe>
X-List-Received-Date: Fri, 13 Aug 2010 12:05:54 -0000

Are we standardizing research results or standardizing what industry
does? Please have a look at http://tools.ietf.org/search/rfc5218

I am happy to discuss the security properties of various aspects but
let's pick it one by one and also consider deployability.=20

* Mutual authentication: We are talking about a three party protocol
here and so what does mutual authentication actually mean.=20

* Channel Binding: Again, a different setup than described in RFC 5056.
Would be nice to know if someone has deployment experience with it.=20

* Authentication: Like with many identity management systems the
authentication part is actually outside the scope of these protocols.
This is true for SAML, OpenID, etc.=20

Ciao
Hannes

> -----Original Message-----
> From: kitten-bounces@ietf.org=20
> [mailto:kitten-bounces@ietf.org] On Behalf Of ext Alexey Melnikov
> Sent: Friday, August 13, 2010 12:37 AM
> To: Sam Hartman
> Cc: kitten@ietf.org
> Subject: Re: Concerns about oauth sasl mechanism
>=20
> Hi Sam,
>=20
> Sam Hartman wrote:
>  [...]
>=20
> >To be secure, I think a mechanism needs to:
> >
> >* Not provide an attacker with a weak password equivelent=20
> (yapp is out)
> >* Provide mutual authentication
> >* Provide channel binding or security layers
> > =20
> >
> Agreed.
> Can we change the proposed OAuth mechanism to satisfy these=20
> requirements?
>=20
> >I understand we've made exceptions for SAML and OpenID. =20
> Perhaps after
> >Oauth is done we'll look at its deployment and make a=20
> similar exception.
> >
> >A lot of my reasoning here is that registering a SASL=20
> mechanism is easy:
> >it's first-come, first-serve. So, what are you actually=20
> getting out of
> >spending IETF security area review cycles and IESG cycles if=20
> at the end
> >of the day the result is not actually secure.
> > =20
> >
> I am a big believer in helping people to avoid damage by designing=20
> something insecure. So I would rather avoid telling people=20
> "go away, we=20
> are not interested".
>=20
> _______________________________________________
> Kitten mailing list
> Kitten@ietf.org
> https://www.ietf.org/mailman/listinfo/kitten
>=20

From hannes.tschofenig@nsn.com  Fri Aug 13 05:19:05 2010
Return-Path: <hannes.tschofenig@nsn.com>
X-Original-To: kitten@core3.amsl.com
Delivered-To: kitten@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 2E6BE3A6946 for <kitten@core3.amsl.com>; Fri, 13 Aug 2010 05:19:05 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -102.511
X-Spam-Level: 
X-Spam-Status: No, score=-102.511 tagged_above=-999 required=5 tests=[AWL=0.088, BAYES_00=-2.599, USER_IN_WHITELIST=-100]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id Yp84UH7xLlvB for <kitten@core3.amsl.com>; Fri, 13 Aug 2010 05:19:04 -0700 (PDT)
Received: from demumfd002.nsn-inter.net (demumfd002.nsn-inter.net [93.183.12.31]) by core3.amsl.com (Postfix) with ESMTP id C90F63A680D for <kitten@ietf.org>; Fri, 13 Aug 2010 05:19:03 -0700 (PDT)
Received: from demuprx016.emea.nsn-intra.net ([10.150.129.55]) by demumfd002.nsn-inter.net (8.12.11.20060308/8.12.11) with ESMTP id o7DCJd9s022997 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=OK); Fri, 13 Aug 2010 14:19:39 +0200
Received: from demuexc025.nsn-intra.net (demuexc025.nsn-intra.net [10.159.32.12]) by demuprx016.emea.nsn-intra.net (8.12.11.20060308/8.12.11) with ESMTP id o7DCJdgM029364; Fri, 13 Aug 2010 14:19:39 +0200
Received: from FIESEXC015.nsn-intra.net ([10.159.0.23]) by demuexc025.nsn-intra.net with Microsoft SMTPSVC(6.0.3790.4675);  Fri, 13 Aug 2010 14:19:26 +0200
X-MimeOLE: Produced By Microsoft Exchange V6.5
Content-class: urn:content-classes:message
MIME-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Content-Transfer-Encoding: quoted-printable
Subject: RE: Concerns about oauth sasl mechanism
Date: Fri, 13 Aug 2010 15:19:25 +0300
Message-ID: <3D3C75174CB95F42AD6BCC56E5555B4502E9BF97@FIESEXC015.nsn-intra.net>
In-Reply-To: <tslocd7tsjr.fsf@mit.edu>
X-MS-Has-Attach: 
X-MS-TNEF-Correlator: 
Thread-Topic: Concerns about oauth sasl mechanism
Thread-Index: Acs6Z5Mp0dk6M+lyQ8a2kt9ZYKsqYgAeGWBw
References: <tsllj8dyo5p.fsf@mit.edu> <4C6460D6.7020800@stpeter.im><tsl1va3v7z5.fsf@mit.edu> <4C64698A.9030405@isode.com> <tslocd7tsjr.fsf@mit.edu>
From: "Tschofenig, Hannes (NSN - FI/Espoo)" <hannes.tschofenig@nsn.com>
To: "ext Sam Hartman" <hartmans-ietf@mit.edu>, "Alexey Melnikov" <alexey.melnikov@isode.com>
X-OriginalArrivalTime: 13 Aug 2010 12:19:26.0650 (UTC) FILETIME=[C58B89A0:01CB3AE1]
Cc: kitten@ietf.org
X-BeenThere: kitten@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Common Authentication Technologies - Next Generation <kitten.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/kitten>
List-Post: <mailto:kitten@ietf.org>
List-Help: <mailto:kitten-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=subscribe>
X-List-Received-Date: Fri, 13 Aug 2010 12:19:05 -0000

Hi Sam,=20


> O, if someone like the OpenID folks or Oauth folks come to us and want
> to do the work to make something meet those security requirements, I'm
> all for it!  That would be great.  However if after discussing what
> someone wants with them, they're unable to design something that is
> secure and that meets their needs then our goals are not=20
> aligned and we
> should get out of each others' way.

I have to comment on that.=20

I am not going to talk about my impression of what went wrong when some
of the OpenID guys came to the IETF and wanted to do the protocol work
here. Instead, I focus on the Oauth case.=20

The Oauth community came to the IETF to do their standardization work
because they hoped they would get really good security advice (and
utilize some of the other protocol experience we have in the IETF).=20

Since the formation of the group I tried to involve various groups in
the IETF to provide their review comments. I gave a presentation at the
SECDIR lunch meeting, I volunteered to give a tutorial at the SAAG
meeting (but instead the presentation about MashSSL was favored by the
Ads and I could only put the slides in the proceedings), I posted mails
to various lists and people. At the last IETF meeting Blaine and I gave
an Oauth tutorial on Friday to involve those folks who have not had time
to read the spec.=20

With all this I got very little feedback from the IETF security
community.=20

The folks in Oauth are not impressed by random statements like "we need
channel bindings" because they want to know "What is it?" and "What are
the benefits?". I think that this is the right attitude.=20

Then, these guys implement and deploy. Many folks on security mailing
lists neither implement nor deploy. They just write documents and
complain. That's cheap.=20

Working in the Web 2.0 environment they want to get something done soon
(rather than in 5 years). Clearly a different culture. Some of the
design constraints therefore come from this working environment. Here is
an example: You cannot sign the entire HTTP message because the Web
programming language does not provide you with the entire HTTP message.
Another example: You have to work with the same origin policy in a web
browser and hence certain encodings and message flow look at bit funny.=20

In a nutshell, those who complain here should sign up at the Oauth list
and do a review.=20

Ciao
Hannes

From cantor.2@osu.edu  Fri Aug 13 08:15:51 2010
Return-Path: <cantor.2@osu.edu>
X-Original-To: kitten@core3.amsl.com
Delivered-To: kitten@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 0C2193A6A0D for <kitten@core3.amsl.com>; Fri, 13 Aug 2010 08:15:51 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -2.55
X-Spam-Level: 
X-Spam-Status: No, score=-2.55 tagged_above=-999 required=5 tests=[AWL=0.049,  BAYES_00=-2.599]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id tEZI-tv+nWAx for <kitten@core3.amsl.com>; Fri, 13 Aug 2010 08:15:50 -0700 (PDT)
Received: from defang19.it.ohio-state.edu (defang19.it.ohio-state.edu [128.146.216.133]) by core3.amsl.com (Postfix) with ESMTP id 24DBD3A6A0C for <kitten@ietf.org>; Fri, 13 Aug 2010 08:15:50 -0700 (PDT)
Received: from defang10.it.ohio-state.edu (defang10.it.ohio-state.edu [128.146.216.79]) by defang19.it.ohio-state.edu (8.13.7/8.13.1) with ESMTP id o7DFGQBX008968; Fri, 13 Aug 2010 11:16:26 -0400
Received: from SNOWDOG ([128.146.178.34]) by defang10.it.ohio-state.edu (8.13.7/8.13.1) with ESMTP id o7DFGPQD026198; Fri, 13 Aug 2010 11:16:26 -0400
From: "Scott Cantor" <cantor.2@osu.edu>
To: "'Simon Josefsson'" <simon@josefsson.org>
References: <tslvd7fs9ra.fsf@mit.edu> <043701cb3a80$866286e0$932794a0$@osu.edu> <87hbiysxbv.fsf@mocca.josefsson.org>
In-Reply-To: <87hbiysxbv.fsf@mocca.josefsson.org>
Subject: RE: Speaking in support of Scott's SAML proposal
Date: Fri, 13 Aug 2010 11:16:29 -0400
Organization: The Ohio State University
Message-ID: <044801cb3afa$8158b320$840a1960$@osu.edu>
MIME-Version: 1.0
Content-Type: text/plain; charset="US-ASCII"
Content-Transfer-Encoding: 7bit
X-Mailer: Microsoft Outlook 14.0
Thread-index: AQHCNyGTwgBvXMFDRVTjG5E2EdVQhAFlrREYAg6JzPSS1eNOMA==
Content-language: en-us
X-CanIt-Geo: ip=128.146.216.79; country=US; region=OH; city=Columbus; latitude=39.9968; longitude=-82.9882; metrocode=535; areacode=614; http://maps.google.com/maps?q=39.9968,-82.9882&z=6
X-CanItPRO-Stream: outbound
X-Scanned-By: CanIt (www . roaringpenguin . com) on 128.146.216.133
Cc: kitten@ietf.org
X-BeenThere: kitten@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Common Authentication Technologies - Next Generation <kitten.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/kitten>
List-Post: <mailto:kitten@ietf.org>
List-Help: <mailto:kitten-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=subscribe>
X-List-Received-Date: Fri, 13 Aug 2010 15:15:51 -0000

> GS2 requires that the GSS-API mechanism that GS2 wraps supports channel
> bindings on the GSS-API level.  This is essential to the GS2 design and
> there is no way to make anything work (securely) without it.  This is
> not channel binding to a TLS session or anything like that, this is just
> the basic support for doing channel binding to application-provided
> data.

Do you mean by this that the GSS exchange is bound to the application layer,
or that the application layer is bound to the GSS? Which is "underneath", in
other words?

The TLS case I mostly understand. Two unauthenticated endpoints with a TLS
session, and you use mutual authentication at a higher layer to attach
identity to the endpoints so that the TLS session becomes usable as an
authenticated session.

I'm not sure of the analogy to what you're describing yet.

> A GS2 mechanism does not have to support channel binding on the SASL
> level, but it is recommended.

Again, when you say "on the SASL level", what's being bound to what?

-- Scott



From Nicolas.Williams@oracle.com  Fri Aug 13 08:23:53 2010
Return-Path: <Nicolas.Williams@oracle.com>
X-Original-To: kitten@core3.amsl.com
Delivered-To: kitten@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 977EB3A6405 for <kitten@core3.amsl.com>; Fri, 13 Aug 2010 08:23:53 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -6.527
X-Spam-Level: 
X-Spam-Status: No, score=-6.527 tagged_above=-999 required=5 tests=[AWL=0.071,  BAYES_00=-2.599, RCVD_IN_DNSWL_MED=-4, UNPARSEABLE_RELAY=0.001]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id eYG0jXiDYj+3 for <kitten@core3.amsl.com>; Fri, 13 Aug 2010 08:23:52 -0700 (PDT)
Received: from rcsinet10.oracle.com (rcsinet10.oracle.com [148.87.113.121]) by core3.amsl.com (Postfix) with ESMTP id A99603A684A for <kitten@ietf.org>; Fri, 13 Aug 2010 08:23:52 -0700 (PDT)
Received: from rcsinet13.oracle.com (rcsinet13.oracle.com [148.87.113.125]) by rcsinet10.oracle.com (Switch-3.4.2/Switch-3.4.2) with ESMTP id o7DFORYL030340 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=OK); Fri, 13 Aug 2010 15:24:29 GMT
Received: from acsmt353.oracle.com (acsmt353.oracle.com [141.146.40.153]) by rcsinet13.oracle.com (Switch-3.4.2/Switch-3.4.1) with ESMTP id o7DD45nC011437; Fri, 13 Aug 2010 15:24:27 GMT
Received: from abhmt008.oracle.com by acsmt354.oracle.com with ESMTP id 494061421281713029; Fri, 13 Aug 2010 08:23:49 -0700
Received: from oracle.com (/129.153.128.104) by default (Oracle Beehive Gateway v4.0) with ESMTP ; Fri, 13 Aug 2010 08:23:48 -0700
Date: Fri, 13 Aug 2010 10:23:44 -0500
From: Nicolas Williams <Nicolas.Williams@oracle.com>
To: "Tschofenig, Hannes (NSN - FI/Espoo)" <hannes.tschofenig@nsn.com>
Subject: Re: Concerns about oauth sasl mechanism
Message-ID: <20100813152341.GE5217@oracle.com>
References: <tsllj8dyo5p.fsf@mit.edu> <3D3C75174CB95F42AD6BCC56E5555B4502E9BF74@FIESEXC015.nsn-intra.net>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <3D3C75174CB95F42AD6BCC56E5555B4502E9BF74@FIESEXC015.nsn-intra.net>
User-Agent: Mutt/1.5.20 (2010-03-02)
Cc: kitten@ietf.org, ext Sam Hartman <hartmans-ietf@mit.edu>
X-BeenThere: kitten@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Common Authentication Technologies - Next Generation <kitten.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/kitten>
List-Post: <mailto:kitten@ietf.org>
List-Help: <mailto:kitten-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=subscribe>
X-List-Received-Date: Fri, 13 Aug 2010 15:23:53 -0000

On Fri, Aug 13, 2010 at 02:59:02PM +0300, Tschofenig, Hannes (NSN - FI/Espoo) wrote:
> I hear complaints about this and that but then nobody cares about
> posting anything to the Oauth mailing list nor to show up at the Oauth
> meeting either. Almost nobody at the OAuth face-to-face meeting has read
> the specs. 

We can't help you with that last part!  ;)

> Having said that I find it a bit strange to require channel bindings.

Is that just because you don't know what that means?

> The funny thing is that I have heard that comment already at one of the
> Oauth meetings. Then, I asked what this actually means. The answer was
> *I don't know*. So, what should I think about these comments?

Well, for one that those OAuth participants haven't "read the specs",
but evidently you knew that (per the above).  If OAuth participants
don't "read the specs" then I am not at all surprised that they don't
know what channel binding means.

> I am happy to have a discussion about channel bindings in the context of
> Oauth (and of Oauth - SASL in particular) and even about how to re-write
> RFC 5056 in a way that it can be understood by humans. Interestingly
> enough I tried to figure out what Channel Bindings would provide us with
> SASL OpenID (together with my co-authors) and we weren't convinced about
> the value!

I too am happy to have a discussion about channel binding.  And I'll be
happy to entertain a re-write of RFC5056 if you like.  Maybe you could
start with some specific complaints?  But what am I to make of comments
that on the one hand indicate you don't understand X and on the other
state that you believe X is not needed?

Yes, you can avoid channel binding in some cases.

Suppose you are using TLS, with server certs, and you believe that you
have a good enough, trustworthy PKI for the server certs, then you don't
need channel binding in the RFC5056 sense.  In that case you'll happen
to be doing a form of channel binding by using the same (or similar)
service/host ID in TLS and OAuth.  Some of us consider this to be weak
security, and too restrictive; it might be good enough for some users.

Nico
-- 

From lear@cisco.com  Fri Aug 13 08:25:58 2010
Return-Path: <lear@cisco.com>
X-Original-To: kitten@core3.amsl.com
Delivered-To: kitten@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id D82353A68E3 for <kitten@core3.amsl.com>; Fri, 13 Aug 2010 08:25:58 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -110.481
X-Spam-Level: 
X-Spam-Status: No, score=-110.481 tagged_above=-999 required=5 tests=[AWL=0.118, BAYES_00=-2.599, RCVD_IN_DNSWL_HI=-8, USER_IN_WHITELIST=-100]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id GFhUudBFQiJc for <kitten@core3.amsl.com>; Fri, 13 Aug 2010 08:25:58 -0700 (PDT)
Received: from rtp-iport-2.cisco.com (rtp-iport-2.cisco.com [64.102.122.149]) by core3.amsl.com (Postfix) with ESMTP id 0D4DA3A6405 for <kitten@ietf.org>; Fri, 13 Aug 2010 08:25:57 -0700 (PDT)
Authentication-Results: rtp-iport-2.cisco.com; dkim=neutral (message not signed) header.i=none
X-IronPort-Anti-Spam-Filtered: true
X-IronPort-Anti-Spam-Result: Av0EAKsAZUxAZnwM/2dsb2JhbACDFZ0kcaIdiVKReIEmgyFzBIlfijY
X-IronPort-AV: E=Sophos;i="4.55,363,1278288000"; d="scan'208";a="147413083"
Received: from rtp-core-1.cisco.com ([64.102.124.12]) by rtp-iport-2.cisco.com with ESMTP; 13 Aug 2010 15:26:34 +0000
Received: from dhcp-10-61-96-21.cisco.com (dhcp-10-61-96-21.cisco.com [10.61.96.21]) by rtp-core-1.cisco.com (8.13.8/8.14.3) with ESMTP id o7DFQXUa026574; Fri, 13 Aug 2010 15:26:34 GMT
Message-ID: <4C656430.3080608@cisco.com>
Date: Fri, 13 Aug 2010 17:26:40 +0200
From: Eliot Lear <lear@cisco.com>
User-Agent: Mozilla/5.0 (Macintosh; U; Intel Mac OS X 10.6; en-US; rv:1.9.2.8) Gecko/20100802 Thunderbird/3.1.2
MIME-Version: 1.0
To: Scott Cantor <cantor.2@osu.edu>
Subject: Re: Speaking in support of Scott's SAML proposal
References: <tslvd7fs9ra.fsf@mit.edu> <043701cb3a80$866286e0$932794a0$@osu.edu>	<87hbiysxbv.fsf@mocca.josefsson.org> <044801cb3afa$8158b320$840a1960$@osu.edu>
In-Reply-To: <044801cb3afa$8158b320$840a1960$@osu.edu>
X-Enigmail-Version: 1.1.1
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 7bit
Cc: kitten@ietf.org, 'Simon Josefsson' <simon@josefsson.org>
X-BeenThere: kitten@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Common Authentication Technologies - Next Generation <kitten.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/kitten>
List-Post: <mailto:kitten@ietf.org>
List-Help: <mailto:kitten-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=subscribe>
X-List-Received-Date: Fri, 13 Aug 2010 15:25:58 -0000

On 8/13/10 5:16 PM, Scott Cantor wrote:
> Again, when you say "on the SASL level", what's being bound to what?

And here is what where we seem to have underlying confusion about
channel bindings in the context of all of this.  Before we go ripping
documents apart or adding additional text, someone needs to answer the
broader question.  Having stepped through this repeatedly with both
Simon, and others, we've always been open to document changes that
secure the protocols and whatever bindings need securing, so long as
someone can demonstrate a problem AND we do not require additional
coding on the IdP (has to integrate with what is used today).

Eliot

From Nicolas.Williams@oracle.com  Fri Aug 13 08:30:33 2010
Return-Path: <Nicolas.Williams@oracle.com>
X-Original-To: kitten@core3.amsl.com
Delivered-To: kitten@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 4459B3A684A for <kitten@core3.amsl.com>; Fri, 13 Aug 2010 08:30:33 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -6.528
X-Spam-Level: 
X-Spam-Status: No, score=-6.528 tagged_above=-999 required=5 tests=[AWL=0.070,  BAYES_00=-2.599, RCVD_IN_DNSWL_MED=-4, UNPARSEABLE_RELAY=0.001]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id koLKstHK-P+F for <kitten@core3.amsl.com>; Fri, 13 Aug 2010 08:30:22 -0700 (PDT)
Received: from rcsinet10.oracle.com (rcsinet10.oracle.com [148.87.113.121]) by core3.amsl.com (Postfix) with ESMTP id E17973A6A03 for <kitten@ietf.org>; Fri, 13 Aug 2010 08:30:20 -0700 (PDT)
Received: from rcsinet15.oracle.com (rcsinet15.oracle.com [148.87.113.117]) by rcsinet10.oracle.com (Switch-3.4.2/Switch-3.4.2) with ESMTP id o7DFUsAj015250 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=OK); Fri, 13 Aug 2010 15:30:55 GMT
Received: from acsmt355.oracle.com (acsmt355.oracle.com [141.146.40.155]) by rcsinet15.oracle.com (Switch-3.4.2/Switch-3.4.1) with ESMTP id o7D9Tbxt015994; Fri, 13 Aug 2010 15:30:54 GMT
Received: from abhmt008.oracle.com by acsmt353.oracle.com with ESMTP id 494086761281713399; Fri, 13 Aug 2010 08:29:59 -0700
Received: from oracle.com (/129.153.128.104) by default (Oracle Beehive Gateway v4.0) with ESMTP ; Fri, 13 Aug 2010 08:29:59 -0700
Date: Fri, 13 Aug 2010 10:29:54 -0500
From: Nicolas Williams <Nicolas.Williams@oracle.com>
To: "Tschofenig, Hannes (NSN - FI/Espoo)" <hannes.tschofenig@nsn.com>
Subject: Re: Concerns about oauth sasl mechanism
Message-ID: <20100813152954.GF5217@oracle.com>
References: <tsllj8dyo5p.fsf@mit.edu> <4C6460D6.7020800@stpeter.im> <tsl1va3v7z5.fsf@mit.edu> <4C64698A.9030405@isode.com> <3D3C75174CB95F42AD6BCC56E5555B4502E9BF85@FIESEXC015.nsn-intra.net>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <3D3C75174CB95F42AD6BCC56E5555B4502E9BF85@FIESEXC015.nsn-intra.net>
User-Agent: Mutt/1.5.20 (2010-03-02)
Cc: kitten@ietf.org, Sam Hartman <hartmans-ietf@mit.edu>
X-BeenThere: kitten@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Common Authentication Technologies - Next Generation <kitten.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/kitten>
List-Post: <mailto:kitten@ietf.org>
List-Help: <mailto:kitten-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=subscribe>
X-List-Received-Date: Fri, 13 Aug 2010 15:30:33 -0000

On Fri, Aug 13, 2010 at 03:06:14PM +0300, Tschofenig, Hannes (NSN - FI/Espoo) wrote:
> Are we standardizing research results or standardizing what industry
> does? Please have a look at http://tools.ietf.org/search/rfc5218
> 
> I am happy to discuss the security properties of various aspects but
> let's pick it one by one and also consider deployability. 
> 
> * Mutual authentication: We are talking about a three party protocol
> here and so what does mutual authentication actually mean. 

"Mutual authentication" is a bit of a term of art.  It can mean either
of two things, as well as both: a) that the two peers "prove" to each
other that they possess the same session keys before proceeding to
exchange sensitive information, b) that the two peers have names that
are authenticated to each other.  (a) comes from Kerberos.  (b) is the
meaning most people would expect.

In general we want both.

> * Channel Binding: Again, a different setup than described in RFC 5056.
> Would be nice to know if someone has deployment experience with it. 

I'm not sure what you're talking about here.

> * Authentication: Like with many identity management systems the
> authentication part is actually outside the scope of these protocols.
> This is true for SAML, OpenID, etc. 

In the sense that these protocols shouldn't require invasive changes to
deployed authentication infrastructure?  Yes.

Nico
-- 

From cantor.2@osu.edu  Fri Aug 13 08:35:57 2010
Return-Path: <cantor.2@osu.edu>
X-Original-To: kitten@core3.amsl.com
Delivered-To: kitten@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 9943A3A6405 for <kitten@core3.amsl.com>; Fri, 13 Aug 2010 08:35:57 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -2.558
X-Spam-Level: 
X-Spam-Status: No, score=-2.558 tagged_above=-999 required=5 tests=[AWL=0.041,  BAYES_00=-2.599]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id Eteu9XHElHs9 for <kitten@core3.amsl.com>; Fri, 13 Aug 2010 08:35:56 -0700 (PDT)
Received: from defang19.it.ohio-state.edu (defang19.it.ohio-state.edu [128.146.216.133]) by core3.amsl.com (Postfix) with ESMTP id 737463A69F7 for <kitten@ietf.org>; Fri, 13 Aug 2010 08:35:56 -0700 (PDT)
Received: from defang6.it.ohio-state.edu (defang6.it.ohio-state.edu [128.146.216.86]) by defang19.it.ohio-state.edu (8.13.7/8.13.1) with ESMTP id o7DFaWVs017625; Fri, 13 Aug 2010 11:36:32 -0400
Received: from SNOWDOG ([128.146.178.34]) by defang6.it.ohio-state.edu (8.13.7/8.13.1) with ESMTP id o7DFaURw014203; Fri, 13 Aug 2010 11:36:30 -0400
From: "Scott Cantor" <cantor.2@osu.edu>
To: "'Simon Josefsson'" <simon@josefsson.org>
References: <tslvd7fs9ra.fsf@mit.edu> <043701cb3a80$866286e0$932794a0$@osu.edu> <87aaoqswnl.fsf@mocca.josefsson.org>
In-Reply-To: <87aaoqswnl.fsf@mocca.josefsson.org>
Subject: RE: Speaking in support of Scott's SAML proposal
Date: Fri, 13 Aug 2010 11:36:33 -0400
Organization: The Ohio State University
Message-ID: <044901cb3afd$5001c020$f0054060$@osu.edu>
MIME-Version: 1.0
Content-Type: text/plain; charset="US-ASCII"
Content-Transfer-Encoding: 7bit
X-Mailer: Microsoft Outlook 14.0
Thread-index: AQHCNyGTwgBvXMFDRVTjG5E2EdVQhAFlrREYAX4fo8aS2mhfoA==
Content-language: en-us
X-CanIt-Geo: ip=128.146.216.86; country=US; region=OH; city=Columbus; latitude=39.9968; longitude=-82.9882; metrocode=535; areacode=614; http://maps.google.com/maps?q=39.9968,-82.9882&z=6
X-CanItPRO-Stream: outbound
X-Scanned-By: CanIt (www . roaringpenguin . com) on 128.146.216.133
Cc: kitten@ietf.org
X-BeenThere: kitten@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Common Authentication Technologies - Next Generation <kitten.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/kitten>
List-Post: <mailto:kitten@ietf.org>
List-Help: <mailto:kitten-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=subscribe>
X-List-Received-Date: Fri, 13 Aug 2010 15:35:57 -0000

> Below are the changes I'd like to see in draft-cantor-ietf-sasl-saml-ec
> to provide initial support for GS2.  If you'd like to co-author the
> document, I'd be happy to edit the XML file to include these changes,
> and any more that will come.

I'm fine with that if you have some time. I haven't made any additional
tweaks at this point, though I had a couple planned after some questions
from Eliot, but I can do that after you take a pass.

> There is more needed than what's shown below, but that can come later.
> The most important outstanding issues I see are:
> 
> * It appears as if we need to define a new GSS-API name type for SAML
>   identifiers.  Your document should use the same SAML GSS-API name type
>   as Klaas/Eliot's SAML mechanism, so possibly this should be done in a
>   separate document that you both can reference?  Or the text could be
>   duplicated, of course, but that seems sub-optimal.  Alternatively,
>   there could be one name type for SAML EC and one for SAML non-EC, but
>   I don't see any reason for that right now, and I see some arguments
>   against.

I think it wants to be a separate document. I seem to recall that there was
some discussion about that in the Moonshot context as well. I haven't had
time yet to dig into GSS-API and understand what the requirements are. I
obviously know SAML naming quite well, but I haven't been prepared to weigh
in until I understand the needs better.

In particular, it's certainly possible, but ultimately rather facile, to
identify the subject of a SAML assertion based on the NameID. But that
NameID isn't always present, and it is exceedingly common for it to be
opaque, or transient, with the real identifier(s) being passed as SAML
Attributes.

> * It should be possible to get (SASL level) channel binding support with
>   SAML, and this should be added.  What is needed here is to transport
>   the channel binding data (or a hash thereof) integrity protected
>   through the SAML assertions.  Is that possible?  My SAML-fu is weak
>   here.

I've been chatting with Nico and Sam in email since last night and getting
some kind of handle on it. It's possible to extend the SAML protocol to
carry CB data that the IdP could process into the assertion. It's also
possible, and may be more elegant, to do something more like the Kerberos
case, in which a key is shared, but the client just encrypts the CB data or
hash itself in the final leg to the RP.

I'm unclear which is better, and I'm unclear how the RP performs a similar
operation, or if it needs to.

> * Support for GSS-API per-message tokens and GSS_Pseudo_random, similar
>   to what is done for SCRAM.  What is needed here is to generate a
>   symmetric key securely out of the SAML negotiation, how can that be
>   done securely with SAML?  Is there already specifications on this?
>   This seems like a generally useful thing to have, so if there is
>   nothing like this today, perhaps we could develop it.

I need to understand what's being generated for whom before I can answer.
The IdP and the RP generally share knowledge of each other's long term keys.
The client rarely has a key, and doing things based on the client's
credentials becomes difficult to standardize because the client's
credentials can be many different things.

If the client has to be involved, my usual thought is to get the client to
supply a key, or input to derive a key, as an extension to the dialog with
the IdP, separate from its actual authentication credentials.

>    SAML EC security contexts always have the mutual_state flag
>    (GSS_C_MUTUAL_FLAG) set to TRUE.  SAML EC does not support credential
>    delegation, therefore SAML EC security contexts alway have the
>    deleg_state flag (GSS_C_DELEG_FLAG) set to FALSE.

We may want to figure out what you mean by delegation. SAML, at least,
supports delegation in a lot of different ways.

> X.1. GSS-API Principal Name Types for SAML EC

As you noted, we definitely need to think about this.

In particular, SAML RPs are always identified by SAML entityIDs (URIs).

-- Scott



From cantor.2@osu.edu  Fri Aug 13 08:58:34 2010
Return-Path: <cantor.2@osu.edu>
X-Original-To: kitten@core3.amsl.com
Delivered-To: kitten@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 5BABB3A6900 for <kitten@core3.amsl.com>; Fri, 13 Aug 2010 08:58:34 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -2.564
X-Spam-Level: 
X-Spam-Status: No, score=-2.564 tagged_above=-999 required=5 tests=[AWL=0.035,  BAYES_00=-2.599]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id KxXiUabPj22V for <kitten@core3.amsl.com>; Fri, 13 Aug 2010 08:58:33 -0700 (PDT)
Received: from defang1.it.ohio-state.edu (defang1.it.ohio-state.edu [128.146.216.81]) by core3.amsl.com (Postfix) with ESMTP id 6AA313A6803 for <kitten@ietf.org>; Fri, 13 Aug 2010 08:58:33 -0700 (PDT)
Received: from defang6.it.ohio-state.edu (defang6.it.ohio-state.edu [128.146.216.86]) by defang1.it.ohio-state.edu (8.13.7/8.13.1) with ESMTP id o7DFx8Pj007603; Fri, 13 Aug 2010 11:59:09 -0400
Received: from SNOWDOG ([128.146.178.34]) by defang6.it.ohio-state.edu (8.13.7/8.13.1) with ESMTP id o7DFx7j3000365; Fri, 13 Aug 2010 11:59:08 -0400
From: "Scott Cantor" <cantor.2@osu.edu>
To: "'Eliot Lear'" <lear@cisco.com>
References: <tslvd7fs9ra.fsf@mit.edu> <043701cb3a80$866286e0$932794a0$@osu.edu> <87hbiysxbv.fsf@mocca.josefsson.org> <044801cb3afa$8158b320$840a1960$@osu.edu> <4C656430.3080608@cisco.com>
In-Reply-To: <4C656430.3080608@cisco.com>
Subject: RE: Speaking in support of Scott's SAML proposal
Date: Fri, 13 Aug 2010 11:59:10 -0400
Organization: The Ohio State University
Message-ID: <044d01cb3b00$78972040$69c560c0$@osu.edu>
MIME-Version: 1.0
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable
X-Mailer: Microsoft Outlook 14.0
Thread-index: AQHCNyGTwgBvXMFDRVTjG5E2EdVQhAFlrREYAg6JzPQCKAlyzwJmwUyZkrF2lwA=
Content-language: en-us
X-CanIt-Geo: ip=128.146.216.86; country=US; region=OH; city=Columbus; latitude=39.9968; longitude=-82.9882; metrocode=535; areacode=614; http://maps.google.com/maps?q=39.9968,-82.9882&z=6
X-CanItPRO-Stream: outbound
X-Scanned-By: CanIt (www . roaringpenguin . com) on 128.146.216.81
Cc: kitten@ietf.org, 'Simon Josefsson' <simon@josefsson.org>
X-BeenThere: kitten@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Common Authentication Technologies - Next Generation <kitten.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/kitten>
List-Post: <mailto:kitten@ietf.org>
List-Help: <mailto:kitten-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=subscribe>
X-List-Received-Date: Fri, 13 Aug 2010 15:58:34 -0000

> On 8/13/10 5:16 PM, Scott Cantor wrote:
> > Again, when you say "on the SASL level", what's being bound to what?
>=20
> And here is what where we seem to have underlying confusion about
> channel bindings in the context of all of this.  Before we go ripping
> documents apart or adding additional text, someone needs to answer the
> broader question.

I agree. For my own purposes, I would like to be able to use the =
client's trust in the IdP to establish trust in the RP given that I =
don't believe commercial TLS is a useful trust model. As a bonus, it =
would be useful to give the RP the ability to connect the TLS endpoint =
it's communicating with to the SAML assertion it's authenticated the =
client with.

If possible, I'd like to do this without requiring the use of XML =
Signature on the client (i.e. a formal signature over the SAML =
response), although I'm not totally averse to it.

> Having stepped through this repeatedly with both
> Simon, and others, we've always been open to document changes that
> secure the protocols and whatever bindings need securing, so long as
> someone can demonstrate a problem AND we do not require additional
> coding on the IdP (has to integrate with what is used today).

Let me again note my difference of opinion on the last point. My IdP can =
change, and it will have to, I don=E2=80=99t see any way around that. I =
was going to point out in my last note that some of the various options =
for injecting CB into the SAML flow are perhaps better or worse for a =
scenario in which a browser is pulled in, but none of them can avoid the =
IdP changing (or even more formally the profiles involved being =
extended, really).

(If there's some magic way to take the weak security guarantees we have =
now with Web SSO and add CB, I'm all ears.)

I think it's fine, by the way, for you to have that constraint, but I =
don't think it's fine if I'm not able to pursue a parallel approach =
without the constraint. There still seems to be this reluctance to have =
two mechanisms on the table, but I don't see how they can be reconciled =
when they have such different goals.

In the SAML community, we don't expect every profile to be identical =
just because it uses SAML. We also think it's better if multiple =
profiles share common foundations, because they usually can share code.
=20
-- Scott



From Nicolas.Williams@oracle.com  Fri Aug 13 09:02:05 2010
Return-Path: <Nicolas.Williams@oracle.com>
X-Original-To: kitten@core3.amsl.com
Delivered-To: kitten@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 51A163A6407 for <kitten@core3.amsl.com>; Fri, 13 Aug 2010 09:02:05 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -6.529
X-Spam-Level: 
X-Spam-Status: No, score=-6.529 tagged_above=-999 required=5 tests=[AWL=0.069,  BAYES_00=-2.599, RCVD_IN_DNSWL_MED=-4, UNPARSEABLE_RELAY=0.001]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id ZB0qGCqE+Ylt for <kitten@core3.amsl.com>; Fri, 13 Aug 2010 09:02:03 -0700 (PDT)
Received: from rcsinet10.oracle.com (rcsinet10.oracle.com [148.87.113.121]) by core3.amsl.com (Postfix) with ESMTP id CC94F3A68F6 for <kitten@ietf.org>; Fri, 13 Aug 2010 09:02:02 -0700 (PDT)
Received: from rcsinet15.oracle.com (rcsinet15.oracle.com [148.87.113.117]) by rcsinet10.oracle.com (Switch-3.4.2/Switch-3.4.2) with ESMTP id o7DG2bNu004011 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=OK); Fri, 13 Aug 2010 16:02:38 GMT
Received: from acsmt355.oracle.com (acsmt355.oracle.com [141.146.40.155]) by rcsinet15.oracle.com (Switch-3.4.2/Switch-3.4.1) with ESMTP id o7DG2bJx007127; Fri, 13 Aug 2010 16:02:37 GMT
Received: from abhmt012.oracle.com by acsmt353.oracle.com with ESMTP id 494204881281715354; Fri, 13 Aug 2010 09:02:34 -0700
Received: from oracle.com (/129.153.128.104) by default (Oracle Beehive Gateway v4.0) with ESMTP ; Fri, 13 Aug 2010 09:02:34 -0700
Date: Fri, 13 Aug 2010 11:02:29 -0500
From: Nicolas Williams <Nicolas.Williams@oracle.com>
To: Scott Cantor <cantor.2@osu.edu>
Subject: Re: Speaking in support of Scott's SAML proposal
Message-ID: <20100813160229.GJ5213@oracle.com>
References: <tslvd7fs9ra.fsf@mit.edu> <043701cb3a80$866286e0$932794a0$@osu.edu> <87hbiysxbv.fsf@mocca.josefsson.org> <044801cb3afa$8158b320$840a1960$@osu.edu>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <044801cb3afa$8158b320$840a1960$@osu.edu>
User-Agent: Mutt/1.5.20 (2010-03-02)
Cc: kitten@ietf.org, 'Simon Josefsson' <simon@josefsson.org>
X-BeenThere: kitten@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Common Authentication Technologies - Next Generation <kitten.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/kitten>
List-Post: <mailto:kitten@ietf.org>
List-Help: <mailto:kitten-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=subscribe>
X-List-Received-Date: Fri, 13 Aug 2010 16:02:06 -0000

On Fri, Aug 13, 2010 at 11:16:29AM -0400, Scott Cantor wrote:
> > GS2 requires that the GSS-API mechanism that GS2 wraps supports channel
> > bindings on the GSS-API level.  This is essential to the GS2 design and
> > there is no way to make anything work (securely) without it.  This is
> > not channel binding to a TLS session or anything like that, this is just
> > the basic support for doing channel binding to application-provided
> > data.
> 
> Do you mean by this that the GSS exchange is bound to the application layer,
> or that the application layer is bound to the GSS? Which is "underneath", in
> other words?

If you read RFC5801 you'll see that GS2 has some plaintext bits to send
"outside" the GSS-API.  For example: the SASL authzid.  There are others
as well.

Now, when we designed GS2 we knew that we need to exchange such
information as authzid and what not, and that such information had to be
"authenticated" (i.e., not modifiable by active attackers when GS2
succeeds).

In the initial iteration of GS2 we used wrap tokens to carry this
information.  But this made GS2 too complicated.  So...  we came up with
a clever alternative: all of this data on the initiator side could be
used as the channel binding input of the GSS-API mechanism.

See RFC5801 for the remaining details.

> The TLS case I mostly understand. Two unauthenticated endpoints with a TLS
> session, and you use mutual authentication at a higher layer to attach
> identity to the endpoints so that the TLS session becomes usable as an
> authenticated session.

Right.  You can use GS2 without channel bindings to TLS.  But GS2
internally uses CB for its own purposes.

> > A GS2 mechanism does not have to support channel binding on the SASL
> > level, but it is recommended.
> 
> Again, when you say "on the SASL level", what's being bound to what?

At the SASL level you'd be binding to a TLS (or IPsec, or whatever) channel.

GS2 is a layer as well, a layer above a GSS-API mechanism.   And GS2
uses the CB input of the GSS-API mechanism in all circumstances.

Thus, if you're designing a GSS-API mechanism and wish it to be usable
within SASL/GS2, then your GSS must provide support for channel binding.

Nico
-- 

From Nicolas.Williams@oracle.com  Fri Aug 13 09:09:08 2010
Return-Path: <Nicolas.Williams@oracle.com>
X-Original-To: kitten@core3.amsl.com
Delivered-To: kitten@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id D02933A684A for <kitten@core3.amsl.com>; Fri, 13 Aug 2010 09:09:08 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -6.53
X-Spam-Level: 
X-Spam-Status: No, score=-6.53 tagged_above=-999 required=5 tests=[AWL=0.068,  BAYES_00=-2.599, RCVD_IN_DNSWL_MED=-4, UNPARSEABLE_RELAY=0.001]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id OsBR+84NQ71l for <kitten@core3.amsl.com>; Fri, 13 Aug 2010 09:09:08 -0700 (PDT)
Received: from rcsinet10.oracle.com (rcsinet10.oracle.com [148.87.113.121]) by core3.amsl.com (Postfix) with ESMTP id CD56E3A6915 for <kitten@ietf.org>; Fri, 13 Aug 2010 09:09:07 -0700 (PDT)
Received: from rcsinet15.oracle.com (rcsinet15.oracle.com [148.87.113.117]) by rcsinet10.oracle.com (Switch-3.4.2/Switch-3.4.2) with ESMTP id o7DG9g56023491 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=OK); Fri, 13 Aug 2010 16:09:43 GMT
Received: from acsmt355.oracle.com (acsmt355.oracle.com [141.146.40.155]) by rcsinet15.oracle.com (Switch-3.4.2/Switch-3.4.1) with ESMTP id o7DG9fq9030386; Fri, 13 Aug 2010 16:09:42 GMT
Received: from abhmt003.oracle.com by acsmt355.oracle.com with ESMTP id 514830131281715678; Fri, 13 Aug 2010 09:07:58 -0700
Received: from oracle.com (/129.153.128.104) by default (Oracle Beehive Gateway v4.0) with ESMTP ; Fri, 13 Aug 2010 09:07:58 -0700
Date: Fri, 13 Aug 2010 11:07:53 -0500
From: Nicolas Williams <Nicolas.Williams@oracle.com>
To: Scott Cantor <cantor.2@osu.edu>
Subject: Re: Speaking in support of Scott's SAML proposal
Message-ID: <20100813160753.GK5213@oracle.com>
References: <tslvd7fs9ra.fsf@mit.edu> <043701cb3a80$866286e0$932794a0$@osu.edu> <87hbiysxbv.fsf@mocca.josefsson.org> <044801cb3afa$8158b320$840a1960$@osu.edu> <4C656430.3080608@cisco.com> <044d01cb3b00$78972040$69c560c0$@osu.edu>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <044d01cb3b00$78972040$69c560c0$@osu.edu>
User-Agent: Mutt/1.5.20 (2010-03-02)
Cc: kitten@ietf.org, 'Simon Josefsson' <simon@josefsson.org>
X-BeenThere: kitten@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Common Authentication Technologies - Next Generation <kitten.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/kitten>
List-Post: <mailto:kitten@ietf.org>
List-Help: <mailto:kitten-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=subscribe>
X-List-Received-Date: Fri, 13 Aug 2010 16:09:08 -0000

On Fri, Aug 13, 2010 at 11:59:10AM -0400, Scott Cantor wrote:
> > On 8/13/10 5:16 PM, Scott Cantor wrote:
> > > Again, when you say "on the SASL level", what's being bound to what?
> > 
> > And here is what where we seem to have underlying confusion about
> > channel bindings in the context of all of this.  Before we go ripping
> > documents apart or adding additional text, someone needs to answer the
> > broader question.
> 
> I agree. For my own purposes, I would like to be able to use the
> client's trust in the IdP to establish trust in the RP given that I
> don't believe commercial TLS is a useful trust model. As a bonus, it
> would be useful to give the RP the ability to connect the TLS endpoint
> it's communicating with to the SAML assertion it's authenticated the
> client with.
> 
> If possible, I'd like to do this without requiring the use of XML
> Signature on the client (i.e. a formal signature over the SAML
> response), although I'm not totally averse to it.

You can do it without XML dsig on the client.

Imagine if in Kerberos we handled channel binding by putting it in an
authorization data element in the Ticket.  The client/initiator would
ask the TGS for a Ticket and would tell the TGS what the CB data are.
The acceptor would extract the CB data (or hash thereof) from the
Ticket, and compare to the CB data passed by the application.

In your case you have the IdP playing a KDC-like role, so you could well
handle CB as I described above.

Nico
-- 

From Nicolas.Williams@oracle.com  Fri Aug 13 09:14:01 2010
Return-Path: <Nicolas.Williams@oracle.com>
X-Original-To: kitten@core3.amsl.com
Delivered-To: kitten@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id CF5763A696E for <kitten@core3.amsl.com>; Fri, 13 Aug 2010 09:14:01 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -6.531
X-Spam-Level: 
X-Spam-Status: No, score=-6.531 tagged_above=-999 required=5 tests=[AWL=0.067,  BAYES_00=-2.599, RCVD_IN_DNSWL_MED=-4, UNPARSEABLE_RELAY=0.001]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id pvAjrK9cnExL for <kitten@core3.amsl.com>; Fri, 13 Aug 2010 09:14:00 -0700 (PDT)
Received: from rcsinet10.oracle.com (rcsinet10.oracle.com [148.87.113.121]) by core3.amsl.com (Postfix) with ESMTP id 9A8713A68F1 for <kitten@ietf.org>; Fri, 13 Aug 2010 09:14:00 -0700 (PDT)
Received: from rcsinet15.oracle.com (rcsinet15.oracle.com [148.87.113.117]) by rcsinet10.oracle.com (Switch-3.4.2/Switch-3.4.2) with ESMTP id o7DGEZoq006346 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=OK); Fri, 13 Aug 2010 16:14:36 GMT
Received: from acsmt354.oracle.com (acsmt354.oracle.com [141.146.40.154]) by rcsinet15.oracle.com (Switch-3.4.2/Switch-3.4.1) with ESMTP id o7D1LZQc009288; Fri, 13 Aug 2010 16:14:35 GMT
Received: from abhmt016.oracle.com by acsmt355.oracle.com with ESMTP id 494249381281716040; Fri, 13 Aug 2010 09:14:00 -0700
Received: from oracle.com (/129.153.128.104) by default (Oracle Beehive Gateway v4.0) with ESMTP ; Fri, 13 Aug 2010 09:13:59 -0700
Date: Fri, 13 Aug 2010 11:13:54 -0500
From: Nicolas Williams <Nicolas.Williams@oracle.com>
To: Scott Cantor <cantor.2@osu.edu>
Subject: Re: Speaking in support of Scott's SAML proposal
Message-ID: <20100813161354.GM5213@oracle.com>
References: <tslvd7fs9ra.fsf@mit.edu> <043701cb3a80$866286e0$932794a0$@osu.edu> <87aaoqswnl.fsf@mocca.josefsson.org> <044901cb3afd$5001c020$f0054060$@osu.edu>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <044901cb3afd$5001c020$f0054060$@osu.edu>
User-Agent: Mutt/1.5.20 (2010-03-02)
Cc: kitten@ietf.org, 'Simon Josefsson' <simon@josefsson.org>
X-BeenThere: kitten@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Common Authentication Technologies - Next Generation <kitten.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/kitten>
List-Post: <mailto:kitten@ietf.org>
List-Help: <mailto:kitten-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=subscribe>
X-List-Received-Date: Fri, 13 Aug 2010 16:14:02 -0000

On Fri, Aug 13, 2010 at 11:36:33AM -0400, Scott Cantor wrote:
> > * It should be possible to get (SASL level) channel binding support with
> >   SAML, and this should be added.  What is needed here is to transport
> >   the channel binding data (or a hash thereof) integrity protected
> >   through the SAML assertions.  Is that possible?  My SAML-fu is weak
> >   here.
> 
> I've been chatting with Nico and Sam in email since last night and getting
> some kind of handle on it. It's possible to extend the SAML protocol to
> carry CB data that the IdP could process into the assertion. It's also
> possible, and may be more elegant, to do something more like the Kerberos
> case, in which a key is shared, but the client just encrypts the CB data or
> hash itself in the final leg to the RP.

Either way is fine.  Note that if you put the CB data into the "ticket"
(the SAML assertion produced by the IdP for the RP's consumption) then
when the CB data are "unique" then you'll end up being unable to cache
"tickets".

(In the Kerberos world we still believe in the utility of caching
tickets, though I don't really think it's that useful anymore.  Ticket
caching is the reason that the CB are in the Authenticator instead of
the Ticket.)

In your case, is there something like an Authenticator?

> > * Support for GSS-API per-message tokens and GSS_Pseudo_random, similar
> >   to what is done for SCRAM.  What is needed here is to generate a
> >   symmetric key securely out of the SAML negotiation, how can that be
> >   done securely with SAML?  Is there already specifications on this?
> >   This seems like a generally useful thing to have, so if there is
> >   nothing like this today, perhaps we could develop it.

SASL/GS2 does not require GSS-API per-message tokens nor
GSS_Pseudo_random().

But many GSS-API applications do, so it's useful to provide these.
Start with CB, then let's work on the rest.

> >    SAML EC security contexts always have the mutual_state flag
> >    (GSS_C_MUTUAL_FLAG) set to TRUE.  SAML EC does not support credential
> >    delegation, therefore SAML EC security contexts alway have the
> >    deleg_state flag (GSS_C_DELEG_FLAG) set to FALSE.
> 
> We may want to figure out what you mean by delegation. SAML, at least,
> supports delegation in a lot of different ways.

Oh, that's a whole 'nother topic.  Let's leave that for another day.

Nico
-- 

From cantor.2@osu.edu  Fri Aug 13 09:46:06 2010
Return-Path: <cantor.2@osu.edu>
X-Original-To: kitten@core3.amsl.com
Delivered-To: kitten@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 9C6643A68CE for <kitten@core3.amsl.com>; Fri, 13 Aug 2010 09:46:06 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -2.568
X-Spam-Level: 
X-Spam-Status: No, score=-2.568 tagged_above=-999 required=5 tests=[AWL=0.031,  BAYES_00=-2.599]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id m8RkrQRUO7Qq for <kitten@core3.amsl.com>; Fri, 13 Aug 2010 09:46:05 -0700 (PDT)
Received: from defang2.it.ohio-state.edu (defang2.it.ohio-state.edu [128.146.216.82]) by core3.amsl.com (Postfix) with ESMTP id A4C843A6802 for <kitten@ietf.org>; Fri, 13 Aug 2010 09:46:04 -0700 (PDT)
Received: from defang6.it.ohio-state.edu (defang6.it.ohio-state.edu [128.146.216.86]) by defang2.it.ohio-state.edu (8.13.7/8.13.1) with ESMTP id o7DGkdXl020253; Fri, 13 Aug 2010 12:46:39 -0400
Received: from SNOWDOG ([128.146.178.34]) by defang6.it.ohio-state.edu (8.13.7/8.13.1) with ESMTP id o7DGkbcJ013349; Fri, 13 Aug 2010 12:46:38 -0400
From: "Scott Cantor" <cantor.2@osu.edu>
To: "'Nicolas Williams'" <Nicolas.Williams@oracle.com>
References: <tslvd7fs9ra.fsf@mit.edu> <043701cb3a80$866286e0$932794a0$@osu.edu> <87aaoqswnl.fsf@mocca.josefsson.org> <044901cb3afd$5001c020$f0054060$@osu.edu> <20100813161354.GM5213@oracle.com>
In-Reply-To: <20100813161354.GM5213@oracle.com>
Subject: RE: Speaking in support of Scott's SAML proposal
Date: Fri, 13 Aug 2010 12:46:41 -0400
Organization: The Ohio State University
Message-ID: <045701cb3b07$1be36b90$53aa42b0$@osu.edu>
MIME-Version: 1.0
Content-Type: text/plain; charset="US-ASCII"
Content-Transfer-Encoding: 7bit
X-Mailer: Microsoft Outlook 14.0
Thread-index: AQHCNyGTwgBvXMFDRVTjG5E2EdVQhAFlrREYAX4fo8YAn22RxwI0ctwtksPgfeA=
Content-language: en-us
X-CanIt-Geo: ip=128.146.216.86; country=US; region=OH; city=Columbus; latitude=39.9968; longitude=-82.9882; metrocode=535; areacode=614; http://maps.google.com/maps?q=39.9968,-82.9882&z=6
X-CanItPRO-Stream: outbound
X-Scanned-By: CanIt (www . roaringpenguin . com) on 128.146.216.82
Cc: kitten@ietf.org, 'Simon Josefsson' <simon@josefsson.org>
X-BeenThere: kitten@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Common Authentication Technologies - Next Generation <kitten.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/kitten>
List-Post: <mailto:kitten@ietf.org>
List-Help: <mailto:kitten-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=subscribe>
X-List-Received-Date: Fri, 13 Aug 2010 16:46:06 -0000

> Either way is fine.  Note that if you put the CB data into the "ticket"
> (the SAML assertion produced by the IdP for the RP's consumption) then
> when the CB data are "unique" then you'll end up being unable to cache
> "tickets".

This is normally not a concern because the assertions are usually short
lived. So far, they're still bearer tokens in the scenarios here (they may
or may not stay that way, but let's assume for now), so they get minted per
login to a RP.

I'm coming around to the idea of having the IdP stick it in, but I think I'm
only missing the final piece of how this authenticates the server/acceptor
end of the channel, and not just the client end.

> In your case, is there something like an Authenticator?

SAML is abstracted from the security layer in which the assertion is
presented. The abstraction is called "subject confirmation". The IdP
includes a description in the assertion of how the client has to prove its
right to use the assertion.

Often, such as in Web SSO, this is a "null proof" consisting of "bearer". I
have it (and perhaps have a particular client address), so I can use it. It
is usually a short lived window in which I can present it but that's it.

It can be something other than bearer, such as "holder of key", in which I
prove key possession by signing something or using TLS. There is a proposed
method of "Kerberos" in which a ticket + Authenticator would accompany the
assertion in the protocol. Etc.

-- Scott



From Nicolas.Williams@oracle.com  Fri Aug 13 09:59:18 2010
Return-Path: <Nicolas.Williams@oracle.com>
X-Original-To: kitten@core3.amsl.com
Delivered-To: kitten@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 6432B3A682E for <kitten@core3.amsl.com>; Fri, 13 Aug 2010 09:59:18 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -6.532
X-Spam-Level: 
X-Spam-Status: No, score=-6.532 tagged_above=-999 required=5 tests=[AWL=0.066,  BAYES_00=-2.599, RCVD_IN_DNSWL_MED=-4, UNPARSEABLE_RELAY=0.001]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id XRRQ3HW2Vo6D for <kitten@core3.amsl.com>; Fri, 13 Aug 2010 09:59:17 -0700 (PDT)
Received: from rcsinet10.oracle.com (rcsinet10.oracle.com [148.87.113.121]) by core3.amsl.com (Postfix) with ESMTP id 8030B3A67EE for <kitten@ietf.org>; Fri, 13 Aug 2010 09:59:17 -0700 (PDT)
Received: from acsinet15.oracle.com (acsinet15.oracle.com [141.146.126.227]) by rcsinet10.oracle.com (Switch-3.4.2/Switch-3.4.2) with ESMTP id o7DGxnLx009042 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=OK); Fri, 13 Aug 2010 16:59:51 GMT
Received: from acsmt355.oracle.com (acsmt355.oracle.com [141.146.40.155]) by acsinet15.oracle.com (Switch-3.4.2/Switch-3.4.1) with ESMTP id o7DCrU8E027209; Fri, 13 Aug 2010 16:59:46 GMT
Received: from abhmt017.oracle.com by acsmt355.oracle.com with ESMTP id 514996361281718699; Fri, 13 Aug 2010 09:58:19 -0700
Received: from oracle.com (/129.153.128.104) by default (Oracle Beehive Gateway v4.0) with ESMTP ; Fri, 13 Aug 2010 09:58:17 -0700
Date: Fri, 13 Aug 2010 11:58:13 -0500
From: Nicolas Williams <Nicolas.Williams@oracle.com>
To: Scott Cantor <cantor.2@osu.edu>
Subject: Re: Speaking in support of Scott's SAML proposal
Message-ID: <20100813165812.GQ5213@oracle.com>
References: <tslvd7fs9ra.fsf@mit.edu> <043701cb3a80$866286e0$932794a0$@osu.edu> <87aaoqswnl.fsf@mocca.josefsson.org> <044901cb3afd$5001c020$f0054060$@osu.edu> <20100813161354.GM5213@oracle.com> <045701cb3b07$1be36b90$53aa42b0$@osu.edu>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <045701cb3b07$1be36b90$53aa42b0$@osu.edu>
User-Agent: Mutt/1.5.20 (2010-03-02)
Cc: kitten@ietf.org, 'Simon Josefsson' <simon@josefsson.org>
X-BeenThere: kitten@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Common Authentication Technologies - Next Generation <kitten.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/kitten>
List-Post: <mailto:kitten@ietf.org>
List-Help: <mailto:kitten-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=subscribe>
X-List-Received-Date: Fri, 13 Aug 2010 16:59:18 -0000

On Fri, Aug 13, 2010 at 12:46:41PM -0400, Scott Cantor wrote:
> > Either way is fine.  Note that if you put the CB data into the "ticket"
> > (the SAML assertion produced by the IdP for the RP's consumption) then
> > when the CB data are "unique" then you'll end up being unable to cache
> > "tickets".
> 
> This is normally not a concern because the assertions are usually short
> lived. So far, they're still bearer tokens in the scenarios here (they may
> or may not stay that way, but let's assume for now), so they get minted per
> login to a RP.
> 
> I'm coming around to the idea of having the IdP stick it in, but I think I'm
> only missing the final piece of how this authenticates the server/acceptor
> end of the channel, and not just the client end.

You need some way to indicate to the initiator that the acceptor checked
that the CB match.  If you don't have enough message legs in your
protocol to do this, then you might have a problem :(

One possibility might be to use the RP's server cert as the CB and have
the IdP do something to confirm that the cert seen by the client is a
known cert for the RP named by the client.  This would mean that you'd
only support 'tls-server-end-point' CB, but it would get you going.

Nico
-- 

From cantor.2@osu.edu  Fri Aug 13 10:14:40 2010
Return-Path: <cantor.2@osu.edu>
X-Original-To: kitten@core3.amsl.com
Delivered-To: kitten@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id DFEAE3A6805 for <kitten@core3.amsl.com>; Fri, 13 Aug 2010 10:14:40 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -2.572
X-Spam-Level: 
X-Spam-Status: No, score=-2.572 tagged_above=-999 required=5 tests=[AWL=0.027,  BAYES_00=-2.599]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id 3HdEI9peak3o for <kitten@core3.amsl.com>; Fri, 13 Aug 2010 10:14:38 -0700 (PDT)
Received: from defang1.it.ohio-state.edu (defang1.it.ohio-state.edu [128.146.216.81]) by core3.amsl.com (Postfix) with ESMTP id 7F8303A682A for <kitten@ietf.org>; Fri, 13 Aug 2010 10:14:38 -0700 (PDT)
Received: from defang6.it.ohio-state.edu (defang6.it.ohio-state.edu [128.146.216.86]) by defang1.it.ohio-state.edu (8.13.7/8.13.1) with ESMTP id o7DHFFj8007055; Fri, 13 Aug 2010 13:15:15 -0400
Received: from SNOWDOG ([128.146.178.34]) by defang6.it.ohio-state.edu (8.13.7/8.13.1) with ESMTP id o7DHFEnt009454; Fri, 13 Aug 2010 13:15:14 -0400
From: "Scott Cantor" <cantor.2@osu.edu>
To: "'Nicolas Williams'" <Nicolas.Williams@oracle.com>
References: <tslvd7fs9ra.fsf@mit.edu> <043701cb3a80$866286e0$932794a0$@osu.edu> <87aaoqswnl.fsf@mocca.josefsson.org> <044901cb3afd$5001c020$f0054060$@osu.edu> <20100813161354.GM5213@oracle.com> <045701cb3b07$1be36b90$53aa42b0$@osu.edu> <20100813165812.GQ5213@oracle.com>
In-Reply-To: <20100813165812.GQ5213@oracle.com>
Subject: RE: Speaking in support of Scott's SAML proposal
Date: Fri, 13 Aug 2010 13:15:14 -0400
Organization: The Ohio State University
Message-ID: <045c01cb3b0b$1a421ad0$4ec65070$@osu.edu>
MIME-Version: 1.0
Content-Type: text/plain; charset="US-ASCII"
Content-Transfer-Encoding: 7bit
X-Mailer: Microsoft Outlook 14.0
Thread-index: AQHCNyGTwgBvXMFDRVTjG5E2EdVQhAFlrREYAX4fo8YAn22RxwI0ctwtAdMLnJQBW7CwBZKqcQsA
Content-language: en-us
X-CanIt-Geo: ip=128.146.216.86; country=US; region=OH; city=Columbus; latitude=39.9968; longitude=-82.9882; metrocode=535; areacode=614; http://maps.google.com/maps?q=39.9968,-82.9882&z=6
X-CanItPRO-Stream: outbound
X-Scanned-By: CanIt (www . roaringpenguin . com) on 128.146.216.81
Cc: kitten@ietf.org, 'Simon Josefsson' <simon@josefsson.org>
X-BeenThere: kitten@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Common Authentication Technologies - Next Generation <kitten.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/kitten>
List-Post: <mailto:kitten@ietf.org>
List-Help: <mailto:kitten-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=subscribe>
X-List-Received-Date: Fri, 13 Aug 2010 17:14:41 -0000

> You need some way to indicate to the initiator that the acceptor checked
> that the CB match.  If you don't have enough message legs in your
> protocol to do this, then you might have a problem :(

The EC SASL mechanism as it stands now assumes that the server is signaling
the client what the outcome was. Is that what you're talking about, or
something more?

> One possibility might be to use the RP's server cert as the CB and have
> the IdP do something to confirm that the cert seen by the client is a
> known cert for the RP named by the client.  This would mean that you'd
> only support 'tls-server-end-point' CB, but it would get you going.

So, what I had in mind was something like this:

1. RP includes its CB data in the SAML AuthnRequest it gives to the client
(exact means TBD). The AuthnRequest is signed with the RP's key.

2. Client adds its CB data to its message to the IdP that includes the
AuthnRequest and its credentials.

3. IdP compares them (along with verifying the request signature, and all
the other bits it does) and if they match, issues the response with the
encrypted assertion containing the CB data (again exact means TBD)

4. RP gets response from client, decrypts the assertion, checks the CB data
in the assertion as it checks the rest of the material, and if they match,
sends back the OK, else an error.

Ok, attack away. If I'm right, this is a simple addition and there's
actually a pretty good way to carry this in SAML that fits very nicely. If
not, back to the drawing board.

-- Scott



From Nicolas.Williams@oracle.com  Fri Aug 13 10:29:19 2010
Return-Path: <Nicolas.Williams@oracle.com>
X-Original-To: kitten@core3.amsl.com
Delivered-To: kitten@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id A3CE13A67EE for <kitten@core3.amsl.com>; Fri, 13 Aug 2010 10:29:19 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -6.532
X-Spam-Level: 
X-Spam-Status: No, score=-6.532 tagged_above=-999 required=5 tests=[AWL=0.066,  BAYES_00=-2.599, RCVD_IN_DNSWL_MED=-4, UNPARSEABLE_RELAY=0.001]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id WhtPIPueUaL8 for <kitten@core3.amsl.com>; Fri, 13 Aug 2010 10:29:18 -0700 (PDT)
Received: from rcsinet10.oracle.com (rcsinet10.oracle.com [148.87.113.121]) by core3.amsl.com (Postfix) with ESMTP id 85F893A698B for <kitten@ietf.org>; Fri, 13 Aug 2010 10:29:18 -0700 (PDT)
Received: from acsinet15.oracle.com (acsinet15.oracle.com [141.146.126.227]) by rcsinet10.oracle.com (Switch-3.4.2/Switch-3.4.2) with ESMTP id o7DHTpqU004681 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=OK); Fri, 13 Aug 2010 17:29:53 GMT
Received: from acsmt354.oracle.com (acsmt354.oracle.com [141.146.40.154]) by acsinet15.oracle.com (Switch-3.4.2/Switch-3.4.1) with ESMTP id o7DBIL78021934; Fri, 13 Aug 2010 17:29:50 GMT
Received: from abhmt012.oracle.com by acsmt355.oracle.com with ESMTP id 494476571281720543; Fri, 13 Aug 2010 10:29:03 -0700
Received: from oracle.com (/129.153.128.104) by default (Oracle Beehive Gateway v4.0) with ESMTP ; Fri, 13 Aug 2010 10:29:02 -0700
Date: Fri, 13 Aug 2010 12:28:58 -0500
From: Nicolas Williams <Nicolas.Williams@oracle.com>
To: Scott Cantor <cantor.2@osu.edu>
Subject: Re: Speaking in support of Scott's SAML proposal
Message-ID: <20100813172857.GS5213@oracle.com>
References: <tslvd7fs9ra.fsf@mit.edu> <043701cb3a80$866286e0$932794a0$@osu.edu> <87aaoqswnl.fsf@mocca.josefsson.org> <044901cb3afd$5001c020$f0054060$@osu.edu> <20100813161354.GM5213@oracle.com> <045701cb3b07$1be36b90$53aa42b0$@osu.edu> <20100813165812.GQ5213@oracle.com> <045c01cb3b0b$1a421ad0$4ec65070$@osu.edu>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <045c01cb3b0b$1a421ad0$4ec65070$@osu.edu>
User-Agent: Mutt/1.5.20 (2010-03-02)
Cc: kitten@ietf.org, 'Simon Josefsson' <simon@josefsson.org>
X-BeenThere: kitten@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Common Authentication Technologies - Next Generation <kitten.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/kitten>
List-Post: <mailto:kitten@ietf.org>
List-Help: <mailto:kitten-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=subscribe>
X-List-Received-Date: Fri, 13 Aug 2010 17:29:19 -0000

On Fri, Aug 13, 2010 at 01:15:14PM -0400, Scott Cantor wrote:
> > You need some way to indicate to the initiator that the acceptor checked
> > that the CB match.  If you don't have enough message legs in your
> > protocol to do this, then you might have a problem :(
> 
> The EC SASL mechanism as it stands now assumes that the server is signaling
> the client what the outcome was. Is that what you're talking about, or
> something more?

The outcome of authentication signal itself needs to be authenticated,
implicitly or explicitly.  Specifically, it must be impossible for an
MITM to make a client think authentication succeeded when the real
server thought it failed.

> > One possibility might be to use the RP's server cert as the CB and have
> > the IdP do something to confirm that the cert seen by the client is a
> > known cert for the RP named by the client.  This would mean that you'd
> > only support 'tls-server-end-point' CB, but it would get you going.
> 
> So, what I had in mind was something like this:
> 
> 1. RP includes its CB data in the SAML AuthnRequest it gives to the client
> (exact means TBD). The AuthnRequest is signed with the RP's key.
> 
> 2. Client adds its CB data to its message to the IdP that includes the
> AuthnRequest and its credentials.
> 
> 3. IdP compares them (along with verifying the request signature, and all
> the other bits it does) and if they match, issues the response with the
> encrypted assertion containing the CB data (again exact means TBD)

Does the client get an indication from the IdP as to whether the CB
check passed?  Presumably the IdP won't issue the assertion if the CB
fail to match!  Assuming that, then, yes, your approach works.

> 4. RP gets response from client, decrypts the assertion, checks the CB data
> in the assertion as it checks the rest of the material, and if they match,
> sends back the OK, else an error.

The RP doesn't have to check the CB data at this point if it trusts the
IdP to have done it.  And if the client doesn't independently pass CB
data to the RP then there's nothing the RP can do but trust that the IdP
did it.

> Ok, attack away. If I'm right, this is a simple addition and there's
> actually a pretty good way to carry this in SAML that fits very nicely. If
> not, back to the drawing board.

I believe your strawman stands.

Nico
-- 

From cantor.2@osu.edu  Fri Aug 13 10:36:12 2010
Return-Path: <cantor.2@osu.edu>
X-Original-To: kitten@core3.amsl.com
Delivered-To: kitten@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id D4D683A67EE for <kitten@core3.amsl.com>; Fri, 13 Aug 2010 10:36:12 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -2.575
X-Spam-Level: 
X-Spam-Status: No, score=-2.575 tagged_above=-999 required=5 tests=[AWL=0.024,  BAYES_00=-2.599]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id 4-5wRkzFk33W for <kitten@core3.amsl.com>; Fri, 13 Aug 2010 10:36:11 -0700 (PDT)
Received: from defang2.it.ohio-state.edu (defang2.it.ohio-state.edu [128.146.216.82]) by core3.amsl.com (Postfix) with ESMTP id 9CF923A67E7 for <kitten@ietf.org>; Fri, 13 Aug 2010 10:36:11 -0700 (PDT)
Received: from defang6.it.ohio-state.edu (defang6.it.ohio-state.edu [128.146.216.86]) by defang2.it.ohio-state.edu (8.13.7/8.13.1) with ESMTP id o7DHallN008840; Fri, 13 Aug 2010 13:36:48 -0400
Received: from SNOWDOG ([128.146.178.34]) by defang6.it.ohio-state.edu (8.13.7/8.13.1) with ESMTP id o7DHakpv030096; Fri, 13 Aug 2010 13:36:47 -0400
From: "Scott Cantor" <cantor.2@osu.edu>
To: "'Tschofenig, Hannes \(NSN - FI/Espoo\)'" <hannes.tschofenig@nsn.com>
References: <tsllj8dyo5p.fsf@mit.edu> <4C6460D6.7020800@stpeter.im> <tsl1va3v7z5.fsf@mit.edu> <4C64698A.9030405@isode.com> <tslocd7tsjr.fsf@mit.edu> <3D3C75174CB95F42AD6BCC56E5555B4502E9BF97@FIESEXC015.nsn-intra.net>
In-Reply-To: <3D3C75174CB95F42AD6BCC56E5555B4502E9BF97@FIESEXC015.nsn-intra.net>
Subject: RE: Concerns about oauth sasl mechanism
Date: Fri, 13 Aug 2010 13:36:49 -0400
Organization: The Ohio State University
Message-ID: <046101cb3b0e$1cca1de0$565e59a0$@osu.edu>
MIME-Version: 1.0
Content-Type: text/plain; charset="US-ASCII"
Content-Transfer-Encoding: 7bit
X-Mailer: Microsoft Outlook 14.0
Thread-index: AQEtpK0+ZMhTGTaNZYkccHtUfLPXDgKpFn1eAjsa3OcCFW1apACo0J5EAVlchLWT0vJIgA==
Content-language: en-us
X-CanIt-Geo: ip=128.146.216.86; country=US; region=OH; city=Columbus; latitude=39.9968; longitude=-82.9882; metrocode=535; areacode=614; http://maps.google.com/maps?q=39.9968,-82.9882&z=6
X-CanItPRO-Stream: outbound
X-Scanned-By: CanIt (www . roaringpenguin . com) on 128.146.216.82
Cc: kitten@ietf.org
X-BeenThere: kitten@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Common Authentication Technologies - Next Generation <kitten.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/kitten>
List-Post: <mailto:kitten@ietf.org>
List-Help: <mailto:kitten-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=subscribe>
X-List-Received-Date: Fri, 13 Aug 2010 17:36:12 -0000

> The folks in Oauth are not impressed by random statements like "we need
> channel bindings" because they want to know "What is it?" and "What are
> the benefits?". I think that this is the right attitude.

I don't think everything has to be useful to Google and a random army of
scripters that want to implement everything from scratch in Javascript
though. If you want a standard, it should be rich enough to handle more than
just what they choose to care about.

And before you say "come to the WG and ask", I don't believe the foundation
is right, so I'm not going to try and build my house extension on it. That's
my prerogative.

> Working in the Web 2.0 environment they want to get something done soon
> (rather than in 5 years).

Here's where that gets you:

http://code.google.com/p/browsersec/wiki/Main

I know some people think that can be reined back in and fixed, but I really
don't.

-- Scott



From cantor.2@osu.edu  Fri Aug 13 11:02:15 2010
Return-Path: <cantor.2@osu.edu>
X-Original-To: kitten@core3.amsl.com
Delivered-To: kitten@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 29B393A67F3 for <kitten@core3.amsl.com>; Fri, 13 Aug 2010 11:02:15 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -2.577
X-Spam-Level: 
X-Spam-Status: No, score=-2.577 tagged_above=-999 required=5 tests=[AWL=0.022,  BAYES_00=-2.599]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id MowY+AF2o96u for <kitten@core3.amsl.com>; Fri, 13 Aug 2010 11:02:14 -0700 (PDT)
Received: from defang2.it.ohio-state.edu (defang2.it.ohio-state.edu [128.146.216.82]) by core3.amsl.com (Postfix) with ESMTP id 27E8A3A635F for <kitten@ietf.org>; Fri, 13 Aug 2010 11:02:12 -0700 (PDT)
Received: from defang10.it.ohio-state.edu (defang10.it.ohio-state.edu [128.146.216.79]) by defang2.it.ohio-state.edu (8.13.7/8.13.1) with ESMTP id o7DI2nA3019905; Fri, 13 Aug 2010 14:02:49 -0400
Received: from SNOWDOG ([128.146.178.34]) by defang10.it.ohio-state.edu (8.13.7/8.13.1) with ESMTP id o7DI2mhB020201; Fri, 13 Aug 2010 14:02:48 -0400
From: "Scott Cantor" <cantor.2@osu.edu>
To: "'Nicolas Williams'" <Nicolas.Williams@oracle.com>
Subject: SAML CB strawman (was RE: Speaking in support of Scott's SAML proposal)
Date: Fri, 13 Aug 2010 14:02:51 -0400
Organization: The Ohio State University
Message-ID: <046b01cb3b11$bf6f3410$3e4d9c30$@osu.edu>
MIME-Version: 1.0
Content-Type: text/plain; charset="US-ASCII"
Content-Transfer-Encoding: 7bit
X-Mailer: Microsoft Outlook 14.0
Thread-index: Acs7DxitGOusT9d2SJSzWM39hVrFjw==
Content-language: en-us
X-CanIt-Geo: ip=128.146.216.79; country=US; region=OH; city=Columbus; latitude=39.9968; longitude=-82.9882; metrocode=535; areacode=614; http://maps.google.com/maps?q=39.9968,-82.9882&z=6
X-CanItPRO-Stream: outbound
X-Scanned-By: CanIt (www . roaringpenguin . com) on 128.146.216.82
Cc: kitten@ietf.org, 'Simon Josefsson' <simon@josefsson.org>
X-BeenThere: kitten@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Common Authentication Technologies - Next Generation <kitten.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/kitten>
List-Post: <mailto:kitten@ietf.org>
List-Help: <mailto:kitten-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=subscribe>
X-List-Received-Date: Fri, 13 Aug 2010 18:02:15 -0000

> The outcome of authentication signal itself needs to be authenticated,
> implicitly or explicitly.  Specifically, it must be impossible for an
> MITM to make a client think authentication succeeded when the real
> server thought it failed.

Yeah, I thought that's what you meant. I think that the goal here is to
implicitly do it through the IdP, which is why we have to sign the
AuthnRequest coming from the RP and get the CB into that. Often in SAML the
request isn't signed but I was leaning towards requiring it anyway in rev of
my draft, because without the web as the context, there's no location-based
authentication of URLs occurring.

> Does the client get an indication from the IdP as to whether the CB
> check passed?  Presumably the IdP won't issue the assertion if the CB
> fail to match!  Assuming that, then, yes, your approach works.

Yes, that's the indication. (And of course, the client probably
authenticates the IdP via TLS, unfortunately, it's still turtles all the way
down. I doubt that's news to anybody.)

> The RP doesn't have to check the CB data at this point if it trusts the
> IdP to have done it.  And if the client doesn't independently pass CB
> data to the RP then there's nothing the RP can do but trust that the IdP
> did it.

The reason I want it in the assertion regardless is so that the RP knows
that the client didn't manufacture an unsigned request without the CB data
and get the IdP to issue the assertion without any regard for that step.
 
> I believe your strawman stands.

If so, there are two main choices for capturing it, SubjectConfirmation and
Conditions. SC is somewhat elegant for this, although if the client isn't
sending the CB data to the RP at the end, it isn't really "confirming"
itself with it. Plus that would formally alter the underlying SAML profiles,
which assume bearer.

Using a new Condition is probably just as good and composes with the SAML
profiles well.

I need to read the RFC(s) to understand the types of TLS CB and get a handle
on how to represent them in the XML, but I think we might have an answer.

Thanks,
-- Scott



From Nicolas.Williams@oracle.com  Fri Aug 13 11:29:16 2010
Return-Path: <Nicolas.Williams@oracle.com>
X-Original-To: kitten@core3.amsl.com
Delivered-To: kitten@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 3D87D3A6850 for <kitten@core3.amsl.com>; Fri, 13 Aug 2010 11:29:16 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -6.533
X-Spam-Level: 
X-Spam-Status: No, score=-6.533 tagged_above=-999 required=5 tests=[AWL=0.065,  BAYES_00=-2.599, RCVD_IN_DNSWL_MED=-4, UNPARSEABLE_RELAY=0.001]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id fov1gGW4teE9 for <kitten@core3.amsl.com>; Fri, 13 Aug 2010 11:29:15 -0700 (PDT)
Received: from rcsinet10.oracle.com (rcsinet10.oracle.com [148.87.113.121]) by core3.amsl.com (Postfix) with ESMTP id 2E0343A6864 for <kitten@ietf.org>; Fri, 13 Aug 2010 11:29:15 -0700 (PDT)
Received: from acsinet15.oracle.com (acsinet15.oracle.com [141.146.126.227]) by rcsinet10.oracle.com (Switch-3.4.2/Switch-3.4.2) with ESMTP id o7DIToLA019984 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=OK); Fri, 13 Aug 2010 18:29:51 GMT
Received: from acsmt354.oracle.com (acsmt354.oracle.com [141.146.40.154]) by acsinet15.oracle.com (Switch-3.4.2/Switch-3.4.1) with ESMTP id o7DCIe9M002470; Fri, 13 Aug 2010 18:29:50 GMT
Received: from abhmt009.oracle.com by acsmt353.oracle.com with ESMTP id 515324201281724120; Fri, 13 Aug 2010 11:28:40 -0700
Received: from oracle.com (/129.153.128.104) by default (Oracle Beehive Gateway v4.0) with ESMTP ; Fri, 13 Aug 2010 11:28:40 -0700
Date: Fri, 13 Aug 2010 13:28:36 -0500
From: Nicolas Williams <Nicolas.Williams@oracle.com>
To: Scott Cantor <cantor.2@osu.edu>
Subject: Re: SAML CB strawman (was RE: Speaking in support of Scott's SAML proposal)
Message-ID: <20100813182836.GJ5217@oracle.com>
References: <046b01cb3b11$bf6f3410$3e4d9c30$@osu.edu>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <046b01cb3b11$bf6f3410$3e4d9c30$@osu.edu>
User-Agent: Mutt/1.5.20 (2010-03-02)
Cc: kitten@ietf.org, 'Simon Josefsson' <simon@josefsson.org>
X-BeenThere: kitten@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Common Authentication Technologies - Next Generation <kitten.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/kitten>
List-Post: <mailto:kitten@ietf.org>
List-Help: <mailto:kitten-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=subscribe>
X-List-Received-Date: Fri, 13 Aug 2010 18:29:16 -0000

On Fri, Aug 13, 2010 at 02:02:51PM -0400, Scott Cantor wrote:
> > The outcome of authentication signal itself needs to be authenticated,
> > implicitly or explicitly.  Specifically, it must be impossible for an
> > MITM to make a client think authentication succeeded when the real
> > server thought it failed.
> 
> Yeah, I thought that's what you meant. I think that the goal here is to
> implicitly do it through the IdP, which is why we have to sign the
> AuthnRequest coming from the RP and get the CB into that. 

Yes, that works.

>                                                           Often in SAML the
> request isn't signed but I was leaning towards requiring it anyway in rev of
> my draft, because without the web as the context, there's no location-based
> authentication of URLs occurring.

It's almost certainly better to do that, yes.

> > Does the client get an indication from the IdP as to whether the CB
> > check passed?  Presumably the IdP won't issue the assertion if the CB
> > fail to match!  Assuming that, then, yes, your approach works.
> 
> Yes, that's the indication. (And of course, the client probably
> authenticates the IdP via TLS, unfortunately, it's still turtles all
> the way down. I doubt that's news to anybody.)

Heh.  But here's the thing: if you reduce the number of places where we
rely on TLS' PKI to just user/device enrolment, then you've greatly
reduced the scope for attack.

> > The RP doesn't have to check the CB data at this point if it trusts the
> > IdP to have done it.  And if the client doesn't independently pass CB
> > data to the RP then there's nothing the RP can do but trust that the IdP
> > did it.
> 
> The reason I want it in the assertion regardless is so that the RP knows
> that the client didn't manufacture an unsigned request without the CB data
> and get the IdP to issue the assertion without any regard for that step.

Oh, sure, I see your point.  Of course, a single bit from the IdP
indicating that it checked the CBs and they matched will do.

> I need to read the RFC(s) to understand the types of TLS CB and get a handle
> on how to represent them in the XML, but I think we might have an answer.

You should represent them as a two item tuple: {CB type, CB data}.  The
CB data are opaque, binary data.

RFCs 5056 and 5929 are the relevant RFCs here.

Typically I think you'll find that getting at the 'tls-server-end-point'
CB is easier than 'tls-unique' because TLS APIs will give you access to
the server's cert and you can compute the CB data from that (well, with
a bit of effort).

Nico
-- 

From cantor.2@osu.edu  Fri Aug 13 11:37:39 2010
Return-Path: <cantor.2@osu.edu>
X-Original-To: kitten@core3.amsl.com
Delivered-To: kitten@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 594923A6850 for <kitten@core3.amsl.com>; Fri, 13 Aug 2010 11:37:39 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -2.579
X-Spam-Level: 
X-Spam-Status: No, score=-2.579 tagged_above=-999 required=5 tests=[AWL=0.020,  BAYES_00=-2.599]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id 86elYtld5j8S for <kitten@core3.amsl.com>; Fri, 13 Aug 2010 11:37:38 -0700 (PDT)
Received: from defang1.it.ohio-state.edu (defang1.it.ohio-state.edu [128.146.216.81]) by core3.amsl.com (Postfix) with ESMTP id 0C6B93A684B for <kitten@ietf.org>; Fri, 13 Aug 2010 11:37:37 -0700 (PDT)
Received: from defang9.it.ohio-state.edu (defang9.it.ohio-state.edu [128.146.216.78]) by defang1.it.ohio-state.edu (8.13.7/8.13.1) with ESMTP id o7DIcEat009213; Fri, 13 Aug 2010 14:38:14 -0400
Received: from SNOWDOG ([128.146.178.34]) by defang9.it.ohio-state.edu (8.13.7/8.13.1) with ESMTP id o7DIcCkt013603; Fri, 13 Aug 2010 14:38:13 -0400
From: "Scott Cantor" <cantor.2@osu.edu>
To: "'Nicolas Williams'" <Nicolas.Williams@oracle.com>
References: <046b01cb3b11$bf6f3410$3e4d9c30$@osu.edu> <20100813182836.GJ5217@oracle.com>
In-Reply-To: <20100813182836.GJ5217@oracle.com>
Subject: RE: SAML CB strawman (was RE: Speaking in support of Scott's SAML proposal)
Date: Fri, 13 Aug 2010 14:38:12 -0400
Organization: The Ohio State University
Message-ID: <047801cb3b16$b21c0090$165401b0$@osu.edu>
MIME-Version: 1.0
Content-Type: text/plain; charset="US-ASCII"
Content-Transfer-Encoding: 7bit
X-Mailer: Microsoft Outlook 14.0
Thread-index: AQHD1pmWw/vqVHiKX5LuFq5Uyh+aSwJi2s9Dkttnn3A=
Content-language: en-us
X-CanIt-Geo: ip=128.146.216.78; country=US; region=OH; city=Columbus; latitude=39.9968; longitude=-82.9882; metrocode=535; areacode=614; http://maps.google.com/maps?q=39.9968,-82.9882&z=6
X-CanItPRO-Stream: outbound
X-Scanned-By: CanIt (www . roaringpenguin . com) on 128.146.216.81
Cc: kitten@ietf.org, 'Simon Josefsson' <simon@josefsson.org>
X-BeenThere: kitten@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Common Authentication Technologies - Next Generation <kitten.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/kitten>
List-Post: <mailto:kitten@ietf.org>
List-Help: <mailto:kitten-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=subscribe>
X-List-Received-Date: Fri, 13 Aug 2010 18:37:39 -0000

> > Yes, that's the indication. (And of course, the client probably
> > authenticates the IdP via TLS, unfortunately, it's still turtles all
> > the way down. I doubt that's news to anybody.)
> 
> Heh.  But here's the thing: if you reduce the number of places where we
> rely on TLS' PKI to just user/device enrolment, then you've greatly
> reduced the scope for attack.

Right. Leap-of-faith works very well here also, since a users IdPs are few,
and stable.

> > The reason I want it in the assertion regardless is so that the RP knows
> > that the client didn't manufacture an unsigned request without the CB
data
> > and get the IdP to issue the assertion without any regard for that step.
> 
> Oh, sure, I see your point.  Of course, a single bit from the IdP
> indicating that it checked the CBs and they matched will do.

A Condition without any data in it is the SAML equivalent of a bit. This is
XML, after all, tens of Unicode characters where a single bit will do.
 
> You should represent them as a two item tuple: {CB type, CB data}.  The
> CB data are opaque, binary data.
> 
> RFCs 5056 and 5929 are the relevant RFCs here.

Yes, thanks.

> Typically I think you'll find that getting at the 'tls-server-end-point'
> CB is easier than 'tls-unique' because TLS APIs will give you access to
> the server's cert and you can compute the CB data from that (well, with
> a bit of effort).

I'm assuming the relevant pieces I'm defining are agnostic about it (thus
the encoding of the type with the data).

-- Scott



From Nicolas.Williams@oracle.com  Fri Aug 13 11:48:30 2010
Return-Path: <Nicolas.Williams@oracle.com>
X-Original-To: kitten@core3.amsl.com
Delivered-To: kitten@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 5BB773A6359 for <kitten@core3.amsl.com>; Fri, 13 Aug 2010 11:48:30 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -6.534
X-Spam-Level: 
X-Spam-Status: No, score=-6.534 tagged_above=-999 required=5 tests=[AWL=0.064,  BAYES_00=-2.599, RCVD_IN_DNSWL_MED=-4, UNPARSEABLE_RELAY=0.001]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id EjaDQ2Sybkwv for <kitten@core3.amsl.com>; Fri, 13 Aug 2010 11:48:28 -0700 (PDT)
Received: from rcsinet10.oracle.com (rcsinet10.oracle.com [148.87.113.121]) by core3.amsl.com (Postfix) with ESMTP id F1E603A6850 for <kitten@ietf.org>; Fri, 13 Aug 2010 11:48:27 -0700 (PDT)
Received: from rcsinet15.oracle.com (rcsinet15.oracle.com [148.87.113.117]) by rcsinet10.oracle.com (Switch-3.4.2/Switch-3.4.2) with ESMTP id o7DIn2A9000849 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=OK); Fri, 13 Aug 2010 18:49:04 GMT
Received: from acsmt353.oracle.com (acsmt353.oracle.com [141.146.40.153]) by rcsinet15.oracle.com (Switch-3.4.2/Switch-3.4.1) with ESMTP id o7DDQ2HK000549; Fri, 13 Aug 2010 18:49:02 GMT
Received: from abhmt017.oracle.com by acsmt353.oracle.com with ESMTP id 494709441281725250; Fri, 13 Aug 2010 11:47:30 -0700
Received: from oracle.com (/129.153.128.104) by default (Oracle Beehive Gateway v4.0) with ESMTP ; Fri, 13 Aug 2010 11:47:30 -0700
Date: Fri, 13 Aug 2010 13:47:20 -0500
From: Nicolas Williams <Nicolas.Williams@oracle.com>
To: Scott Cantor <cantor.2@osu.edu>
Subject: Re: SAML CB strawman (was RE: Speaking in support of Scott's SAML proposal)
Message-ID: <20100813184720.GA14409@oracle.com>
References: <046b01cb3b11$bf6f3410$3e4d9c30$@osu.edu> <20100813182836.GJ5217@oracle.com> <047801cb3b16$b21c0090$165401b0$@osu.edu>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <047801cb3b16$b21c0090$165401b0$@osu.edu>
User-Agent: Mutt/1.5.20 (2010-03-02)
Cc: kitten@ietf.org, 'Simon Josefsson' <simon@josefsson.org>
X-BeenThere: kitten@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Common Authentication Technologies - Next Generation <kitten.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/kitten>
List-Post: <mailto:kitten@ietf.org>
List-Help: <mailto:kitten-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=subscribe>
X-List-Received-Date: Fri, 13 Aug 2010 18:48:30 -0000

On Fri, Aug 13, 2010 at 02:38:12PM -0400, Scott Cantor wrote:
> > > Yes, that's the indication. (And of course, the client probably
> > > authenticates the IdP via TLS, unfortunately, it's still turtles all
> > > the way down. I doubt that's news to anybody.)
> > 
> > Heh.  But here's the thing: if you reduce the number of places where we
> > rely on TLS' PKI to just user/device enrolment, then you've greatly
> > reduced the scope for attack.
> 
> Right. Leap-of-faith works very well here also, since a users IdPs are few,
> and stable.

Exactly.

> > Typically I think you'll find that getting at the 'tls-server-end-point'
> > CB is easier than 'tls-unique' because TLS APIs will give you access to
> > the server's cert and you can compute the CB data from that (well, with
> > a bit of effort).
> 
> I'm assuming the relevant pieces I'm defining are agnostic about it (thus
> the encoding of the type with the data).

They must be.

From simon@josefsson.org  Fri Aug 13 14:57:34 2010
Return-Path: <simon@josefsson.org>
X-Original-To: kitten@core3.amsl.com
Delivered-To: kitten@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id AE2E53A6835 for <kitten@core3.amsl.com>; Fri, 13 Aug 2010 14:57:34 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -101.676
X-Spam-Level: 
X-Spam-Status: No, score=-101.676 tagged_above=-999 required=5 tests=[AWL=-0.936, BAYES_20=-0.74, USER_IN_WHITELIST=-100]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id VfGch7G0SKqh for <kitten@core3.amsl.com>; Fri, 13 Aug 2010 14:57:33 -0700 (PDT)
Received: from yxa-v.extundo.com (yxa-v.extundo.com [83.241.177.39]) by core3.amsl.com (Postfix) with ESMTP id C8DBC3A67B1 for <kitten@ietf.org>; Fri, 13 Aug 2010 14:57:32 -0700 (PDT)
Received: from mocca (c80-216-27-64.bredband.comhem.se [80.216.27.64]) (authenticated bits=0) by yxa-v.extundo.com (8.14.3/8.14.3/Debian-5+lenny1) with ESMTP id o7DLw1n9013009 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES128-SHA bits=128 verify=NOT) for <kitten@ietf.org>; Fri, 13 Aug 2010 23:58:03 +0200
X-Hashcash: 1:22:100813:kitten@ietf.org::cBid2ijrE0csjXlx:9sGl
From: Simon Josefsson <simon@josefsson.org>
To: kitten@ietf.org
Subject: SAML GSS-API name type
OpenPGP: id=B565716F; url=http://josefsson.org/key.txt
Date: Fri, 13 Aug 2010 23:58:01 +0200
Message-ID: <87fwyiyy3a.fsf@mocca.josefsson.org>
User-Agent: Gnus/5.110011 (No Gnus v0.11) Emacs/23.1 (gnu/linux)
MIME-Version: 1.0
Content-Type: multipart/mixed; boundary="=-=-="
X-Virus-Scanned: clamav-milter 0.96.1 at yxa-v
X-Virus-Status: Clean
X-BeenThere: kitten@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Common Authentication Technologies - Next Generation <kitten.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/kitten>
List-Post: <mailto:kitten@ietf.org>
List-Help: <mailto:kitten-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=subscribe>
X-List-Received-Date: Fri, 13 Aug 2010 21:57:34 -0000

--=-=-=

I posted this privately to Scott but I believe wider discussion would be
useful, so I'm posting my strawman specification on a SAML GSS-API name
type.  What do people think of it?

The open issues I can see are:

* It is not clear to me if it has to be restricted to SAML RPs.  It
  could be broadened by saying "SAML end-points" since the entityId is
  also used to identify IdPs (right?).

  I prefer to be conservative about broadening the scope until there is
  justification for it, though.  GSS-API normally only concerns clients
  and servers, and I don't believe an IdP could be regarded as either.

  The scope could be broadened even further by defining a GSS_C_NT_URI
  name type instead, which would hold a URI for arbitrary usage.  It
  could be used for both OpenID and SAML.  This seems a bit risky
  though, but reduces implementation costs.

* More details about SAML entityID canonicalization is required.

There may be more issues to consider...

/Simon

--=-=-=
Content-Disposition: inline;
 filename=draft-josefsson-gssapi-saml-nametype.txt




Network Working Group                                       S. Josefsson
Internet-Draft                                                    SJD AB
Intended status: Informational                           August 13, 2010
Expires: February 14, 2011


               GSS-API Name Type For a SAML Relying Party
                draft-josefsson-gssapi-saml-nametype-00

Abstract

   This document specify a new GSS-API name type and associated syntax
   intended for identifying a SAML Relying Party.

Status of this Memo

   This Internet-Draft is submitted in full conformance with the
   provisions of BCP 78 and BCP 79.

   Internet-Drafts are working documents of the Internet Engineering
   Task Force (IETF).  Note that other groups may also distribute
   working documents as Internet-Drafts.  The list of current Internet-
   Drafts is at http://datatracker.ietf.org/drafts/current/.

   Internet-Drafts are draft documents valid for a maximum of six months
   and may be updated, replaced, or obsoleted by other documents at any
   time.  It is inappropriate to use Internet-Drafts as reference
   material or to cite them other than as "work in progress."

   This Internet-Draft will expire on February 14, 2011.

Copyright Notice

   Copyright (c) 2010 IETF Trust and the persons identified as the
   document authors.  All rights reserved.

   This document is subject to BCP 78 and the IETF Trust's Legal
   Provisions Relating to IETF Documents
   (http://trustee.ietf.org/license-info) in effect on the date of
   publication of this document.  Please review these documents
   carefully, as they describe your rights and restrictions with respect
   to this document.  Code Components extracted from this document must
   include Simplified BSD License text as described in Section 4.e of
   the Trust Legal Provisions and are provided without warranty as
   described in the Simplified BSD License.






Josefsson               Expires February 14, 2011               [Page 1]

Internet-Draft          SAML RP GSS-API Name Type            August 2010


Table of Contents

   1.  Introduction  . . . . . . . . . . . . . . . . . . . . . . . . . 3
   2.  SAML Relying Party Name Form  . . . . . . . . . . . . . . . . . 3
   3.  Copying conditions  . . . . . . . . . . . . . . . . . . . . . . 3
   4.  Security Considerations . . . . . . . . . . . . . . . . . . . . 3
   5.  IANA Considerations . . . . . . . . . . . . . . . . . . . . . . 3
   6.  Normative References  . . . . . . . . . . . . . . . . . . . . . 3
   Author's Address  . . . . . . . . . . . . . . . . . . . . . . . . . 4










































Josefsson               Expires February 14, 2011               [Page 2]

Internet-Draft          SAML RP GSS-API Name Type            August 2010


1.  Introduction

   The Generic Security Service Application Program Interface (GSS-API)
   [RFC2743] specify name types for use in various situations, but none
   is applicable to Security Assertion Markup Language (SAML) Relying
   Partys which are identified by a entityID Uniform Resource Identifier
   (URI) [RFC3986].  This document specify a name type and associated
   syntax for use when identifying SAML Relying Party end-points.


2.  SAML Relying Party Name Form

   This name form shall be represented by the Object Identifier
   1.3.6.1.4.1.11591.4.7.

   The recommended symbolic name for this type is "GSS_C_NT_SAML_RP".

   This name type is used to represent services associated with SAML
   Relying Partys.  The name form syntax is a URI.  For example:

        https://idp.example.org/SAML2

   XXX: How does canonicalization work?


3.  Copying conditions

   This document should be considered a Code Component, and is thus
   effectively available under the BSD license.


4.  Security Considerations

   TBW.

   XXX: Canonicalization needs to be secure -- no insecure DNS.


5.  IANA Considerations

   None.


6.  Normative References

   [RFC2743]  Linn, J., "Generic Security Service Application Program
              Interface Version 2, Update 1", RFC 2743, January 2000.




Josefsson               Expires February 14, 2011               [Page 3]

Internet-Draft          SAML RP GSS-API Name Type            August 2010


   [RFC3986]  Berners-Lee, T., Fielding, R., and L. Masinter, "Uniform
              Resource Identifier (URI): Generic Syntax", STD 66,
              RFC 3986, January 2005.


Author's Address

   Simon Josefsson
   SJD AB
   Hagagatan 24
   Stockholm  113 47
   SE

   Email: simon@josefsson.org
   URI:   http://josefsson.org/




































Josefsson               Expires February 14, 2011               [Page 4]


--=-=-=--

From Nicolas.Williams@oracle.com  Fri Aug 13 15:05:51 2010
Return-Path: <Nicolas.Williams@oracle.com>
X-Original-To: kitten@core3.amsl.com
Delivered-To: kitten@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id D01293A6A02 for <kitten@core3.amsl.com>; Fri, 13 Aug 2010 15:05:51 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -6.535
X-Spam-Level: 
X-Spam-Status: No, score=-6.535 tagged_above=-999 required=5 tests=[AWL=0.063,  BAYES_00=-2.599, RCVD_IN_DNSWL_MED=-4, UNPARSEABLE_RELAY=0.001]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id kS77dqnBftQ1 for <kitten@core3.amsl.com>; Fri, 13 Aug 2010 15:05:50 -0700 (PDT)
Received: from rcsinet10.oracle.com (rcsinet10.oracle.com [148.87.113.121]) by core3.amsl.com (Postfix) with ESMTP id B28643A6A01 for <kitten@ietf.org>; Fri, 13 Aug 2010 15:05:50 -0700 (PDT)
Received: from rcsinet15.oracle.com (rcsinet15.oracle.com [148.87.113.117]) by rcsinet10.oracle.com (Switch-3.4.2/Switch-3.4.2) with ESMTP id o7DM6Phs021291 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=OK); Fri, 13 Aug 2010 22:06:27 GMT
Received: from acsmt354.oracle.com (acsmt354.oracle.com [141.146.40.154]) by rcsinet15.oracle.com (Switch-3.4.2/Switch-3.4.1) with ESMTP id o7D8e4Kn015249; Fri, 13 Aug 2010 22:06:25 GMT
Received: from abhmt007.oracle.com by acsmt354.oracle.com with ESMTP id 515862031281737152; Fri, 13 Aug 2010 15:05:52 -0700
Received: from oracle.com (/129.153.128.104) by default (Oracle Beehive Gateway v4.0) with ESMTP ; Fri, 13 Aug 2010 15:05:52 -0700
Date: Fri, 13 Aug 2010 17:05:47 -0500
From: Nicolas Williams <Nicolas.Williams@oracle.com>
To: Simon Josefsson <simon@josefsson.org>
Subject: Re: SAML GSS-API name type
Message-ID: <20100813220547.GE14409@oracle.com>
References: <87fwyiyy3a.fsf@mocca.josefsson.org>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <87fwyiyy3a.fsf@mocca.josefsson.org>
User-Agent: Mutt/1.5.20 (2010-03-02)
Cc: kitten@ietf.org
X-BeenThere: kitten@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Common Authentication Technologies - Next Generation <kitten.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/kitten>
List-Post: <mailto:kitten@ietf.org>
List-Help: <mailto:kitten-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=subscribe>
X-List-Received-Date: Fri, 13 Aug 2010 22:05:51 -0000

On Fri, Aug 13, 2010 at 11:58:01PM +0200, Simon Josefsson wrote:
> I posted this privately to Scott but I believe wider discussion would be
> useful, so I'm posting my strawman specification on a SAML GSS-API name
> type.  What do people think of it?
> 
> The open issues I can see are:
> 
> * It is not clear to me if it has to be restricted to SAML RPs.  It
>   could be broadened by saying "SAML end-points" since the entityId is
>   also used to identify IdPs (right?).
> 
>   I prefer to be conservative about broadening the scope until there is
>   justification for it, though.  GSS-API normally only concerns clients
>   and servers, and I don't believe an IdP could be regarded as either.

I don't agree.  See Sam's and my concerns about the naming extensions
draft.  One of them is that name attributes need to have an associated
"issuer name" meta-attribute.

Name of Kerberos realm, name of IdP, name of CA, ... -- these are all
"issuer names", and they are very much relevant.  As is "trust path",
which I'd envisioned as a name attribute -- trust paths are... sequences
of issuer names.

>   The scope could be broadened even further by defining a GSS_C_NT_URI
>   name type instead, which would hold a URI for arbitrary usage.  It
>   could be used for both OpenID and SAML.  This seems a bit risky
>   though, but reduces implementation costs.

For names to be useable by users there must be a user-friendly name
syntax and a realistic process for canonicalizing user-input names.

You can use anything you want for canonical names.  For example, OpenID
URLs would be fine.  But you can't expect users to type those in in any
UIs (heh; imagine users shortening OpenID URLs with URL shortening
services!)

> * More details about SAML entityID canonicalization is required.

Yep!

Nico
-- 

From cantor.2@osu.edu  Fri Aug 13 15:21:32 2010
Return-Path: <cantor.2@osu.edu>
X-Original-To: kitten@core3.amsl.com
Delivered-To: kitten@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id C17D63A6940 for <kitten@core3.amsl.com>; Fri, 13 Aug 2010 15:21:32 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -2.58
X-Spam-Level: 
X-Spam-Status: No, score=-2.58 tagged_above=-999 required=5 tests=[AWL=0.019,  BAYES_00=-2.599]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id nBLIgAodd+IU for <kitten@core3.amsl.com>; Fri, 13 Aug 2010 15:21:29 -0700 (PDT)
Received: from defang2.it.ohio-state.edu (defang2.it.ohio-state.edu [128.146.216.82]) by core3.amsl.com (Postfix) with ESMTP id 0636D3A6969 for <kitten@ietf.org>; Fri, 13 Aug 2010 15:21:28 -0700 (PDT)
Received: from defang9.it.ohio-state.edu (defang9.it.ohio-state.edu [128.146.216.78]) by defang2.it.ohio-state.edu (8.13.7/8.13.1) with ESMTP id o7DMM5vx030187; Fri, 13 Aug 2010 18:22:05 -0400
Received: from SNOWDOG ([128.146.178.34]) by defang9.it.ohio-state.edu (8.13.7/8.13.1) with ESMTP id o7DMM4kH004555; Fri, 13 Aug 2010 18:22:04 -0400
From: "Scott Cantor" <cantor.2@osu.edu>
To: "'Simon Josefsson'" <simon@josefsson.org>, <kitten@ietf.org>
References: <87fwyiyy3a.fsf@mocca.josefsson.org>
In-Reply-To: <87fwyiyy3a.fsf@mocca.josefsson.org>
Subject: RE: SAML GSS-API name type
Date: Fri, 13 Aug 2010 18:22:07 -0400
Organization: The Ohio State University
Message-ID: <04ad01cb3b35$f7c046f0$e740d4d0$@osu.edu>
MIME-Version: 1.0
Content-Type: text/plain; charset="US-ASCII"
Content-Transfer-Encoding: 7bit
X-Mailer: Microsoft Outlook 14.0
Thread-index: AQIYQJ6U2n8rcwBxL/cPUKI4oSIo1ZJF6aHw
Content-language: en-us
X-CanIt-Geo: ip=128.146.216.78; country=US; region=OH; city=Columbus; latitude=39.9968; longitude=-82.9882; metrocode=535; areacode=614; http://maps.google.com/maps?q=39.9968,-82.9882&z=6
X-CanItPRO-Stream: outbound
X-Scanned-By: CanIt (www . roaringpenguin . com) on 128.146.216.82
X-BeenThere: kitten@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Common Authentication Technologies - Next Generation <kitten.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/kitten>
List-Post: <mailto:kitten@ietf.org>
List-Help: <mailto:kitten-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=subscribe>
X-List-Received-Date: Fri, 13 Aug 2010 22:21:33 -0000

> I posted this privately to Scott but I believe wider discussion would be
> useful, so I'm posting my strawman specification on a SAML GSS-API name
> type.  What do people think of it?

Simon, I just sent you a response, you can forward that to the list if you
like.

> The open issues I can see are:
> 
> * It is not clear to me if it has to be restricted to SAML RPs.  It
>   could be broadened by saying "SAML end-points" since the entityId is
>   also used to identify IdPs (right?).

It should be cast as "this is how to represent a SAML entityID", formally
referencing the definition of that concept in SAML core, 8.3. It isn't
necessary to define when that would be used, that's a SAML issue.

>   The scope could be broadened even further by defining a GSS_C_NT_URI
>   name type instead, which would hold a URI for arbitrary usage.  It
>   could be used for both OpenID and SAML.  This seems a bit risky
>   though, but reduces implementation costs.

It's important to communicate the semantics of names, not just the syntax.
OpenIDs and SAML entityIDs are not at all the same, so they shouldn't be
represented as such.

> * More details about SAML entityID canonicalization is required.

To give me a leg up, could you point a newbie at the relevant RFCs to help
me understand the underlying requirements?

-- Scott



From simon@josefsson.org  Fri Aug 13 15:30:36 2010
Return-Path: <simon@josefsson.org>
X-Original-To: kitten@core3.amsl.com
Delivered-To: kitten@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 126F73A6922 for <kitten@core3.amsl.com>; Fri, 13 Aug 2010 15:30:36 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -102.662
X-Spam-Level: 
X-Spam-Status: No, score=-102.662 tagged_above=-999 required=5 tests=[AWL=-0.063, BAYES_00=-2.599, USER_IN_WHITELIST=-100]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id 0V3wG8Qk+e6H for <kitten@core3.amsl.com>; Fri, 13 Aug 2010 15:30:33 -0700 (PDT)
Received: from yxa-v.extundo.com (yxa-v.extundo.com [83.241.177.39]) by core3.amsl.com (Postfix) with ESMTP id 12DE83A69BD for <kitten@ietf.org>; Fri, 13 Aug 2010 15:30:31 -0700 (PDT)
Received: from mocca (c80-216-27-64.bredband.comhem.se [80.216.27.64]) (authenticated bits=0) by yxa-v.extundo.com (8.14.3/8.14.3/Debian-5+lenny1) with ESMTP id o7DMV2HW014895 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES128-SHA bits=128 verify=NOT); Sat, 14 Aug 2010 00:31:04 +0200
From: Simon Josefsson <simon@josefsson.org>
To: "Scott Cantor" <cantor.2@osu.edu>
Subject: Re: SAML GSS-API name type
References: <87fwyiyy3a.fsf@mocca.josefsson.org> <04ad01cb3b35$f7c046f0$e740d4d0$@osu.edu>
OpenPGP: id=B565716F; url=http://josefsson.org/key.txt
X-Hashcash: 1:22:100813:cantor.2@osu.edu::dOpSd0sH+0HSa2cs:7OOr
X-Hashcash: 1:22:100813:kitten@ietf.org::JAt77+qv9Ebd2qij:ImNw
Date: Sat, 14 Aug 2010 00:31:02 +0200
In-Reply-To: <04ad01cb3b35$f7c046f0$e740d4d0$@osu.edu> (Scott Cantor's message of "Fri, 13 Aug 2010 18:22:07 -0400")
Message-ID: <87sk2ixhzt.fsf@mocca.josefsson.org>
User-Agent: Gnus/5.110011 (No Gnus v0.11) Emacs/23.1 (gnu/linux)
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
X-Virus-Scanned: clamav-milter 0.96.1 at yxa-v
X-Virus-Status: Clean
Cc: kitten@ietf.org
X-BeenThere: kitten@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Common Authentication Technologies - Next Generation <kitten.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/kitten>
List-Post: <mailto:kitten@ietf.org>
List-Help: <mailto:kitten-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=subscribe>
X-List-Received-Date: Fri, 13 Aug 2010 22:30:36 -0000

"Scott Cantor" <cantor.2@osu.edu> writes:

>> I posted this privately to Scott but I believe wider discussion would be
>> useful, so I'm posting my strawman specification on a SAML GSS-API name
>> type.  What do people think of it?
>
> Simon, I just sent you a response, you can forward that to the list if you
> like.

I'll quote one relevant part below.

>> The open issues I can see are:
>> 
>> * It is not clear to me if it has to be restricted to SAML RPs.  It
>>   could be broadened by saying "SAML end-points" since the entityId is
>>   also used to identify IdPs (right?).
>
> It should be cast as "this is how to represent a SAML entityID", formally
> referencing the definition of that concept in SAML core, 8.3. It isn't
> necessary to define when that would be used, that's a SAML issue.

I like this approach better.

>>   The scope could be broadened even further by defining a GSS_C_NT_URI
>>   name type instead, which would hold a URI for arbitrary usage.  It
>>   could be used for both OpenID and SAML.  This seems a bit risky
>>   though, but reduces implementation costs.
>
> It's important to communicate the semantics of names, not just the syntax.
> OpenIDs and SAML entityIDs are not at all the same, so they shouldn't be
> represented as such.

That is my preference as well.

>> * More details about SAML entityID canonicalization is required.
>
> To give me a leg up, could you point a newbie at the relevant RFCs to help
> me understand the underlying requirements?

Check RFC 2743 and search for "canonical".

But you said privately:

  I'll take a look. Can you give me a brief summary of what the goals of
  the c14n process here are? In general, SAML doesn't expect
  implementations to do any normalization of URIs when used as
  identifiers. They're compared as strings, so only the character
  encoding has to be addressed to compare them as binary data.

This suggests that there is no URI canonicalization process to speak of
in SAML.  Every URI is a canonical SAML entityID?  This is quite
different from OpenID which definitely specify how URIs are
canonicalized, see:

http://openid.net/specs/openid-authentication-2_0.html#normalization

This is another reason why the OpenID and SAML name types should be
specified separately for GSS-API.

/Simon

From simon@josefsson.org  Fri Aug 13 15:34:44 2010
Return-Path: <simon@josefsson.org>
X-Original-To: kitten@core3.amsl.com
Delivered-To: kitten@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 723683A67E3 for <kitten@core3.amsl.com>; Fri, 13 Aug 2010 15:34:44 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -102.658
X-Spam-Level: 
X-Spam-Status: No, score=-102.658 tagged_above=-999 required=5 tests=[AWL=-0.059, BAYES_00=-2.599, USER_IN_WHITELIST=-100]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id xDtUw9c-KvdN for <kitten@core3.amsl.com>; Fri, 13 Aug 2010 15:34:42 -0700 (PDT)
Received: from yxa-v.extundo.com (yxa-v.extundo.com [83.241.177.39]) by core3.amsl.com (Postfix) with ESMTP id 03E2F3A6852 for <kitten@ietf.org>; Fri, 13 Aug 2010 15:34:41 -0700 (PDT)
Received: from mocca (c80-216-27-64.bredband.comhem.se [80.216.27.64]) (authenticated bits=0) by yxa-v.extundo.com (8.14.3/8.14.3/Debian-5+lenny1) with ESMTP id o7DMZ66n015071 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES128-SHA bits=128 verify=NOT); Sat, 14 Aug 2010 00:35:10 +0200
From: Simon Josefsson <simon@josefsson.org>
To: Nicolas Williams <Nicolas.Williams@oracle.com>
Subject: Re: SAML GSS-API name type
References: <87fwyiyy3a.fsf@mocca.josefsson.org> <20100813220547.GE14409@oracle.com>
OpenPGP: id=B565716F; url=http://josefsson.org/key.txt
X-Hashcash: 1:22:100813:nicolas.williams@oracle.com::uUM2S+P1x3EWa7Qw:vEc
X-Hashcash: 1:22:100813:kitten@ietf.org::COqq7jFK0Doq0uLV:bd+H
Date: Sat, 14 Aug 2010 00:35:05 +0200
In-Reply-To: <20100813220547.GE14409@oracle.com> (Nicolas Williams's message of "Fri, 13 Aug 2010 17:05:47 -0500")
Message-ID: <87ocd6xht2.fsf@mocca.josefsson.org>
User-Agent: Gnus/5.110011 (No Gnus v0.11) Emacs/23.1 (gnu/linux)
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
X-Virus-Scanned: clamav-milter 0.96.1 at yxa-v
X-Virus-Status: Clean
Cc: kitten@ietf.org
X-BeenThere: kitten@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Common Authentication Technologies - Next Generation <kitten.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/kitten>
List-Post: <mailto:kitten@ietf.org>
List-Help: <mailto:kitten-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=subscribe>
X-List-Received-Date: Fri, 13 Aug 2010 22:34:45 -0000

Nicolas Williams <Nicolas.Williams@oracle.com> writes:

> On Fri, Aug 13, 2010 at 11:58:01PM +0200, Simon Josefsson wrote:
>> I posted this privately to Scott but I believe wider discussion would be
>> useful, so I'm posting my strawman specification on a SAML GSS-API name
>> type.  What do people think of it?
>> 
>> The open issues I can see are:
>> 
>> * It is not clear to me if it has to be restricted to SAML RPs.  It
>>   could be broadened by saying "SAML end-points" since the entityId is
>>   also used to identify IdPs (right?).
>> 
>>   I prefer to be conservative about broadening the scope until there is
>>   justification for it, though.  GSS-API normally only concerns clients
>>   and servers, and I don't believe an IdP could be regarded as either.
>
> I don't agree.  See Sam's and my concerns about the naming extensions
> draft.  One of them is that name attributes need to have an associated
> "issuer name" meta-attribute.
>
> Name of Kerberos realm, name of IdP, name of CA, ... -- these are all
> "issuer names", and they are very much relevant.  As is "trust path",
> which I'd envisioned as a name attribute -- trust paths are... sequences
> of issuer names.

If the document uses Scott's suggestion to just encode a SAML entityID,
does that address your concern?  I'm not sure I understand what you are
getting at here.

>>   The scope could be broadened even further by defining a GSS_C_NT_URI
>>   name type instead, which would hold a URI for arbitrary usage.  It
>>   could be used for both OpenID and SAML.  This seems a bit risky
>>   though, but reduces implementation costs.
>
> For names to be useable by users there must be a user-friendly name
> syntax and a realistic process for canonicalizing user-input names.
>
> You can use anything you want for canonical names.  For example, OpenID
> URLs would be fine.  But you can't expect users to type those in in any
> UIs (heh; imagine users shortening OpenID URLs with URL shortening
> services!)

I'm not sure SAML has a concept of user-friendly names that we need to
care about?  It would certainly be simpler to just let the
canonicalization process for this name type do nothing.  If I understand
correctly, if the SAML entityIDs doesn't match completely,
authentication will not work.

>> * More details about SAML entityID canonicalization is required.
>
> Yep!

Right now I'm hoping it will be simple.

/Simon

From cantor.2@osu.edu  Fri Aug 13 15:52:04 2010
Return-Path: <cantor.2@osu.edu>
X-Original-To: kitten@core3.amsl.com
Delivered-To: kitten@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id EF5D53A6999 for <kitten@core3.amsl.com>; Fri, 13 Aug 2010 15:52:03 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -3.082
X-Spam-Level: 
X-Spam-Status: No, score=-3.082 tagged_above=-999 required=5 tests=[AWL=0.517,  BAYES_00=-2.599, RCVD_IN_DNSWL_LOW=-1]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id kQIL0NzZrl3O for <kitten@core3.amsl.com>; Fri, 13 Aug 2010 15:52:03 -0700 (PDT)
Received: from defang20.it.ohio-state.edu (defang20.it.ohio-state.edu [128.146.216.134]) by core3.amsl.com (Postfix) with ESMTP id 18A013A6995 for <kitten@ietf.org>; Fri, 13 Aug 2010 15:52:03 -0700 (PDT)
Received: from defang9.it.ohio-state.edu (defang9.it.ohio-state.edu [128.146.216.78]) by defang20.it.ohio-state.edu (8.13.7/8.13.1) with ESMTP id o7DMqdeG017828; Fri, 13 Aug 2010 18:52:39 -0400
Received: from SNOWDOG ([128.146.178.34]) by defang9.it.ohio-state.edu (8.13.7/8.13.1) with ESMTP id o7DMqc57015932; Fri, 13 Aug 2010 18:52:39 -0400
From: "Scott Cantor" <cantor.2@osu.edu>
To: "'Simon Josefsson'" <simon@josefsson.org>
References: <87fwyiyy3a.fsf@mocca.josefsson.org> <04ad01cb3b35$f7c046f0$e740d4d0$@osu.edu> <87sk2ixhzt.fsf@mocca.josefsson.org>
In-Reply-To: <87sk2ixhzt.fsf@mocca.josefsson.org>
Subject: RE: SAML GSS-API name type
Date: Fri, 13 Aug 2010 18:52:42 -0400
Organization: The Ohio State University
Message-ID: <04b301cb3b3a$3d29c280$b77d4780$@osu.edu>
MIME-Version: 1.0
Content-Type: text/plain; charset="US-ASCII"
Content-Transfer-Encoding: 7bit
X-Mailer: Microsoft Outlook 14.0
Thread-index: AQIYQJ6U2n8rcwBxL/cPUKI4oSIo1QIu/D8IAhYOhGGSI8mi0A==
Content-language: en-us
X-CanIt-Geo: ip=128.146.216.78; country=US; region=OH; city=Columbus; latitude=39.9968; longitude=-82.9882; metrocode=535; areacode=614; http://maps.google.com/maps?q=39.9968,-82.9882&z=6
X-CanItPRO-Stream: outbound
X-Scanned-By: CanIt (www . roaringpenguin . com) on 128.146.216.134
Cc: kitten@ietf.org
X-BeenThere: kitten@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Common Authentication Technologies - Next Generation <kitten.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/kitten>
List-Post: <mailto:kitten@ietf.org>
List-Help: <mailto:kitten-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=subscribe>
X-List-Received-Date: Fri, 13 Aug 2010 22:52:04 -0000

> This suggests that there is no URI canonicalization process to speak of
> in SAML.  Every URI is a canonical SAML entityID?

Yes, basically. The actual normative language explaining that is obscure and
a bit wrongly worded as it turns out, but the intent is that strings and
URIs are always compared byte for byte (assuming the same encoding) unless
something specifies differently in a profile or particular construct.

-- Scott



From cantor.2@osu.edu  Fri Aug 13 16:00:06 2010
Return-Path: <cantor.2@osu.edu>
X-Original-To: kitten@core3.amsl.com
Delivered-To: kitten@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 22B843A6A2D for <kitten@core3.amsl.com>; Fri, 13 Aug 2010 16:00:06 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -2.616
X-Spam-Level: 
X-Spam-Status: No, score=-2.616 tagged_above=-999 required=5 tests=[AWL=-0.017, BAYES_00=-2.599]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id Vl8lP9HEwQxL for <kitten@core3.amsl.com>; Fri, 13 Aug 2010 16:00:05 -0700 (PDT)
Received: from defang19.it.ohio-state.edu (defang19.it.ohio-state.edu [128.146.216.133]) by core3.amsl.com (Postfix) with ESMTP id 13FD43A69BF for <kitten@ietf.org>; Fri, 13 Aug 2010 15:59:55 -0700 (PDT)
Received: from defang10.it.ohio-state.edu (defang10.it.ohio-state.edu [128.146.216.79]) by defang19.it.ohio-state.edu (8.13.7/8.13.1) with ESMTP id o7DN0W9v003472; Fri, 13 Aug 2010 19:00:32 -0400
Received: from SNOWDOG ([128.146.178.34]) by defang10.it.ohio-state.edu (8.13.7/8.13.1) with ESMTP id o7DN0U8v013025; Fri, 13 Aug 2010 19:00:31 -0400
From: "Scott Cantor" <cantor.2@osu.edu>
To: "'Simon Josefsson'" <simon@josefsson.org>, "'Nicolas Williams'" <Nicolas.Williams@oracle.com>
References: <87fwyiyy3a.fsf@mocca.josefsson.org> <20100813220547.GE14409@oracle.com> <87ocd6xht2.fsf@mocca.josefsson.org>
In-Reply-To: <87ocd6xht2.fsf@mocca.josefsson.org>
Subject: RE: SAML GSS-API name type
Date: Fri, 13 Aug 2010 19:00:34 -0400
Organization: The Ohio State University
Message-ID: <04ba01cb3b3b$56a05b10$03e11130$@osu.edu>
MIME-Version: 1.0
Content-Type: text/plain; charset="US-ASCII"
Content-Transfer-Encoding: 7bit
X-Mailer: Microsoft Outlook 14.0
Thread-index: AQIYQJ6U2n8rcwBxL/cPUKI4oSIo1QJmYzg4AfPVN72SIyIWQA==
Content-language: en-us
X-CanIt-Geo: ip=128.146.216.79; country=US; region=OH; city=Columbus; latitude=39.9968; longitude=-82.9882; metrocode=535; areacode=614; http://maps.google.com/maps?q=39.9968,-82.9882&z=6
X-CanItPRO-Stream: outbound
X-Scanned-By: CanIt (www . roaringpenguin . com) on 128.146.216.133
Cc: kitten@ietf.org
X-BeenThere: kitten@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Common Authentication Technologies - Next Generation <kitten.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/kitten>
List-Post: <mailto:kitten@ietf.org>
List-Help: <mailto:kitten-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=subscribe>
X-List-Received-Date: Fri, 13 Aug 2010 23:00:06 -0000

> I'm not sure SAML has a concept of user-friendly names that we need to
> care about?

Not in scope of SAML, that's a higher order issue that would be captured in
the definition of the NameID Format or Attribute definition.

In the specific case of server names (loosely defining an entityID), they
aren't meant that way, and one doesn't tend to name entities with crazy URL
patterns that would be inclined to cause errors in matching to avoid
problems.

> It would certainly be simpler to just let the
> canonicalization process for this name type do nothing.  If I understand
> correctly, if the SAML entityIDs doesn't match completely,
> authentication will not work.

If the scenario is that some kind of expected peer name is specified by the
client, then yes. The matching would be "make sure the encoding is in
common, then compare byte for byte".

-- Scott



From Nicolas.Williams@oracle.com  Fri Aug 13 16:03:12 2010
Return-Path: <Nicolas.Williams@oracle.com>
X-Original-To: kitten@core3.amsl.com
Delivered-To: kitten@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 9B30B3A69D1 for <kitten@core3.amsl.com>; Fri, 13 Aug 2010 16:03:12 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -6.536
X-Spam-Level: 
X-Spam-Status: No, score=-6.536 tagged_above=-999 required=5 tests=[AWL=0.062,  BAYES_00=-2.599, RCVD_IN_DNSWL_MED=-4, UNPARSEABLE_RELAY=0.001]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id sVkRY+jtJXnY for <kitten@core3.amsl.com>; Fri, 13 Aug 2010 16:03:11 -0700 (PDT)
Received: from rcsinet10.oracle.com (rcsinet10.oracle.com [148.87.113.121]) by core3.amsl.com (Postfix) with ESMTP id A5FA83A69BF for <kitten@ietf.org>; Fri, 13 Aug 2010 16:03:11 -0700 (PDT)
Received: from acsinet15.oracle.com (acsinet15.oracle.com [141.146.126.227]) by rcsinet10.oracle.com (Switch-3.4.2/Switch-3.4.2) with ESMTP id o7DN3kOu006646 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=OK); Fri, 13 Aug 2010 23:03:47 GMT
Received: from acsmt353.oracle.com (acsmt353.oracle.com [141.146.40.153]) by acsinet15.oracle.com (Switch-3.4.2/Switch-3.4.1) with ESMTP id o7DHwErt022617; Fri, 13 Aug 2010 23:03:45 GMT
Received: from abhmt006.oracle.com by acsmt355.oracle.com with ESMTP id 515969661281740517; Fri, 13 Aug 2010 16:01:57 -0700
Received: from oracle.com (/129.153.128.104) by default (Oracle Beehive Gateway v4.0) with ESMTP ; Fri, 13 Aug 2010 16:01:57 -0700
Date: Fri, 13 Aug 2010 18:01:50 -0500
From: Nicolas Williams <Nicolas.Williams@oracle.com>
To: Simon Josefsson <simon@josefsson.org>
Subject: Re: SAML GSS-API name type
Message-ID: <20100813230150.GF14409@oracle.com>
References: <87fwyiyy3a.fsf@mocca.josefsson.org> <20100813220547.GE14409@oracle.com> <87ocd6xht2.fsf@mocca.josefsson.org>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <87ocd6xht2.fsf@mocca.josefsson.org>
User-Agent: Mutt/1.5.20 (2010-03-02)
Cc: kitten@ietf.org
X-BeenThere: kitten@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Common Authentication Technologies - Next Generation <kitten.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/kitten>
List-Post: <mailto:kitten@ietf.org>
List-Help: <mailto:kitten-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=subscribe>
X-List-Received-Date: Fri, 13 Aug 2010 23:03:12 -0000

On Sat, Aug 14, 2010 at 12:35:05AM +0200, Simon Josefsson wrote:
> Nicolas Williams <Nicolas.Williams@oracle.com> writes:
> > Name of Kerberos realm, name of IdP, name of CA, ... -- these are all
> > "issuer names", and they are very much relevant.  As is "trust path",
> > which I'd envisioned as a name attribute -- trust paths are... sequences
> > of issuer names.
> 
> If the document uses Scott's suggestion to just encode a SAML entityID,
> does that address your concern?  I'm not sure I understand what you are
> getting at here.

Yes, it does.

> I'm not sure SAML has a concept of user-friendly names that we need to
> care about?  It would certainly be simpler to just let the
> canonicalization process for this name type do nothing.  If I understand
> correctly, if the SAML entityIDs doesn't match completely,
> authentication will not work.

But the GSS-API does have "a concept of user-friendly names that we need
to care about", at least for host-based service names (OK, that's
probably not an issue here).

> >> * More details about SAML entityID canonicalization is required.
> >
> > Yep!
> 
> Right now I'm hoping it will be simple.

I think so: just deal with names much as Kerberos implementations do.

This should work:

	gss_canonicalize_name(saml_mech_oid,
	    gss_import_name(GSS_C_NO_OID, "<entityID>"))

so should:

	gss_canonicalize_name(saml_mech_oid,
	    gss_import_name(GSS_C_NT_SAML_ENTITYID, "<entityID>"))

Ideally there will be some user-friendly <entityID> syntax that will be
accepted and canonicalized.

This must work:

	gss_canonicalize_name(saml_mech_oid,
	    gss_import_name(GSS_C_NT_HOSTBASED_SERVICE,
	    "ldap@foo.bar.example"))

gss_display_name() applied to an MN of this mechanism should produce a
raw SAML entityID.

Whether this works:

	gss_canonicalize_name(saml_mech_oid,
		gss_import_name(GSS_C_NT_USERNAME, "joe"))

is a local problem (since usernames are local).

Nico
-- 

From simon@josefsson.org  Fri Aug 13 16:11:38 2010
Return-Path: <simon@josefsson.org>
X-Original-To: kitten@core3.amsl.com
Delivered-To: kitten@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id D40FE3A6A1E for <kitten@core3.amsl.com>; Fri, 13 Aug 2010 16:11:38 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -102.656
X-Spam-Level: 
X-Spam-Status: No, score=-102.656 tagged_above=-999 required=5 tests=[AWL=-0.057, BAYES_00=-2.599, USER_IN_WHITELIST=-100]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id 57GCKHZJvXDB for <kitten@core3.amsl.com>; Fri, 13 Aug 2010 16:11:36 -0700 (PDT)
Received: from yxa-v.extundo.com (yxa-v.extundo.com [83.241.177.39]) by core3.amsl.com (Postfix) with ESMTP id 8814F3A69D1 for <kitten@ietf.org>; Fri, 13 Aug 2010 16:11:35 -0700 (PDT)
Received: from mocca (c80-216-27-64.bredband.comhem.se [80.216.27.64]) (authenticated bits=0) by yxa-v.extundo.com (8.14.3/8.14.3/Debian-5+lenny1) with ESMTP id o7DNC6gO017127 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES128-SHA bits=128 verify=NOT); Sat, 14 Aug 2010 01:12:08 +0200
From: Simon Josefsson <simon@josefsson.org>
To: "Scott Cantor" <cantor.2@osu.edu>
Subject: Re: SAML GSS-API name type
References: <87fwyiyy3a.fsf@mocca.josefsson.org> <04ad01cb3b35$f7c046f0$e740d4d0$@osu.edu> <87sk2ixhzt.fsf@mocca.josefsson.org> <04b301cb3b3a$3d29c280$b77d4780$@osu.edu>
OpenPGP: id=B565716F; url=http://josefsson.org/key.txt
X-Hashcash: 1:22:100813:cantor.2@osu.edu::bYIrcho/PNcew+lB:4h4a
X-Hashcash: 1:22:100813:kitten@ietf.org::qbn02OvIEGNAFZWj:4/fa
Date: Sat, 14 Aug 2010 01:12:06 +0200
In-Reply-To: <04b301cb3b3a$3d29c280$b77d4780$@osu.edu> (Scott Cantor's message of "Fri, 13 Aug 2010 18:52:42 -0400")
Message-ID: <87d3tmxg3d.fsf@mocca.josefsson.org>
User-Agent: Gnus/5.110011 (No Gnus v0.11) Emacs/23.1 (gnu/linux)
MIME-Version: 1.0
Content-Type: multipart/mixed; boundary="=-=-="
X-Virus-Scanned: clamav-milter 0.96.1 at yxa-v
X-Virus-Status: Clean
Cc: kitten@ietf.org
X-BeenThere: kitten@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Common Authentication Technologies - Next Generation <kitten.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/kitten>
List-Post: <mailto:kitten@ietf.org>
List-Help: <mailto:kitten-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=subscribe>
X-List-Received-Date: Fri, 13 Aug 2010 23:11:39 -0000

--=-=-=

"Scott Cantor" <cantor.2@osu.edu> writes:

>> This suggests that there is no URI canonicalization process to speak of
>> in SAML.  Every URI is a canonical SAML entityID?
>
> Yes, basically. The actual normative language explaining that is obscure and
> a bit wrongly worded as it turns out, but the intent is that strings and
> URIs are always compared byte for byte (assuming the same encoding) unless
> something specifies differently in a profile or particular construct.

Sounds good.  Updated document below.

/Simon

--=-=-=
Content-Disposition: inline;
 filename=draft-josefsson-gssapi-saml-nametype.txt




Network Working Group                                       S. Josefsson
Internet-Draft                                                    SJD AB
Intended status: Informational                           August 14, 2010
Expires: February 15, 2011


                SAML Entity Identifier GSS-API Name Type
                draft-josefsson-gssapi-saml-nametype-00

Abstract

   This document specify a new GSS-API name type and associated syntax
   to represent SAML Entity Identifiers.

Status of this Memo

   This Internet-Draft is submitted in full conformance with the
   provisions of BCP 78 and BCP 79.

   Internet-Drafts are working documents of the Internet Engineering
   Task Force (IETF).  Note that other groups may also distribute
   working documents as Internet-Drafts.  The list of current Internet-
   Drafts is at http://datatracker.ietf.org/drafts/current/.

   Internet-Drafts are draft documents valid for a maximum of six months
   and may be updated, replaced, or obsoleted by other documents at any
   time.  It is inappropriate to use Internet-Drafts as reference
   material or to cite them other than as "work in progress."

   This Internet-Draft will expire on February 15, 2011.

Copyright Notice

   Copyright (c) 2010 IETF Trust and the persons identified as the
   document authors.  All rights reserved.

   This document is subject to BCP 78 and the IETF Trust's Legal
   Provisions Relating to IETF Documents
   (http://trustee.ietf.org/license-info) in effect on the date of
   publication of this document.  Please review these documents
   carefully, as they describe your rights and restrictions with respect
   to this document.  Code Components extracted from this document must
   include Simplified BSD License text as described in Section 4.e of
   the Trust Legal Provisions and are provided without warranty as
   described in the Simplified BSD License.






Josefsson               Expires February 15, 2011               [Page 1]

Internet-Draft  SAML Entity Identifier GSS-API Name Type     August 2010


Table of Contents

   1.  Introduction  . . . . . . . . . . . . . . . . . . . . . . . . . 3
   2.  SAML Entity Identifier Name Form  . . . . . . . . . . . . . . . 3
   3.  Acknowledgements  . . . . . . . . . . . . . . . . . . . . . . . 3
   4.  Copying conditions  . . . . . . . . . . . . . . . . . . . . . . 3
   5.  Security Considerations . . . . . . . . . . . . . . . . . . . . 3
   6.  IANA Considerations . . . . . . . . . . . . . . . . . . . . . . 3
   7.  Normative References  . . . . . . . . . . . . . . . . . . . . . 4
   Author's Address  . . . . . . . . . . . . . . . . . . . . . . . . . 4









































Josefsson               Expires February 15, 2011               [Page 2]

Internet-Draft  SAML Entity Identifier GSS-API Name Type     August 2010


1.  Introduction

   The Generic Security Service Application Program Interface (GSS-API)
   [RFC2743] specify name types for use in various situations, but none
   is applicable to Security Assertion Markup Language (SAML)
   [OASIS.saml-core-2.0-os] that uses Uniform Resource Identifier (URI)
   [RFC3986] syntax for Entity Identifier.  This document specify a new
   GSS-API name type and form to represent SAML Entity Identifier.


2.  SAML Entity Identifier Name Form

   This name form shall be represented by the Object Identifier
   1.3.6.1.4.1.11591.4.7.

   The recommended symbolic name for this type is
   "GSS_C_NT_SAML_ENTITYID".

   This name type is used to represent services associated with a SAML
   Entity Identifier.  The name form syntax is a URI.  For example:

        https://idp.example.org/SAML2

   There is no canonicalization process, every valid URI input is a
   valid mechanism name form.


3.  Acknowledgements

   The document was created as a result of discussions with Scott Cantor
   and Nicolas Williams.  TBA.


4.  Copying conditions

   This document should be considered a Code Component, and is thus
   effectively available under the BSD license.


5.  Security Considerations

   TBW.


6.  IANA Considerations

   None.




Josefsson               Expires February 15, 2011               [Page 3]

Internet-Draft  SAML Entity Identifier GSS-API Name Type     August 2010


7.  Normative References

   [RFC2743]  Linn, J., "Generic Security Service Application Program
              Interface Version 2, Update 1", RFC 2743, January 2000.

   [RFC3986]  Berners-Lee, T., Fielding, R., and L. Masinter, "Uniform
              Resource Identifier (URI): Generic Syntax", STD 66,
              RFC 3986, January 2005.

   [OASIS.saml-core-2.0-os]
              Cantor, S., Kemp, J., Philpott, R., and E. Maler,
              "Assertions and Protocol for the OASIS Security Assertion
              Markup Language (SAML) V2.0", OASIS Standard saml-core-
              2.0-os, March 2005.


Author's Address

   Simon Josefsson
   SJD AB
   Hagagatan 24
   Stockholm  113 47
   SE

   Email: simon@josefsson.org
   URI:   http://josefsson.org/

























Josefsson               Expires February 15, 2011               [Page 4]


--=-=-=--

From cantor.2@osu.edu  Fri Aug 13 16:16:52 2010
Return-Path: <cantor.2@osu.edu>
X-Original-To: kitten@core3.amsl.com
Delivered-To: kitten@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id AD86A3A67C0 for <kitten@core3.amsl.com>; Fri, 13 Aug 2010 16:16:52 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -2.615
X-Spam-Level: 
X-Spam-Status: No, score=-2.615 tagged_above=-999 required=5 tests=[AWL=-0.016, BAYES_00=-2.599]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id BMm1Qbe7aY7K for <kitten@core3.amsl.com>; Fri, 13 Aug 2010 16:16:51 -0700 (PDT)
Received: from defang2.it.ohio-state.edu (defang2.it.ohio-state.edu [128.146.216.82]) by core3.amsl.com (Postfix) with ESMTP id BF9E13A69D1 for <kitten@ietf.org>; Fri, 13 Aug 2010 16:16:51 -0700 (PDT)
Received: from defang10.it.ohio-state.edu (defang10.it.ohio-state.edu [128.146.216.79]) by defang2.it.ohio-state.edu (8.13.7/8.13.1) with ESMTP id o7DNHSPp018125; Fri, 13 Aug 2010 19:17:28 -0400
Received: from SNOWDOG ([128.146.178.34]) by defang10.it.ohio-state.edu (8.13.7/8.13.1) with ESMTP id o7DNHR4l007622; Fri, 13 Aug 2010 19:17:27 -0400
From: "Scott Cantor" <cantor.2@osu.edu>
To: "'Nicolas Williams'" <Nicolas.Williams@oracle.com>, "'Simon Josefsson'" <simon@josefsson.org>
References: <87fwyiyy3a.fsf@mocca.josefsson.org> <20100813220547.GE14409@oracle.com> <87ocd6xht2.fsf@mocca.josefsson.org> <20100813230150.GF14409@oracle.com>
In-Reply-To: <20100813230150.GF14409@oracle.com>
Subject: RE: SAML GSS-API name type
Date: Fri, 13 Aug 2010 19:17:26 -0400
Organization: The Ohio State University
Message-ID: <04bb01cb3b3d$b4952c30$1dbf8490$@osu.edu>
MIME-Version: 1.0
Content-Type: text/plain; charset="US-ASCII"
Content-Transfer-Encoding: 7bit
X-Mailer: Microsoft Outlook 14.0
Thread-index: AQIYQJ6U2n8rcwBxL/cPUKI4oSIo1QJmYzg4AfPVN70Cwol3DpINEbsw
Content-language: en-us
X-CanIt-Geo: ip=128.146.216.79; country=US; region=OH; city=Columbus; latitude=39.9968; longitude=-82.9882; metrocode=535; areacode=614; http://maps.google.com/maps?q=39.9968,-82.9882&z=6
X-CanItPRO-Stream: outbound
X-Scanned-By: CanIt (www . roaringpenguin . com) on 128.146.216.82
Cc: kitten@ietf.org
X-BeenThere: kitten@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Common Authentication Technologies - Next Generation <kitten.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/kitten>
List-Post: <mailto:kitten@ietf.org>
List-Help: <mailto:kitten-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=subscribe>
X-List-Received-Date: Fri, 13 Aug 2010 23:16:52 -0000

> But the GSS-API does have "a concept of user-friendly names that we need
> to care about", at least for host-based service names (OK, that's
> probably not an issue here).

The point is that there is no way in SAML to go from something that's not an
entityID to an entityID. So if you want to specify a SAML entity, you have
to know the entityID, exactly. Anything else, DNS-based or otherwise, would
require new work to standardize somewhere.

> Ideally there will be some user-friendly <entityID> syntax that will be
> accepted and canonicalized.

There is no syntax other than the URI itself. Change the case, new ID. Add a
slash, new ID. Add a default port designation, new ID.

> This must work:
> 
> 	gss_canonicalize_name(saml_mech_oid,
> 	    gss_import_name(GSS_C_NT_HOSTBASED_SERVICE,
> 	    "ldap@foo.bar.example"))
> 
> gss_display_name() applied to an MN of this mechanism should produce a
> raw SAML entityID.

I can't think of any simple way offhand to make that work, so that's
probably new work somewhere. And I don't foresee any obvious way short of
brute force mapping. EntityIDs aren't locations, and they can't be assumed
to correspond to DNS information in a host or service address. They're
logical names that are associated with service endpoints using SAML metadata
in a one to many way.

-- Scott



From simon@josefsson.org  Fri Aug 13 16:17:08 2010
Return-Path: <simon@josefsson.org>
X-Original-To: kitten@core3.amsl.com
Delivered-To: kitten@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 952D93A68FC for <kitten@core3.amsl.com>; Fri, 13 Aug 2010 16:17:08 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -102.653
X-Spam-Level: 
X-Spam-Status: No, score=-102.653 tagged_above=-999 required=5 tests=[AWL=-0.054, BAYES_00=-2.599, USER_IN_WHITELIST=-100]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id U2Go824U21Nx for <kitten@core3.amsl.com>; Fri, 13 Aug 2010 16:17:06 -0700 (PDT)
Received: from yxa-v.extundo.com (yxa-v.extundo.com [83.241.177.39]) by core3.amsl.com (Postfix) with ESMTP id 778523A67C0 for <kitten@ietf.org>; Fri, 13 Aug 2010 16:17:06 -0700 (PDT)
Received: from mocca (c80-216-27-64.bredband.comhem.se [80.216.27.64]) (authenticated bits=0) by yxa-v.extundo.com (8.14.3/8.14.3/Debian-5+lenny1) with ESMTP id o7DNHcvp017394 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES128-SHA bits=128 verify=NOT); Sat, 14 Aug 2010 01:17:40 +0200
From: Simon Josefsson <simon@josefsson.org>
To: Nicolas Williams <Nicolas.Williams@oracle.com>
Subject: Re: SAML GSS-API name type
References: <87fwyiyy3a.fsf@mocca.josefsson.org> <20100813220547.GE14409@oracle.com> <87ocd6xht2.fsf@mocca.josefsson.org> <20100813230150.GF14409@oracle.com>
OpenPGP: id=B565716F; url=http://josefsson.org/key.txt
X-Hashcash: 1:22:100813:kitten@ietf.org::AmKZe3WHhkZEzafi:1/1M
X-Hashcash: 1:22:100813:nicolas.williams@oracle.com::7NYVsBpjHXzdsv22:3oxw
Date: Sat, 14 Aug 2010 01:17:38 +0200
In-Reply-To: <20100813230150.GF14409@oracle.com> (Nicolas Williams's message of "Fri, 13 Aug 2010 18:01:50 -0500")
Message-ID: <878w4axfu5.fsf@mocca.josefsson.org>
User-Agent: Gnus/5.110011 (No Gnus v0.11) Emacs/23.1 (gnu/linux)
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
X-Virus-Scanned: clamav-milter 0.96.1 at yxa-v
X-Virus-Status: Clean
Cc: kitten@ietf.org
X-BeenThere: kitten@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Common Authentication Technologies - Next Generation <kitten.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/kitten>
List-Post: <mailto:kitten@ietf.org>
List-Help: <mailto:kitten-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=subscribe>
X-List-Received-Date: Fri, 13 Aug 2010 23:17:08 -0000

Nicolas Williams <Nicolas.Williams@oracle.com> writes:

> On Sat, Aug 14, 2010 at 12:35:05AM +0200, Simon Josefsson wrote:
>> Nicolas Williams <Nicolas.Williams@oracle.com> writes:
>> > Name of Kerberos realm, name of IdP, name of CA, ... -- these are all
>> > "issuer names", and they are very much relevant.  As is "trust path",
>> > which I'd envisioned as a name attribute -- trust paths are... sequences
>> > of issuer names.
>> 
>> If the document uses Scott's suggestion to just encode a SAML entityID,
>> does that address your concern?  I'm not sure I understand what you are
>> getting at here.
>
> Yes, it does.

Good.

>> I'm not sure SAML has a concept of user-friendly names that we need to
>> care about?  It would certainly be simpler to just let the
>> canonicalization process for this name type do nothing.  If I understand
>> correctly, if the SAML entityIDs doesn't match completely,
>> authentication will not work.
>
> But the GSS-API does have "a concept of user-friendly names that we need
> to care about", at least for host-based service names (OK, that's
> probably not an issue here).

Ah, I see.  Let's see if anything needs to be said about that.

>> >> * More details about SAML entityID canonicalization is required.
>> >
>> > Yep!
>> 
>> Right now I'm hoping it will be simple.
>
> I think so: just deal with names much as Kerberos implementations do.
>
> This should work:
>
> 	gss_canonicalize_name(saml_mech_oid,
> 	    gss_import_name(GSS_C_NO_OID, "<entityID>"))
>
> so should:
>
> 	gss_canonicalize_name(saml_mech_oid,
> 	    gss_import_name(GSS_C_NT_SAML_ENTITYID, "<entityID>"))

This should work fine now.

> Ideally there will be some user-friendly <entityID> syntax that will be
> accepted and canonicalized.

This doesn't work because I don't know what a user-friendly syntax would
be here.

> This must work:
>
> 	gss_canonicalize_name(saml_mech_oid,
> 	    gss_import_name(GSS_C_NT_HOSTBASED_SERVICE,
> 	    "ldap@foo.bar.example"))

What would gss_display_name on the output of that result in?
'ldap://foo.bar.example/'?

If we want that to work, some more text around this is necessary.  It
seems a simple string transformation is sufficient, right?

  service@hostname   =>  service://hostname/

> gss_display_name() applied to an MN of this mechanism should produce a
> raw SAML entityID.

That should also work now, I believe.

> Whether this works:
>
> 	gss_canonicalize_name(saml_mech_oid,
> 		gss_import_name(GSS_C_NT_USERNAME, "joe"))
>
> is a local problem (since usernames are local).

I'm not sure there is a use-case for making that work, so leaving it
undefined seems simplest.

Thanks,
/Simon

From simon@josefsson.org  Fri Aug 13 16:18:48 2010
Return-Path: <simon@josefsson.org>
X-Original-To: kitten@core3.amsl.com
Delivered-To: kitten@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 6E9953A68FC for <kitten@core3.amsl.com>; Fri, 13 Aug 2010 16:18:48 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -102.651
X-Spam-Level: 
X-Spam-Status: No, score=-102.651 tagged_above=-999 required=5 tests=[AWL=-0.052, BAYES_00=-2.599, USER_IN_WHITELIST=-100]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id tvstCVaDHz4C for <kitten@core3.amsl.com>; Fri, 13 Aug 2010 16:18:47 -0700 (PDT)
Received: from yxa-v.extundo.com (yxa-v.extundo.com [83.241.177.39]) by core3.amsl.com (Postfix) with ESMTP id 3DDEC3A69D1 for <kitten@ietf.org>; Fri, 13 Aug 2010 16:18:47 -0700 (PDT)
Received: from mocca (c80-216-27-64.bredband.comhem.se [80.216.27.64]) (authenticated bits=0) by yxa-v.extundo.com (8.14.3/8.14.3/Debian-5+lenny1) with ESMTP id o7DNJHEh017410 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES128-SHA bits=128 verify=NOT); Sat, 14 Aug 2010 01:19:19 +0200
From: Simon Josefsson <simon@josefsson.org>
To: "Scott Cantor" <cantor.2@osu.edu>
Subject: Re: SAML GSS-API name type
References: <87fwyiyy3a.fsf@mocca.josefsson.org> <20100813220547.GE14409@oracle.com> <87ocd6xht2.fsf@mocca.josefsson.org> <20100813230150.GF14409@oracle.com> <04bb01cb3b3d$b4952c30$1dbf8490$@osu.edu>
OpenPGP: id=B565716F; url=http://josefsson.org/key.txt
X-Hashcash: 1:22:100813:kitten@ietf.org::h8qmNWiI6WiBDFLU:3kx2
X-Hashcash: 1:22:100813:cantor.2@osu.edu::ihzLBArdjxSEoFrO:BEjW
X-Hashcash: 1:22:100813:nicolas.williams@oracle.com::/vXgmD52ASnXsC8T:bfn6
Date: Sat, 14 Aug 2010 01:19:17 +0200
In-Reply-To: <04bb01cb3b3d$b4952c30$1dbf8490$@osu.edu> (Scott Cantor's message of "Fri, 13 Aug 2010 19:17:26 -0400")
Message-ID: <874oeyxfre.fsf@mocca.josefsson.org>
User-Agent: Gnus/5.110011 (No Gnus v0.11) Emacs/23.1 (gnu/linux)
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
X-Virus-Scanned: clamav-milter 0.96.1 at yxa-v
X-Virus-Status: Clean
Cc: kitten@ietf.org
X-BeenThere: kitten@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Common Authentication Technologies - Next Generation <kitten.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/kitten>
List-Post: <mailto:kitten@ietf.org>
List-Help: <mailto:kitten-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=subscribe>
X-List-Received-Date: Fri, 13 Aug 2010 23:18:48 -0000

"Scott Cantor" <cantor.2@osu.edu> writes:

>> This must work:
>> 
>> 	gss_canonicalize_name(saml_mech_oid,
>> 	    gss_import_name(GSS_C_NT_HOSTBASED_SERVICE,
>> 	    "ldap@foo.bar.example"))
>> 
>> gss_display_name() applied to an MN of this mechanism should produce a
>> raw SAML entityID.
>
> I can't think of any simple way offhand to make that work, so that's
> probably new work somewhere. And I don't foresee any obvious way short of
> brute force mapping. EntityIDs aren't locations, and they can't be assumed
> to correspond to DNS information in a host or service address. They're
> logical names that are associated with service endpoints using SAML metadata
> in a one to many way.

Good point.  I believe there is no point in making the above work at
all.  Disregard my previous suggestions on string transformations.

/Simon

From cantor.2@osu.edu  Fri Aug 13 16:22:20 2010
Return-Path: <cantor.2@osu.edu>
X-Original-To: kitten@core3.amsl.com
Delivered-To: kitten@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id E331B3A685B for <kitten@core3.amsl.com>; Fri, 13 Aug 2010 16:22:20 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -2.614
X-Spam-Level: 
X-Spam-Status: No, score=-2.614 tagged_above=-999 required=5 tests=[AWL=-0.015, BAYES_00=-2.599]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id JDSHB+yJUMB1 for <kitten@core3.amsl.com>; Fri, 13 Aug 2010 16:22:20 -0700 (PDT)
Received: from defang1.it.ohio-state.edu (defang1.it.ohio-state.edu [128.146.216.81]) by core3.amsl.com (Postfix) with ESMTP id E4F2A3A67C0 for <kitten@ietf.org>; Fri, 13 Aug 2010 16:22:19 -0700 (PDT)
Received: from defang6.it.ohio-state.edu (defang6.it.ohio-state.edu [128.146.216.86]) by defang1.it.ohio-state.edu (8.13.7/8.13.1) with ESMTP id o7DNMu72024728; Fri, 13 Aug 2010 19:22:56 -0400
Received: from SNOWDOG ([128.146.178.34]) by defang6.it.ohio-state.edu (8.13.7/8.13.1) with ESMTP id o7DNMsRj015941; Fri, 13 Aug 2010 19:22:54 -0400
From: "Scott Cantor" <cantor.2@osu.edu>
To: "'Simon Josefsson'" <simon@josefsson.org>, "'Nicolas Williams'" <Nicolas.Williams@oracle.com>
References: <87fwyiyy3a.fsf@mocca.josefsson.org> <20100813220547.GE14409@oracle.com> <87ocd6xht2.fsf@mocca.josefsson.org> <20100813230150.GF14409@oracle.com> <878w4axfu5.fsf@mocca.josefsson.org>
In-Reply-To: <878w4axfu5.fsf@mocca.josefsson.org>
Subject: RE: SAML GSS-API name type
Date: Fri, 13 Aug 2010 19:22:58 -0400
Organization: The Ohio State University
Message-ID: <04be01cb3b3e$78142b70$683c8250$@osu.edu>
MIME-Version: 1.0
Content-Type: text/plain; charset="US-ASCII"
Content-Transfer-Encoding: 7bit
X-Mailer: Microsoft Outlook 14.0
Thread-index: AQIYQJ6U2n8rcwBxL/cPUKI4oSIo1QJmYzg4AfPVN70Cwol3DgH+s2N+kf0fBUA=
Content-language: en-us
X-CanIt-Geo: ip=128.146.216.86; country=US; region=OH; city=Columbus; latitude=39.9968; longitude=-82.9882; metrocode=535; areacode=614; http://maps.google.com/maps?q=39.9968,-82.9882&z=6
X-CanItPRO-Stream: outbound
X-Scanned-By: CanIt (www . roaringpenguin . com) on 128.146.216.81
Cc: kitten@ietf.org
X-BeenThere: kitten@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Common Authentication Technologies - Next Generation <kitten.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/kitten>
List-Post: <mailto:kitten@ietf.org>
List-Help: <mailto:kitten-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=subscribe>
X-List-Received-Date: Fri, 13 Aug 2010 23:22:21 -0000

> If we want that to work, some more text around this is necessary.  It
> seems a simple string transformation is sufficient, right?
> 
>   service@hostname   =>  service://hostname/

My previous note is a long way of saying that won't work. There is no
guarantee at all what the entityID of a service using SAML that happens to
run at service@hostname is. If there were, they would be locations, and not
names. Big difference.

In particular, many services will share an entityID, including services that
run on many different systems with different hostnames. The whole purpose of
an entityID is to abstract out of those names into something persistent and
appropriately granular (or non-granular as the case may be).

If this kind of mapping is a requirement, a lot of very thorny issues are
raised that don't have great answers.

-- Scott



From Nicolas.Williams@oracle.com  Fri Aug 13 16:25:35 2010
Return-Path: <Nicolas.Williams@oracle.com>
X-Original-To: kitten@core3.amsl.com
Delivered-To: kitten@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 6206E3A6A20 for <kitten@core3.amsl.com>; Fri, 13 Aug 2010 16:25:35 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -6.536
X-Spam-Level: 
X-Spam-Status: No, score=-6.536 tagged_above=-999 required=5 tests=[AWL=0.062,  BAYES_00=-2.599, RCVD_IN_DNSWL_MED=-4, UNPARSEABLE_RELAY=0.001]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id N+gNjTrEOdO2 for <kitten@core3.amsl.com>; Fri, 13 Aug 2010 16:25:34 -0700 (PDT)
Received: from rcsinet10.oracle.com (rcsinet10.oracle.com [148.87.113.121]) by core3.amsl.com (Postfix) with ESMTP id 4A5E93A699A for <kitten@ietf.org>; Fri, 13 Aug 2010 16:25:34 -0700 (PDT)
Received: from rcsinet15.oracle.com (rcsinet15.oracle.com [148.87.113.117]) by rcsinet10.oracle.com (Switch-3.4.2/Switch-3.4.2) with ESMTP id o7DNQ9le000693 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=OK); Fri, 13 Aug 2010 23:26:11 GMT
Received: from acsmt354.oracle.com (acsmt354.oracle.com [141.146.40.154]) by rcsinet15.oracle.com (Switch-3.4.2/Switch-3.4.1) with ESMTP id o7D8e4MD015249; Fri, 13 Aug 2010 23:26:09 GMT
Received: from abhmt004.oracle.com by acsmt353.oracle.com with ESMTP id 516007411281741857; Fri, 13 Aug 2010 16:24:17 -0700
Received: from oracle.com (/129.153.128.104) by default (Oracle Beehive Gateway v4.0) with ESMTP ; Fri, 13 Aug 2010 16:24:16 -0700
Date: Fri, 13 Aug 2010 18:24:12 -0500
From: Nicolas Williams <Nicolas.Williams@oracle.com>
To: Scott Cantor <cantor.2@osu.edu>
Subject: Re: SAML GSS-API name type
Message-ID: <20100813232411.GG14409@oracle.com>
References: <87fwyiyy3a.fsf@mocca.josefsson.org> <20100813220547.GE14409@oracle.com> <87ocd6xht2.fsf@mocca.josefsson.org> <20100813230150.GF14409@oracle.com> <04bb01cb3b3d$b4952c30$1dbf8490$@osu.edu>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <04bb01cb3b3d$b4952c30$1dbf8490$@osu.edu>
User-Agent: Mutt/1.5.20 (2010-03-02)
Cc: kitten@ietf.org, 'Simon Josefsson' <simon@josefsson.org>
X-BeenThere: kitten@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Common Authentication Technologies - Next Generation <kitten.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/kitten>
List-Post: <mailto:kitten@ietf.org>
List-Help: <mailto:kitten-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=subscribe>
X-List-Received-Date: Fri, 13 Aug 2010 23:25:35 -0000

On Fri, Aug 13, 2010 at 07:17:26PM -0400, Scott Cantor wrote:
> > But the GSS-API does have "a concept of user-friendly names that we need
> > to care about", at least for host-based service names (OK, that's
> > probably not an issue here).
> 
> The point is that there is no way in SAML to go from something that's not an
> entityID to an entityID. So if you want to specify a SAML entity, you have
> to know the entityID, exactly. Anything else, DNS-based or otherwise, would
> require new work to standardize somewhere.

OK.

> > This must work:
> > 
> > 	gss_canonicalize_name(saml_mech_oid,
> > 	    gss_import_name(GSS_C_NT_HOSTBASED_SERVICE,
> > 	    "ldap@foo.bar.example"))
> > 
> > gss_display_name() applied to an MN of this mechanism should produce a
> > raw SAML entityID.
> 
> I can't think of any simple way offhand to make that work, so that's
> probably new work somewhere. And I don't foresee any obvious way short of
> brute force mapping. EntityIDs aren't locations, and they can't be assumed
> to correspond to DNS information in a host or service address. They're
> logical names that are associated with service endpoints using SAML metadata
> in a one to many way.

I'd settle for this working:

    target = gss_import_name(GSS_C_NT_HOSTBASED_SERVICE,
	"ldap@foo.bar.example");

    gss_init_sec_context(..., target, ...);

and then have gss_inquire_context() output an MN for the acceptor for
which gss_display_name() gets you the SAML entityID.

But if even that won't work then you're kinda stuck.

When a user says "ssh somehost" and the SSHv2 negotiation uses GSS-API
for authentication, then the client will import "host@somehost" as the
target name and will pass that to gss_init_sec_context().

You could pass GSS_C_NO_NAME as the target, but then all the channel
binding in the world won't help you make sure you're not talking to an
MITM since there's nothing you can do to make sure that the GSS acceptor
you're talking to is the one that the user wanted to talk to.

What does it mean to authenticate an entity whose name is meaningless to
and unrecognizable by users??  IMO: nothing.

Nico
-- 

From Nicolas.Williams@oracle.com  Fri Aug 13 16:36:43 2010
Return-Path: <Nicolas.Williams@oracle.com>
X-Original-To: kitten@core3.amsl.com
Delivered-To: kitten@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id AEF843A6A0C for <kitten@core3.amsl.com>; Fri, 13 Aug 2010 16:36:29 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -6.537
X-Spam-Level: 
X-Spam-Status: No, score=-6.537 tagged_above=-999 required=5 tests=[AWL=0.061,  BAYES_00=-2.599, RCVD_IN_DNSWL_MED=-4, UNPARSEABLE_RELAY=0.001]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id 3thbwgSg8mGF for <kitten@core3.amsl.com>; Fri, 13 Aug 2010 16:35:46 -0700 (PDT)
Received: from rcsinet10.oracle.com (rcsinet10.oracle.com [148.87.113.121]) by core3.amsl.com (Postfix) with ESMTP id 99FE43A6A42 for <kitten@ietf.org>; Fri, 13 Aug 2010 16:35:21 -0700 (PDT)
Received: from acsinet15.oracle.com (acsinet15.oracle.com [141.146.126.227]) by rcsinet10.oracle.com (Switch-3.4.2/Switch-3.4.2) with ESMTP id o7DNZbc8009791 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=OK); Fri, 13 Aug 2010 23:35:39 GMT
Received: from acsmt353.oracle.com (acsmt353.oracle.com [141.146.40.153]) by acsinet15.oracle.com (Switch-3.4.2/Switch-3.4.1) with ESMTP id o7DLlOeP018818; Fri, 13 Aug 2010 23:35:37 GMT
Received: from abhmt018.oracle.com by acsmt353.oracle.com with ESMTP id 495392801281742469; Fri, 13 Aug 2010 16:34:29 -0700
Received: from oracle.com (/129.153.128.104) by default (Oracle Beehive Gateway v4.0) with ESMTP ; Fri, 13 Aug 2010 16:34:29 -0700
Date: Fri, 13 Aug 2010 18:34:22 -0500
From: Nicolas Williams <Nicolas.Williams@oracle.com>
To: Scott Cantor <cantor.2@osu.edu>
Subject: Re: SAML GSS-API name type
Message-ID: <20100813233422.GH14409@oracle.com>
References: <87fwyiyy3a.fsf@mocca.josefsson.org> <20100813220547.GE14409@oracle.com> <87ocd6xht2.fsf@mocca.josefsson.org> <20100813230150.GF14409@oracle.com> <878w4axfu5.fsf@mocca.josefsson.org> <04be01cb3b3e$78142b70$683c8250$@osu.edu>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <04be01cb3b3e$78142b70$683c8250$@osu.edu>
User-Agent: Mutt/1.5.20 (2010-03-02)
Cc: kitten@ietf.org, 'Simon Josefsson' <simon@josefsson.org>
X-BeenThere: kitten@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Common Authentication Technologies - Next Generation <kitten.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/kitten>
List-Post: <mailto:kitten@ietf.org>
List-Help: <mailto:kitten-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=subscribe>
X-List-Received-Date: Fri, 13 Aug 2010 23:36:43 -0000

On Fri, Aug 13, 2010 at 07:22:58PM -0400, Scott Cantor wrote:
> > If we want that to work, some more text around this is necessary.  It
> > seems a simple string transformation is sufficient, right?
> > 
> >   service@hostname   =>  service://hostname/
> 
> My previous note is a long way of saying that won't work. There is no
> guarantee at all what the entityID of a service using SAML that happens to
> run at service@hostname is. If there were, they would be locations, and not
> names. Big difference.
> 
> In particular, many services will share an entityID, including services that
> run on many different systems with different hostnames. The whole purpose of
> an entityID is to abstract out of those names into something persistent and
> appropriately granular (or non-granular as the case may be).
> 
> If this kind of mapping is a requirement, a lot of very thorny issues are
> raised that don't have great answers.

Well, as I just stated, and will repeat here: there is no point to
authenticating entities whose names are meaningless to, and
unrecognizable by humans.

Yes, we could have an option for the ssh command that let's you specify
the acceptor's name separately from its hostname, and then you'd
cut-n-paste from somewhere.  Or maybe ssh could do leap-of-faith
learning of hostname->entityID mappings just for the SAML mechanism.
But I don't think that's terribly usable, it's rather complicated
for users, and, IMO, pointless.

The same applies to other GSS-API applications.

Let's backtrack.

How do current non-GSS implementations of this SAML mechanism interface
the user to entityIDs?

There must be some answer.  If that answer is that we trust the TLS
server cert PKI to authenticate the host, that the entityID is then
learned from the server, and that therefore it doesn't really matter
what the entityID is... then the current mechanism is weak.

But there is hope yet!

Can we impose naming conventions on these entityIDs so that the results
are meaningful to users?

If not, and if no one else has a better idea, then consider giving up.

Nico
-- 

From simon@josefsson.org  Fri Aug 13 16:41:43 2010
Return-Path: <simon@josefsson.org>
X-Original-To: kitten@core3.amsl.com
Delivered-To: kitten@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 4DF113A67C0 for <kitten@core3.amsl.com>; Fri, 13 Aug 2010 16:41:43 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -102.649
X-Spam-Level: 
X-Spam-Status: No, score=-102.649 tagged_above=-999 required=5 tests=[AWL=-0.050, BAYES_00=-2.599, USER_IN_WHITELIST=-100]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id tmaSzHilSfrI for <kitten@core3.amsl.com>; Fri, 13 Aug 2010 16:41:41 -0700 (PDT)
Received: from yxa-v.extundo.com (yxa-v.extundo.com [83.241.177.39]) by core3.amsl.com (Postfix) with ESMTP id 0B6CE3A67B3 for <kitten@ietf.org>; Fri, 13 Aug 2010 16:41:38 -0700 (PDT)
Received: from mocca (c80-216-27-64.bredband.comhem.se [80.216.27.64]) (authenticated bits=0) by yxa-v.extundo.com (8.14.3/8.14.3/Debian-5+lenny1) with ESMTP id o7DNfuHY018663 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES128-SHA bits=128 verify=NOT); Sat, 14 Aug 2010 01:42:03 +0200
From: Simon Josefsson <simon@josefsson.org>
To: Nicolas Williams <Nicolas.Williams@oracle.com>
Subject: Re: SAML GSS-API name type
References: <87fwyiyy3a.fsf@mocca.josefsson.org> <20100813220547.GE14409@oracle.com> <87ocd6xht2.fsf@mocca.josefsson.org> <20100813230150.GF14409@oracle.com> <04bb01cb3b3d$b4952c30$1dbf8490$@osu.edu> <20100813232411.GG14409@oracle.com>
OpenPGP: id=B565716F; url=http://josefsson.org/key.txt
X-Hashcash: 1:22:100813:cantor.2@osu.edu::rNwIPMPyWwPSloPx:GkYe
X-Hashcash: 1:22:100813:kitten@ietf.org::sRWzFjZA8AkVIb8r:P2/6
X-Hashcash: 1:22:100813:nicolas.williams@oracle.com::5oFsSMtd30W/qNA2:FBt6
Date: Sat, 14 Aug 2010 01:41:55 +0200
In-Reply-To: <20100813232411.GG14409@oracle.com> (Nicolas Williams's message of "Fri, 13 Aug 2010 18:24:12 -0500")
Message-ID: <877hjuw058.fsf@mocca.josefsson.org>
User-Agent: Gnus/5.110011 (No Gnus v0.11) Emacs/23.1 (gnu/linux)
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
X-Virus-Scanned: clamav-milter 0.96.1 at yxa-v
X-Virus-Status: Clean
Cc: kitten@ietf.org
X-BeenThere: kitten@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Common Authentication Technologies - Next Generation <kitten.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/kitten>
List-Post: <mailto:kitten@ietf.org>
List-Help: <mailto:kitten-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=subscribe>
X-List-Received-Date: Fri, 13 Aug 2010 23:41:43 -0000

Nicolas Williams <Nicolas.Williams@oracle.com> writes:

> I'd settle for this working:
>
>     target = gss_import_name(GSS_C_NT_HOSTBASED_SERVICE,
> 	"ldap@foo.bar.example");
>
>     gss_init_sec_context(..., target, ...);
>
> and then have gss_inquire_context() output an MN for the acceptor for
> which gss_display_name() gets you the SAML entityID.
>
> But if even that won't work then you're kinda stuck.
>
> When a user says "ssh somehost" and the SSHv2 negotiation uses GSS-API
> for authentication, then the client will import "host@somehost" as the
> target name and will pass that to gss_init_sec_context().
>
> You could pass GSS_C_NO_NAME as the target, but then all the channel
> binding in the world won't help you make sure you're not talking to an
> MITM since there's nothing you can do to make sure that the GSS acceptor
> you're talking to is the one that the user wanted to talk to.
>
> What does it mean to authenticate an entity whose name is meaningless to
> and unrecognizable by users??  IMO: nothing.

Reading section 8.3 of SAML-Core [1] makes me believe we have a problem.
None of the names defined there is a normal host based name.  It has
e-mail addresses, X.509 subject names, Kerberos Principals and some
other stuff, but no host based names.

I think "ssh somehost" is a good use-case to have in mind when we think
about how the SAML GSS-API mechanism will work.

/Simon

[1] http://docs.oasis-open.org/security/saml/v2.0/saml-core-2.0-os.pdf

From cantor.2@osu.edu  Fri Aug 13 16:42:51 2010
Return-Path: <cantor.2@osu.edu>
X-Original-To: kitten@core3.amsl.com
Delivered-To: kitten@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 57A933A67C0 for <kitten@core3.amsl.com>; Fri, 13 Aug 2010 16:42:51 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -2.613
X-Spam-Level: 
X-Spam-Status: No, score=-2.613 tagged_above=-999 required=5 tests=[AWL=-0.014, BAYES_00=-2.599]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id 199Avys1UklB for <kitten@core3.amsl.com>; Fri, 13 Aug 2010 16:42:50 -0700 (PDT)
Received: from defang19.it.ohio-state.edu (defang19.it.ohio-state.edu [128.146.216.133]) by core3.amsl.com (Postfix) with ESMTP id 61ED33A67B3 for <kitten@ietf.org>; Fri, 13 Aug 2010 16:42:50 -0700 (PDT)
Received: from defang9.it.ohio-state.edu (defang9.it.ohio-state.edu [128.146.216.78]) by defang19.it.ohio-state.edu (8.13.7/8.13.1) with ESMTP id o7DNhRLo019191; Fri, 13 Aug 2010 19:43:27 -0400
Received: from SNOWDOG ([128.146.178.34]) by defang9.it.ohio-state.edu (8.13.7/8.13.1) with ESMTP id o7DNhQXc013879; Fri, 13 Aug 2010 19:43:26 -0400
From: "Scott Cantor" <cantor.2@osu.edu>
To: "'Nicolas Williams'" <Nicolas.Williams@oracle.com>
References: <87fwyiyy3a.fsf@mocca.josefsson.org> <20100813220547.GE14409@oracle.com> <87ocd6xht2.fsf@mocca.josefsson.org> <20100813230150.GF14409@oracle.com> <04bb01cb3b3d$b4952c30$1dbf8490$@osu.edu> <20100813232411.GG14409@oracle.com>
In-Reply-To: <20100813232411.GG14409@oracle.com>
Subject: RE: SAML GSS-API name type
Date: Fri, 13 Aug 2010 19:43:26 -0400
Organization: The Ohio State University
Message-ID: <04bf01cb3b41$558391b0$008ab510$@osu.edu>
MIME-Version: 1.0
Content-Type: text/plain; charset="US-ASCII"
Content-Transfer-Encoding: 7bit
X-Mailer: Microsoft Outlook 14.0
Thread-index: AQIYQJ6U2n8rcwBxL/cPUKI4oSIo1QJmYzg4AfPVN70Cwol3DgDttiZcAfAiANyR9ij4QA==
Content-language: en-us
X-CanIt-Geo: ip=128.146.216.78; country=US; region=OH; city=Columbus; latitude=39.9968; longitude=-82.9882; metrocode=535; areacode=614; http://maps.google.com/maps?q=39.9968,-82.9882&z=6
X-CanItPRO-Stream: outbound
X-Scanned-By: CanIt (www . roaringpenguin . com) on 128.146.216.133
Cc: kitten@ietf.org, 'Simon Josefsson' <simon@josefsson.org>
X-BeenThere: kitten@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Common Authentication Technologies - Next Generation <kitten.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/kitten>
List-Post: <mailto:kitten@ietf.org>
List-Help: <mailto:kitten-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=subscribe>
X-List-Received-Date: Fri, 13 Aug 2010 23:42:51 -0000

> You could pass GSS_C_NO_NAME as the target, but then all the channel
> binding in the world won't help you make sure you're not talking to an
> MITM since there's nothing you can do to make sure that the GSS acceptor
> you're talking to is the one that the user wanted to talk to.

I take your point, but if the client needs to know that name, it will
probably have to be provisioned into the request process alongside the
location-based designation of the service, or there would have to be a
mapping process defined.

It's doable, but it would probably want to be an implementation detail of
the client, not something the GSS layer did for it.

Since the end result of the CB is to know that the TLS server is the entity
that the IdP issued the assertion to, I don't think it would be impossible
to rely on distributed discovery (host-meta/XRD/etc) to just ask it for its
entityID. If it lied, the IdP would catch it since it couldn't generate a
valid SAML request.

-- Scott



From Nicolas.Williams@oracle.com  Fri Aug 13 16:50:00 2010
Return-Path: <Nicolas.Williams@oracle.com>
X-Original-To: kitten@core3.amsl.com
Delivered-To: kitten@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id D4BC83A67B3 for <kitten@core3.amsl.com>; Fri, 13 Aug 2010 16:50:00 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -6.538
X-Spam-Level: 
X-Spam-Status: No, score=-6.538 tagged_above=-999 required=5 tests=[AWL=0.060,  BAYES_00=-2.599, RCVD_IN_DNSWL_MED=-4, UNPARSEABLE_RELAY=0.001]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id IXP0WbJHg9Am for <kitten@core3.amsl.com>; Fri, 13 Aug 2010 16:49:59 -0700 (PDT)
Received: from rcsinet10.oracle.com (rcsinet10.oracle.com [148.87.113.121]) by core3.amsl.com (Postfix) with ESMTP id 998B53A67C0 for <kitten@ietf.org>; Fri, 13 Aug 2010 16:49:59 -0700 (PDT)
Received: from rcsinet13.oracle.com (rcsinet13.oracle.com [148.87.113.125]) by rcsinet10.oracle.com (Switch-3.4.2/Switch-3.4.2) with ESMTP id o7DNoYPR023958 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=OK); Fri, 13 Aug 2010 23:50:36 GMT
Received: from acsmt354.oracle.com (acsmt354.oracle.com [141.146.40.154]) by rcsinet13.oracle.com (Switch-3.4.2/Switch-3.4.1) with ESMTP id o7DCX1fR028739; Fri, 13 Aug 2010 23:50:34 GMT
Received: from abhmt009.oracle.com by acsmt354.oracle.com with ESMTP id 516042301281743313; Fri, 13 Aug 2010 16:48:33 -0700
Received: from oracle.com (/129.153.128.104) by default (Oracle Beehive Gateway v4.0) with ESMTP ; Fri, 13 Aug 2010 16:48:33 -0700
Date: Fri, 13 Aug 2010 18:48:28 -0500
From: Nicolas Williams <Nicolas.Williams@oracle.com>
To: Simon Josefsson <simon@josefsson.org>
Subject: Re: SAML GSS-API name type
Message-ID: <20100813234828.GL14409@oracle.com>
References: <87fwyiyy3a.fsf@mocca.josefsson.org> <20100813220547.GE14409@oracle.com> <87ocd6xht2.fsf@mocca.josefsson.org> <20100813230150.GF14409@oracle.com> <04bb01cb3b3d$b4952c30$1dbf8490$@osu.edu> <20100813232411.GG14409@oracle.com> <877hjuw058.fsf@mocca.josefsson.org>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <877hjuw058.fsf@mocca.josefsson.org>
User-Agent: Mutt/1.5.20 (2010-03-02)
Cc: kitten@ietf.org
X-BeenThere: kitten@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Common Authentication Technologies - Next Generation <kitten.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/kitten>
List-Post: <mailto:kitten@ietf.org>
List-Help: <mailto:kitten-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=subscribe>
X-List-Received-Date: Fri, 13 Aug 2010 23:50:01 -0000

On Sat, Aug 14, 2010 at 01:41:55AM +0200, Simon Josefsson wrote:
> Reading section 8.3 of SAML-Core [1] makes me believe we have a problem.
> None of the names defined there is a normal host based name.  It has
> e-mail addresses, X.509 subject names, Kerberos Principals and some
> other stuff, but no host based names.

If it has Kerberos principal and X.509 subject names then in a pinch we
can use those, even when Kerberos or PKIX are not used under the covers,
to represent host-based service names.  That'd be by convention.

> I think "ssh somehost" is a good use-case to have in mind when we think
> about how the SAML GSS-API mechanism will work.

Yes.

> [1] http://docs.oasis-open.org/security/saml/v2.0/saml-core-2.0-os.pdf

We might be able to impose naming conventions on Persistent Identifiers.

From Nicolas.Williams@oracle.com  Fri Aug 13 16:53:22 2010
Return-Path: <Nicolas.Williams@oracle.com>
X-Original-To: kitten@core3.amsl.com
Delivered-To: kitten@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id EC5A83A67B3 for <kitten@core3.amsl.com>; Fri, 13 Aug 2010 16:53:22 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -6.538
X-Spam-Level: 
X-Spam-Status: No, score=-6.538 tagged_above=-999 required=5 tests=[AWL=0.060,  BAYES_00=-2.599, RCVD_IN_DNSWL_MED=-4, UNPARSEABLE_RELAY=0.001]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id uX-YvlnMpcr8 for <kitten@core3.amsl.com>; Fri, 13 Aug 2010 16:53:22 -0700 (PDT)
Received: from rcsinet10.oracle.com (rcsinet10.oracle.com [148.87.113.121]) by core3.amsl.com (Postfix) with ESMTP id 0E9D93A6876 for <kitten@ietf.org>; Fri, 13 Aug 2010 16:53:22 -0700 (PDT)
Received: from acsinet15.oracle.com (acsinet15.oracle.com [141.146.126.227]) by rcsinet10.oracle.com (Switch-3.4.2/Switch-3.4.2) with ESMTP id o7DNruk4027896 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=OK); Fri, 13 Aug 2010 23:53:58 GMT
Received: from acsmt355.oracle.com (acsmt355.oracle.com [141.146.40.155]) by acsinet15.oracle.com (Switch-3.4.2/Switch-3.4.1) with ESMTP id o7DD4513028768; Fri, 13 Aug 2010 23:53:56 GMT
Received: from abhmt010.oracle.com by acsmt353.oracle.com with ESMTP id 495422611281743633; Fri, 13 Aug 2010 16:53:53 -0700
Received: from oracle.com (/129.153.128.104) by default (Oracle Beehive Gateway v4.0) with ESMTP ; Fri, 13 Aug 2010 16:53:53 -0700
Date: Fri, 13 Aug 2010 18:53:48 -0500
From: Nicolas Williams <Nicolas.Williams@oracle.com>
To: Scott Cantor <cantor.2@osu.edu>
Subject: Re: SAML GSS-API name type
Message-ID: <20100813235348.GM14409@oracle.com>
References: <87fwyiyy3a.fsf@mocca.josefsson.org> <20100813220547.GE14409@oracle.com> <87ocd6xht2.fsf@mocca.josefsson.org> <20100813230150.GF14409@oracle.com> <04bb01cb3b3d$b4952c30$1dbf8490$@osu.edu> <20100813232411.GG14409@oracle.com> <04bf01cb3b41$558391b0$008ab510$@osu.edu>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <04bf01cb3b41$558391b0$008ab510$@osu.edu>
User-Agent: Mutt/1.5.20 (2010-03-02)
Cc: kitten@ietf.org, 'Simon Josefsson' <simon@josefsson.org>
X-BeenThere: kitten@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Common Authentication Technologies - Next Generation <kitten.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/kitten>
List-Post: <mailto:kitten@ietf.org>
List-Help: <mailto:kitten-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=subscribe>
X-List-Received-Date: Fri, 13 Aug 2010 23:53:23 -0000

On Fri, Aug 13, 2010 at 07:43:26PM -0400, Scott Cantor wrote:
> > You could pass GSS_C_NO_NAME as the target, but then all the channel
> > binding in the world won't help you make sure you're not talking to an
> > MITM since there's nothing you can do to make sure that the GSS acceptor
> > you're talking to is the one that the user wanted to talk to.
> 
> I take your point, but if the client needs to know that name, it will
> probably have to be provisioned into the request process alongside the
> location-based designation of the service, or there would have to be a
> mapping process defined.
> 
> It's doable, but it would probably want to be an implementation detail of
> the client, not something the GSS layer did for it.

Any canonicalization mechanism has to be implemented in the mechanism
itself, not in the application.

In other words: ssh should require no special knowledge of SAML service
name canonicalization in order to use this mechanism.

> Since the end result of the CB is to know that the TLS server is the entity
> that the IdP issued the assertion to, I don't think it would be impossible
> to rely on distributed discovery (host-meta/XRD/etc) to just ask it for its
> entityID. If it lied, the IdP would catch it since it couldn't generate a
> valid SAML request.

The RP's entityID clearly must be known to or resolvable by the IdP,
since otherwise the IdP won't be able to find a key in which to encrypt
the assertion.

The IdP can then be expected to resolve host-based service names, or at
least hostnames, to RP entityIDs.  And if you can impose suitable naming
conventions, then you should be able to avoid having to add a name
resolution service.  OTOH, imposing a naming convention would mean
having to rename or re-enrol existing services in order to make them
useable through this mechanism.

I strongly recommend exploring the naming convention approach.

Nico
-- 

From cantor.2@osu.edu  Fri Aug 13 17:00:18 2010
Return-Path: <cantor.2@osu.edu>
X-Original-To: kitten@core3.amsl.com
Delivered-To: kitten@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 49D123A683C for <kitten@core3.amsl.com>; Fri, 13 Aug 2010 17:00:18 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -2.612
X-Spam-Level: 
X-Spam-Status: No, score=-2.612 tagged_above=-999 required=5 tests=[AWL=-0.013, BAYES_00=-2.599]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id pwEOJUP5Erw5 for <kitten@core3.amsl.com>; Fri, 13 Aug 2010 17:00:17 -0700 (PDT)
Received: from defang19.it.ohio-state.edu (defang19.it.ohio-state.edu [128.146.216.133]) by core3.amsl.com (Postfix) with ESMTP id 5A1AD3A682F for <kitten@ietf.org>; Fri, 13 Aug 2010 17:00:15 -0700 (PDT)
Received: from defang6.it.ohio-state.edu (defang6.it.ohio-state.edu [128.146.216.86]) by defang19.it.ohio-state.edu (8.13.7/8.13.1) with ESMTP id o7E00qNJ025458; Fri, 13 Aug 2010 20:00:52 -0400
Received: from SNOWDOG ([128.146.178.34]) by defang6.it.ohio-state.edu (8.13.7/8.13.1) with ESMTP id o7E00oJd028856; Fri, 13 Aug 2010 20:00:51 -0400
From: "Scott Cantor" <cantor.2@osu.edu>
To: "'Nicolas Williams'" <Nicolas.Williams@oracle.com>
References: <87fwyiyy3a.fsf@mocca.josefsson.org> <20100813220547.GE14409@oracle.com> <87ocd6xht2.fsf@mocca.josefsson.org> <20100813230150.GF14409@oracle.com> <878w4axfu5.fsf@mocca.josefsson.org> <04be01cb3b3e$78142b70$683c8250$@osu.edu> <20100813233422.GH14409@oracle.com>
In-Reply-To: <20100813233422.GH14409@oracle.com>
Subject: RE: SAML GSS-API name type
Date: Fri, 13 Aug 2010 20:00:54 -0400
Organization: The Ohio State University
Message-ID: <04c201cb3b43$c46b11f0$4d4135d0$@osu.edu>
MIME-Version: 1.0
Content-Type: text/plain; charset="US-ASCII"
Content-Transfer-Encoding: 7bit
X-Mailer: Microsoft Outlook 14.0
Thread-index: AQIYQJ6U2n8rcwBxL/cPUKI4oSIo1QJmYzg4AfPVN70Cwol3DgH+s2N+AbJndrYCCJg5Q5HfTiAw
Content-language: en-us
X-CanIt-Geo: ip=128.146.216.86; country=US; region=OH; city=Columbus; latitude=39.9968; longitude=-82.9882; metrocode=535; areacode=614; http://maps.google.com/maps?q=39.9968,-82.9882&z=6
X-CanItPRO-Stream: outbound
X-Scanned-By: CanIt (www . roaringpenguin . com) on 128.146.216.133
Cc: kitten@ietf.org, 'Simon Josefsson' <simon@josefsson.org>
X-BeenThere: kitten@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Common Authentication Technologies - Next Generation <kitten.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/kitten>
List-Post: <mailto:kitten@ietf.org>
List-Help: <mailto:kitten-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=subscribe>
X-List-Received-Date: Sat, 14 Aug 2010 00:00:18 -0000

> Yes, we could have an option for the ssh command that let's you specify
> the acceptor's name separately from its hostname, and then you'd
> cut-n-paste from somewhere.  Or maybe ssh could do leap-of-faith
> learning of hostname->entityID mappings just for the SAML mechanism.
> But I don't think that's terribly usable, it's rather complicated
> for users, and, IMO, pointless.

I'm not sure I have a better suggestion for you. I would claim that
experience suggests that trying to map between hostnames to credentialed
identities doesn't actually have a very successful track record outside of
perhaps Kerberos. I would hope nobody would claim DN-based server matching
is anything but a total fiasco.

> How do current non-GSS implementations of this SAML mechanism interface
> the user to entityIDs?

There are none, so it hasn't come up. The proposals on the table as is do
not perform CB, and they assume authentication of the server is basically
out of scope. That's what we were trying to add today.

> There must be some answer.  If that answer is that we trust the TLS
> server cert PKI to authenticate the host, that the entityID is then
> learned from the server, and that therefore it doesn't really matter
> what the entityID is... then the current mechanism is weak.

The entityID is never even seen (or at least noted) by the client in
traditional practice, since the client is normally a browser.

So yes, it's all TLS today.

> Can we impose naming conventions on these entityIDs so that the results
> are meaningful to users?

They won't be very meaningful no matter what we do. That's the lesson of
OpenID...taking concepts that aren't URLs and making them URLs doesn't make
them understandable to users.

-- Scott



From cantor.2@osu.edu  Fri Aug 13 17:10:17 2010
Return-Path: <cantor.2@osu.edu>
X-Original-To: kitten@core3.amsl.com
Delivered-To: kitten@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 0E0633A6969 for <kitten@core3.amsl.com>; Fri, 13 Aug 2010 17:10:17 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -2.612
X-Spam-Level: 
X-Spam-Status: No, score=-2.612 tagged_above=-999 required=5 tests=[AWL=-0.013, BAYES_00=-2.599]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id oXVNis5G17Gl for <kitten@core3.amsl.com>; Fri, 13 Aug 2010 17:10:16 -0700 (PDT)
Received: from defang1.it.ohio-state.edu (defang1.it.ohio-state.edu [128.146.216.81]) by core3.amsl.com (Postfix) with ESMTP id 5BF7D3A6925 for <kitten@ietf.org>; Fri, 13 Aug 2010 17:10:14 -0700 (PDT)
Received: from defang10.it.ohio-state.edu (defang10.it.ohio-state.edu [128.146.216.79]) by defang1.it.ohio-state.edu (8.13.7/8.13.1) with ESMTP id o7E0AoXw009881; Fri, 13 Aug 2010 20:10:51 -0400
Received: from SNOWDOG ([128.146.178.34]) by defang10.it.ohio-state.edu (8.13.7/8.13.1) with ESMTP id o7E0Ao5g014910; Fri, 13 Aug 2010 20:10:50 -0400
From: "Scott Cantor" <cantor.2@osu.edu>
To: "'Nicolas Williams'" <Nicolas.Williams@oracle.com>
References: <87fwyiyy3a.fsf@mocca.josefsson.org> <20100813220547.GE14409@oracle.com> <87ocd6xht2.fsf@mocca.josefsson.org> <20100813230150.GF14409@oracle.com> <04bb01cb3b3d$b4952c30$1dbf8490$@osu.edu> <20100813232411.GG14409@oracle.com> <04bf01cb3b41$558391b0$008ab510$@osu.edu> <20100813235348.GM14409@oracle.com>
In-Reply-To: <20100813235348.GM14409@oracle.com>
Subject: RE: SAML GSS-API name type
Date: Fri, 13 Aug 2010 20:10:53 -0400
Organization: The Ohio State University
Message-ID: <04c301cb3b45$297d97b0$7c78c710$@osu.edu>
MIME-Version: 1.0
Content-Type: text/plain; charset="US-ASCII"
Content-Transfer-Encoding: 7bit
X-Mailer: Microsoft Outlook 14.0
Thread-index: AQIYQJ6U2n8rcwBxL/cPUKI4oSIo1QJmYzg4AfPVN70Cwol3DgDttiZcAfAiANwCxyxSjAGUyFmPkdNR3TA=
Content-language: en-us
X-CanIt-Geo: ip=128.146.216.79; country=US; region=OH; city=Columbus; latitude=39.9968; longitude=-82.9882; metrocode=535; areacode=614; http://maps.google.com/maps?q=39.9968,-82.9882&z=6
X-CanItPRO-Stream: outbound
X-Scanned-By: CanIt (www . roaringpenguin . com) on 128.146.216.81
Cc: kitten@ietf.org, 'Simon Josefsson' <simon@josefsson.org>
X-BeenThere: kitten@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Common Authentication Technologies - Next Generation <kitten.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/kitten>
List-Post: <mailto:kitten@ietf.org>
List-Help: <mailto:kitten-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=subscribe>
X-List-Received-Date: Sat, 14 Aug 2010 00:10:17 -0000

> The RP's entityID clearly must be known to or resolvable by the IdP,
> since otherwise the IdP won't be able to find a key in which to encrypt
> the assertion.

Yes, it is.

> The IdP can then be expected to resolve host-based service names, or at
> least hostnames, to RP entityIDs. And if you can impose suitable naming
> conventions, then you should be able to avoid having to add a name
> resolution service.

Metadata is already used for this purpose in a web context. I think what has
to happen is to define some additional conventions by which the client
supplies the "service name" that it's trying to access to the IdP in a way
that's normally used to specify web locations. The IdP could then use that
to verify that the service name and the entityID lineup.

There's actually some text in my draft that mentions the field in question,
but I chose not to think too hard about it in the first draft I did.

This aspect of the problem doesn't come up with the "use a browser"
mechanism, since the IdP is only dealing with Web SSO. But there's also no
binding of any kind to the service name on the SASL side, and of course no
CB in that mechanism anyway right now.

> I strongly recommend exploring the naming convention approach.

It isn't impossible, I need to think about it.

Some of this came up in the Moonshot conversation, and the issues of service
vs. entity granularity are very sticky.

-- Scott



From Nicolas.Williams@oracle.com  Fri Aug 13 17:13:06 2010
Return-Path: <Nicolas.Williams@oracle.com>
X-Original-To: kitten@core3.amsl.com
Delivered-To: kitten@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 1A16D3A6987 for <kitten@core3.amsl.com>; Fri, 13 Aug 2010 17:13:06 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -6.539
X-Spam-Level: 
X-Spam-Status: No, score=-6.539 tagged_above=-999 required=5 tests=[AWL=0.059,  BAYES_00=-2.599, RCVD_IN_DNSWL_MED=-4, UNPARSEABLE_RELAY=0.001]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id LsYtnYetM6rz for <kitten@core3.amsl.com>; Fri, 13 Aug 2010 17:13:04 -0700 (PDT)
Received: from rcsinet10.oracle.com (rcsinet10.oracle.com [148.87.113.121]) by core3.amsl.com (Postfix) with ESMTP id 967DB3A6899 for <kitten@ietf.org>; Fri, 13 Aug 2010 17:13:03 -0700 (PDT)
Received: from acsinet15.oracle.com (acsinet15.oracle.com [141.146.126.227]) by rcsinet10.oracle.com (Switch-3.4.2/Switch-3.4.2) with ESMTP id o7E0DdJd018668 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=OK); Sat, 14 Aug 2010 00:13:40 GMT
Received: from acsmt355.oracle.com (acsmt355.oracle.com [141.146.40.155]) by acsinet15.oracle.com (Switch-3.4.2/Switch-3.4.1) with ESMTP id o7DJFvxO016722; Sat, 14 Aug 2010 00:13:39 GMT
Received: from abhmt009.oracle.com by acsmt355.oracle.com with ESMTP id 495450151281744779; Fri, 13 Aug 2010 17:12:59 -0700
Received: from oracle.com (/129.153.128.104) by default (Oracle Beehive Gateway v4.0) with ESMTP ; Fri, 13 Aug 2010 17:12:59 -0700
Date: Fri, 13 Aug 2010 19:12:54 -0500
From: Nicolas Williams <Nicolas.Williams@oracle.com>
To: Scott Cantor <cantor.2@osu.edu>
Subject: Re: SAML GSS-API name type
Message-ID: <20100814001254.GO14409@oracle.com>
References: <87fwyiyy3a.fsf@mocca.josefsson.org> <20100813220547.GE14409@oracle.com> <87ocd6xht2.fsf@mocca.josefsson.org> <20100813230150.GF14409@oracle.com> <878w4axfu5.fsf@mocca.josefsson.org> <04be01cb3b3e$78142b70$683c8250$@osu.edu> <20100813233422.GH14409@oracle.com> <04c201cb3b43$c46b11f0$4d4135d0$@osu.edu>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <04c201cb3b43$c46b11f0$4d4135d0$@osu.edu>
User-Agent: Mutt/1.5.20 (2010-03-02)
Cc: kitten@ietf.org, 'Simon Josefsson' <simon@josefsson.org>
X-BeenThere: kitten@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Common Authentication Technologies - Next Generation <kitten.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/kitten>
List-Post: <mailto:kitten@ietf.org>
List-Help: <mailto:kitten-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=subscribe>
X-List-Received-Date: Sat, 14 Aug 2010 00:13:06 -0000

On Fri, Aug 13, 2010 at 08:00:54PM -0400, Scott Cantor wrote:
> > Yes, we could have an option for the ssh command that let's you specify
> > the acceptor's name separately from its hostname, and then you'd
> > cut-n-paste from somewhere.  Or maybe ssh could do leap-of-faith
> > learning of hostname->entityID mappings just for the SAML mechanism.
> > But I don't think that's terribly usable, it's rather complicated
> > for users, and, IMO, pointless.
> 
> I'm not sure I have a better suggestion for you. I would claim that
> experience suggests that trying to map between hostnames to credentialed
> identities doesn't actually have a very successful track record outside of
> perhaps Kerberos. I would hope nobody would claim DN-based server matching
> is anything but a total fiasco.

Kerberos worked because Kerberos principal names were sufficiently
flexible to begin with.

PKIX didn't because TBSCertificate uses Name for the subject and Name
uses X.500 naming, with no good naming convention for hosts at the time
that that was needed.  If PKIX has used GeneralName for the subject name
from the get-go, then we'd be using dNSName and everyone would be happy.

> > How do current non-GSS implementations of this SAML mechanism interface
> > the user to entityIDs?
> 
> There are none, so it hasn't come up. The proposals on the table as is do
> not perform CB, and they assume authentication of the server is basically
> out of scope. That's what we were trying to add today.

Well, then that will have to change or you will have to give up on this.

> The entityID is never even seen (or at least noted) by the client in
> traditional practice, since the client is normally a browser.
> 
> So yes, it's all TLS today.

:(

> > Can we impose naming conventions on these entityIDs so that the results
> > are meaningful to users?
> 
> They won't be very meaningful no matter what we do.

Why not?  The mechanism would reject name forms that don't match the
convention.  And hopefully you could make sure that the naming
convention doesn't result in collisions with entityIDs currently in use.

>                                                     That's the lesson of
> OpenID...taking concepts that aren't URLs and making them URLs doesn't make
> them understandable to users.

I agree about OpenID, but I believe that you _can_ have URL naming
conventions that users can understand.  For example, users understand
mailto URLs.  And they understand the hostname part of HTTP URLs (well,
ignore confusion relating to port numbers and username/password parts).

The parts of such a naming convention that would be superfluous to users
need not be any more confusing than, say, the realm name part of
Kerberos principal names (or the use of '/' to separate principal name
components).

Nico
-- 

From cantor.2@osu.edu  Fri Aug 13 17:36:08 2010
Return-Path: <cantor.2@osu.edu>
X-Original-To: kitten@core3.amsl.com
Delivered-To: kitten@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id D97EF3A6A4A for <kitten@core3.amsl.com>; Fri, 13 Aug 2010 17:36:08 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -2.611
X-Spam-Level: 
X-Spam-Status: No, score=-2.611 tagged_above=-999 required=5 tests=[AWL=-0.012, BAYES_00=-2.599]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id itPApfP1hA9M for <kitten@core3.amsl.com>; Fri, 13 Aug 2010 17:36:08 -0700 (PDT)
Received: from defang2.it.ohio-state.edu (defang2.it.ohio-state.edu [128.146.216.82]) by core3.amsl.com (Postfix) with ESMTP id E0A813A6A4D for <kitten@ietf.org>; Fri, 13 Aug 2010 17:36:07 -0700 (PDT)
Received: from defang6.it.ohio-state.edu (defang6.it.ohio-state.edu [128.146.216.86]) by defang2.it.ohio-state.edu (8.13.7/8.13.1) with ESMTP id o7E0aino014759; Fri, 13 Aug 2010 20:36:44 -0400
Received: from SNOWDOG ([128.146.178.34]) by defang6.it.ohio-state.edu (8.13.7/8.13.1) with ESMTP id o7E0agIm019182; Fri, 13 Aug 2010 20:36:42 -0400
From: "Scott Cantor" <cantor.2@osu.edu>
To: "'Nicolas Williams'" <Nicolas.Williams@oracle.com>
References: <87fwyiyy3a.fsf@mocca.josefsson.org> <20100813220547.GE14409@oracle.com> <87ocd6xht2.fsf@mocca.josefsson.org> <20100813230150.GF14409@oracle.com> <878w4axfu5.fsf@mocca.josefsson.org> <04be01cb3b3e$78142b70$683c8250$@osu.edu> <20100813233422.GH14409@oracle.com> <04c201cb3b43$c46b11f0$4d4135d0$@osu.edu> <20100814001254.GO14409@oracle.com>
In-Reply-To: <20100814001254.GO14409@oracle.com>
Subject: RE: SAML GSS-API name type
Date: Fri, 13 Aug 2010 20:36:42 -0400
Organization: The Ohio State University
Message-ID: <04ca01cb3b48$c74a98f0$55dfcad0$@osu.edu>
MIME-Version: 1.0
Content-Type: text/plain; charset="US-ASCII"
Content-Transfer-Encoding: 7bit
X-Mailer: Microsoft Outlook 14.0
Thread-index: AQIYQJ6U2n8rcwBxL/cPUKI4oSIo1QJmYzg4AfPVN70Cwol3DgH+s2N+AbJndrYCCJg5QwFYvfvqAgTp2lCRxGtVQA==
Content-language: en-us
X-CanIt-Geo: ip=128.146.216.86; country=US; region=OH; city=Columbus; latitude=39.9968; longitude=-82.9882; metrocode=535; areacode=614; http://maps.google.com/maps?q=39.9968,-82.9882&z=6
X-CanItPRO-Stream: outbound
X-Scanned-By: CanIt (www . roaringpenguin . com) on 128.146.216.82
Cc: kitten@ietf.org, 'Simon Josefsson' <simon@josefsson.org>
X-BeenThere: kitten@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Common Authentication Technologies - Next Generation <kitten.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/kitten>
List-Post: <mailto:kitten@ietf.org>
List-Help: <mailto:kitten-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=subscribe>
X-List-Received-Date: Sat, 14 Aug 2010 00:36:09 -0000

> Why not?  The mechanism would reject name forms that don't match the
> convention.  And hopefully you could make sure that the naming
> convention doesn't result in collisions with entityIDs currently in use.

We could, but the proliferation of entityIDs (and therefore the metadata to
go with them) would be rather large. The point of entityIDs was to *not*
assume a correspondence with service endpoints, not to impose it.

I think there's a way out of this by following the line of thought I had
earlier. I think it's a mistake in general to try and stick an entityID in
front of a user. Stick with service names.

The client wants to talk to server.example.org using ssh. The server returns
a signed AuthnRequest with its service name included in the material that
the client can pull off and read. It obviously should match what the client
thinks it is.
 
The IdP should be able to obtain that from the AuthnRequest and verify, much
as it does today with web locations, that the entity asking for the
assertion is in control of that service name. It also returns the service
name it's seeing in the headers it returns to the client.

The client can cross-check and verify they match and it knows that IdP
believed it was legit, since it returned the assertion, and you have the CB,
so you can bubble that service name up as the legitimate name.

The entityID never comes into play for the client, which is how it's
designed to work today.

None of this is very "off" from the existing uses of the fields in question
in SAML. It's new work, but the CB already make it new work, and it's not
"different work" from what is implemented today.

I think the only trick is that I'd probably want to turn the service names
into URLs for transport in the messages, but *that* can be a simple
transform.

-- Scott



From Nicolas.Williams@oracle.com  Fri Aug 13 18:43:39 2010
Return-Path: <Nicolas.Williams@oracle.com>
X-Original-To: kitten@core3.amsl.com
Delivered-To: kitten@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 2DB713A6A62 for <kitten@core3.amsl.com>; Fri, 13 Aug 2010 18:43:39 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -6.54
X-Spam-Level: 
X-Spam-Status: No, score=-6.54 tagged_above=-999 required=5 tests=[AWL=0.058,  BAYES_00=-2.599, RCVD_IN_DNSWL_MED=-4, UNPARSEABLE_RELAY=0.001]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id heekUD2eKIIg for <kitten@core3.amsl.com>; Fri, 13 Aug 2010 18:43:38 -0700 (PDT)
Received: from rcsinet10.oracle.com (rcsinet10.oracle.com [148.87.113.121]) by core3.amsl.com (Postfix) with ESMTP id 4AEBD3A6A4F for <kitten@ietf.org>; Fri, 13 Aug 2010 18:43:38 -0700 (PDT)
Received: from rcsinet13.oracle.com (rcsinet13.oracle.com [148.87.113.125]) by rcsinet10.oracle.com (Switch-3.4.2/Switch-3.4.2) with ESMTP id o7E1iD3A014638 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=OK); Sat, 14 Aug 2010 01:44:14 GMT
Received: from acsmt355.oracle.com (acsmt355.oracle.com [141.146.40.155]) by rcsinet13.oracle.com (Switch-3.4.2/Switch-3.4.1) with ESMTP id o7DJvrvC012129; Sat, 14 Aug 2010 01:44:13 GMT
Received: from abhmt004.oracle.com by acsmt354.oracle.com with ESMTP id 516202361281750189; Fri, 13 Aug 2010 18:43:09 -0700
Received: from oracle.com (/129.153.128.104) by default (Oracle Beehive Gateway v4.0) with ESMTP ; Fri, 13 Aug 2010 18:43:09 -0700
Date: Fri, 13 Aug 2010 20:43:04 -0500
From: Nicolas Williams <Nicolas.Williams@oracle.com>
To: Scott Cantor <cantor.2@osu.edu>
Subject: Re: SAML GSS-API name type
Message-ID: <20100814014304.GR14409@oracle.com>
References: <87fwyiyy3a.fsf@mocca.josefsson.org> <20100813220547.GE14409@oracle.com> <87ocd6xht2.fsf@mocca.josefsson.org> <20100813230150.GF14409@oracle.com> <878w4axfu5.fsf@mocca.josefsson.org> <04be01cb3b3e$78142b70$683c8250$@osu.edu> <20100813233422.GH14409@oracle.com> <04c201cb3b43$c46b11f0$4d4135d0$@osu.edu> <20100814001254.GO14409@oracle.com> <04ca01cb3b48$c74a98f0$55dfcad0$@osu.edu>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <04ca01cb3b48$c74a98f0$55dfcad0$@osu.edu>
User-Agent: Mutt/1.5.20 (2010-03-02)
Cc: kitten@ietf.org, 'Simon Josefsson' <simon@josefsson.org>
X-BeenThere: kitten@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Common Authentication Technologies - Next Generation <kitten.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/kitten>
List-Post: <mailto:kitten@ietf.org>
List-Help: <mailto:kitten-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=subscribe>
X-List-Received-Date: Sat, 14 Aug 2010 01:43:39 -0000

On Fri, Aug 13, 2010 at 08:36:42PM -0400, Scott Cantor wrote:
> I think there's a way out of this by following the line of thought I had
> earlier. I think it's a mistake in general to try and stick an entityID in
> front of a user. Stick with service names.
> 
> The client wants to talk to server.example.org using ssh. The server returns
> a signed AuthnRequest with its service name included in the material that
> the client can pull off and read. It obviously should match what the client
> thinks it is.
>  
> The IdP should be able to obtain that from the AuthnRequest and verify, much
> as it does today with web locations, that the entity asking for the
> assertion is in control of that service name. It also returns the service
> name it's seeing in the headers it returns to the client.
> 
> The client can cross-check and verify they match and it knows that IdP
> believed it was legit, since it returned the assertion, and you have the CB,
> so you can bubble that service name up as the legitimate name.
> 
> The entityID never comes into play for the client, which is how it's
> designed to work today.

Yes, this works.

> None of this is very "off" from the existing uses of the fields in question
> in SAML. It's new work, but the CB already make it new work, and it's not
> "different work" from what is implemented today.

OK.

> I think the only trick is that I'd probably want to turn the service names
> into URLs for transport in the messages, but *that* can be a simple
> transform.

:)

Have a good weekend,

Nico
-- 

From simon@josefsson.org  Sat Aug 14 02:35:46 2010
Return-Path: <simon@josefsson.org>
X-Original-To: kitten@core3.amsl.com
Delivered-To: kitten@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 853383A67CF for <kitten@core3.amsl.com>; Sat, 14 Aug 2010 02:35:46 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -102.347
X-Spam-Level: 
X-Spam-Status: No, score=-102.347 tagged_above=-999 required=5 tests=[AWL=-0.348, BAYES_00=-2.599, J_CHICKENPOX_12=0.6, USER_IN_WHITELIST=-100]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id d4ZrmxvC4ql5 for <kitten@core3.amsl.com>; Sat, 14 Aug 2010 02:35:45 -0700 (PDT)
Received: from yxa-v.extundo.com (yxa-v.extundo.com [83.241.177.39]) by core3.amsl.com (Postfix) with ESMTP id 5B08C3A6784 for <kitten@ietf.org>; Sat, 14 Aug 2010 02:35:44 -0700 (PDT)
Received: from mocca (c80-216-27-64.bredband.comhem.se [80.216.27.64]) (authenticated bits=0) by yxa-v.extundo.com (8.14.3/8.14.3/Debian-5+lenny1) with ESMTP id o7E9aDOe019536 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES128-SHA bits=128 verify=NOT); Sat, 14 Aug 2010 11:36:14 +0200
From: Simon Josefsson <simon@josefsson.org>
To: "Scott Cantor" <cantor.2@osu.edu>
Subject: Re: SAML GSS-API name type
References: <87fwyiyy3a.fsf@mocca.josefsson.org> <20100813220547.GE14409@oracle.com> <87ocd6xht2.fsf@mocca.josefsson.org> <20100813230150.GF14409@oracle.com> <878w4axfu5.fsf@mocca.josefsson.org> <04be01cb3b3e$78142b70$683c8250$@osu.edu> <20100813233422.GH14409@oracle.com> <04c201cb3b43$c46b11f0$4d4135d0$@osu.edu> <20100814001254.GO14409@oracle.com> <04ca01cb3b48$c74a98f0$55dfcad0$@osu.edu>
OpenPGP: id=B565716F; url=http://josefsson.org/key.txt
X-Hashcash: 1:22:100814:kitten@ietf.org::VHTEe4Tv223pHPPB:91S3
X-Hashcash: 1:22:100814:nicolas.williams@oracle.com::L8ac5aw52k8Jm/Cd:BB4H
X-Hashcash: 1:22:100814:cantor.2@osu.edu::5ReULMfmdrwuqMdO:SXCW
Date: Sat, 14 Aug 2010 11:36:12 +0200
In-Reply-To: <04ca01cb3b48$c74a98f0$55dfcad0$@osu.edu> (Scott Cantor's message of "Fri, 13 Aug 2010 20:36:42 -0400")
Message-ID: <8762zdv8mr.fsf@mocca.josefsson.org>
User-Agent: Gnus/5.110011 (No Gnus v0.11) Emacs/23.1 (gnu/linux)
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
X-Virus-Scanned: clamav-milter 0.96.1 at yxa-v
X-Virus-Status: Clean
Cc: kitten@ietf.org
X-BeenThere: kitten@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Common Authentication Technologies - Next Generation <kitten.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/kitten>
List-Post: <mailto:kitten@ietf.org>
List-Help: <mailto:kitten-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=subscribe>
X-List-Received-Date: Sat, 14 Aug 2010 09:35:46 -0000

"Scott Cantor" <cantor.2@osu.edu> writes:

>> Why not?  The mechanism would reject name forms that don't match the
>> convention.  And hopefully you could make sure that the naming
>> convention doesn't result in collisions with entityIDs currently in use.
>
> We could, but the proliferation of entityIDs (and therefore the metadata to
> go with them) would be rather large. The point of entityIDs was to *not*
> assume a correspondence with service endpoints, not to impose it.
>
> I think there's a way out of this by following the line of thought I had
> earlier. I think it's a mistake in general to try and stick an entityID in
> front of a user. Stick with service names.
>
> The client wants to talk to server.example.org using ssh. The server returns
> a signed AuthnRequest with its service name included in the material that
> the client can pull off and read. It obviously should match what the client
> thinks it is.
>  
> The IdP should be able to obtain that from the AuthnRequest and verify, much
> as it does today with web locations, that the entity asking for the
> assertion is in control of that service name. It also returns the service
> name it's seeing in the headers it returns to the client.
>
> The client can cross-check and verify they match and it knows that IdP
> believed it was legit, since it returned the assertion, and you have the CB,
> so you can bubble that service name up as the legitimate name.
>
> The entityID never comes into play for the client, which is how it's
> designed to work today.
>
> None of this is very "off" from the existing uses of the fields in question
> in SAML. It's new work, but the CB already make it new work, and it's not
> "different work" from what is implemented today.
>
> I think the only trick is that I'd probably want to turn the service names
> into URLs for transport in the messages, but *that* can be a simple
> transform.

This sounds more robust than what we have discussed earlier.  Thanks for
explaining.

Wrt GSS-API naming, perhaps the above mean there is no strong need for a
new name type?  GSS-API applications can continue to use
GSS_C_NT_HOSTBASED_SERVICE a'la user@hostname?  The SAML mechanism will
discover the right SAML endpoint internally.  If application for some
reason wants to learn the SAML entityId used, it could be retrievable as
a name attribute.

We could also have a GSS_C_NT_SAML_ENTITYID that is the "canonical" SAML
mechanism name, and a GSS_C_NT_HOSTBASED_SERVICE is canonicalized into a
GSS_C_NT_SAML_ENTITYID by the SAML mechanism.  However, I'm not sure
when that would be useful.

/Simon

From cantor.2@osu.edu  Sat Aug 14 10:46:41 2010
Return-Path: <cantor.2@osu.edu>
X-Original-To: kitten@core3.amsl.com
Delivered-To: kitten@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id A1DD93A6834 for <kitten@core3.amsl.com>; Sat, 14 Aug 2010 10:46:41 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -2.199
X-Spam-Level: 
X-Spam-Status: No, score=-2.199 tagged_above=-999 required=5 tests=[AWL=-0.200, BAYES_00=-2.599, J_CHICKENPOX_12=0.6]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id 5GgqEsgGJRmK for <kitten@core3.amsl.com>; Sat, 14 Aug 2010 10:46:40 -0700 (PDT)
Received: from hrndva-omtalb.mail.rr.com (hrndva-omtalb.mail.rr.com [71.74.56.124]) by core3.amsl.com (Postfix) with ESMTP id BE8B53A6802 for <kitten@ietf.org>; Sat, 14 Aug 2010 10:46:39 -0700 (PDT)
X-Authority-Analysis: v=1.1 cv=mWMe3HXt1P3ELXxrmMgI9QKdSRnwjSxQbBh3ay0KAh4= c=1 sm=0 a=kj9zAlcOel0A:10 a=/Wh0N9815gtYTLoiNwER9g==:17 a=PxSzmH4GzewqpoBHlQkA:9 a=NBoQUgmZ1ZOQnUJhuyKQZGExaIgA:4 a=CjuIK1q_8ugA:10 a=/Wh0N9815gtYTLoiNwER9g==:117
X-Cloudmark-Score: 0
X-Originating-IP: 24.210.116.83
Received: from [24.210.116.83] ([24.210.116.83:27305] helo=SNOWDOG) by hrndva-oedge03.mail.rr.com (envelope-from <cantor.2@osu.edu>) (ecelerity 2.2.2.39 r()) with ESMTP id 52/8F-21482-2A6D66C4; Sat, 14 Aug 2010 17:47:15 +0000
From: "Scott Cantor" <cantor.2@osu.edu>
To: "'Simon Josefsson'" <simon@josefsson.org>
References: <87fwyiyy3a.fsf@mocca.josefsson.org> <20100813220547.GE14409@oracle.com> <87ocd6xht2.fsf@mocca.josefsson.org> <20100813230150.GF14409@oracle.com> <878w4axfu5.fsf@mocca.josefsson.org> <04be01cb3b3e$78142b70$683c8250$@osu.edu> <20100813233422.GH14409@oracle.com> <04c201cb3b43$c46b11f0$4d4135d0$@osu.edu> <20100814001254.GO14409@oracle.com> <04ca01cb3b48$c74a98f0$55dfcad0$@osu.edu> <8762zdv8mr.fsf@mocca.josefsson.org>
In-Reply-To: <8762zdv8mr.fsf@mocca.josefsson.org>
Subject: RE: SAML GSS-API name type
Date: Sat, 14 Aug 2010 13:47:12 -0400
Organization: The Ohio State University
Message-ID: <04db01cb3bd8$bda20dc0$38e62940$@osu.edu>
MIME-Version: 1.0
Content-Type: text/plain; charset="US-ASCII"
Content-Transfer-Encoding: 7bit
X-Mailer: Microsoft Outlook 14.0
Thread-index: AQIYQJ6U2n8rcwBxL/cPUKI4oSIo1QJmYzg4AfPVN70Cwol3DgH+s2N+AbJndrYCCJg5QwFYvfvqAgTp2lACKPyragLMQWZ0kZ3h+7A=
Content-language: en-us
Cc: kitten@ietf.org
X-BeenThere: kitten@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Common Authentication Technologies - Next Generation <kitten.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/kitten>
List-Post: <mailto:kitten@ietf.org>
List-Help: <mailto:kitten-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=subscribe>
X-List-Received-Date: Sat, 14 Aug 2010 17:46:41 -0000

> Wrt GSS-API naming, perhaps the above mean there is no strong need for a
> new name type?

Well, we have different issues in play. I don't think there needs to be a
new name type to represent services to clients, no. I think we still have to
look at the name type problem for representing users to services.

> GSS-API applications can continue to use GSS_C_NT_HOSTBASED_SERVICE a'la
> user@hostname?

I think so (though is that really how it's done? The username as part of the
service name?). Is the service type or protocol (ssh? nfs?) represented in
any way?

> The SAML mechanism will
> discover the right SAML endpoint internally.  If application for some
> reason wants to learn the SAML entityId used, it could be retrievable as
> a name attribute.

It's unlikely that the client would ever need it, and it's totally unlikely
that the service would need its own. The entityID of the IdP may well need
to be expressible that way, and it would be a name attribute in some sense,
and definitely not the "name" of the client, since it's not. But that's a
broader discussion, probably with Moonshot/abfab involved I think.

> We could also have a GSS_C_NT_SAML_ENTITYID that is the "canonical" SAML
> mechanism name, and a GSS_C_NT_HOSTBASED_SERVICE is canonicalized into a
> GSS_C_NT_SAML_ENTITYID by the SAML mechanism.  However, I'm not sure
> when that would be useful.

If conversion was needed, it would be via the same mapping process that IdPs
will have to use, namely SAML metadata. I don't anticipate any strong need
for it in clients and servers. Particularly if the point of these canonical
names is for user consumption/comparison/recognition.

-- Scott



From root@core3.amsl.com  Sat Aug 14 20:45:02 2010
Return-Path: <root@core3.amsl.com>
X-Original-To: kitten@ietf.org
Delivered-To: kitten@core3.amsl.com
Received: by core3.amsl.com (Postfix, from userid 0) id 4528E3A6818; Sat, 14 Aug 2010 20:45:01 -0700 (PDT)
From: Internet-Drafts@ietf.org
To: i-d-announce@ietf.org
Subject: I-D Action:draft-ietf-kitten-sasl-openid-00.txt 
Content-Type: Multipart/Mixed; Boundary="NextPart"
Mime-Version: 1.0
Message-Id: <20100815034502.4528E3A6818@core3.amsl.com>
Date: Sat, 14 Aug 2010 20:45:02 -0700 (PDT)
Cc: kitten@ietf.org
X-BeenThere: kitten@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Common Authentication Technologies - Next Generation <kitten.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/kitten>
List-Post: <mailto:kitten@ietf.org>
List-Help: <mailto:kitten-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=subscribe>
X-List-Received-Date: Sun, 15 Aug 2010 03:45:02 -0000

--NextPart

A New Internet-Draft is available from the on-line Internet-Drafts directories.
This draft is a work item of the Common Authentication Technology Next Generation Working Group of the IETF.


	Title           : A SASL & GSS-API Mechanism for OpenID
	Author(s)       : E. Lear, et al.
	Filename        : draft-ietf-kitten-sasl-openid-00.txt
	Pages           : 21
	Date            : 2010-08-13

OpenID has found its usage on the Internet for Web Single Sign-On.
Simple Authentication and Security Layer (SASL) and the Generic
Security Service Application Program Interface (GSS-API) are
application frameworks to generalize authentication.  This memo
specifies a SASL and GSS-API mechanism for OpenID that allows the
integration of existing OpenID Identity Providers with applications
using SASL and GSS-API.

A URL for this Internet-Draft is:
http://www.ietf.org/internet-drafts/draft-ietf-kitten-sasl-openid-00.txt

Internet-Drafts are also available by anonymous FTP at:
ftp://ftp.ietf.org/internet-drafts/

Below is the data which will enable a MIME compliant mail reader
implementation to automatically retrieve the ASCII version of the
Internet-Draft.

--NextPart
Content-Type: Message/External-body;
	name="draft-ietf-kitten-sasl-openid-00.txt";
	site="ftp.ietf.org";
	access-type="anon-ftp";
	directory="internet-drafts"

Content-Type: text/plain
Content-ID: <2010-08-14203804.I-D@ietf.org>


--NextPart--

From shawn.emery@oracle.com  Sat Aug 14 23:42:59 2010
Return-Path: <shawn.emery@oracle.com>
X-Original-To: kitten@core3.amsl.com
Delivered-To: kitten@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 271C03A69B2 for <kitten@core3.amsl.com>; Sat, 14 Aug 2010 23:42:59 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -5.317
X-Spam-Level: 
X-Spam-Status: No, score=-5.317 tagged_above=-999 required=5 tests=[AWL=-1.319, BAYES_50=0.001, HTML_MESSAGE=0.001, RCVD_IN_DNSWL_MED=-4]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id RKi1LXj1YdKg for <kitten@core3.amsl.com>; Sat, 14 Aug 2010 23:42:57 -0700 (PDT)
Received: from rcsinet10.oracle.com (rcsinet10.oracle.com [148.87.113.121]) by core3.amsl.com (Postfix) with ESMTP id CE3123A69D0 for <kitten@ietf.org>; Sat, 14 Aug 2010 23:42:57 -0700 (PDT)
Received: from rcsinet13.oracle.com (rcsinet13.oracle.com [148.87.113.125]) by rcsinet10.oracle.com (Switch-3.4.2/Switch-3.4.2) with ESMTP id o7F6hWhm022087 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=OK) for <kitten@ietf.org>; Sun, 15 Aug 2010 06:43:33 GMT
Received: from acsmt354.oracle.com (acsmt354.oracle.com [141.146.40.154]) by rcsinet13.oracle.com (Switch-3.4.2/Switch-3.4.1) with ESMTP id o7F6hWhB023298 for <kitten@ietf.org>; Sun, 15 Aug 2010 06:43:32 GMT
Received: from abhmt017.oracle.com by acsmt355.oracle.com with ESMTP id 517670511281854527; Sat, 14 Aug 2010 23:42:07 -0700
Received: from [10.7.250.156] (/10.7.250.156) by default (Oracle Beehive Gateway v4.0) with ESMTP ; Sat, 14 Aug 2010 23:42:07 -0700
Message-ID: <4C678C3C.3060207@oracle.com>
Date: Sun, 15 Aug 2010 00:42:04 -0600
From: Shawn Emery <shawn.emery@oracle.com>
User-Agent: Mozilla/5.0 (X11; U; SunOS i86pc; en-US; rv:1.9.2.4) Gecko/20100610 Lightning/1.0b2 Thunderbird/3.1
MIME-Version: 1.0
To: "kitten@ietf.org" <kitten@ietf.org>
Content-Type: multipart/alternative; boundary="------------020804050806040804030200"
Subject: [kitten] draft-ietf-kitten-gssapi-extensions-iana
X-BeenThere: kitten@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Common Authentication Technologies - Next Generation <kitten.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/kitten>
List-Post: <mailto:kitten@ietf.org>
List-Help: <mailto:kitten-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=subscribe>
X-List-Received-Date: Sun, 15 Aug 2010 06:42:59 -0000

This is a multi-part message in MIME format.
--------------020804050806040804030200
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit


Per the kitten session discussion at IETF 78, the WG needs to reach 
consensus on what name registry type IANA will create for the GSS-API 
name space.  The two types proposed were:

single GSS-API name-space registry for all languages
registry per programming language

Love had previously stated on the last consensus call that he prefers a 
registry for each language and these registries should contain mappings 
to the base specification.

The time-out of the consensus call is in two weeks (8/29).

Shawn.
--

--------------020804050806040804030200
Content-Type: text/html; charset=ISO-8859-1
Content-Transfer-Encoding: 7bit

<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN">
<html>
  <head>

    <meta http-equiv="content-type" content="text/html; charset=ISO-8859-1">
  </head>
  <body text="#000000" bgcolor="#ffffff">
    <font size="+1"><br>
      <small>Per the kitten session discussion at IETF 78, the WG needs
        to reach consensus on what name registry type IANA will create
        for the GSS-API name space.&nbsp; The two types proposed were:<br>
        <br>
      </small></font><font size="+1"><small>single GSS-API name-space
        registry for all languages<br>
        registry per programming language</small><br>
    </font><br>
    Love had previously stated on the last consensus call that he
    prefers a registry for each language and these registries should
    contain mappings to the base specification.<br>
    <br>
    The time-out of the consensus call is in two weeks (8/29).<br>
    <br>
    Shawn.<br>
    --<br>
  </body>
</html>

--------------020804050806040804030200--

From hannes.tschofenig@gmx.net  Sun Aug 15 11:09:10 2010
Return-Path: <hannes.tschofenig@gmx.net>
X-Original-To: kitten@core3.amsl.com
Delivered-To: kitten@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id A74C73A6403 for <kitten@core3.amsl.com>; Sun, 15 Aug 2010 11:09:10 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -102.599
X-Spam-Level: 
X-Spam-Status: No, score=-102.599 tagged_above=-999 required=5 tests=[AWL=0.000, BAYES_00=-2.599, USER_IN_WHITELIST=-100]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id t-y0OwD63uJ2 for <kitten@core3.amsl.com>; Sun, 15 Aug 2010 11:09:09 -0700 (PDT)
Received: from mail.gmx.net (mailout-de.gmx.net [213.165.64.23]) by core3.amsl.com (Postfix) with SMTP id 4AAAB3A682D for <kitten@ietf.org>; Sun, 15 Aug 2010 11:09:08 -0700 (PDT)
Received: (qmail invoked by alias); 15 Aug 2010 18:09:44 -0000
Received: from a88-115-222-204.elisa-laajakaista.fi (EHLO [192.168.1.5]) [88.115.222.204] by mail.gmx.net (mp059) with SMTP; 15 Aug 2010 20:09:44 +0200
X-Authenticated: #29516787
X-Provags-ID: V01U2FsdGVkX1/M3RZKZ6LMWfRHBFcJLgP1uco8+esqXiDxSpT6yJ Cy+VLKVONBFWt+
Message-ID: <4C682D66.7050707@gmx.net>
Date: Sun, 15 Aug 2010 21:09:42 +0300
From: Hannes Tschofenig <hannes.tschofenig@gmx.net>
User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.9.1.11) Gecko/20100711 Thunderbird/3.0.6
MIME-Version: 1.0
To: Nicolas Williams <Nicolas.Williams@oracle.com>
References: <tsllj8dyo5p.fsf@mit.edu>	<3D3C75174CB95F42AD6BCC56E5555B4502E9BF74@FIESEXC015.nsn-intra.net> <20100813152341.GE5217@oracle.com>
In-Reply-To: <20100813152341.GE5217@oracle.com>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit
X-Y-GMX-Trusted: 0
Cc: kitten@ietf.org, ext Sam Hartman <hartmans-ietf@mit.edu>
Subject: Re: [kitten] Concerns about oauth sasl mechanism
X-BeenThere: kitten@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Common Authentication Technologies - Next Generation <kitten.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/kitten>
List-Post: <mailto:kitten@ietf.org>
List-Help: <mailto:kitten-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=subscribe>
X-List-Received-Date: Sun, 15 Aug 2010 18:09:10 -0000

I have asked my OpenID-SASL co-authors whether I can forward the private 
mail exchange we had about the topic of channel bindings. I hope to be 
able to forward it to the list soon.

Ciao
Hannes

On 8/13/2010 6:23 PM, Nicolas Williams wrote:
> On Fri, Aug 13, 2010 at 02:59:02PM +0300, Tschofenig, Hannes (NSN - FI/Espoo) wrote:
>    
>> I hear complaints about this and that but then nobody cares about
>> posting anything to the Oauth mailing list nor to show up at the Oauth
>> meeting either. Almost nobody at the OAuth face-to-face meeting has read
>> the specs.
>>      
> We can't help you with that last part!  ;)
>
>    
>> Having said that I find it a bit strange to require channel bindings.
>>      
> Is that just because you don't know what that means?
>
>    
>> The funny thing is that I have heard that comment already at one of the
>> Oauth meetings. Then, I asked what this actually means. The answer was
>> *I don't know*. So, what should I think about these comments?
>>      
> Well, for one that those OAuth participants haven't "read the specs",
> but evidently you knew that (per the above).  If OAuth participants
> don't "read the specs" then I am not at all surprised that they don't
> know what channel binding means.
>
>    
>> I am happy to have a discussion about channel bindings in the context of
>> Oauth (and of Oauth - SASL in particular) and even about how to re-write
>> RFC 5056 in a way that it can be understood by humans. Interestingly
>> enough I tried to figure out what Channel Bindings would provide us with
>> SASL OpenID (together with my co-authors) and we weren't convinced about
>> the value!
>>      
> I too am happy to have a discussion about channel binding.  And I'll be
> happy to entertain a re-write of RFC5056 if you like.  Maybe you could
> start with some specific complaints?  But what am I to make of comments
> that on the one hand indicate you don't understand X and on the other
> state that you believe X is not needed?
>
> Yes, you can avoid channel binding in some cases.
>
> Suppose you are using TLS, with server certs, and you believe that you
> have a good enough, trustworthy PKI for the server certs, then you don't
> need channel binding in the RFC5056 sense.  In that case you'll happen
> to be doing a form of channel binding by using the same (or similar)
> service/host ID in TLS and OAuth.  Some of us consider this to be weak
> security, and too restrictive; it might be good enough for some users.
>
> Nico
>    


From hannes.tschofenig@gmx.net  Sun Aug 15 11:25:52 2010
Return-Path: <hannes.tschofenig@gmx.net>
X-Original-To: kitten@core3.amsl.com
Delivered-To: kitten@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 91F363A6403 for <kitten@core3.amsl.com>; Sun, 15 Aug 2010 11:25:52 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -102.599
X-Spam-Level: 
X-Spam-Status: No, score=-102.599 tagged_above=-999 required=5 tests=[AWL=0.000, BAYES_00=-2.599, USER_IN_WHITELIST=-100]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id lirdvy45TAiQ for <kitten@core3.amsl.com>; Sun, 15 Aug 2010 11:25:50 -0700 (PDT)
Received: from mail.gmx.net (mailout-de.gmx.net [213.165.64.22]) by core3.amsl.com (Postfix) with SMTP id 265393A6838 for <kitten@ietf.org>; Sun, 15 Aug 2010 11:25:49 -0700 (PDT)
Received: (qmail invoked by alias); 15 Aug 2010 18:26:25 -0000
Received: from a88-115-222-204.elisa-laajakaista.fi (EHLO [192.168.1.5]) [88.115.222.204] by mail.gmx.net (mp018) with SMTP; 15 Aug 2010 20:26:25 +0200
X-Authenticated: #29516787
X-Provags-ID: V01U2FsdGVkX19zFZ5QJIJzxVgg3SkS8WSOK2L6VaHaw6BDYbz0zW MAS2BoTtjWmkX9
Message-ID: <4C683150.6000003@gmx.net>
Date: Sun, 15 Aug 2010 21:26:25 +0300
From: Hannes Tschofenig <hannes.tschofenig@gmx.net>
User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.9.1.11) Gecko/20100711 Thunderbird/3.0.6
MIME-Version: 1.0
To: Nicolas Williams <Nicolas.Williams@oracle.com>
References: <tsllj8dyo5p.fsf@mit.edu> <4C6460D6.7020800@stpeter.im>	<tsl1va3v7z5.fsf@mit.edu> <4C64698A.9030405@isode.com>	<3D3C75174CB95F42AD6BCC56E5555B4502E9BF85@FIESEXC015.nsn-intra.net> <20100813152954.GF5217@oracle.com>
In-Reply-To: <20100813152954.GF5217@oracle.com>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit
X-Y-GMX-Trusted: 0
Cc: kitten@ietf.org, Sam Hartman <hartmans-ietf@mit.edu>
Subject: Re: [kitten] Concerns about oauth sasl mechanism
X-BeenThere: kitten@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Common Authentication Technologies - Next Generation <kitten.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/kitten>
List-Post: <mailto:kitten@ietf.org>
List-Help: <mailto:kitten-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=subscribe>
X-List-Received-Date: Sun, 15 Aug 2010 18:25:52 -0000

Hi Nico,

On 8/13/2010 6:29 PM, Nicolas Williams wrote:
> On Fri, Aug 13, 2010 at 03:06:14PM +0300, Tschofenig, Hannes (NSN - FI/Espoo) wrote:
>    
>> Are we standardizing research results or standardizing what industry
>> does? Please have a look at http://tools.ietf.org/search/rfc5218
>>
>> I am happy to discuss the security properties of various aspects but
>> let's pick it one by one and also consider deployability.
>>
>> * Mutual authentication: We are talking about a three party protocol
>> here and so what does mutual authentication actually mean.
>>      
> "Mutual authentication" is a bit of a term of art.  It can mean either
> of two things, as well as both: a) that the two peers "prove" to each
> other that they possess the same session keys before proceeding to
> exchange sensitive information, b) that the two peers have names that
> are authenticated to each other.  (a) comes from Kerberos.  (b) is the
> meaning most people would expect.
>
> In general we want both.
>    
I know the definition of mutual authentication. In fact, I prefer the 
one from the Handbook of Applied Cryptography:
http://www.cacr.math.uwaterloo.ca/hac/about/chap10.pdf
"
Entity authentication is the process whereby one party is assured 
(through acquisition
of corroborative evidence) of the identity of a second party involved in 
a protocol,
and that the second has actually participated (i.e., is active at, or 
immediately prior to, the
time the evidence is acquired).
"

Note that for some of the discussions it might be useful to 
differentiate between entity authentication and key confirmation.

"Key confirmation is the property whereby one party is assured that a second
(possibly unidentified) party actually has possession of a particular 
secret key.
"
http://www.cacr.math.uwaterloo.ca/hac/about/chap12.pdf




Anyway, what I meant is that with OpenID-SASL, OAuth-SASL, and with the 
SAML-SASL approach as well we have three parties in the picture, not 
just two. Hence, there is the question between whom authentication 
really happens.

For OpenID-SASL, for example, mutual authentication happens between the 
end host / user and the identity provider, and then there is also 
authentication of the relying party to the end host. Finally, depending 
on the specific profile being used there is also authentication between 
the identity provider and the relying party.


>    
>> * Channel Binding: Again, a different setup than described in RFC 5056.
>> Would be nice to know if someone has deployment experience with it.
>>      
> I'm not sure what you're talking about here.
>
>    

Do we have some implementation and deployment experience with some of 
the channel binding approaches?

>> * Authentication: Like with many identity management systems the
>> authentication part is actually outside the scope of these protocols.
>> This is true for SAML, OpenID, etc.
>>      
> In the sense that these protocols shouldn't require invasive changes to
> deployed authentication infrastructure?  Yes.
>    
Some of the protocols it was the intention to leave the authentication 
part between the user and the identity provider outside the scope of the 
standardized protocol. This has certain implication. It allows a variety 
of different authentication infrastructures to be used without the need 
to update the standard in a way that is suitable for the identity 
providers environment. Typically, the user interface experience is 
different between each identity provider and sometimes this is actually 
used for the purpose of branding. On the other hand it makes a security 
analysis more difficult. For the OpenID-SASL design, for example, it 
meant that we had to think about how to integrate this flexible 
authentication approach into the design since we didn't want to change 
the way OpenID works nor to change the deployed OpenID infrastructure 
(by, for example, mandating one specific authentication procedure). It 
also means that we had to deal with different protocol interactions, in 
our case SASL and HTTP-Browser interactions.

Ciao
Hannes

> Nico
>    


From hannes.tschofenig@gmx.net  Sun Aug 15 11:35:08 2010
Return-Path: <hannes.tschofenig@gmx.net>
X-Original-To: kitten@core3.amsl.com
Delivered-To: kitten@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id D935C3A68BF for <kitten@core3.amsl.com>; Sun, 15 Aug 2010 11:35:08 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -102.599
X-Spam-Level: 
X-Spam-Status: No, score=-102.599 tagged_above=-999 required=5 tests=[AWL=0.000, BAYES_00=-2.599, USER_IN_WHITELIST=-100]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id FnB+veRra3zh for <kitten@core3.amsl.com>; Sun, 15 Aug 2010 11:35:08 -0700 (PDT)
Received: from mail.gmx.net (mailout-de.gmx.net [213.165.64.23]) by core3.amsl.com (Postfix) with SMTP id 75F9B3A6875 for <kitten@ietf.org>; Sun, 15 Aug 2010 11:35:07 -0700 (PDT)
Received: (qmail invoked by alias); 15 Aug 2010 18:35:42 -0000
Received: from a88-115-222-204.elisa-laajakaista.fi (EHLO [192.168.1.5]) [88.115.222.204] by mail.gmx.net (mp069) with SMTP; 15 Aug 2010 20:35:42 +0200
X-Authenticated: #29516787
X-Provags-ID: V01U2FsdGVkX18HPy0Y9RcHVyVWirNyzIyAllfUbM7hxLwba3l5di 5/9Yr22KQyIwDH
Message-ID: <4C68337F.8040008@gmx.net>
Date: Sun, 15 Aug 2010 21:35:43 +0300
From: Hannes Tschofenig <hannes.tschofenig@gmx.net>
User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.9.1.11) Gecko/20100711 Thunderbird/3.0.6
MIME-Version: 1.0
To: Scott Cantor <cantor.2@osu.edu>
References: <tsllj8dyo5p.fsf@mit.edu> <4C6460D6.7020800@stpeter.im>	<tsl1va3v7z5.fsf@mit.edu> <4C64698A.9030405@isode.com>	<tslocd7tsjr.fsf@mit.edu>	<3D3C75174CB95F42AD6BCC56E5555B4502E9BF97@FIESEXC015.nsn-intra.net> <046101cb3b0e$1cca1de0$565e59a0$@osu.edu>
In-Reply-To: <046101cb3b0e$1cca1de0$565e59a0$@osu.edu>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit
X-Y-GMX-Trusted: 0
Cc: kitten@ietf.org
Subject: Re: [kitten] Concerns about oauth sasl mechanism
X-BeenThere: kitten@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Common Authentication Technologies - Next Generation <kitten.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/kitten>
List-Post: <mailto:kitten@ietf.org>
List-Help: <mailto:kitten-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=subscribe>
X-List-Received-Date: Sun, 15 Aug 2010 18:35:08 -0000

Hey Scott,

I understand your approach and the thinking behind it.

There are obviously trade-offs and it is hard to find the right balance 
between deployability and security.
In the IETF we had some good examples of "picking up deployed stuff" and 
some examples of where we thought it would be a good idea and then the 
industry decided against it.

I also understand that some of the communities have a somewhat different 
way of working. There are sometimes quite big gaps (in thinking) between 
the different communities. I should know that best given that I had the 
chance to look a bit deeper into the privacy work recently and noticed 
this massive imbalance between the approach people from the research 
community take in comparison with deployment happening on the Internet 
pushed by some Web2.0 companies.

Anyway, we are not going to reach a conclusion on a philosophical level. 
Hence, let us move on and figure out what the channel stuff would get us 
for the protocols in discussion.

Ciao
Hannes

On 8/13/2010 8:36 PM, Scott Cantor wrote:
>> The folks in Oauth are not impressed by random statements like "we need
>> channel bindings" because they want to know "What is it?" and "What are
>> the benefits?". I think that this is the right attitude.
>>      
> I don't think everything has to be useful to Google and a random army of
> scripters that want to implement everything from scratch in Javascript
> though. If you want a standard, it should be rich enough to handle more than
> just what they choose to care about.
>
> And before you say "come to the WG and ask", I don't believe the foundation
> is right, so I'm not going to try and build my house extension on it. That's
> my prerogative.
>
>    
>> Working in the Web 2.0 environment they want to get something done soon
>> (rather than in 5 years).
>>      
> Here's where that gets you:
>
> http://code.google.com/p/browsersec/wiki/Main
>
> I know some people think that can be reined back in and fixed, but I really
> don't.
>
> -- Scott
>
>
> _______________________________________________
> Kitten mailing list
> Kitten@ietf.org
> https://www.ietf.org/mailman/listinfo/kitten
>
>    


From cantor.2@osu.edu  Sun Aug 15 15:52:51 2010
Return-Path: <cantor.2@osu.edu>
X-Original-To: kitten@core3.amsl.com
Delivered-To: kitten@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 124BE3A6800 for <kitten@core3.amsl.com>; Sun, 15 Aug 2010 15:52:51 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -1.704
X-Spam-Level: 
X-Spam-Status: No, score=-1.704 tagged_above=-999 required=5 tests=[AWL=-0.594, BAYES_05=-1.11]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id h7UD4Xpaxzjx for <kitten@core3.amsl.com>; Sun, 15 Aug 2010 15:52:50 -0700 (PDT)
Received: from hrndva-omtalb.mail.rr.com (hrndva-omtalb.mail.rr.com [71.74.56.124]) by core3.amsl.com (Postfix) with ESMTP id 124023A67A2 for <kitten@ietf.org>; Sun, 15 Aug 2010 15:52:48 -0700 (PDT)
X-Authority-Analysis: v=1.1 cv=FLDyhxrUEZZDQYGWQdnXfbzWjjll+fIPMqQTcKt1R7E= c=1 sm=0 a=PMct6vZm66EA:10 a=kj9zAlcOel0A:10 a=/Wh0N9815gtYTLoiNwER9g==:17 a=48vgC7mUAAAA:8 a=R6I5MG2CzWC2o1FX_L8A:9 a=OvEEGMmAOfqZq4TskrUGmtjCoCsA:4 a=CjuIK1q_8ugA:10 a=xQERjPQv1MMA:10 a=/Wh0N9815gtYTLoiNwER9g==:117
X-Cloudmark-Score: 0
X-Originating-IP: 24.210.116.83
Received: from [24.210.116.83] ([24.210.116.83:1164] helo=SNOWDOG) by hrndva-oedge01.mail.rr.com (envelope-from <cantor.2@osu.edu>) (ecelerity 2.2.2.39 r()) with ESMTP id C2/EC-24738-3EF686C4; Sun, 15 Aug 2010 22:53:24 +0000
From: "Scott Cantor" <cantor.2@osu.edu>
To: "'Hannes Tschofenig'" <hannes.tschofenig@gmx.net>
References: <tsllj8dyo5p.fsf@mit.edu> <4C6460D6.7020800@stpeter.im> <tsl1va3v7z5.fsf@mit.edu> <4C64698A.9030405@isode.com> <tslocd7tsjr.fsf@mit.edu> <3D3C75174CB95F42AD6BCC56E5555B4502E9BF97@FIESEXC015.nsn-intra.net> <046101cb3b0e$1cca1de0$565e59a0$@osu.edu> <4C68337F.8040008@gmx.net>
In-Reply-To: <4C68337F.8040008@gmx.net>
Date: Sun, 15 Aug 2010 18:53:23 -0400
Organization: The Ohio State University
Message-ID: <000901cb3ccc$aa732960$ff597c20$@osu.edu>
MIME-Version: 1.0
Content-Type: text/plain; charset="US-ASCII"
Content-Transfer-Encoding: 7bit
X-Mailer: Microsoft Outlook 14.0
Thread-index: AQEtpK0+ZMhTGTaNZYkccHtUfLPXDgKpFn1eAjsa3OcCFW1apACo0J5EAVlchLUCNX9OWAIKoh3Jk7RuzSA=
Content-language: en-us
Cc: kitten@ietf.org
Subject: Re: [kitten] Concerns about oauth sasl mechanism
X-BeenThere: kitten@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Common Authentication Technologies - Next Generation <kitten.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/kitten>
List-Post: <mailto:kitten@ietf.org>
List-Help: <mailto:kitten-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=subscribe>
X-List-Received-Date: Sun, 15 Aug 2010 22:52:51 -0000

> Anyway, we are not going to reach a conclusion on a philosophical level.
> Hence, let us move on and figure out what the channel stuff would get us
> for the protocols in discussion.

If you want to see what I think it buys you, you can find that here:

http://www.ietf.org/mail-archive/web/kitten/current/msg01943.html

(Sorry for the line runons, I guess the soft wrapping in my client isn't
friendly to this particular archive.)

I think it fundamentally comes back to whether one thinks commercial SSL is
nonsense or not. I think it is. I know a lot of Web 2.0ers claim it's not.

-- Scott



From Nicolas.Williams@oracle.com  Sun Aug 15 21:17:25 2010
Return-Path: <Nicolas.Williams@oracle.com>
X-Original-To: kitten@core3.amsl.com
Delivered-To: kitten@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 613D13A6944 for <kitten@core3.amsl.com>; Sun, 15 Aug 2010 21:17:25 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -6.54
X-Spam-Level: 
X-Spam-Status: No, score=-6.54 tagged_above=-999 required=5 tests=[AWL=0.058,  BAYES_00=-2.599, RCVD_IN_DNSWL_MED=-4, UNPARSEABLE_RELAY=0.001]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id qQD3C7sJcXye for <kitten@core3.amsl.com>; Sun, 15 Aug 2010 21:17:24 -0700 (PDT)
Received: from rcsinet10.oracle.com (rcsinet10.oracle.com [148.87.113.121]) by core3.amsl.com (Postfix) with ESMTP id 386A93A67E5 for <kitten@ietf.org>; Sun, 15 Aug 2010 21:17:24 -0700 (PDT)
Received: from acsinet15.oracle.com (acsinet15.oracle.com [141.146.126.227]) by rcsinet10.oracle.com (Switch-3.4.2/Switch-3.4.2) with ESMTP id o7G4HuJ7024064 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=OK); Mon, 16 Aug 2010 04:17:58 GMT
Received: from acsmt354.oracle.com (acsmt354.oracle.com [141.146.40.154]) by acsinet15.oracle.com (Switch-3.4.2/Switch-3.4.1) with ESMTP id o7FEFYsN002547; Mon, 16 Aug 2010 04:17:56 GMT
Received: from abhmt017.oracle.com by acsmt355.oracle.com with ESMTP id 518980881281932275; Sun, 15 Aug 2010 21:17:55 -0700
Received: from oracle.com (/129.153.128.104) by default (Oracle Beehive Gateway v4.0) with ESMTP ; Sun, 15 Aug 2010 21:17:55 -0700
Date: Sun, 15 Aug 2010 23:17:50 -0500
From: Nicolas Williams <Nicolas.Williams@oracle.com>
To: Hannes Tschofenig <hannes.tschofenig@gmx.net>
Message-ID: <20100816041750.GT14409@oracle.com>
References: <tsllj8dyo5p.fsf@mit.edu> <4C6460D6.7020800@stpeter.im> <tsl1va3v7z5.fsf@mit.edu> <4C64698A.9030405@isode.com> <3D3C75174CB95F42AD6BCC56E5555B4502E9BF85@FIESEXC015.nsn-intra.net> <20100813152954.GF5217@oracle.com> <4C683150.6000003@gmx.net>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <4C683150.6000003@gmx.net>
User-Agent: Mutt/1.5.20 (2010-03-02)
Cc: kitten@ietf.org, Sam Hartman <hartmans-ietf@mit.edu>
Subject: Re: [kitten] Concerns about oauth sasl mechanism
X-BeenThere: kitten@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Common Authentication Technologies - Next Generation <kitten.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/kitten>
List-Post: <mailto:kitten@ietf.org>
List-Help: <mailto:kitten-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=subscribe>
X-List-Received-Date: Mon, 16 Aug 2010 04:17:25 -0000

On Sun, Aug 15, 2010 at 09:26:25PM +0300, Hannes Tschofenig wrote:
> Anyway, what I meant is that with OpenID-SASL, OAuth-SASL, and with
> the SAML-SASL approach as well we have three parties in the picture,
> not just two. Hence, there is the question between whom
> authentication really happens.

The GSS-API concerns itself with just the initiator and acceptor.  These
are typically the client and the server.  Infrastructure services are
not exposed through the API (well, the naming extensions work will
expose them, but that's a different issue).

So, when we say mutual authentication in this context, then, we mean
that the initiator (client) and acceptor (server) both have a _name_,
and that each name is "authenticated" to the other.  A name that's
meaningful to users.  Typically the initator name is the name of a user.
And typically the acceptor name is a host-based service name.

We also want key confirmation.

> For OpenID-SASL, for example, mutual authentication happens between
> the end host / user and the identity provider, and then there is
> also authentication of the relying party to the end host. Finally,
> depending on the specific profile being used there is also
> authentication between the identity provider and the relying party.

You lost me.  What are all the entities here?

 - user
 - IdP
 - service being accessed by the user
 - host hosting the service

(I'd rather stay away from the term "relying party".)

Now, who is authenticated, and to whom?

> Do we have some implementation and deployment experience with some
> of the channel binding approaches?

Microsoft has deployed IWA.  That's, effectively, HTTP/Negotiate, but
with HTTPS (i.e., TLS), and with tls-unique or tls-server-end-point CB.

> >>This is true for SAML, OpenID, etc.
> >
> >In the sense that these protocols shouldn't require invasive changes to
> >deployed authentication infrastructure?  Yes.
> 
> Some of the protocols it was the intention to leave the
> authentication part between the user and the identity provider
> outside the scope of the standardized protocol. This has certain
> implication. It allows a variety of different authentication
> infrastructures to be used without the need to update the standard
> in a way that is suitable for the identity providers environment.

This is fine.  See the discussion of Scott Cantor's ECP SAML mechanism.

> Typically, the user interface experience is different between each
> identity provider and sometimes this is actually used for the
> purpose of branding. On the other hand it makes a security analysis
> more difficult.  [...]

Trusted third parties change the security analysis, but don't
necessarily make it more difficult.  Mostly you end up having to explain
that a trusted third party can become a passive or active attacker.

(For example, Kerberos KDCs can recover the plaintext of all
client<->server conversations where the server is in the same realm as
the KDC and the conversations are not further protected.  PKIX CAs can
MITM you.)

In fact, Kerberos is a perfect example.  The Kerberos GSS-API mechanism
doesn't concern itself at all with how the user was pre-authenticated to
the KDC.  The IdP in your OpenID design, like the IdP in Scott Cantor's,
can authenticate users in any way it wants.

Nico
-- 

From hartmans@mit.edu  Mon Aug 16 05:56:42 2010
Return-Path: <hartmans@mit.edu>
X-Original-To: kitten@core3.amsl.com
Delivered-To: kitten@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 970123A6855 for <kitten@core3.amsl.com>; Mon, 16 Aug 2010 05:56:42 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -102.723
X-Spam-Level: 
X-Spam-Status: No, score=-102.723 tagged_above=-999 required=5 tests=[AWL=-0.458, BAYES_00=-2.599, IP_NOT_FRIENDLY=0.334, USER_IN_WHITELIST=-100]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id NQq2DQ612jmG for <kitten@core3.amsl.com>; Mon, 16 Aug 2010 05:56:41 -0700 (PDT)
Received: from mail.suchdamage.org (permutation-city.suchdamage.org [69.25.196.28]) by core3.amsl.com (Postfix) with ESMTP id C145C3A659B for <kitten@ietf.org>; Mon, 16 Aug 2010 05:56:41 -0700 (PDT)
Received: from carter-zimmerman.suchdamage.org (carter-zimmerman.suchdamage.org [69.25.196.178]) (using TLSv1 with cipher DHE-RSA-AES256-SHA (256/256 bits)) (Client CN "laptop", Issuer "laptop" (not verified)) by mail.suchdamage.org (Postfix) with ESMTPS id BCF36200B4; Mon, 16 Aug 2010 08:57:17 -0400 (EDT)
Received: by carter-zimmerman.suchdamage.org (Postfix, from userid 8042) id 98F59412A; Mon, 16 Aug 2010 08:57:15 -0400 (EDT)
From: Sam Hartman <hartmans-ietf@mit.edu>
To: "Tschofenig\, Hannes \(NSN - FI\/Espoo\)" <hannes.tschofenig@nsn.com>
References: <tsllj8dyo5p.fsf@mit.edu> <3D3C75174CB95F42AD6BCC56E5555B4502E9BF74@FIESEXC015.nsn-intra.net>
Date: Mon, 16 Aug 2010 08:57:15 -0400
In-Reply-To: <3D3C75174CB95F42AD6BCC56E5555B4502E9BF74@FIESEXC015.nsn-intra.net> (Hannes Tschofenig's message of "Fri, 13 Aug 2010 14:59:02 +0300")
Message-ID: <tslbp92pvf8.fsf@mit.edu>
User-Agent: Gnus/5.110009 (No Gnus v0.9) Emacs/22.3 (gnu/linux)
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Cc: kitten@ietf.org, ext Sam Hartman <hartmans-ietf@mit.edu>
Subject: Re: [kitten] Concerns about oauth sasl mechanism
X-BeenThere: kitten@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Common Authentication Technologies - Next Generation <kitten.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/kitten>
List-Post: <mailto:kitten@ietf.org>
List-Help: <mailto:kitten-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=subscribe>
X-List-Received-Date: Mon, 16 Aug 2010 12:56:42 -0000

>>>>> ""Tschofenig," == "Tschofenig, Hannes (NSN <- FI/Espoo)" <hannes.tschofenig@nsn.com>> writes:

    "Tschofenig,> Hi Sam, "I haven't followed xyz but here are my
    "Tschofenig,> comments" is not a good start.

I have followed SASL.  I feel qualified to speak to the SASL working
group about what mechanisms I believe it should standardize.


I think that as someone proposing to add standards-track work to
kitten's charter, you have an obligation to understand the community
you're joining, and to a fairly large extent its previous discussions
especially as they relate to scope, style and the like.


Certainly if I were going to show up at oauth and make a statement about
channel binding there--I'd do a lot better than "You should have channel
binding; I don't know what it is."

My statement is that if this working group is going to standardize a
SASL mechanism it should have channel bindings or security layers.

Your tone has put me off explaining how we got where we are today in
kitten on the list. I have no desire to participate in a list
conversation that has already reached this level of defensiveness.  In
another context, such as a phone call, I'd be happy to try and work
through things with you.  You may find others who will have this
conversation with you on-list.

From hartmans@mit.edu  Mon Aug 16 06:21:06 2010
Return-Path: <hartmans@mit.edu>
X-Original-To: kitten@core3.amsl.com
Delivered-To: kitten@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 99E6A3A688E for <kitten@core3.amsl.com>; Mon, 16 Aug 2010 06:21:06 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -102.709
X-Spam-Level: 
X-Spam-Status: No, score=-102.709 tagged_above=-999 required=5 tests=[AWL=-0.444, BAYES_00=-2.599, IP_NOT_FRIENDLY=0.334, USER_IN_WHITELIST=-100]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id dpzOFllGfI8P for <kitten@core3.amsl.com>; Mon, 16 Aug 2010 06:21:05 -0700 (PDT)
Received: from mail.suchdamage.org (permutation-city.suchdamage.org [69.25.196.28]) by core3.amsl.com (Postfix) with ESMTP id C1F6F3A6837 for <kitten@ietf.org>; Mon, 16 Aug 2010 06:21:05 -0700 (PDT)
Received: from carter-zimmerman.suchdamage.org (carter-zimmerman.suchdamage.org [69.25.196.178]) (using TLSv1 with cipher DHE-RSA-AES256-SHA (256/256 bits)) (Client CN "laptop", Issuer "laptop" (not verified)) by mail.suchdamage.org (Postfix) with ESMTPS id 858C4203CB for <kitten@ietf.org>; Mon, 16 Aug 2010 09:21:39 -0400 (EDT)
Received: by carter-zimmerman.suchdamage.org (Postfix, from userid 8042) id 83DE3412A; Mon, 16 Aug 2010 09:21:37 -0400 (EDT)
From: Sam Hartman <hartmans-ietf@mit.edu>
To: kitten@ietf.org
Cc: 
References: <tsllj8dyo5p.fsf@mit.edu> <4C6460D6.7020800@stpeter.im> <tsl1va3v7z5.fsf@mit.edu> <4C64698A.9030405@isode.com> <3D3C75174CB95F42AD6BCC56E5555B4502E9BF85@FIESEXC015.nsn-intra.net> <20100813152954.GF5217@oracle.com> <4C683150.6000003@gmx.net> <20100816041750.GT14409@oracle.com>
Date: Mon, 16 Aug 2010 09:21:37 -0400
In-Reply-To: <20100816041750.GT14409@oracle.com> (Nicolas Williams's message of "Sun, 15 Aug 2010 23:17:50 -0500")
Message-ID: <tslvd7aofq6.fsf_-_@mit.edu>
User-Agent: Gnus/5.110009 (No Gnus v0.9) Emacs/22.3 (gnu/linux)
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Subject: [kitten] Mutual auth
X-BeenThere: kitten@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Common Authentication Technologies - Next Generation <kitten.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/kitten>
List-Post: <mailto:kitten@ietf.org>
List-Help: <mailto:kitten-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=subscribe>
X-List-Received-Date: Mon, 16 Aug 2010 13:21:06 -0000

>>>>> "Nicolas" == Nicolas Williams <Nicolas.Williams@oracle.com> writes:
    Nicolas> then, we mean that the initiator (client) and acceptor
    Nicolas> (server) both have a _name_, and that each name is
    Nicolas> "authenticated" to the other.  A name that's meaningful to
    Nicolas> users.  Typically the initator name is the name of a user.
    Nicolas> And typically the acceptor name is a host-based service
    Nicolas> name.


Hmm.  I'm not sure that's really true.  In particular in scram, I
thought it was more like the client has a name and the server was the
guy who actually had those credentials.  Regardless of whether this is
permissible in GSS-API, SASL's definition of mutual authentication is
permissive enough that you can have mutual authentication if you happen
to share long-term credentials and prove that you have the same
credentials on both sides.

I'm not sure that you can extend this to a multi-party protocol without
introducing naming.

From klaas@cisco.com  Mon Aug 16 06:28:05 2010
Return-Path: <klaas@cisco.com>
X-Original-To: kitten@core3.amsl.com
Delivered-To: kitten@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 0ECBD3A6855 for <kitten@core3.amsl.com>; Mon, 16 Aug 2010 06:28:05 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -10.268
X-Spam-Level: 
X-Spam-Status: No, score=-10.268 tagged_above=-999 required=5 tests=[AWL=0.331, BAYES_00=-2.599, RCVD_IN_DNSWL_HI=-8]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id fS7Q8aSmEiVI for <kitten@core3.amsl.com>; Mon, 16 Aug 2010 06:28:04 -0700 (PDT)
Received: from rtp-iport-1.cisco.com (rtp-iport-1.cisco.com [64.102.122.148]) by core3.amsl.com (Postfix) with ESMTP id DECE13A659B for <kitten@ietf.org>; Mon, 16 Aug 2010 06:28:03 -0700 (PDT)
Authentication-Results: rtp-iport-1.cisco.com; dkim=neutral (message not signed) header.i=none
X-IronPort-Anti-Spam-Filtered: true
X-IronPort-Anti-Spam-Result: An8FAL7ZaExAZnwM/2dsb2JhbACTMI0ScaAEmyyFOwSJYg
X-IronPort-AV: E=Sophos;i="4.55,376,1278288000"; d="scan'208";a="148162043"
Received: from rtp-core-1.cisco.com ([64.102.124.12]) by rtp-iport-1.cisco.com with ESMTP; 16 Aug 2010 13:28:39 +0000
Received: from macmini.wierenga.net (ams-kwiereng-8715.cisco.com [10.55.220.246]) by rtp-core-1.cisco.com (8.13.8/8.14.3) with ESMTP id o7GDSdco010140 for <kitten@ietf.org>; Mon, 16 Aug 2010 13:28:39 GMT
Message-ID: <4C693D06.8050800@cisco.com>
Date: Mon, 16 Aug 2010 15:28:38 +0200
From: Klaas Wierenga <klaas@cisco.com>
User-Agent: Mozilla/5.0 (Macintosh; U; Intel Mac OS X 10.6; en-GB; rv:1.9.1.11) Gecko/20100711 Thunderbird/3.0.6
MIME-Version: 1.0
To: kitten@ietf.org
References: <tsllj8dyo5p.fsf@mit.edu> <4C6460D6.7020800@stpeter.im>	<tsl1va3v7z5.fsf@mit.edu> <4C64698A.9030405@isode.com>	<3D3C75174CB95F42AD6BCC56E5555B4502E9BF85@FIESEXC015.nsn-intra.net>	<20100813152954.GF5217@oracle.com> <4C683150.6000003@gmx.net>	<20100816041750.GT14409@oracle.com> <tslvd7aofq6.fsf_-_@mit.edu>
In-Reply-To: <tslvd7aofq6.fsf_-_@mit.edu>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit
Subject: Re: [kitten] Mutual auth
X-BeenThere: kitten@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Common Authentication Technologies - Next Generation <kitten.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/kitten>
List-Post: <mailto:kitten@ietf.org>
List-Help: <mailto:kitten-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=subscribe>
X-List-Received-Date: Mon, 16 Aug 2010 13:28:05 -0000

On 8/16/10 3:21 PM, Sam Hartman wrote:

Sam,

>>>>>> "Nicolas" == Nicolas Williams<Nicolas.Williams@oracle.com>  writes:
>      Nicolas>  then, we mean that the initiator (client) and acceptor
>      Nicolas>  (server) both have a _name_, and that each name is
>      Nicolas>  "authenticated" to the other.  A name that's meaningful to
>      Nicolas>  users.  Typically the initator name is the name of a user.
>      Nicolas>  And typically the acceptor name is a host-based service
>      Nicolas>  name.
>
>
> Hmm.  I'm not sure that's really true.  In particular in scram, I
> thought it was more like the client has a name and the server was the
> guy who actually had those credentials.  Regardless of whether this is
> permissible in GSS-API, SASL's definition of mutual authentication is
> permissive enough that you can have mutual authentication if you happen
> to share long-term credentials and prove that you have the same
> credentials on both sides.
>
> I'm not sure that you can extend this to a multi-party protocol without
> introducing naming.

me neither, otoh, I am not conviced that we need to extend this to a 
multi-party protocol. So far it seems that we are still talking about an 
initiator and an acceptor only in the SASL-sense. Outside SASL there may 
be tons of stuff going on, with all kinds of either or not mutual 
authentications, but that is here not really relevant, or am I missing 
something?

Klaas

From Nicolas.Williams@oracle.com  Mon Aug 16 07:21:31 2010
Return-Path: <Nicolas.Williams@oracle.com>
X-Original-To: kitten@core3.amsl.com
Delivered-To: kitten@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 553783A686A for <kitten@core3.amsl.com>; Mon, 16 Aug 2010 07:21:31 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -6.541
X-Spam-Level: 
X-Spam-Status: No, score=-6.541 tagged_above=-999 required=5 tests=[AWL=0.057,  BAYES_00=-2.599, RCVD_IN_DNSWL_MED=-4, UNPARSEABLE_RELAY=0.001]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id XDsj6zAqX-VN for <kitten@core3.amsl.com>; Mon, 16 Aug 2010 07:21:30 -0700 (PDT)
Received: from rcsinet10.oracle.com (rcsinet10.oracle.com [148.87.113.121]) by core3.amsl.com (Postfix) with ESMTP id 3C5253A6829 for <kitten@ietf.org>; Mon, 16 Aug 2010 07:21:30 -0700 (PDT)
Received: from acsinet15.oracle.com (acsinet15.oracle.com [141.146.126.227]) by rcsinet10.oracle.com (Switch-3.4.2/Switch-3.4.2) with ESMTP id o7GEM4EX025810 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=OK); Mon, 16 Aug 2010 14:22:06 GMT
Received: from acsmt355.oracle.com (acsmt355.oracle.com [141.146.40.155]) by acsinet15.oracle.com (Switch-3.4.2/Switch-3.4.1) with ESMTP id o7FD27qW010912; Mon, 16 Aug 2010 14:22:03 GMT
Received: from abhmt019.oracle.com by acsmt353.oracle.com with ESMTP id 499362971281968449; Mon, 16 Aug 2010 07:20:49 -0700
Received: from oracle.com (/129.153.128.104) by default (Oracle Beehive Gateway v4.0) with ESMTP ; Mon, 16 Aug 2010 07:20:48 -0700
Date: Mon, 16 Aug 2010 09:20:43 -0500
From: Nicolas Williams <Nicolas.Williams@oracle.com>
To: Sam Hartman <hartmans-ietf@mit.edu>
Message-ID: <20100816142042.GA14409@oracle.com>
References: <tsllj8dyo5p.fsf@mit.edu> <4C6460D6.7020800@stpeter.im> <tsl1va3v7z5.fsf@mit.edu> <4C64698A.9030405@isode.com> <3D3C75174CB95F42AD6BCC56E5555B4502E9BF85@FIESEXC015.nsn-intra.net> <20100813152954.GF5217@oracle.com> <4C683150.6000003@gmx.net> <20100816041750.GT14409@oracle.com> <tslvd7aofq6.fsf_-_@mit.edu>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <tslvd7aofq6.fsf_-_@mit.edu>
User-Agent: Mutt/1.5.20 (2010-03-02)
Cc: kitten@ietf.org
Subject: Re: [kitten] Mutual auth
X-BeenThere: kitten@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Common Authentication Technologies - Next Generation <kitten.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/kitten>
List-Post: <mailto:kitten@ietf.org>
List-Help: <mailto:kitten-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=subscribe>
X-List-Received-Date: Mon, 16 Aug 2010 14:21:31 -0000

On Mon, Aug 16, 2010 at 09:21:37AM -0400, Sam Hartman wrote:
> >>>>> "Nicolas" == Nicolas Williams <Nicolas.Williams@oracle.com> writes:
>     Nicolas> then, we mean that the initiator (client) and acceptor
>     Nicolas> (server) both have a _name_, and that each name is
>     Nicolas> "authenticated" to the other.  A name that's meaningful to
>     Nicolas> users.  Typically the initator name is the name of a user.
>     Nicolas> And typically the acceptor name is a host-based service
>     Nicolas> name.
> 
> Hmm.  I'm not sure that's really true.  In particular in scram, I
> thought it was more like the client has a name and the server was the
> guy who actually had those credentials.  Regardless of whether this is
> permissible in GSS-API, SASL's definition of mutual authentication is
> permissive enough that you can have mutual authentication if you happen
> to share long-term credentials and prove that you have the same
> credentials on both sides.
> 
> I'm not sure that you can extend this to a multi-party protocol without
> introducing naming.

As a matter of fact, we concluded that SCRAM does name the acceptor.
Implementations can lookup passwords in a table indexed by {initiator
name, acceptor name}, or prompt the user for a password for the given
{initiator name, acceptor name}.

Nico
-- 

From hannes.tschofenig@nsn.com  Mon Aug 16 07:40:02 2010
Return-Path: <hannes.tschofenig@nsn.com>
X-Original-To: kitten@core3.amsl.com
Delivered-To: kitten@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 5AC063A67FE for <kitten@core3.amsl.com>; Mon, 16 Aug 2010 07:40:02 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -100.995
X-Spam-Level: 
X-Spam-Status: No, score=-100.995 tagged_above=-999 required=5 tests=[AWL=-1.410, BAYES_40=-0.185, J_CHICKENPOX_62=0.6, USER_IN_WHITELIST=-100]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id i7N-diRJNnuM for <kitten@core3.amsl.com>; Mon, 16 Aug 2010 07:39:57 -0700 (PDT)
Received: from demumfd002.nsn-inter.net (demumfd002.nsn-inter.net [93.183.12.31]) by core3.amsl.com (Postfix) with ESMTP id 6EE3B3A6860 for <kitten@ietf.org>; Mon, 16 Aug 2010 07:39:56 -0700 (PDT)
Received: from demuprx017.emea.nsn-intra.net ([10.150.129.56]) by demumfd002.nsn-inter.net (8.12.11.20060308/8.12.11) with ESMTP id o7GEeVL3006126 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=OK) for <kitten@ietf.org>; Mon, 16 Aug 2010 16:40:31 +0200
Received: from demuexc023.nsn-intra.net (demuexc023.nsn-intra.net [10.150.128.36]) by demuprx017.emea.nsn-intra.net (8.12.11.20060308/8.12.11) with ESMTP id o7GEeUu3008132 for <kitten@ietf.org>; Mon, 16 Aug 2010 16:40:30 +0200
Received: from FIESEXC015.nsn-intra.net ([10.159.0.23]) by demuexc023.nsn-intra.net with Microsoft SMTPSVC(6.0.3790.4675);  Mon, 16 Aug 2010 16:40:11 +0200
X-MimeOLE: Produced By Microsoft Exchange V6.5
Content-class: urn:content-classes:message
MIME-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Content-Transfer-Encoding: quoted-printable
Date: Mon, 16 Aug 2010 17:40:08 +0300
Message-ID: <3D3C75174CB95F42AD6BCC56E5555B4502E9C89F@FIESEXC015.nsn-intra.net>
X-MS-Has-Attach: 
X-MS-TNEF-Correlator: 
Thread-Topic: Channel Bindings Discussion
Thread-Index: Acs9UOw6e0mNP3gvRaOv4s4dUzSQaQ==
From: "Tschofenig, Hannes (NSN - FI/Espoo)" <hannes.tschofenig@nsn.com>
To: <kitten@ietf.org>
X-OriginalArrivalTime: 16 Aug 2010 14:40:11.0685 (UTC) FILETIME=[EE6ADD50:01CB3D50]
Subject: [kitten] Channel Bindings Discussion
X-BeenThere: kitten@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Common Authentication Technologies - Next Generation <kitten.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/kitten>
List-Post: <mailto:kitten@ietf.org>
List-Help: <mailto:kitten-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=subscribe>
X-List-Received-Date: Mon, 16 Aug 2010 14:40:02 -0000

Hi all,=20

As mentioned recently, the OpenID - SASL I-D authors (Simon, Eliot,
Hank, and myself) discussed the topic of channel bindings offlist. Since
the topic resurfaced on the mailing list we thought it would be useful
to share our private conversation with the rest of the group.=20

Ciao
Hannes

---------------------

From: Hannes Tschofenig <Hannes.Tschofenig@gmx.net>
Subject: OpenID & SASL: Channel Bindings
To: Simon Josefsson <simon@josefsson.org>, Eliot Lear <lear@cisco.com>,
        hmauldin@cisco.com
Date: Thu, 08 Jul 2010 14:47:50 +0300

Hi all,

I had a brief chat with Eliot two days ago and we ran into the topic
of "channel bindings". I thought I should share my opinion with you.

First, I have to say that I find the writeup by Nico in RFC 5056 hard
to understand. I re-read it today and it still confuses me. I also
have the history of the EAP work in my head, which (due to the name
collision) makes everything just more complicated.

Ignoring what I said in the previous paragraph, let me start slightly
differently to understand what you guys worry about (based on the
figure from the draft):

                                    +-----------+
                                    |           |
                                   >|  Relying  |
                                  / |  Party    |
                                //  |           |
                              //    +-----------+
                            //            ^
                   OpenID //           +--|--+
                        //             | O|  |
                       /             S | p|  |
                     //              A | e|  |
                   //                S | n|  |
                 //                  L | I|  |
               //                      | D|  |
             </                        +--|--+
      +------------+                      v
      |            |                 +----------+
      |  OpenID    |   OpenID        |          |
      |  Provider  |<--------------->|  Client  |
      |            |                 |          |
      +------------+                 +----------+

We assume that TLS is used between the client and the relying party with
server-side autnentication.
The certificate used for this purpose will show some identity, let us
call it RP_TLS_X.

Then, we might have TLS between the Client and the OpenID provider but
that largely depends on the authentication mechanism being used.

Furthermore, we have authentication of the RP to the OP (and vice versa)
at the OpenID "application layer". The RP will have some identity, let
us call it RP_OpenID_Y.

Now, I believe you guys worry about the linkage between the TLS channel
(used between the client and the RP) and the OpenID identity used by the
RP.

Is this correct?

Ciao
Hannes



From: Simon Josefsson <simon@josefsson.org>
Subject: Re: OpenID & SASL: Channel Bindings
To: Hannes Tschofenig <Hannes.Tschofenig@gmx.net>
Cc: Eliot Lear <lear@cisco.com>,  hmauldin@cisco.com
Date: Thu, 08 Jul 2010 15:03:31 +0200

Hannes Tschofenig <Hannes.Tschofenig@gmx.net> writes:

> Hi all,
>
> I had a brief chat with Eliot two days ago and we ran into the topic
> of "channel bindings". I thought I should share my opinion with you.
>
> First, I have to say that I find the writeup by Nico in RFC 5056 hard
> to understand. I re-read it today and it still confuses me. I also
> have the history of the EAP work in my head, which (due to the name
> collision) makes everything just more complicated.

Hi.  Yes, I have also found that document difficult to understand, and I
participated in the discussions around it.  I don't envy anyone else
trying to understand it.

> Ignoring what I said in the previous paragraph, let me start slightly
> differently to understand what you guys worry about (based on the
> figure from the draft):

Let me also say that my answer below is what _I_ worry about: I don't
know for what purpose Nico would like to see support for channel
bindings.

>
>                                     +-----------+
>                                     |           |
>                                    >|  Relying  |
>                                   / |  Party    |
>                                 //  |           |
>                               //    +-----------+
>                             //            ^
>                    OpenID //           +--|--+
>                         //             | O|  |
>                        /             S | p|  |
>                      //              A | e|  |
>                    //                S | n|  |
>                  //                  L | I|  |
>                //                      | D|  |
>              </                        +--|--+
>       +------------+                      v
>       |            |                 +----------+
>       |  OpenID    |   OpenID        |          |
>       |  Provider  |<--------------->|  Client  |
>       |            |                 |          |
>       +------------+                 +----------+
>
> We assume that TLS is used between the client and the relying party
> with server-side autnentication.  The certificate used for this
> purpose will show some identity, let us call it RP_TLS_X.

Right.

> Then, we might have TLS between the Client and the OpenID provider but
> that largely depends on the authentication mechanism being used.

Yes.  Mentally, I try to ignore what happens outside of the Client<->RP
communication because it is authentication mechanism specific.  Whether
it uses TLS or not is not relevant.  So to me, the only relevant TLS
channel to worry about is the Client<->RP channel.

> Furthermore, we have authentication of the RP to the OP (and vice
> versa) at the OpenID "application layer". The RP will have some
> identity, let us call it RP_OpenID_Y.
>
> Now, I believe you guys worry about the linkage between the TLS
> channel (used between the client and the RP) and the OpenID identity
> used by the RP.
>
> Is this correct?

I'm not sure that is really the same that I worry about.  I worry about
a MITM between Client and RP.  There are two ways to detect the MITM:

1) Require server certificate validation by the client.  This is the
traditional solution, but in practice it is quite weak because clients
either trust a gazillion of CAs out there (which is a bad idea) that
easily allows a MITM to acquire a fraudulent certificate or the end-user
manually inspects the certificate and clicks OK (in practice probably
even worse given that end-users can't be trusted with these decisions in
general).  There is no automatic and secure way for clients to learn the
trust anchor used by the server.

2) Both client and RP hashes (or somehow mixes in) the channel binding
data into the authentication process in a way that if the channel
binding data does NOT match, authentication will fail.

SASL PLAIN uses approach 1).  SASL SCRAM uses approach 2).  For SCRAM,
achieving 2 is easy because there is a shared secret between client and
server, and each peer can HMAC its channel binding and the peer can
verify the HMAC and then any MITM is detected.

I would like to have support for channel bindings in the OpenID (and
SAML) SASL mechanism because I believe 2) provides stronger protection
than 1).  However, I also realize that given how OpenID works, it _may_
be impossible without modifying the IdP.  But I don't think we have
proven that yet.  Maybe we can use the returns_to URL to store the
channel binding data, and allow Client+RP to parse and validate that.

Is this clearer?

/Simon

From: "Hannes Tschofenig" <Hannes.Tschofenig@gmx.net>
Subject: Re: OpenID & SASL: Channel Bindings
To: Simon Josefsson <simon@josefsson.org>
Cc: hmauldin@cisco.com, lear@cisco.com
Date: Fri, 09 Jul 2010 10:56:52 +0200

Hi Simon,=20

I noticed that you and Eliot had exchanged some emails about this topic
already (and they all ended up in my spam folder). I will have to browse
through them. I will also have to read SASL SCRAM to be sure I
understand your concern.=20

However, with the concern you describe it seems that you are more
looking into something related to cryptographic bindings, see
http://www.springerlink.com/content/8qv4qncf4fk106pg/.=20

I doubt that you will easily get around verifying the server provided
certificate but you could  check whether the identities provided at the
different layers match, namely RP_TLS_X and RP_OpenID_Y. Doing this
would at least ensure that the end points are the same, which is what
you are verifying when you hash data from both layers together.=20

A separate, but important aspect, is to figure out what are the
pre-conditions for a MITM attack to work and whether the threat is high
enough to justify a solution.=20

Ciao
Hannes

-------- Original-Nachricht --------
> Datum: Thu, 08 Jul 2010 15:03:31 +0200
> Von: Simon Josefsson <simon@josefsson.org>
> An: Hannes Tschofenig <Hannes.Tschofenig@gmx.net>
> CC: Eliot Lear <lear@cisco.com>, hmauldin@cisco.com
> Betreff: Re: OpenID & SASL: Channel Bindings

> Hannes Tschofenig <Hannes.Tschofenig@gmx.net> writes:
>=20
> > Hi all,
> >
> > I had a brief chat with Eliot two days ago and we ran into the topic
> > of "channel bindings". I thought I should share my opinion with you.
> >
> > First, I have to say that I find the writeup by Nico in RFC 5056
hard
> > to understand. I re-read it today and it still confuses me. I also
> > have the history of the EAP work in my head, which (due to the name
> > collision) makes everything just more complicated.
>=20
> Hi.  Yes, I have also found that document difficult to understand, and
I
> participated in the discussions around it.  I don't envy anyone else
> trying to understand it.
>=20
> > Ignoring what I said in the previous paragraph, let me start
slightly
> > differently to understand what you guys worry about (based on the
> > figure from the draft):
>=20
> Let me also say that my answer below is what _I_ worry about: I don't
> know for what purpose Nico would like to see support for channel
> bindings.
>=20
> >
> >                                     +-----------+
> >                                     |           |
> >                                    >|  Relying  |
> >                                   / |  Party    |
> >                                 //  |           |
> >                               //    +-----------+
> >                             //            ^
> >                    OpenID //           +--|--+
> >                         //             | O|  |
> >                        /             S | p|  |
> >                      //              A | e|  |
> >                    //                S | n|  |
> >                  //                  L | I|  |
> >                //                      | D|  |
> >              </                        +--|--+
> >       +------------+                      v
> >       |            |                 +----------+
> >       |  OpenID    |   OpenID        |          |
> >       |  Provider  |<--------------->|  Client  |
> >       |            |                 |          |
> >       +------------+                 +----------+
> >
> > We assume that TLS is used between the client and the relying party
> > with server-side autnentication.  The certificate used for this
> > purpose will show some identity, let us call it RP_TLS_X.
>=20
> Right.
>=20
> > Then, we might have TLS between the Client and the OpenID provider
but
> > that largely depends on the authentication mechanism being used.
>=20
> Yes.  Mentally, I try to ignore what happens outside of the
Client<->RP
> communication because it is authentication mechanism specific.
Whether
> it uses TLS or not is not relevant.  So to me, the only relevant TLS
> channel to worry about is the Client<->RP channel.
>=20
> > Furthermore, we have authentication of the RP to the OP (and vice
> > versa) at the OpenID "application layer". The RP will have some
> > identity, let us call it RP_OpenID_Y.
> >
> > Now, I believe you guys worry about the linkage between the TLS
> > channel (used between the client and the RP) and the OpenID identity
> > used by the RP.
> >
> > Is this correct?
>=20
> I'm not sure that is really the same that I worry about.  I worry
about
> a MITM between Client and RP.  There are two ways to detect the MITM:
>=20
> 1) Require server certificate validation by the client.  This is the
> traditional solution, but in practice it is quite weak because clients
> either trust a gazillion of CAs out there (which is a bad idea) that
> easily allows a MITM to acquire a fraudulent certificate or the
end-user
> manually inspects the certificate and clicks OK (in practice probably
> even worse given that end-users can't be trusted with these decisions
in
> general).  There is no automatic and secure way for clients to learn
the
> trust anchor used by the server.
>=20
> 2) Both client and RP hashes (or somehow mixes in) the channel binding
> data into the authentication process in a way that if the channel
> binding data does NOT match, authentication will fail.
>=20
> SASL PLAIN uses approach 1).  SASL SCRAM uses approach 2).  For SCRAM,
> achieving 2 is easy because there is a shared secret between client
and
> server, and each peer can HMAC its channel binding and the peer can
> verify the HMAC and then any MITM is detected.
>=20
> I would like to have support for channel bindings in the OpenID (and
> SAML) SASL mechanism because I believe 2) provides stronger protection
> than 1).  However, I also realize that given how OpenID works, it
_may_
> be impossible without modifying the IdP.  But I don't think we have
> proven that yet.  Maybe we can use the returns_to URL to store the
> channel binding data, and allow Client+RP to parse and validate that.
>=20
> Is this clearer?
>=20
> /Simon

From: Simon Josefsson <simon@josefsson.org>
Subject: Re: OpenID & SASL: Channel Bindings
To: "Hannes Tschofenig" <Hannes.Tschofenig@gmx.net>
Cc: hmauldin@cisco.com,  lear@cisco.com
Date: Mon, 12 Jul 2010 14:38:45 +0200

I don't think RP_TLS_X and RP_OpenID_Y will be the same in practice.
Typically RP_TLS_X will be a SMTP/XMPP server certificate, and
RP_OpenID_Y would be a URL.  If comparison actually is useful (I'm not
sure of that), the document needs to describe how it is done.

Hm.  This discussion reminds me of the OAuth session fixation attack.
Translating the attack into our environment it would work like this:
Client X wants to send an e-mail using OpenID authentication.  Client X
is tricked into talking to attacker Y.  Attacker opens another channel
to real server Z and starts authentication, and forwards the requests
back to Client X.  Client X performs the authentication, which results
in attacker Y is authenticated as Client X to real server Z.

However, this attack is pre-empted by requiring the Client X to validate
that the server it talks to is a trusted server using TLS server
certificate validation.  I don't think it will be particulary strong in
practice -- SMTP clients rarely have trust anchors configured by the
users, but at best rely on pre-configured CAs from the software
manufacturer (which is weak security).  But at least we can claim it is
not a protocol weakness.

I could very well be missing your point... thoughts?

/Simon

"Hannes Tschofenig" <Hannes.Tschofenig@gmx.net> writes:

> Hi Simon,
>
> I noticed that you and Eliot had exchanged some emails about this
> topic already (and they all ended up in my spam folder). I will have
> to browse through them. I will also have to read SASL SCRAM to be sure
> I understand your concern.
>
> However, with the concern you describe it seems that you are more
> looking into something related to cryptographic bindings, see
> http://www.springerlink.com/content/8qv4qncf4fk106pg/.
>
> I doubt that you will easily get around verifying the server provided
> certificate but you could check whether the identities provided at the
> different layers match, namely RP_TLS_X and RP_OpenID_Y. Doing this
> would at least ensure that the end points are the same, which is what
> you are verifying when you hash data from both layers together.
>
> A separate, but important aspect, is to figure out what are the
> pre-conditions for a MITM attack to work and whether the threat is
> high enough to justify a solution.
>
> Ciao Hannes
>
> -------- Original-Nachricht --------
>> Datum: Thu, 08 Jul 2010 15:03:31 +0200 Von: Simon Josefsson
>> <simon@josefsson.org> An: Hannes Tschofenig
>> <Hannes.Tschofenig@gmx.net> CC: Eliot Lear <lear@cisco.com>,
>> hmauldin@cisco.com Betreff: Re: OpenID & SASL: Channel Bindings
>
>> Hannes Tschofenig <Hannes.Tschofenig@gmx.net> writes:
>>=20
>> > Hi all,
>> > I had a brief chat with Eliot two days ago and we ran into the
>> > topic of "channel bindings". I thought I should share my opinion
>> > with you.
>> > First, I have to say that I find the writeup by Nico in RFC 5056
>> > hard to understand. I re-read it today and it still confuses me. I
>> > also have the history of the EAP work in my head, which (due to
>> > the name collision) makes everything just more complicated.
>>  Hi.  Yes, I have also found that document difficult to understand,
>> and I participated in the discussions around it.  I don't envy
>> anyone else trying to understand it.
>>=20
>> > Ignoring what I said in the previous paragraph, let me start
>> > slightly differently to understand what you guys worry about
>> > (based on the figure from the draft):
>>  Let me also say that my answer below is what _I_ worry about: I
don't
>> know for what purpose Nico would like to see support for channel
>> bindings.
>>=20
>> > +-----------+ | | >| Relying | / | Party | // | | // +-----------+
>> > // ^ OpenID // +--|--+ // | O| | / S | p| | // A | e| | // S | n|
>> > | // L | I| | // | D| | </ +--|--+ +------------+ v
>> >       |            | +----------+ OpenID | OpenID | | Provider
>> >       | |<--------------->| Client | | | |
>> >       +------------+ +----------+
>> > We assume that TLS is used between the client and the relying
>> > party with server-side autnentication.  The certificate used for
>> > this purpose will show some identity, let us call it RP_TLS_X.
>>  Right.
>>=20
>> > Then, we might have TLS between the Client and the OpenID provider
>> > but that largely depends on the authentication mechanism being
>> > used.
>>  Yes.  Mentally, I try to ignore what happens outside of the
>> Client<->RP communication because it is authentication mechanism
>> specific.  Whether it uses TLS or not is not relevant.  So to me,
>> the only relevant TLS channel to worry about is the Client<->RP
>> channel.
>>=20
>> > Furthermore, we have authentication of the RP to the OP (and vice
>> > versa) at the OpenID "application layer". The RP will have some
>> > identity, let us call it RP_OpenID_Y.
>> > Now, I believe you guys worry about the linkage between the TLS
>> > channel (used between the client and the RP) and the OpenID
>> > identity used by the RP.
>> > Is this correct?
>>  I'm not sure that is really the same that I worry about.  I worry
>> about a MITM between Client and RP.  There are two ways to detect
>> the MITM:
>>=20
>> 1) Require server certificate validation by the client.  This is the
>> traditional solution, but in practice it is quite weak because
>> clients either trust a gazillion of CAs out there (which is a bad
>> idea) that easily allows a MITM to acquire a fraudulent certificate
>> or the end-user manually inspects the certificate and clicks OK (in
>> practice probably even worse given that end-users can't be trusted
>> with these decisions in general).  There is no automatic and secure
>> way for clients to learn the trust anchor used by the server.
>>=20
>> 2) Both client and RP hashes (or somehow mixes in) the channel
>> binding data into the authentication process in a way that if the
>> channel binding data does NOT match, authentication will fail.
>>=20
>> SASL PLAIN uses approach 1).  SASL SCRAM uses approach 2).  For
>> SCRAM, achieving 2 is easy because there is a shared secret between
>> client and server, and each peer can HMAC its channel binding and
>> the peer can verify the HMAC and then any MITM is detected.
>>=20
>> I would like to have support for channel bindings in the OpenID (and
>> SAML) SASL mechanism because I believe 2) provides stronger
>> protection than 1).  However, I also realize that given how OpenID
>> works, it _may_ be impossible without modifying the IdP.  But I don't
>> think we have proven that yet.  Maybe we can use the returns_to URL
>> to store the channel binding data, and allow Client+RP to parse and
>> validate that.
>>=20
>> Is this clearer?
>>=20
>> /Simon

From: "Hannes Tschofenig" <Hannes.Tschofenig@gmx.net>
Subject: Re: OpenID & SASL: Channel Bindings
To: Simon Josefsson <simon@josefsson.org>
Cc: lear@cisco.com, hmauldin@cisco.com
Date: Tue, 20 Jul 2010 11:53:46 +0200

Hi Simon,=20


> I don't think RP_TLS_X and RP_OpenID_Y will be the same in practice.
> Typically RP_TLS_X will be a SMTP/XMPP server certificate, and
> RP_OpenID_Y would be a URL.  If comparison actually is useful (I'm not
> sure of that), the document needs to describe how it is done.

I also think that this is rather uncommon. While it sounds interesting
in theory it might not be worth going along the path of verifying that
these two identities match.=20

>=20
> Hm.  This discussion reminds me of the OAuth session fixation attack.
> Translating the attack into our environment it would work like this:
> Client X wants to send an e-mail using OpenID authentication.  Client
X
> is tricked into talking to attacker Y.  Attacker opens another channel
> to real server Z and starts authentication, and forwards the requests
> back to Client X.  Client X performs the authentication, which results
> in attacker Y is authenticated as Client X to real server Z.
>=20
> However, this attack is pre-empted by requiring the Client X to
validate
> that the server it talks to is a trusted server using TLS server
> certificate validation.  I don't think it will be particulary strong
in
> practice -- SMTP clients rarely have trust anchors configured by the
> users, but at best rely on pre-configured CAs from the software
> manufacturer (which is weak security).  But at least we can claim it
is
> not a protocol weakness.

I agree with you. The attack will fail if the client indeed knows with
whom they should be talking to, have a certain sets of certs
preconfigured, or use a specific trust anchor only.=20

I still believe it is important to describe this type of attack in the
document clearly to make sure that system architects and developers
understand the assumptions being made.=20

>=20
> I could very well be missing your point... thoughts?
>=20
> /Simon

Ciao
Hannes
>=20
> "Hannes Tschofenig" <Hannes.Tschofenig@gmx.net> writes:
>=20
> > Hi Simon,
> >
> > I noticed that you and Eliot had exchanged some emails about this
> > topic already (and they all ended up in my spam folder). I will have
> > to browse through them. I will also have to read SASL SCRAM to be
sure
> > I understand your concern.
> >
> > However, with the concern you describe it seems that you are more
> > looking into something related to cryptographic bindings, see
> > http://www.springerlink.com/content/8qv4qncf4fk106pg/.
> >
> > I doubt that you will easily get around verifying the server
provided
> > certificate but you could check whether the identities provided at
the
> > different layers match, namely RP_TLS_X and RP_OpenID_Y. Doing this
> > would at least ensure that the end points are the same, which is
what
> > you are verifying when you hash data from both layers together.
> >
> > A separate, but important aspect, is to figure out what are the
> > pre-conditions for a MITM attack to work and whether the threat is
> > high enough to justify a solution.
> >
> > Ciao Hannes
> >
> > -------- Original-Nachricht --------
> >> Datum: Thu, 08 Jul 2010 15:03:31 +0200 Von: Simon Josefsson
> >> <simon@josefsson.org> An: Hannes Tschofenig
> >> <Hannes.Tschofenig@gmx.net> CC: Eliot Lear <lear@cisco.com>,
> >> hmauldin@cisco.com Betreff: Re: OpenID & SASL: Channel Bindings
> >
> >> Hannes Tschofenig <Hannes.Tschofenig@gmx.net> writes:
> >>=20
> >> > Hi all,
> >> > I had a brief chat with Eliot two days ago and we ran into the
> >> > topic of "channel bindings". I thought I should share my opinion
> >> > with you.
> >> > First, I have to say that I find the writeup by Nico in RFC 5056
> >> > hard to understand. I re-read it today and it still confuses me.
I
> >> > also have the history of the EAP work in my head, which (due to
> >> > the name collision) makes everything just more complicated.
> >>  Hi.  Yes, I have also found that document difficult to understand,
> >> and I participated in the discussions around it.  I don't envy
> >> anyone else trying to understand it.
> >>=20
> >> > Ignoring what I said in the previous paragraph, let me start
> >> > slightly differently to understand what you guys worry about
> >> > (based on the figure from the draft):
> >>  Let me also say that my answer below is what _I_ worry about: I
don't
> >> know for what purpose Nico would like to see support for channel
> >> bindings.
> >>=20
> >> > +-----------+ | | >| Relying | / | Party | // | | //
+-----------+
> >> > // ^ OpenID // +--|--+ // | O| | / S | p| | // A | e| | // S | n|
> >> > | // L | I| | // | D| | </ +--|--+ +------------+ v
> >> >       |            | +----------+ OpenID | OpenID | | Provider
> >> >       | |<--------------->| Client | | | |
> >> >       +------------+ +----------+
> >> > We assume that TLS is used between the client and the relying
> >> > party with server-side autnentication.  The certificate used for
> >> > this purpose will show some identity, let us call it RP_TLS_X.
> >>  Right.
> >>=20
> >> > Then, we might have TLS between the Client and the OpenID
provider
> >> > but that largely depends on the authentication mechanism being
> >> > used.
> >>  Yes.  Mentally, I try to ignore what happens outside of the
> >> Client<->RP communication because it is authentication mechanism
> >> specific.  Whether it uses TLS or not is not relevant.  So to me,
> >> the only relevant TLS channel to worry about is the Client<->RP
> >> channel.
> >>=20
> >> > Furthermore, we have authentication of the RP to the OP (and vice
> >> > versa) at the OpenID "application layer". The RP will have some
> >> > identity, let us call it RP_OpenID_Y.
> >> > Now, I believe you guys worry about the linkage between the TLS
> >> > channel (used between the client and the RP) and the OpenID
> >> > identity used by the RP.
> >> > Is this correct?
> >>  I'm not sure that is really the same that I worry about.  I worry
> >> about a MITM between Client and RP.  There are two ways to detect
> >> the MITM:
> >>=20
> >> 1) Require server certificate validation by the client.  This is
the
> >> traditional solution, but in practice it is quite weak because
> >> clients either trust a gazillion of CAs out there (which is a bad
> >> idea) that easily allows a MITM to acquire a fraudulent certificate
> >> or the end-user manually inspects the certificate and clicks OK (in
> >> practice probably even worse given that end-users can't be trusted
> >> with these decisions in general).  There is no automatic and secure
> >> way for clients to learn the trust anchor used by the server.
> >>=20
> >> 2) Both client and RP hashes (or somehow mixes in) the channel
> >> binding data into the authentication process in a way that if the
> >> channel binding data does NOT match, authentication will fail.
> >>=20
> >> SASL PLAIN uses approach 1).  SASL SCRAM uses approach 2).  For
> >> SCRAM, achieving 2 is easy because there is a shared secret between
> >> client and server, and each peer can HMAC its channel binding and
> >> the peer can verify the HMAC and then any MITM is detected.
> >>=20
> >> I would like to have support for channel bindings in the OpenID
(and
> >> SAML) SASL mechanism because I believe 2) provides stronger
> >> protection than 1).  However, I also realize that given how OpenID
> >> works, it _may_ be impossible without modifying the IdP.  But I
don't
> >> think we have proven that yet.  Maybe we can use the returns_to URL
> >> to store the channel binding data, and allow Client+RP to parse and
> >> validate that.
> >>=20
> >> Is this clearer?
> >>=20
> >> /Simon

From: "Hannes Tschofenig" <Hannes.Tschofenig@gmx.net>
Subject: Re: OpenID & SASL: Channel Bindings
To: Simon Josefsson <simon@josefsson.org>
Cc: hmauldin@cisco.com, lear@cisco.com
Date: Sun, 11 Jul 2010 11:24:34 +0200

Hi Simon,=20

now I read SASL SCRAM and I browsed my way through your prior email
exchange with Eliot.=20

SASL SCRAM has a very different motivation for linking the challenge -
response authentication exchange with an underlying TLS (or similar
exchange).=20

First, SASL SCRAM provides a lot of options and that's not particularly
useful for security; although this approach is quite fashionable today.=20
There are at least the following approaches:=20

(1) Challenge Response without any underlying security=20
(2) Challenge Response with underlying TLS (without client- and
server-side exchange, such as a anonymous DH exchange)
(3) Challenge Response with TLS including server-side authentication.=20

We will not further discuss item #1.=20

Now, the assumption for linking the TLS exchange with the SASL
authentication seems to focus on item #2 to ensure that security does
not downgrade to the one of item #1 when an active adversary is present.
The SASL SCRAM specification does not talk about item #3 because it does
not really make sense given that the design could be done entirely
differently if you assume server-side PK-based authentication, i.e. you
can accomplish much better properties if you assume that you can utilize
a PK-based infrastructure.=20

Now, the main question is how this translates to our problem.=20
The usage of OpenID does not really follow the same patterns. Instead
nobody seems to have a problem with the usage of server-side
authentication using TLS nor do we assume that the client and the
relying party use a shared secret either (which would totally void the
entire OpenID / WebSSO model). Finally, I hope that in the SASL SCRAM
model the client does not talk to a random server it has never met
before (at least the server has to be from the same administrative
domain as the password database). This is, however, the model with
OpenID.

In a nutshell, I don't believe the SASL SCRAM channel binding makes a
lot of sense in our use case unless we change some of the fundamental
assumptions of how OpenID is supposed to work.=20

Hence, I would rather go along the road to brainstorming on how to
compare the identities of the two authentication exchanges (only the
RP's identities of course), as I mentioned in my previous mail. Funny
enough even there it is not clear whether we actually want to have the
two identities to match since it would rule out load balancers where the
TLS exchange terminates earlier and the OpenID exchange starts somewhere
in the backend.=20

I hope this makes some sense.=20

Ciao
Hannes

-------- Original-Nachricht --------
> Datum: Thu, 08 Jul 2010 15:03:31 +0200
> Von: Simon Josefsson <simon@josefsson.org>
> An: Hannes Tschofenig <Hannes.Tschofenig@gmx.net>
> CC: Eliot Lear <lear@cisco.com>, hmauldin@cisco.com
> Betreff: Re: OpenID & SASL: Channel Bindings

> Hannes Tschofenig <Hannes.Tschofenig@gmx.net> writes:
>=20
> > Hi all,
> >
> > I had a brief chat with Eliot two days ago and we ran into the topic
> > of "channel bindings". I thought I should share my opinion with you.
> >
> > First, I have to say that I find the writeup by Nico in RFC 5056
hard
> > to understand. I re-read it today and it still confuses me. I also
> > have the history of the EAP work in my head, which (due to the name
> > collision) makes everything just more complicated.
>=20
> Hi.  Yes, I have also found that document difficult to understand, and
I
> participated in the discussions around it.  I don't envy anyone else
> trying to understand it.
>=20
> > Ignoring what I said in the previous paragraph, let me start
slightly
> > differently to understand what you guys worry about (based on the
> > figure from the draft):
>=20
> Let me also say that my answer below is what _I_ worry about: I don't
> know for what purpose Nico would like to see support for channel
> bindings.
>=20
> >
> >                                     +-----------+
> >                                     |           |
> >                                    >|  Relying  |
> >                                   / |  Party    |
> >                                 //  |           |
> >                               //    +-----------+
> >                             //            ^
> >                    OpenID //           +--|--+
> >                         //             | O|  |
> >                        /             S | p|  |
> >                      //              A | e|  |
> >                    //                S | n|  |
> >                  //                  L | I|  |
> >                //                      | D|  |
> >              </                        +--|--+
> >       +------------+                      v
> >       |            |                 +----------+
> >       |  OpenID    |   OpenID        |          |
> >       |  Provider  |<--------------->|  Client  |
> >       |            |                 |          |
> >       +------------+                 +----------+
> >
> > We assume that TLS is used between the client and the relying party
> > with server-side autnentication.  The certificate used for this
> > purpose will show some identity, let us call it RP_TLS_X.
>=20
> Right.
>=20
> > Then, we might have TLS between the Client and the OpenID provider
but
> > that largely depends on the authentication mechanism being used.
>=20
> Yes.  Mentally, I try to ignore what happens outside of the
Client<->RP
> communication because it is authentication mechanism specific.
Whether
> it uses TLS or not is not relevant.  So to me, the only relevant TLS
> channel to worry about is the Client<->RP channel.
>=20
> > Furthermore, we have authentication of the RP to the OP (and vice
> > versa) at the OpenID "application layer". The RP will have some
> > identity, let us call it RP_OpenID_Y.
> >
> > Now, I believe you guys worry about the linkage between the TLS
> > channel (used between the client and the RP) and the OpenID identity
> > used by the RP.
> >
> > Is this correct?
>=20
> I'm not sure that is really the same that I worry about.  I worry
about
> a MITM between Client and RP.  There are two ways to detect the MITM:
>=20
> 1) Require server certificate validation by the client.  This is the
> traditional solution, but in practice it is quite weak because clients
> either trust a gazillion of CAs out there (which is a bad idea) that
> easily allows a MITM to acquire a fraudulent certificate or the
end-user
> manually inspects the certificate and clicks OK (in practice probably
> even worse given that end-users can't be trusted with these decisions
in
> general).  There is no automatic and secure way for clients to learn
the
> trust anchor used by the server.
>=20
> 2) Both client and RP hashes (or somehow mixes in) the channel binding
> data into the authentication process in a way that if the channel
> binding data does NOT match, authentication will fail.
>=20
> SASL PLAIN uses approach 1).  SASL SCRAM uses approach 2).  For SCRAM,
> achieving 2 is easy because there is a shared secret between client
and
> server, and each peer can HMAC its channel binding and the peer can
> verify the HMAC and then any MITM is detected.
>=20
> I would like to have support for channel bindings in the OpenID (and
> SAML) SASL mechanism because I believe 2) provides stronger protection
> than 1).  However, I also realize that given how OpenID works, it
_may_
> be impossible without modifying the IdP.  But I don't think we have
> proven that yet.  Maybe we can use the returns_to URL to store the
> channel binding data, and allow Client+RP to parse and validate that.
>=20
> Is this clearer?
>=20
> /Simon

From: Simon Josefsson <simon@josefsson.org>
Subject: Re: OpenID & SASL: Channel Bindings
To: "Hannes Tschofenig" <Hannes.Tschofenig@gmx.net>
Cc: hmauldin@cisco.com,  lear@cisco.com
Date: Mon, 12 Jul 2010 14:27:09 +0200

Hi Hannes,

I agree with your analysis, which is why I suggested adding the
following section to the document (and Eliot added it for -01):

1.2.  Applicability

   Because this mechanism transports information that should not be
   controlled by an attacker, the OpenID mechanism MUST only be used
   over channels protected by TLS [RFC5246], and the client MUST
   successfully validate the server certificate, or similar integrity
   protected and authenticated channels.

And further:

   The "gs2-cb- flag" MUST be "n" because channel binding is not
   supported by this mechanism.

This puts OpenID-SASL in the (3) case.  Doesn't this solve your concern?

I still haven't ruled out the possibility that we can negotiate channel
binding through OpenID, if that is really is required.  I'll try to
describe one way to do this, so we can all analyze it further.  If
possible, this would support both (2) and (3) approaches, which is more
flexible, but it isn't clear if that flexibility is actually needed.  I
prefer being conservative in adding features, so if it is not needed,
let's not add it, at least not initially.

I'll try to read your other e-mail about matching RP identities.  I'm
not sure I follow that part.

/Simon

"Hannes Tschofenig" <Hannes.Tschofenig@gmx.net> writes:

> Hi Simon,
>
> now I read SASL SCRAM and I browsed my way through your prior email
> exchange with Eliot.
>
> SASL SCRAM has a very different motivation for linking the challenge -
> response authentication exchange with an underlying TLS (or similar
> exchange).
>
> First, SASL SCRAM provides a lot of options and that's not
> particularly useful for security; although this approach is quite
> fashionable today.  There are at least the following approaches:
>
> (1) Challenge Response without any underlying security (2) Challenge
> Response with underlying TLS (without client- and server-side
> exchange, such as a anonymous DH exchange) (3) Challenge Response with
> TLS including server-side authentication.
>
> We will not further discuss item #1.
>
> Now, the assumption for linking the TLS exchange with the SASL
> authentication seems to focus on item #2 to ensure that security does
> not downgrade to the one of item #1 when an active adversary is
> present. The SASL SCRAM specification does not talk about item #3
> because it does not really make sense given that the design could be
> done entirely differently if you assume server-side PK-based
> authentication, i.e. you can accomplish much better properties if you
> assume that you can utilize a PK-based infrastructure.
>
> Now, the main question is how this translates to our problem.  The
> usage of OpenID does not really follow the same patterns. Instead
> nobody seems to have a problem with the usage of server-side
> authentication using TLS nor do we assume that the client and the
> relying party use a shared secret either (which would totally void the
> entire OpenID / WebSSO model). Finally, I hope that in the SASL SCRAM
> model the client does not talk to a random server it has never met
> before (at least the server has to be from the same administrative
> domain as the password database). This is, however, the model with
> OpenID.
>
> In a nutshell, I don't believe the SASL SCRAM channel binding makes a
> lot of sense in our use case unless we change some of the fundamental
> assumptions of how OpenID is supposed to work.
>
> Hence, I would rather go along the road to brainstorming on how to
> compare the identities of the two authentication exchanges (only the
> RP's identities of course), as I mentioned in my previous mail. Funny
> enough even there it is not clear whether we actually want to have the
> two identities to match since it would rule out load balancers where
> the TLS exchange terminates earlier and the OpenID exchange starts
> somewhere in the backend.
>
> I hope this makes some sense.
>
> Ciao Hannes
>
> -------- Original-Nachricht --------
>> Datum: Thu, 08 Jul 2010 15:03:31 +0200 Von: Simon Josefsson
>> <simon@josefsson.org> An: Hannes Tschofenig
>> <Hannes.Tschofenig@gmx.net> CC: Eliot Lear <lear@cisco.com>,
>> hmauldin@cisco.com Betreff: Re: OpenID & SASL: Channel Bindings
>
>> Hannes Tschofenig <Hannes.Tschofenig@gmx.net> writes:
>>=20
>> > Hi all,
>> > I had a brief chat with Eliot two days ago and we ran into the
>> > topic of "channel bindings". I thought I should share my opinion
>> > with you.
>> > First, I have to say that I find the writeup by Nico in RFC 5056
>> > hard to understand. I re-read it today and it still confuses me. I
>> > also have the history of the EAP work in my head, which (due to
>> > the name collision) makes everything just more complicated.
>>  Hi.  Yes, I have also found that document difficult to understand,
>> and I participated in the discussions around it.  I don't envy
>> anyone else trying to understand it.
>>=20
>> > Ignoring what I said in the previous paragraph, let me start
>> > slightly differently to understand what you guys worry about
>> > (based on the figure from the draft):
>>  Let me also say that my answer below is what _I_ worry about: I
don't
>> know for what purpose Nico would like to see support for channel
>> bindings.
>>=20
>> > +-----------+ | | >| Relying | / | Party | // | | // +-----------+
>> > // ^ OpenID // +--|--+ // | O| | / S | p| | // A | e| | // S | n|
>> > | // L | I| | // | D| | </ +--|--+ +------------+ v
>> >       |            | +----------+ OpenID | OpenID | | Provider
>> >       | |<--------------->| Client | | | |
>> >       +------------+ +----------+
>> > We assume that TLS is used between the client and the relying
>> > party with server-side autnentication.  The certificate used for
>> > this purpose will show some identity, let us call it RP_TLS_X.
>>  Right.
>>=20
>> > Then, we might have TLS between the Client and the OpenID provider
>> > but that largely depends on the authentication mechanism being
>> > used.
>>  Yes.  Mentally, I try to ignore what happens outside of the
>> Client<->RP communication because it is authentication mechanism
>> specific.  Whether it uses TLS or not is not relevant.  So to me,
>> the only relevant TLS channel to worry about is the Client<->RP
>> channel.
>>=20
>> > Furthermore, we have authentication of the RP to the OP (and vice
>> > versa) at the OpenID "application layer". The RP will have some
>> > identity, let us call it RP_OpenID_Y.
>> > Now, I believe you guys worry about the linkage between the TLS
>> > channel (used between the client and the RP) and the OpenID
>> > identity used by the RP.
>> > Is this correct?
>>  I'm not sure that is really the same that I worry about.  I worry
>> about a MITM between Client and RP.  There are two ways to detect
>> the MITM:
>>=20
>> 1) Require server certificate validation by the client.  This is the
>> traditional solution, but in practice it is quite weak because
>> clients either trust a gazillion of CAs out there (which is a bad
>> idea) that easily allows a MITM to acquire a fraudulent certificate
>> or the end-user manually inspects the certificate and clicks OK (in
>> practice probably even worse given that end-users can't be trusted
>> with these decisions in general).  There is no automatic and secure
>> way for clients to learn the trust anchor used by the server.
>>=20
>> 2) Both client and RP hashes (or somehow mixes in) the channel
>> binding data into the authentication process in a way that if the
>> channel binding data does NOT match, authentication will fail.
>>=20
>> SASL PLAIN uses approach 1).  SASL SCRAM uses approach 2).  For
>> SCRAM, achieving 2 is easy because there is a shared secret between
>> client and server, and each peer can HMAC its channel binding and
>> the peer can verify the HMAC and then any MITM is detected.
>>=20
>> I would like to have support for channel bindings in the OpenID (and
>> SAML) SASL mechanism because I believe 2) provides stronger
>> protection than 1).  However, I also realize that given how OpenID
>> works, it _may_ be impossible without modifying the IdP.  But I don't
>> think we have proven that yet.  Maybe we can use the returns_to URL
>> to store the channel binding data, and allow Client+RP to parse and
>> validate that.
>>=20
>> Is this clearer?
>>=20
>> /Simon

From: Eliot Lear <lear@cisco.com>
Subject: Re: OpenID & SASL: Channel Bindings
To: Simon Josefsson <simon@josefsson.org>
CC: Hannes Tschofenig <Hannes.Tschofenig@gmx.net>, hmauldin@cisco.com
Date: Mon, 12 Jul 2010 14:37:26 +0200

 Simon, Hannes,

Excellent discussion.  I think we're all in agreement.  There is now a
-02 fyi which corrects one missed OPENID2.0.  I would like to propose
meeting Sunday night.  Does this work for everyone?

Eliot

On 7/12/10 2:27 PM, Simon Josefsson wrote:
> Hi Hannes,
>
> I agree with your analysis, which is why I suggested adding the
> following section to the document (and Eliot added it for -01):
>
> 1.2.  Applicability
>
>    Because this mechanism transports information that should not be
>    controlled by an attacker, the OpenID mechanism MUST only be used
>    over channels protected by TLS [RFC5246], and the client MUST
>    successfully validate the server certificate, or similar integrity
>    protected and authenticated channels.
>
> And further:
>
>    The "gs2-cb- flag" MUST be "n" because channel binding is not
>    supported by this mechanism.
>
> This puts OpenID-SASL in the (3) case.  Doesn't this solve your
concern?
>
> I still haven't ruled out the possibility that we can negotiate
channel
> binding through OpenID, if that is really is required.  I'll try to
> describe one way to do this, so we can all analyze it further.  If
> possible, this would support both (2) and (3) approaches, which is
more
> flexible, but it isn't clear if that flexibility is actually needed.
I
> prefer being conservative in adding features, so if it is not needed,
> let's not add it, at least not initially.
>
> I'll try to read your other e-mail about matching RP identities.  I'm
> not sure I follow that part.
>
> /Simon
>
> "Hannes Tschofenig" <Hannes.Tschofenig@gmx.net> writes:
>
>> Hi Simon,
>>
>> now I read SASL SCRAM and I browsed my way through your prior email
>> exchange with Eliot.
>>
>> SASL SCRAM has a very different motivation for linking the challenge
-
>> response authentication exchange with an underlying TLS (or similar
>> exchange).
>>
>> First, SASL SCRAM provides a lot of options and that's not
>> particularly useful for security; although this approach is quite
>> fashionable today.  There are at least the following approaches:
>>
>> (1) Challenge Response without any underlying security (2) Challenge
>> Response with underlying TLS (without client- and server-side
>> exchange, such as a anonymous DH exchange) (3) Challenge Response
with
>> TLS including server-side authentication.
>>
>> We will not further discuss item #1.
>>
>> Now, the assumption for linking the TLS exchange with the SASL
>> authentication seems to focus on item #2 to ensure that security does
>> not downgrade to the one of item #1 when an active adversary is
>> present. The SASL SCRAM specification does not talk about item #3
>> because it does not really make sense given that the design could be
>> done entirely differently if you assume server-side PK-based
>> authentication, i.e. you can accomplish much better properties if you
>> assume that you can utilize a PK-based infrastructure.
>>
>> Now, the main question is how this translates to our problem.  The
>> usage of OpenID does not really follow the same patterns. Instead
>> nobody seems to have a problem with the usage of server-side
>> authentication using TLS nor do we assume that the client and the
>> relying party use a shared secret either (which would totally void
the
>> entire OpenID / WebSSO model). Finally, I hope that in the SASL SCRAM
>> model the client does not talk to a random server it has never met
>> before (at least the server has to be from the same administrative
>> domain as the password database). This is, however, the model with
>> OpenID.
>>
>> In a nutshell, I don't believe the SASL SCRAM channel binding makes a
>> lot of sense in our use case unless we change some of the fundamental
>> assumptions of how OpenID is supposed to work.
>>
>> Hence, I would rather go along the road to brainstorming on how to
>> compare the identities of the two authentication exchanges (only the
>> RP's identities of course), as I mentioned in my previous mail. Funny
>> enough even there it is not clear whether we actually want to have
the
>> two identities to match since it would rule out load balancers where
>> the TLS exchange terminates earlier and the OpenID exchange starts
>> somewhere in the backend.
>>
>> I hope this makes some sense.
>>
>> Ciao Hannes
>>
>> -------- Original-Nachricht --------
>>> Datum: Thu, 08 Jul 2010 15:03:31 +0200 Von: Simon Josefsson
>>> <simon@josefsson.org> An: Hannes Tschofenig
>>> <Hannes.Tschofenig@gmx.net> CC: Eliot Lear <lear@cisco.com>,
>>> hmauldin@cisco.com Betreff: Re: OpenID & SASL: Channel Bindings
>>> Hannes Tschofenig <Hannes.Tschofenig@gmx.net> writes:
>>>
>>>> Hi all,
>>>> I had a brief chat with Eliot two days ago and we ran into the
>>>> topic of "channel bindings". I thought I should share my opinion
>>>> with you.
>>>> First, I have to say that I find the writeup by Nico in RFC 5056
>>>> hard to understand. I re-read it today and it still confuses me. I
>>>> also have the history of the EAP work in my head, which (due to
>>>> the name collision) makes everything just more complicated.
>>>  Hi.  Yes, I have also found that document difficult to understand,
>>> and I participated in the discussions around it.  I don't envy
>>> anyone else trying to understand it.
>>>
>>>> Ignoring what I said in the previous paragraph, let me start
>>>> slightly differently to understand what you guys worry about
>>>> (based on the figure from the draft):
>>>  Let me also say that my answer below is what _I_ worry about: I
don't
>>> know for what purpose Nico would like to see support for channel
>>> bindings.
>>>
>>>> +-----------+ | | >| Relying | / | Party | // | | // +-----------+
>>>> // ^ OpenID // +--|--+ // | O| | / S | p| | // A | e| | // S | n|
>>>> | // L | I| | // | D| | </ +--|--+ +------------+ v
>>>>       |            | +----------+ OpenID | OpenID | | Provider
>>>>       | |<--------------->| Client | | | |
>>>>       +------------+ +----------+
>>>> We assume that TLS is used between the client and the relying
>>>> party with server-side autnentication.  The certificate used for
>>>> this purpose will show some identity, let us call it RP_TLS_X.
>>>  Right.
>>>
>>>> Then, we might have TLS between the Client and the OpenID provider
>>>> but that largely depends on the authentication mechanism being
>>>> used.
>>>  Yes.  Mentally, I try to ignore what happens outside of the
>>> Client<->RP communication because it is authentication mechanism
>>> specific.  Whether it uses TLS or not is not relevant.  So to me,
>>> the only relevant TLS channel to worry about is the Client<->RP
>>> channel.
>>>
>>>> Furthermore, we have authentication of the RP to the OP (and vice
>>>> versa) at the OpenID "application layer". The RP will have some
>>>> identity, let us call it RP_OpenID_Y.
>>>> Now, I believe you guys worry about the linkage between the TLS
>>>> channel (used between the client and the RP) and the OpenID
>>>> identity used by the RP.
>>>> Is this correct?
>>>  I'm not sure that is really the same that I worry about.  I worry
>>> about a MITM between Client and RP.  There are two ways to detect
>>> the MITM:
>>>
>>> 1) Require server certificate validation by the client.  This is the
>>> traditional solution, but in practice it is quite weak because
>>> clients either trust a gazillion of CAs out there (which is a bad
>>> idea) that easily allows a MITM to acquire a fraudulent certificate
>>> or the end-user manually inspects the certificate and clicks OK (in
>>> practice probably even worse given that end-users can't be trusted
>>> with these decisions in general).  There is no automatic and secure
>>> way for clients to learn the trust anchor used by the server.
>>>
>>> 2) Both client and RP hashes (or somehow mixes in) the channel
>>> binding data into the authentication process in a way that if the
>>> channel binding data does NOT match, authentication will fail.
>>>
>>> SASL PLAIN uses approach 1).  SASL SCRAM uses approach 2).  For
>>> SCRAM, achieving 2 is easy because there is a shared secret between
>>> client and server, and each peer can HMAC its channel binding and
>>> the peer can verify the HMAC and then any MITM is detected.
>>>
>>> I would like to have support for channel bindings in the OpenID (and
>>> SAML) SASL mechanism because I believe 2) provides stronger
>>> protection than 1).  However, I also realize that given how OpenID
>>> works, it _may_ be impossible without modifying the IdP.  But I
don't
>>> think we have proven that yet.  Maybe we can use the returns_to URL
>>> to store the channel binding data, and allow Client+RP to parse and
>>> validate that.
>>>
>>> Is this clearer?
>>>
>>> /Simon

From: "Hannes Tschofenig" <Hannes.Tschofenig@gmx.net>
Subject: Re: OpenID & SASL: Channel Bindings
To: Simon Josefsson <simon@josefsson.org>
Cc: lear@cisco.com, hmauldin@cisco.com
Date: Tue, 20 Jul 2010 11:45:23 +0200

This text is good for me.=20

-------- Original-Nachricht --------
> Datum: Mon, 12 Jul 2010 14:27:09 +0200
> Von: Simon Josefsson <simon@josefsson.org>
> An: "Hannes Tschofenig" <Hannes.Tschofenig@gmx.net>
> CC: hmauldin@cisco.com, lear@cisco.com
> Betreff: Re: OpenID & SASL: Channel Bindings

> Hi Hannes,
>=20
> I agree with your analysis, which is why I suggested adding the
> following section to the document (and Eliot added it for -01):
>=20
> 1.2.  Applicability
>=20
>    Because this mechanism transports information that should not be
>    controlled by an attacker, the OpenID mechanism MUST only be used
>    over channels protected by TLS [RFC5246], and the client MUST
>    successfully validate the server certificate, or similar integrity
>    protected and authenticated channels.
>=20
> And further:
>=20
>    The "gs2-cb- flag" MUST be "n" because channel binding is not
>    supported by this mechanism.
>=20
> This puts OpenID-SASL in the (3) case.  Doesn't this solve your
concern?
>=20
> I still haven't ruled out the possibility that we can negotiate
channel
> binding through OpenID, if that is really is required.  I'll try to
> describe one way to do this, so we can all analyze it further.  If
> possible, this would support both (2) and (3) approaches, which is
more
> flexible, but it isn't clear if that flexibility is actually needed.
I
> prefer being conservative in adding features, so if it is not needed,
> let's not add it, at least not initially.
>=20
> I'll try to read your other e-mail about matching RP identities.  I'm
> not sure I follow that part.
>=20
> /Simon
>=20
> "Hannes Tschofenig" <Hannes.Tschofenig@gmx.net> writes:
>=20
> > Hi Simon,
> >
> > now I read SASL SCRAM and I browsed my way through your prior email
> > exchange with Eliot.
> >
> > SASL SCRAM has a very different motivation for linking the challenge
-
> > response authentication exchange with an underlying TLS (or similar
> > exchange).
> >
> > First, SASL SCRAM provides a lot of options and that's not
> > particularly useful for security; although this approach is quite
> > fashionable today.  There are at least the following approaches:
> >
> > (1) Challenge Response without any underlying security (2) Challenge
> > Response with underlying TLS (without client- and server-side
> > exchange, such as a anonymous DH exchange) (3) Challenge Response
with
> > TLS including server-side authentication.
> >
> > We will not further discuss item #1.
> >
> > Now, the assumption for linking the TLS exchange with the SASL
> > authentication seems to focus on item #2 to ensure that security
does
> > not downgrade to the one of item #1 when an active adversary is
> > present. The SASL SCRAM specification does not talk about item #3
> > because it does not really make sense given that the design could be
> > done entirely differently if you assume server-side PK-based
> > authentication, i.e. you can accomplish much better properties if
you
> > assume that you can utilize a PK-based infrastructure.
> >
> > Now, the main question is how this translates to our problem.  The
> > usage of OpenID does not really follow the same patterns. Instead
> > nobody seems to have a problem with the usage of server-side
> > authentication using TLS nor do we assume that the client and the
> > relying party use a shared secret either (which would totally void
the
> > entire OpenID / WebSSO model). Finally, I hope that in the SASL
SCRAM
> > model the client does not talk to a random server it has never met
> > before (at least the server has to be from the same administrative
> > domain as the password database). This is, however, the model with
> > OpenID.
> >
> > In a nutshell, I don't believe the SASL SCRAM channel binding makes
a
> > lot of sense in our use case unless we change some of the
fundamental
> > assumptions of how OpenID is supposed to work.
> >
> > Hence, I would rather go along the road to brainstorming on how to
> > compare the identities of the two authentication exchanges (only the
> > RP's identities of course), as I mentioned in my previous mail.
Funny
> > enough even there it is not clear whether we actually want to have
the
> > two identities to match since it would rule out load balancers where
> > the TLS exchange terminates earlier and the OpenID exchange starts
> > somewhere in the backend.
> >
> > I hope this makes some sense.
> >
> > Ciao Hannes
> >
> > -------- Original-Nachricht --------
> >> Datum: Thu, 08 Jul 2010 15:03:31 +0200 Von: Simon Josefsson
> >> <simon@josefsson.org> An: Hannes Tschofenig
> >> <Hannes.Tschofenig@gmx.net> CC: Eliot Lear <lear@cisco.com>,
> >> hmauldin@cisco.com Betreff: Re: OpenID & SASL: Channel Bindings
> >
> >> Hannes Tschofenig <Hannes.Tschofenig@gmx.net> writes:
> >>=20
> >> > Hi all,
> >> > I had a brief chat with Eliot two days ago and we ran into the
> >> > topic of "channel bindings". I thought I should share my opinion
> >> > with you.
> >> > First, I have to say that I find the writeup by Nico in RFC 5056
> >> > hard to understand. I re-read it today and it still confuses me.
I
> >> > also have the history of the EAP work in my head, which (due to
> >> > the name collision) makes everything just more complicated.
> >>  Hi.  Yes, I have also found that document difficult to understand,
> >> and I participated in the discussions around it.  I don't envy
> >> anyone else trying to understand it.
> >>=20
> >> > Ignoring what I said in the previous paragraph, let me start
> >> > slightly differently to understand what you guys worry about
> >> > (based on the figure from the draft):
> >>  Let me also say that my answer below is what _I_ worry about: I
don't
> >> know for what purpose Nico would like to see support for channel
> >> bindings.
> >>=20
> >> > +-----------+ | | >| Relying | / | Party | // | | //
+-----------+
> >> > // ^ OpenID // +--|--+ // | O| | / S | p| | // A | e| | // S | n|
> >> > | // L | I| | // | D| | </ +--|--+ +------------+ v
> >> >       |            | +----------+ OpenID | OpenID | | Provider
> >> >       | |<--------------->| Client | | | |
> >> >       +------------+ +----------+
> >> > We assume that TLS is used between the client and the relying
> >> > party with server-side autnentication.  The certificate used for
> >> > this purpose will show some identity, let us call it RP_TLS_X.
> >>  Right.
> >>=20
> >> > Then, we might have TLS between the Client and the OpenID
provider
> >> > but that largely depends on the authentication mechanism being
> >> > used.
> >>  Yes.  Mentally, I try to ignore what happens outside of the
> >> Client<->RP communication because it is authentication mechanism
> >> specific.  Whether it uses TLS or not is not relevant.  So to me,
> >> the only relevant TLS channel to worry about is the Client<->RP
> >> channel.
> >>=20
> >> > Furthermore, we have authentication of the RP to the OP (and vice
> >> > versa) at the OpenID "application layer". The RP will have some
> >> > identity, let us call it RP_OpenID_Y.
> >> > Now, I believe you guys worry about the linkage between the TLS
> >> > channel (used between the client and the RP) and the OpenID
> >> > identity used by the RP.
> >> > Is this correct?
> >>  I'm not sure that is really the same that I worry about.  I worry
> >> about a MITM between Client and RP.  There are two ways to detect
> >> the MITM:
> >>=20
> >> 1) Require server certificate validation by the client.  This is
the
> >> traditional solution, but in practice it is quite weak because
> >> clients either trust a gazillion of CAs out there (which is a bad
> >> idea) that easily allows a MITM to acquire a fraudulent certificate
> >> or the end-user manually inspects the certificate and clicks OK (in
> >> practice probably even worse given that end-users can't be trusted
> >> with these decisions in general).  There is no automatic and secure
> >> way for clients to learn the trust anchor used by the server.
> >>=20
> >> 2) Both client and RP hashes (or somehow mixes in) the channel
> >> binding data into the authentication process in a way that if the
> >> channel binding data does NOT match, authentication will fail.
> >>=20
> >> SASL PLAIN uses approach 1).  SASL SCRAM uses approach 2).  For
> >> SCRAM, achieving 2 is easy because there is a shared secret between
> >> client and server, and each peer can HMAC its channel binding and
> >> the peer can verify the HMAC and then any MITM is detected.
> >>=20
> >> I would like to have support for channel bindings in the OpenID
(and
> >> SAML) SASL mechanism because I believe 2) provides stronger
> >> protection than 1).  However, I also realize that given how OpenID
> >> works, it _may_ be impossible without modifying the IdP.  But I
don't
> >> think we have proven that yet.  Maybe we can use the returns_to URL
> >> to store the channel binding data, and allow Client+RP to parse and
> >> validate that.
> >>=20
> >> Is this clearer?
> >>=20
> >> /Simon


From Nicolas.Williams@oracle.com  Mon Aug 16 07:42:48 2010
Return-Path: <Nicolas.Williams@oracle.com>
X-Original-To: kitten@core3.amsl.com
Delivered-To: kitten@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id DE4A63A67FA for <kitten@core3.amsl.com>; Mon, 16 Aug 2010 07:42:48 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -6.542
X-Spam-Level: 
X-Spam-Status: No, score=-6.542 tagged_above=-999 required=5 tests=[AWL=0.056,  BAYES_00=-2.599, RCVD_IN_DNSWL_MED=-4, UNPARSEABLE_RELAY=0.001]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id aX8HTX-bJjYi for <kitten@core3.amsl.com>; Mon, 16 Aug 2010 07:42:48 -0700 (PDT)
Received: from rcsinet10.oracle.com (rcsinet10.oracle.com [148.87.113.121]) by core3.amsl.com (Postfix) with ESMTP id 015763A67AF for <kitten@ietf.org>; Mon, 16 Aug 2010 07:42:47 -0700 (PDT)
Received: from acsinet15.oracle.com (acsinet15.oracle.com [141.146.126.227]) by rcsinet10.oracle.com (Switch-3.4.2/Switch-3.4.2) with ESMTP id o7GEhL98032690 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=OK); Mon, 16 Aug 2010 14:43:23 GMT
Received: from acsmt354.oracle.com (acsmt354.oracle.com [141.146.40.154]) by acsinet15.oracle.com (Switch-3.4.2/Switch-3.4.1) with ESMTP id o7G31pSu005199; Mon, 16 Aug 2010 14:43:21 GMT
Received: from abhmt006.oracle.com by acsmt355.oracle.com with ESMTP id 499432341281969712; Mon, 16 Aug 2010 07:41:52 -0700
Received: from oracle.com (/129.153.128.104) by default (Oracle Beehive Gateway v4.0) with ESMTP ; Mon, 16 Aug 2010 07:41:51 -0700
Date: Mon, 16 Aug 2010 09:41:46 -0500
From: Nicolas Williams <Nicolas.Williams@oracle.com>
To: Sam Hartman <hartmans-ietf@mit.edu>
Message-ID: <20100816144144.GB14409@oracle.com>
References: <tsllj8dyo5p.fsf@mit.edu> <4C6460D6.7020800@stpeter.im> <tsl1va3v7z5.fsf@mit.edu> <4C64698A.9030405@isode.com> <3D3C75174CB95F42AD6BCC56E5555B4502E9BF85@FIESEXC015.nsn-intra.net> <20100813152954.GF5217@oracle.com> <4C683150.6000003@gmx.net> <20100816041750.GT14409@oracle.com> <tslvd7aofq6.fsf_-_@mit.edu> <20100816142042.GA14409@oracle.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <20100816142042.GA14409@oracle.com>
User-Agent: Mutt/1.5.20 (2010-03-02)
Cc: kitten@ietf.org
Subject: Re: [kitten] Mutual auth
X-BeenThere: kitten@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Common Authentication Technologies - Next Generation <kitten.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/kitten>
List-Post: <mailto:kitten@ietf.org>
List-Help: <mailto:kitten-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=subscribe>
X-List-Received-Date: Mon, 16 Aug 2010 14:42:49 -0000

On Mon, Aug 16, 2010 at 09:20:43AM -0500, Nicolas Williams wrote:
> On Mon, Aug 16, 2010 at 09:21:37AM -0400, Sam Hartman wrote:
> > I'm not sure that you can extend this to a multi-party protocol without
> > introducing naming.

I forgot to answer this.

First, in SCRAM we do name the acceptor, but because the acceptor
doesn't have any kind of credentials aside from those shared with the
initiator... the acceptor isn't authenticated as strongly as the
initiator.

Perhaps if we had insisted that the acceptor name be added to the
password salting process, then we could say that the acceptor is
strongly authenticated (as strongly as the initiator anyways).  But
there was a lot of resistance to putting in anything like an acceptor
name, or realm (or domain) in particular.

Now, to extend PSK-like mechanisms to a multi-party protocol...  I don't
see how to do it without adding the acceptor name into the salting
process.  If we do then we definitely have acceptor naming.  If we don't
then you've no idea who an initiator's peer is.

Nico
-- 

From Nicolas.Williams@oracle.com  Mon Aug 16 08:09:02 2010
Return-Path: <Nicolas.Williams@oracle.com>
X-Original-To: kitten@core3.amsl.com
Delivered-To: kitten@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id AE71E3A67FD for <kitten@core3.amsl.com>; Mon, 16 Aug 2010 08:09:02 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -6.542
X-Spam-Level: 
X-Spam-Status: No, score=-6.542 tagged_above=-999 required=5 tests=[AWL=0.056,  BAYES_00=-2.599, RCVD_IN_DNSWL_MED=-4, UNPARSEABLE_RELAY=0.001]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id 14Qe+lkRjvp0 for <kitten@core3.amsl.com>; Mon, 16 Aug 2010 08:09:01 -0700 (PDT)
Received: from rcsinet10.oracle.com (rcsinet10.oracle.com [148.87.113.121]) by core3.amsl.com (Postfix) with ESMTP id 55A353A67AF for <kitten@ietf.org>; Mon, 16 Aug 2010 08:09:01 -0700 (PDT)
Received: from rcsinet13.oracle.com (rcsinet13.oracle.com [148.87.113.125]) by rcsinet10.oracle.com (Switch-3.4.2/Switch-3.4.2) with ESMTP id o7GF9awO006058 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=OK); Mon, 16 Aug 2010 15:09:37 GMT
Received: from acsmt355.oracle.com (acsmt355.oracle.com [141.146.40.155]) by rcsinet13.oracle.com (Switch-3.4.2/Switch-3.4.1) with ESMTP id o7GELEbe028772; Mon, 16 Aug 2010 15:09:35 GMT
Received: from abhmt019.oracle.com by acsmt354.oracle.com with ESMTP id 520757331281971372; Mon, 16 Aug 2010 08:09:32 -0700
Received: from oracle.com (/129.153.128.104) by default (Oracle Beehive Gateway v4.0) with ESMTP ; Mon, 16 Aug 2010 08:09:31 -0700
Date: Mon, 16 Aug 2010 10:09:24 -0500
From: Nicolas Williams <Nicolas.Williams@oracle.com>
To: "Tschofenig, Hannes (NSN - FI/Espoo)" <hannes.tschofenig@nsn.com>
Message-ID: <20100816150915.GC14409@oracle.com>
References: <3D3C75174CB95F42AD6BCC56E5555B4502E9C89F@FIESEXC015.nsn-intra.net>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <3D3C75174CB95F42AD6BCC56E5555B4502E9C89F@FIESEXC015.nsn-intra.net>
User-Agent: Mutt/1.5.20 (2010-03-02)
Cc: kitten@ietf.org
Subject: Re: [kitten] Channel Bindings Discussion
X-BeenThere: kitten@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Common Authentication Technologies - Next Generation <kitten.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/kitten>
List-Post: <mailto:kitten@ietf.org>
List-Help: <mailto:kitten-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=subscribe>
X-List-Received-Date: Mon, 16 Aug 2010 15:09:02 -0000

On Mon, Aug 16, 2010 at 05:40:08PM +0300, Tschofenig, Hannes (NSN - FI/Espoo) wrote:
> I had a brief chat with Eliot two days ago and we ran into the topic
> of "channel bindings". I thought I should share my opinion with you.
> 
> First, I have to say that I find the writeup by Nico in RFC 5056 hard
> to understand. I re-read it today and it still confuses me. I also
> have the history of the EAP work in my head, which (due to the name
> collision) makes everything just more complicated.

Certainly the EAP terminology collision doesn't help.  You do have to
understand a littled bit about the GSS-API to understand RFC5056, even
though RFC5056 was written to be more generic than that.  Specifically
you have to understand that the only participants that matter for
RFC5056 are the initiator (client) and acceptor (server) -- any
infrastructure services are abstrated out.

You seem to be confused about what "application layer" means in the
RFC5056 context.  If you understand that only the client<->server
communications are in scope, then you can see that "application layer"
refers to "the layer above the secure channel" -- in this case, "the
layer above HTTPS / TLS".  See below.

> Ignoring what I said in the previous paragraph, let me start slightly
> differently to understand what you guys worry about (based on the
> figure from the draft):
> 
>                                     +-----------+
>                                     |           |
>                                    >|  Relying  |
>                                   / |  Party    |
>                                 //  |           |
>                               //    +-----------+
>                             //            ^
>                    OpenID //           +--|--+
>                         //             | O|  |
>                        /             S | p|  |
>                      //              A | e|  |
>                    //                S | n|  |
>                  //                  L | I|  |
>                //                      | D|  |
>              </                        +--|--+
>       +------------+                      v
>       |            |                 +----------+
>       |  OpenID    |   OpenID        |          |
>       |  Provider  |<--------------->|  Client  |
>       |            |                 |          |
>       +------------+                 +----------+
> 
> We assume that TLS is used between the client and the relying party with
> server-side autnentication.
> The certificate used for this purpose will show some identity, let us
> call it RP_TLS_X.

The question is: does OpenID rely on TLS for authenticating the RP to
the client?  If so then channel binding will be irrelevant to you,
unless you change OpenID to separately authenticate the server.

Now, from the above picture I would think that the OpenID Provider
certainly can authenticate the RP and then tell the client that the RP
is authenticated (either directly or via the RP).

BTW, "relying party" is a very bad term to use here because the client
can be a relying party too (if it has to validate the server's TLS
cert).  In a PKIX sense, to me, "relying party" is any party that
validates certificates.  Whereas in the above figure you're using
"relying party" in the sense of "party that relies on assertions made
by an IdP".

I'd go further: the fact that you use "relying party" to refer to the
server tells me that there may be a fundamental problem with OpenID:
that it concerns itself only with authenticating users to services, but
not services to users.  But services have to be authenticated to users!

> Then, we might have TLS between the Client and the OpenID provider but
> that largely depends on the authentication mechanism being used.

Sure.  This part is not relevant.  As I explained above, RFC5056
concerns itself only with the client<->server interaction.  The security
of client<->infrastructure and server<->infrastructure communications is
outside the scopre of RFC5056, though they are very much relevant to the
security analysis of the whole system, of course.

> Furthermore, we have authentication of the RP to the OP (and vice versa)
> at the OpenID "application layer". The RP will have some identity, let
> us call it RP_OpenID_Y.

Ah, I see where RFC5056 has you confused.  The only "application layer"
that is relevant here is between the client and the server.

> Now, I believe you guys worry about the linkage between the TLS channel
> (used between the client and the RP) and the OpenID identity used by the
> RP.
> 
> Is this correct?

Yes, that is correct.

Imagine that the RP is using a self-signed certificate.  Or that the RP
doesn't even have a certificate for use in TLS and therefore negotiates
anonymous DH cipher suites only.  With channel binding we can still make
sure that there is no MITM.

Now, suppose you say "but the RP uses the same certificate to
authenticate itself to the client in TLS as it does to authenticate
itself to the OpenID Provider" [*].  Fair enough, but, the OpenID
Provider can do what is variously called "TLS pinning", "leap-of-faith",
... -- it can keep a table mapping service names to certificates, and
thus provide somewhat stronger authentication than the commercial TLS
PKI can.  Or perhaps the OpenID provider will have a trust relationship
with the RP based on something other than the commercial TLS PKI, and
then it can simply "know" the RP's cert(s).

So you see, channel binding can add value above and beyond that which is
added by the commercial TLS PKI.

Nico

[*] Ignoring, for now, any key hygiene issues.  Most server certs for
    use in TLS are meant for encryption, for RSA key transport.  While
    in all likelihood the server would authenticate itself to the
    provider using an RSA signature.  Since one should not sign and
    encrypt with the same public key pair... presumably the server would
    have different PKI credentials for these two purposes.

From mrex@sap.com  Mon Aug 16 08:20:26 2010
Return-Path: <mrex@sap.com>
X-Original-To: kitten@core3.amsl.com
Delivered-To: kitten@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id CC8BE3A65A6 for <kitten@core3.amsl.com>; Mon, 16 Aug 2010 08:20:26 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -9.8
X-Spam-Level: 
X-Spam-Status: No, score=-9.8 tagged_above=-999 required=5 tests=[AWL=0.449, BAYES_00=-2.599, HELO_EQ_DE=0.35, RCVD_IN_DNSWL_HI=-8]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id MTOuVyG90zk0 for <kitten@core3.amsl.com>; Mon, 16 Aug 2010 08:20:25 -0700 (PDT)
Received: from smtpde01.sap-ag.de (smtpde01.sap-ag.de [155.56.68.170]) by core3.amsl.com (Postfix) with ESMTP id 6810E3A63C9 for <kitten@ietf.org>; Mon, 16 Aug 2010 08:20:24 -0700 (PDT)
Received: from mail.sap.corp by smtpde01.sap-ag.de (26) with ESMTP id o7GFKwSf023161 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=OK); Mon, 16 Aug 2010 17:20:59 +0200 (MEST)
From: Martin Rex <mrex@sap.com>
Message-Id: <201008161520.o7GFKweX025917@fs4113.wdf.sap.corp>
To: Nicolas.Williams@oracle.com (Nicolas Williams)
Date: Mon, 16 Aug 2010 17:20:58 +0200 (MEST)
In-Reply-To: <20100813233422.GH14409@oracle.com> from "Nicolas Williams" at Aug 13, 10 06:34:22 pm
MIME-Version: 1.0
Content-Type: text/plain; charset=ISO-8859-1
Content-Transfer-Encoding: 8bit
X-Scanner: Virus Scanner virwal07
X-SAP: out
Cc: kitten@ietf.org, simon@josefsson.org
Subject: Re: [kitten] SAML GSS-API name type
X-BeenThere: kitten@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
Reply-To: mrex@sap.com
List-Id: Common Authentication Technologies - Next Generation <kitten.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/kitten>
List-Post: <mailto:kitten@ietf.org>
List-Help: <mailto:kitten-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=subscribe>
X-List-Received-Date: Mon, 16 Aug 2010 15:20:26 -0000

Nicolas Williams wrote:
> 
> Well, as I just stated, and will repeat here: there is no point to
> authenticating entities whose names are meaningless to, and
> unrecognizable by humans.

If the idea is about **authentication** then this is exactly
about how ist MUST work.

There are tons of applications out there that don't give a damn
about authenticating the server.  They're using a very cheap and
weak server endpoint identification scheme at best.  Basically
anything derives the pattern for matching server endpoints straight
from the network endpoint address does not perform authentication
of the server.

The only Kerberos-derivative protocol that I'm aware of that is
designed to perform secure server authentication is DCE when using
the secure directory service to resolve the network endpoint for
a server, *AND* an unrelated(!!) secure identity for the server
along with it.


The pupose of the name canonicalization is for creation and
maintenance of ACLs by application.  As an service, you usually
want to be sure the you have exact matches for you ACL, not just
some vague pattern matching, and you want efficient ACL lookups,
not linear searches matching against patterns.


A real secure mutual authentication will let you do this symmetrically
in both ways, i.e. the authentication of the server should be done
in the same quality, and servers should be able to use their
server accepting credentials (and the canonical identity of that
credential) to perform client requests and be authenticated similar
to how users are authenticated with their canonical name.


If you look at Kerberos 5 through GSS-API, the only thing that
a GSS-API based application sees is a Kerberos 5 principal name
(involving a Kerberos Realm).  Crossrealm trust and traversal
of KDCs is entirely irrelevant at the GSS-API level because it
is not visible there.


Aside from Kerberos, there are a number of (mostly proprietary)
variants of SPKM (The Simple Public Key mechanism) out there,
GSS-API mechanism based on X.509 certs and PKI.  Most of those
that I know emit exactly and only the subject DName of the certificate,
This means that they work only with a tightly controlled PKI with
a single CA, or of more than one CA is involved, then with a
set of CAs where the Naming of End Entities and Servers is
guaranteed to be non-colliding, either through administrative
procedures or a single RootCA with name constraints (and working
support for name constraints in all involved implementations).


Concerning the naming of servers:

I do consider the SSH leap-of-faith approach on first connect a
workable concept.  And it is defintely more secure than anything
in the area of "hostbased service names" or "TLS server endpoint matching".

Servers in SSH are identified by their public keys, and that's what
is being stored _and_ used in every client-side ACL for secure
(re-)authentication.  traditionally a file called "$HOME/.ssh/known_hosts".
End users don't have to memorize public keys, instead they normally
use a local secure lookup of the server identity, and for convenience,
they often use the hostname as the lookup key.  But there is a significant
difference between looking up a peer's canonical identity in a secure(!)
local directory (like your $HOME/.ssh/known_hosts) and performing
an extremely weak endpoint identification such as doing lazy pattern
matching as with TLS server certificates or doing the traditional
insecure canonicalization of the servername through DNS as done
by Kerberos 5 as documented in rfc-2743 for hostbased service names:

http://tools.ietf.org/html/rfc2743#page-86


> 
> Yes, we could have an option for the ssh command that let's you specify
> the acceptor's name separately from its hostname, and then you'd
> cut-n-paste from somewhere.  Or maybe ssh could do leap-of-faith
> learning of hostname->entityID mappings just for the SAML mechanism.
> But I don't think that's terribly usable, it's rather complicated
> for users, and, IMO, pointless.

That would be very usable and very sensible.

Web Browsers are doing leap-of-faith on each and every connect
unless you flush all of their trust anchors and go through the
painful scary page for accepting a server's certificate for
each and every server.  And some browsers were broken in that
they didn't memorize for which server a user accepted the server
certificate and would trust it being served from any host
after accepting it.  I don't know which browsers suffered from
this and whether/when this was patched.  It definitely was a
serious bug.



-Martin

From cantor.2@osu.edu  Mon Aug 16 08:24:39 2010
Return-Path: <cantor.2@osu.edu>
X-Original-To: kitten@core3.amsl.com
Delivered-To: kitten@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id E22413A6765 for <kitten@core3.amsl.com>; Mon, 16 Aug 2010 08:24:39 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -2.787
X-Spam-Level: 
X-Spam-Status: No, score=-2.787 tagged_above=-999 required=5 tests=[AWL=-0.188, BAYES_00=-2.599]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id 703X3FQkQl5E for <kitten@core3.amsl.com>; Mon, 16 Aug 2010 08:24:39 -0700 (PDT)
Received: from defang1.it.ohio-state.edu (defang1.it.ohio-state.edu [128.146.216.81]) by core3.amsl.com (Postfix) with ESMTP id F0CFB3A63C9 for <kitten@ietf.org>; Mon, 16 Aug 2010 08:24:38 -0700 (PDT)
Received: from defang6.it.ohio-state.edu (defang6.it.ohio-state.edu [128.146.216.86]) by defang1.it.ohio-state.edu (8.13.7/8.13.1) with ESMTP id o7GFPEdG022752; Mon, 16 Aug 2010 11:25:14 -0400
Received: from SNOWDOG (SNOWDOG.dyn.cio.osu.edu [164.107.161.86]) by defang6.it.ohio-state.edu (8.13.7/8.13.1) with ESMTP id o7GFPE1x015176; Mon, 16 Aug 2010 11:25:14 -0400
From: "Scott Cantor" <cantor.2@osu.edu>
To: "'Nicolas Williams'" <Nicolas.Williams@oracle.com>, "'Tschofenig, Hannes \(NSN - FI/Espoo\)'" <hannes.tschofenig@nsn.com>
References: <3D3C75174CB95F42AD6BCC56E5555B4502E9C89F@FIESEXC015.nsn-intra.net> <20100816150915.GC14409@oracle.com>
In-Reply-To: <20100816150915.GC14409@oracle.com>
Date: Mon, 16 Aug 2010 11:25:14 -0400
Organization: The Ohio State University
Message-ID: <006901cb3d57$39acdbf0$ad0693d0$@osu.edu>
MIME-Version: 1.0
Content-Type: text/plain; charset="US-ASCII"
Content-Transfer-Encoding: 7bit
X-Mailer: Microsoft Outlook 14.0
Thread-index: AQIms3cDeEFuGw/20EOge7duUn13bwKXmvTzkhiIuaA=
Content-language: en-us
X-CanIt-Geo: ip=128.146.216.86; country=US; region=OH; city=Columbus; latitude=39.9968; longitude=-82.9882; metrocode=535; areacode=614; http://maps.google.com/maps?q=39.9968,-82.9882&z=6
X-CanItPRO-Stream: outbound
X-Scanned-By: CanIt (www . roaringpenguin . com) on 128.146.216.81
Cc: kitten@ietf.org
Subject: Re: [kitten] Channel Bindings Discussion
X-BeenThere: kitten@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Common Authentication Technologies - Next Generation <kitten.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/kitten>
List-Post: <mailto:kitten@ietf.org>
List-Help: <mailto:kitten-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=subscribe>
X-List-Received-Date: Mon, 16 Aug 2010 15:24:40 -0000

> The question is: does OpenID rely on TLS for authenticating the RP to
> the client?  If so then channel binding will be irrelevant to you,
> unless you change OpenID to separately authenticate the server.

Cookie/redirect Web SSO mechanisms designed around a browser (let's ignore
SPNEGO) are essentially incapable of this. In particular, the authentication
of the RP to the client via TLS is not "in scope" of the SSO protocol itself
and is not leveraged, it's merely assumed as a way of avoiding the problem.

The protocol can (SAML does, OpenID does not, mostly) expect the IdP to
provide some protection by validating response URLs, but the actual proof
that the client is visiting the right URL is treated as a TLS problem, not
something the IdP/OP gets involved with.

(This is why I'm interested in the concept of CB, per last weekend's
discussion. But you can't do it with a browser.)

> > Furthermore, we have authentication of the RP to the OP (and vice versa)
> > at the OpenID "application layer". The RP will have some identity, let
> > us call it RP_OpenID_Y.

That's news to me (so I'd like to clarify if I'm misled on that). SAML does,
but I was under the impression the only such authentication expected in
OpenID is some simple whitelisting of URLs. In particular, I thought the
whole point was to avoid credentialing RPs.
 
(Nothing to do with the topic, though, just curious.)

-- Scott



From ghudson@mit.edu  Mon Aug 16 09:17:00 2010
Return-Path: <ghudson@mit.edu>
X-Original-To: kitten@core3.amsl.com
Delivered-To: kitten@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 4B4423A6801 for <kitten@core3.amsl.com>; Mon, 16 Aug 2010 09:17:00 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -4.599
X-Spam-Level: 
X-Spam-Status: No, score=-4.599 tagged_above=-999 required=5 tests=[AWL=-2.000, BAYES_00=-2.599]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id UeTMRaDjyxSq for <kitten@core3.amsl.com>; Mon, 16 Aug 2010 09:16:59 -0700 (PDT)
Received: from dmz-mailsec-scanner-1.mit.edu (DMZ-MAILSEC-SCANNER-1.MIT.EDU [18.9.25.12]) by core3.amsl.com (Postfix) with ESMTP id E74A93A67FE for <kitten@ietf.org>; Mon, 16 Aug 2010 09:16:56 -0700 (PDT)
X-AuditID: 1209190c-b7c0dae000000a0c-77-4c69649edff3
Received: from mailhub-auth-3.mit.edu (MAILHUB-AUTH-3.MIT.EDU [18.9.21.43]) by dmz-mailsec-scanner-1.mit.edu (Symantec Brightmail Gateway) with SMTP id 6D.93.02572.E94696C4; Mon, 16 Aug 2010 12:17:34 -0400 (EDT)
Received: from outgoing.mit.edu (OUTGOING-AUTH.MIT.EDU [18.7.22.103]) by mailhub-auth-3.mit.edu (8.13.8/8.9.2) with ESMTP id o7GGHVsh002119;  Mon, 16 Aug 2010 12:17:32 -0400
Received: from [10.0.0.102] (c-24-61-11-81.hsd1.ma.comcast.net [24.61.11.81]) (authenticated bits=0) (User authenticated as ghudson@ATHENA.MIT.EDU) by outgoing.mit.edu (8.13.6/8.12.4) with ESMTP id o7GGHUxE013381 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=NOT); Mon, 16 Aug 2010 12:17:31 -0400 (EDT)
From: Greg Hudson <ghudson@MIT.EDU>
To: "mrex@sap.com" <mrex@sap.com>
In-Reply-To: <201008161520.o7GFKweX025917@fs4113.wdf.sap.corp>
References: <201008161520.o7GFKweX025917@fs4113.wdf.sap.corp>
Content-Type: text/plain; charset="UTF-8"
Date: Mon, 16 Aug 2010 12:17:29 -0400
Message-ID: <1281975449.8066.976.camel@ray>
Mime-Version: 1.0
X-Mailer: Evolution 2.28.3 
Content-Transfer-Encoding: 7bit
X-Brightmail-Tracker: AAAAARWWQ4A=
Cc: "kitten@ietf.org" <kitten@ietf.org>
Subject: Re: [kitten] SAML GSS-API name type
X-BeenThere: kitten@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Common Authentication Technologies - Next Generation <kitten.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/kitten>
List-Post: <mailto:kitten@ietf.org>
List-Help: <mailto:kitten-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=subscribe>
X-List-Received-Date: Mon, 16 Aug 2010 16:17:00 -0000

On Mon, 2010-08-16 at 11:20 -0400, Martin Rex wrote:
> I do consider the SSH leap-of-faith approach on first connect a
> workable concept.  And it is defintely more secure than anything
> in the area of "hostbased service names" or "TLS server endpoint matching".

I've heard from a lot of people who find the SSH approach at least as
weak as the commercial TLS approach, and for the same reason: it asks
users to make a decision based on parameters they can't understand and
information they don't know.  In environments with a lot of hosts which
are frequently reinstalled, it's the same as no service authentication
at all.

> But there is a significant
> difference between looking up a peer's canonical identity in a secure(!)
> local directory (like your $HOME/.ssh/known_hosts) and performing
> an extremely weak endpoint identification such as doing lazy pattern
> matching as with TLS server certificates or doing the traditional
> insecure canonicalization of the servername through DNS as done
> by Kerberos 5 as documented in rfc-2743 for hostbased service names:

There are certainly more secure approaches to service name
canonicalization than insecure DNS.  I believe Microsoft AD deployments
use one such.



From Nicolas.Williams@oracle.com  Mon Aug 16 10:07:37 2010
Return-Path: <Nicolas.Williams@oracle.com>
X-Original-To: kitten@core3.amsl.com
Delivered-To: kitten@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 15FB53A6846 for <kitten@core3.amsl.com>; Mon, 16 Aug 2010 10:07:37 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -6.543
X-Spam-Level: 
X-Spam-Status: No, score=-6.543 tagged_above=-999 required=5 tests=[AWL=0.055,  BAYES_00=-2.599, RCVD_IN_DNSWL_MED=-4, UNPARSEABLE_RELAY=0.001]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id 7Y7VqZLD0nng for <kitten@core3.amsl.com>; Mon, 16 Aug 2010 10:07:36 -0700 (PDT)
Received: from rcsinet10.oracle.com (rcsinet10.oracle.com [148.87.113.121]) by core3.amsl.com (Postfix) with ESMTP id E22AA3A6869 for <kitten@ietf.org>; Mon, 16 Aug 2010 10:07:35 -0700 (PDT)
Received: from acsinet15.oracle.com (acsinet15.oracle.com [141.146.126.227]) by rcsinet10.oracle.com (Switch-3.4.2/Switch-3.4.2) with ESMTP id o7GH89hp015635 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=OK); Mon, 16 Aug 2010 17:08:10 GMT
Received: from acsmt355.oracle.com (acsmt355.oracle.com [141.146.40.155]) by acsinet15.oracle.com (Switch-3.4.2/Switch-3.4.1) with ESMTP id o7GFYcnm025983; Mon, 16 Aug 2010 17:08:08 GMT
Received: from abhmt001.oracle.com by acsmt355.oracle.com with ESMTP id 521211691281978382; Mon, 16 Aug 2010 10:06:22 -0700
Received: from oracle.com (/129.153.128.104) by default (Oracle Beehive Gateway v4.0) with ESMTP ; Mon, 16 Aug 2010 10:06:20 -0700
Date: Mon, 16 Aug 2010 12:06:14 -0500
From: Nicolas Williams <Nicolas.Williams@oracle.com>
To: Scott Cantor <cantor.2@osu.edu>
Message-ID: <20100816170613.GF14409@oracle.com>
References: <3D3C75174CB95F42AD6BCC56E5555B4502E9C89F@FIESEXC015.nsn-intra.net> <20100816150915.GC14409@oracle.com> <006901cb3d57$39acdbf0$ad0693d0$@osu.edu>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <006901cb3d57$39acdbf0$ad0693d0$@osu.edu>
User-Agent: Mutt/1.5.20 (2010-03-02)
Cc: kitten@ietf.org
Subject: Re: [kitten] Channel Bindings Discussion
X-BeenThere: kitten@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Common Authentication Technologies - Next Generation <kitten.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/kitten>
List-Post: <mailto:kitten@ietf.org>
List-Help: <mailto:kitten-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=subscribe>
X-List-Received-Date: Mon, 16 Aug 2010 17:07:37 -0000

On Mon, Aug 16, 2010 at 11:25:14AM -0400, Scott Cantor wrote:
> > The question is: does OpenID rely on TLS for authenticating the RP to
> > the client?  If so then channel binding will be irrelevant to you,
> > unless you change OpenID to separately authenticate the server.
> 
> Cookie/redirect Web SSO mechanisms designed around a browser (let's ignore
> SPNEGO) are essentially incapable of this. In particular, the authentication
> of the RP to the client via TLS is not "in scope" of the SSO protocol itself
> and is not leveraged, it's merely assumed as a way of avoiding the problem.
> 
> The protocol can (SAML does, OpenID does not, mostly) expect the IdP to
> provide some protection by validating response URLs, but the actual proof
> that the client is visiting the right URL is treated as a TLS problem, not
> something the IdP/OP gets involved with.
> 
> (This is why I'm interested in the concept of CB, per last weekend's
> discussion. But you can't do it with a browser.)

Well, maybe browsers today can't do it.

In any case what browsers can do and what non-browsers can do are
different things.  We shouldn't condemn the latter to the limitations of
the former; hopefully the former can catch up to the latter.  (I'm sure
you agree, but I'm stating this for the benefit of others.)

> > > Furthermore, we have authentication of the RP to the OP (and vice versa)
> > > at the OpenID "application layer". The RP will have some identity, let
> > > us call it RP_OpenID_Y.
> 
> That's news to me (so I'd like to clarify if I'm misled on that). SAML does,
> but I was under the impression the only such authentication expected in
> OpenID is some simple whitelisting of URLs. In particular, I thought the
> whole point was to avoid credentialing RPs.
>  
> (Nothing to do with the topic, though, just curious.)

I could see the OpenID provider using TLS client certs to authenticate
the server...  Yes, it'd still be the commercial TLS PKI, but the OpenID
is in a better position to confirm keys off-line or to do leap-of-faith.

Nico
-- 

From cantor.2@osu.edu  Mon Aug 16 10:21:19 2010
Return-Path: <cantor.2@osu.edu>
X-Original-To: kitten@core3.amsl.com
Delivered-To: kitten@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 39FE83A68E0 for <kitten@core3.amsl.com>; Mon, 16 Aug 2010 10:21:19 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -3.266
X-Spam-Level: 
X-Spam-Status: No, score=-3.266 tagged_above=-999 required=5 tests=[AWL=0.333,  BAYES_00=-2.599, RCVD_IN_DNSWL_LOW=-1]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id 0u-V9uLZLZuv for <kitten@core3.amsl.com>; Mon, 16 Aug 2010 10:21:18 -0700 (PDT)
Received: from defang20.it.ohio-state.edu (defang20.it.ohio-state.edu [128.146.216.134]) by core3.amsl.com (Postfix) with ESMTP id 43D5B3A68CD for <kitten@ietf.org>; Mon, 16 Aug 2010 10:21:16 -0700 (PDT)
Received: from defang9.it.ohio-state.edu (defang9.it.ohio-state.edu [128.146.216.78]) by defang20.it.ohio-state.edu (8.13.7/8.13.1) with ESMTP id o7GHLpun026349; Mon, 16 Aug 2010 13:21:51 -0400
Received: from SNOWDOG (SNOWDOG.dyn.cio.osu.edu [164.107.161.86]) by defang9.it.ohio-state.edu (8.13.7/8.13.1) with ESMTP id o7GHLpQ9021083; Mon, 16 Aug 2010 13:21:51 -0400
From: "Scott Cantor" <cantor.2@osu.edu>
To: "'Nicolas Williams'" <Nicolas.Williams@oracle.com>
References: <3D3C75174CB95F42AD6BCC56E5555B4502E9C89F@FIESEXC015.nsn-intra.net> <20100816150915.GC14409@oracle.com> <006901cb3d57$39acdbf0$ad0693d0$@osu.edu> <20100816170613.GF14409@oracle.com>
In-Reply-To: <20100816170613.GF14409@oracle.com>
Date: Mon, 16 Aug 2010 13:21:52 -0400
Organization: The Ohio State University
Message-ID: <000701cb3d67$84a012c0$8de03840$@osu.edu>
MIME-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Content-Transfer-Encoding: 7bit
X-Mailer: Microsoft Outlook 14.0
Thread-Index: AQIms3cDeEFuGw/20EOge7duUn13bwKXmvTzAkLHIpABoz0xkJH5ec8Q
Content-Language: en-us
X-CanIt-Geo: ip=128.146.216.78; country=US; region=OH; city=Columbus; latitude=39.9968; longitude=-82.9882; metrocode=535; areacode=614; http://maps.google.com/maps?q=39.9968,-82.9882&z=6
X-CanItPRO-Stream: outbound
X-Scanned-By: CanIt (www . roaringpenguin . com) on 128.146.216.134
Cc: kitten@ietf.org
Subject: Re: [kitten] Channel Bindings Discussion
X-BeenThere: kitten@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Common Authentication Technologies - Next Generation <kitten.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/kitten>
List-Post: <mailto:kitten@ietf.org>
List-Help: <mailto:kitten-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=subscribe>
X-List-Received-Date: Mon, 16 Aug 2010 17:21:19 -0000

> Well, maybe browsers today can't do it.

That's all I meant.

> In any case what browsers can do and what non-browsers can do are
> different things.  We shouldn't condemn the latter to the limitations of
> the former; hopefully the former can catch up to the latter.  (I'm sure
> you agree, but I'm stating this for the benefit of others.)

Yes, but I also think (but couldn't say for sure) that tying the SASL
mechanism to a browser presents fairly similar constraints. This is exactly
why I'm trying to insist that at least for SAML, there be a completely
non-browser mechanism available and not only the original proposal. For
OpenID there is no choice, because OpenID has no ECP equivalent to leverage.

> I could see the OpenID provider using TLS client certs to authenticate
> the server...  Yes, it'd still be the commercial TLS PKI, but the OpenID
> is in a better position to confirm keys off-line or to do leap-of-faith.

I'm not asking about what's theoretically possible in such SSO protocols,
but what the OpenID protocol is defined to be. Believe it or not, the lack
of such a step is part of its raison d'etre.

-- Scott



From cantor.2@osu.edu  Mon Aug 16 10:23:56 2010
Return-Path: <cantor.2@osu.edu>
X-Original-To: kitten@core3.amsl.com
Delivered-To: kitten@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id DCCCA3A67FB for <kitten@core3.amsl.com>; Mon, 16 Aug 2010 10:23:56 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -3.299
X-Spam-Level: 
X-Spam-Status: No, score=-3.299 tagged_above=-999 required=5 tests=[AWL=0.300,  BAYES_00=-2.599, RCVD_IN_DNSWL_LOW=-1]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id 429BGPHLU7b6 for <kitten@core3.amsl.com>; Mon, 16 Aug 2010 10:23:56 -0700 (PDT)
Received: from defang20.it.ohio-state.edu (defang20.it.ohio-state.edu [128.146.216.134]) by core3.amsl.com (Postfix) with ESMTP id E84103A686B for <kitten@ietf.org>; Mon, 16 Aug 2010 10:23:55 -0700 (PDT)
Received: from defang6.it.ohio-state.edu (defang6.it.ohio-state.edu [128.146.216.86]) by defang20.it.ohio-state.edu (8.13.7/8.13.1) with ESMTP id o7GHOVjk027560; Mon, 16 Aug 2010 13:24:31 -0400
Received: from SNOWDOG (SNOWDOG.dyn.cio.osu.edu [164.107.161.86]) by defang6.it.ohio-state.edu (8.13.7/8.13.1) with ESMTP id o7GHOVjt002645; Mon, 16 Aug 2010 13:24:31 -0400
From: "Scott Cantor" <cantor.2@osu.edu>
To: "'Nicolas Williams'" <Nicolas.Williams@oracle.com>
References: <3D3C75174CB95F42AD6BCC56E5555B4502E9C89F@FIESEXC015.nsn-intra.net> <20100816150915.GC14409@oracle.com> <006901cb3d57$39acdbf0$ad0693d0$@osu.edu> <20100816170613.GF14409@oracle.com>
In-Reply-To: <20100816170613.GF14409@oracle.com>
Date: Mon, 16 Aug 2010 13:24:32 -0400
Organization: The Ohio State University
Message-ID: <000801cb3d67$e3c7f790$ab57e6b0$@osu.edu>
MIME-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Content-Transfer-Encoding: 7bit
X-Mailer: Microsoft Outlook 14.0
Thread-Index: AQIms3cDeEFuGw/20EOge7duUn13bwKXmvTzAkLHIpABoz0xkJH5ezZg
Content-Language: en-us
X-CanIt-Geo: ip=128.146.216.86; country=US; region=OH; city=Columbus; latitude=39.9968; longitude=-82.9882; metrocode=535; areacode=614; http://maps.google.com/maps?q=39.9968,-82.9882&z=6
X-CanItPRO-Stream: outbound
X-Scanned-By: CanIt (www . roaringpenguin . com) on 128.146.216.134
Cc: kitten@ietf.org
Subject: Re: [kitten] Channel Bindings Discussion
X-BeenThere: kitten@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Common Authentication Technologies - Next Generation <kitten.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/kitten>
List-Post: <mailto:kitten@ietf.org>
List-Help: <mailto:kitten-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=subscribe>
X-List-Received-Date: Mon, 16 Aug 2010 17:23:57 -0000

> Well, maybe browsers today can't do it.

BTW, it might be apropos to note that the name "Moonshot" in the Moonshot
proposal comes from a statement I made on a list that *if* you're going to
change the client (that is, change the browser), and your solution is
predilected on getting browser and/or OS vendors to actually move the ball,
there's little point in taking halfway steps. Design a better solution and
build it, i.e. shoot for the moon.

-- Scott



From klaas@cisco.com  Mon Aug 16 10:36:20 2010
Return-Path: <klaas@cisco.com>
X-Original-To: kitten@core3.amsl.com
Delivered-To: kitten@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 862D33A69F1 for <kitten@core3.amsl.com>; Mon, 16 Aug 2010 10:36:20 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -10.285
X-Spam-Level: 
X-Spam-Status: No, score=-10.285 tagged_above=-999 required=5 tests=[AWL=0.314, BAYES_00=-2.599, RCVD_IN_DNSWL_HI=-8]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id bkhRZ3OtchFm for <kitten@core3.amsl.com>; Mon, 16 Aug 2010 10:36:19 -0700 (PDT)
Received: from rtp-iport-2.cisco.com (rtp-iport-2.cisco.com [64.102.122.149]) by core3.amsl.com (Postfix) with ESMTP id 6DFE83A6846 for <kitten@ietf.org>; Mon, 16 Aug 2010 10:36:19 -0700 (PDT)
Authentication-Results: rtp-iport-2.cisco.com; dkim=neutral (message not signed) header.i=none
X-IronPort-Anti-Spam-Filtered: true
X-IronPort-Anti-Spam-Result: An8FAKITaUxAZnwM/2dsb2JhbACTNo0ScaMom2OFOwSJYg
X-IronPort-AV: E=Sophos;i="4.55,377,1278288000"; d="scan'208";a="148439806"
Received: from rtp-core-1.cisco.com ([64.102.124.12]) by rtp-iport-2.cisco.com with ESMTP; 16 Aug 2010 17:36:55 +0000
Received: from macmini.wierenga.net (ams-kwiereng-8715.cisco.com [10.55.220.246]) by rtp-core-1.cisco.com (8.13.8/8.14.3) with ESMTP id o7GHasnl022194 for <kitten@ietf.org>; Mon, 16 Aug 2010 17:36:54 GMT
Message-ID: <4C697736.1070005@cisco.com>
Date: Mon, 16 Aug 2010 19:36:54 +0200
From: Klaas Wierenga <klaas@cisco.com>
User-Agent: Mozilla/5.0 (Macintosh; U; Intel Mac OS X 10.6; en-GB; rv:1.9.2.8) Gecko/20100802 Thunderbird/3.1.2
MIME-Version: 1.0
To: kitten@ietf.org
References: <3D3C75174CB95F42AD6BCC56E5555B4502E9C89F@FIESEXC015.nsn-intra.net>	<20100816150915.GC14409@oracle.com>	<006901cb3d57$39acdbf0$ad0693d0$@osu.edu>	<20100816170613.GF14409@oracle.com> <000701cb3d67$84a012c0$8de03840$@osu.edu>
In-Reply-To: <000701cb3d67$84a012c0$8de03840$@osu.edu>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit
Subject: Re: [kitten] Channel Bindings Discussion
X-BeenThere: kitten@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Common Authentication Technologies - Next Generation <kitten.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/kitten>
List-Post: <mailto:kitten@ietf.org>
List-Help: <mailto:kitten-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=subscribe>
X-List-Received-Date: Mon, 16 Aug 2010 17:36:20 -0000

On 8/16/10 7:21 PM, Scott Cantor wrote:
>> Well, maybe browsers today can't do it.
>
> That's all I meant.
>
>> In any case what browsers can do and what non-browsers can do are
>> different things.  We shouldn't condemn the latter to the limitations of
>> the former; hopefully the former can catch up to the latter.  (I'm sure
>> you agree, but I'm stating this for the benefit of others.)
>
> Yes, but I also think (but couldn't say for sure) that tying the SASL
> mechanism to a browser presents fairly similar constraints. This is exactly
> why I'm trying to insist that at least for SAML, there be a completely
> non-browser mechanism available and not only the original proposal. For

for the record, I support that, and it is the reason why I asked Scott 
to write it down. I'd like to see an ECP mechanism get standardized and 
deployed.

> OpenID there is no choice, because OpenID has no ECP equivalent to leverage.

I believe OpenID has limited applicability, for low-value (for whatever 
metric of value) transactions in the browser context, the SASL method 
should not make this worse, but will also not make it any better.

>> I could see the OpenID provider using TLS client certs to authenticate
>> the server...  Yes, it'd still be the commercial TLS PKI, but the OpenID
>> is in a better position to confirm keys off-line or to do leap-of-faith.
>
> I'm not asking about what's theoretically possible in such SSO protocols,
> but what the OpenID protocol is defined to be. Believe it or not, the lack
> of such a step is part of its raison d'etre.

Klaas

From cantor.2@osu.edu  Mon Aug 16 10:44:22 2010
Return-Path: <cantor.2@osu.edu>
X-Original-To: kitten@core3.amsl.com
Delivered-To: kitten@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id E845D3A6846 for <kitten@core3.amsl.com>; Mon, 16 Aug 2010 10:44:22 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -2.826
X-Spam-Level: 
X-Spam-Status: No, score=-2.826 tagged_above=-999 required=5 tests=[AWL=-0.227, BAYES_00=-2.599]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id 131m18icYbk6 for <kitten@core3.amsl.com>; Mon, 16 Aug 2010 10:44:22 -0700 (PDT)
Received: from defang2.it.ohio-state.edu (defang2.it.ohio-state.edu [128.146.216.82]) by core3.amsl.com (Postfix) with ESMTP id D0DE33A69E8 for <kitten@ietf.org>; Mon, 16 Aug 2010 10:44:13 -0700 (PDT)
Received: from defang6.it.ohio-state.edu (defang6.it.ohio-state.edu [128.146.216.86]) by defang2.it.ohio-state.edu (8.13.7/8.13.1) with ESMTP id o7GHimcG026339; Mon, 16 Aug 2010 13:44:48 -0400
Received: from SNOWDOG (SNOWDOG.dyn.cio.osu.edu [164.107.161.86]) by defang6.it.ohio-state.edu (8.13.7/8.13.1) with ESMTP id o7GHimkO024584; Mon, 16 Aug 2010 13:44:48 -0400
From: "Scott Cantor" <cantor.2@osu.edu>
To: "'Klaas Wierenga'" <klaas@cisco.com>, <kitten@ietf.org>
References: <3D3C75174CB95F42AD6BCC56E5555B4502E9C89F@FIESEXC015.nsn-intra.net> <20100816150915.GC14409@oracle.com> <006901cb3d57$39acdbf0$ad0693d0$@osu.edu> <20100816170613.GF14409@oracle.com> <000701cb3d67$84a012c0$8de03840$@osu.edu> <4C697736.1070005@cisco.com>
In-Reply-To: <4C697736.1070005@cisco.com>
Date: Mon, 16 Aug 2010 13:44:49 -0400
Organization: The Ohio State University
Message-ID: <000901cb3d6a$b942edb0$2bc8c910$@osu.edu>
MIME-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Content-Transfer-Encoding: 7bit
X-Mailer: Microsoft Outlook 14.0
Thread-Index: AQIms3cDeEFuGw/20EOge7duUn13bwKXmvTzAkLHIpABoz0xkAIxkV5qAbHmVXGR2mToMA==
Content-Language: en-us
X-CanIt-Geo: ip=128.146.216.86; country=US; region=OH; city=Columbus; latitude=39.9968; longitude=-82.9882; metrocode=535; areacode=614; http://maps.google.com/maps?q=39.9968,-82.9882&z=6
X-CanItPRO-Stream: outbound
X-Scanned-By: CanIt (www . roaringpenguin . com) on 128.146.216.82
Subject: Re: [kitten] Channel Bindings Discussion
X-BeenThere: kitten@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Common Authentication Technologies - Next Generation <kitten.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/kitten>
List-Post: <mailto:kitten@ietf.org>
List-Help: <mailto:kitten-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=subscribe>
X-List-Received-Date: Mon, 16 Aug 2010 17:44:23 -0000

> for the record, I support that, and it is the reason why I asked Scott
> to write it down. I'd like to see an ECP mechanism get standardized and
> deployed.

Which I appreciate. ;-)

> I believe OpenID has limited applicability, for low-value (for whatever
> metric of value) transactions in the browser context, the SASL method
> should not make this worse, but will also not make it any better.

I don't want to get into the applicability discussion, because in my
experience, "lightweight" reinvention does not historically stay in the box
that its originators might think it should. I'm just looking purely at the
technical side. To the extent that the technical side is squishy ("some
people implement it this way, even though the spec doesn't really define
that"), that's its own indictment.

-- Scott



From mrex@sap.com  Mon Aug 16 10:49:19 2010
Return-Path: <mrex@sap.com>
X-Original-To: kitten@core3.amsl.com
Delivered-To: kitten@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 1E2F23A692C for <kitten@core3.amsl.com>; Mon, 16 Aug 2010 10:49:19 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -9.81
X-Spam-Level: 
X-Spam-Status: No, score=-9.81 tagged_above=-999 required=5 tests=[AWL=0.439,  BAYES_00=-2.599, HELO_EQ_DE=0.35, RCVD_IN_DNSWL_HI=-8]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id WL-VQ1utBpqQ for <kitten@core3.amsl.com>; Mon, 16 Aug 2010 10:49:18 -0700 (PDT)
Received: from smtpde02.sap-ag.de (smtpde02.sap-ag.de [155.56.68.140]) by core3.amsl.com (Postfix) with ESMTP id CD1BC3A69F6 for <kitten@ietf.org>; Mon, 16 Aug 2010 10:49:17 -0700 (PDT)
Received: from mail.sap.corp by smtpde02.sap-ag.de (26) with ESMTP id o7GHnqAc029454 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=OK); Mon, 16 Aug 2010 19:49:52 +0200 (MEST)
From: Martin Rex <mrex@sap.com>
Message-Id: <201008161749.o7GHnpCc004587@fs4113.wdf.sap.corp>
To: ghudson@MIT.EDU (Greg Hudson)
Date: Mon, 16 Aug 2010 19:49:51 +0200 (MEST)
In-Reply-To: <1281975449.8066.976.camel@ray> from "Greg Hudson" at Aug 16, 10 12:17:29 pm
MIME-Version: 1.0
Content-Type: text/plain; charset=ISO-8859-1
Content-Transfer-Encoding: 8bit
X-Scanner: Virus Scanner virwal07
X-SAP: out
Cc: kitten@ietf.org
Subject: Re: [kitten] SAML GSS-API name type
X-BeenThere: kitten@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
Reply-To: mrex@sap.com
List-Id: Common Authentication Technologies - Next Generation <kitten.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/kitten>
List-Post: <mailto:kitten@ietf.org>
List-Help: <mailto:kitten-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=subscribe>
X-List-Received-Date: Mon, 16 Aug 2010 17:49:19 -0000

Greg Hudson wrote:
> 
> On Mon, 2010-08-16 at 11:20 -0400, Martin Rex wrote:
> > I do consider the SSH leap-of-faith approach on first connect a
> > workable concept.  And it is defintely more secure than anything
> > in the area of "hostbased service names" or "TLS server endpoint matching".
> 
> I've heard from a lot of people who find the SSH approach at least as
> weak as the commercial TLS approach, and for the same reason: it asks
> users to make a decision based on parameters they can't understand and
> information they don't know.  In environments with a lot of hosts which
> are frequently reinstalled, it's the same as no service authentication
> at all.

There are many possible solutions to every key distribution problem.

Even when re-installing, carrying over the host key is quite possible.
For those admins that would like to replace the host key on each re-install,
a redistribution of the new key through a central known_hosts file to
end-users machines would be another alternative (e.g. /etc/ssh/known_hosts).
No admin is required to create the worst possible user experience.

The security model currently shipping with Web-Browsers is based
on faith, and the entire leap-of-faith is repeated for each and
every connect.  Which browser currently offers you to manually
verify the organization identified by the certficate for a
particular web site and mark this server cert as "trusted",
and gives you a visual clue about this whenever you visit that
site again?  I'm not currently aware of a browser that does this,
although this feature could obviously cut down the number of
leap-of-faiths of everyone's everyday usage considerably,
and limit the vulnerability to very few inital leap-of-faith
situations.

Afraid of weak md5-based signature algorithms?  no more.
Afraid of integer-overflow in AVA-OIDs of Subject DNames
because of poor vetting of DName contents?  no more.
Afraid of '\0' characters in CN= of server certs?  no more.

This list goes on and on, and there are likely going to be
new things added every year that would be much harder to exploit
if users were able to manually confirm server certificates and
browsers would memorize this and warn about changes.


But this might interfere with the business models of CAs, and
TLS server certificate and TLS in Web Browsers is first of all about
business models, rather than about end users and true security.


> 
> > But there is a significant
> > difference between looking up a peer's canonical identity in a secure(!)
> > local directory (like your $HOME/.ssh/known_hosts) and performing
> > an extremely weak endpoint identification such as doing lazy pattern
> > matching as with TLS server certificates or doing the traditional
> > insecure canonicalization of the servername through DNS as done
> > by Kerberos 5 as documented in rfc-2743 for hostbased service names:
> 
> There are certainly more secure approaches to service name
> canonicalization than insecure DNS.  I believe Microsoft AD deployments
> use one such.

Yes.  But Microsoft deployments have the other drawback that running
multiple instances of a distributed service with Microsoft Kerberos
authentication is a security problem (vulnerable to replay attacks).

Unless you use user2user authentication exclusively, which avoids
the necessity to share replay caches cross hosts -- but that option
may not be available (AFAIK, user2user does not work with rfc4559
HTTP Negotiate).  Although Microsoft has been shipping an implementation
of user2user authentication for use with SSPI since Windows 2000, it
has not yet been standardized for the Kerberos 5 GSS-API mechanism.

Hostbases service names are a hack to accomodate optimizations in the
Kerberos authentication exchange.  Sharing replay caches across multiple
backend machines is too difficult, so a necessity to distinguish
instances of a service on different hosts with different credentials
and therefore different hostbased service names.

You have much more freedom in the backend architecture of a distributed
service when it doesn't matter for the client to which particular
instance of a service it is talking, because they're all authenticating
with the exact same identity.  Doing NAT-style load-balancing or
giving out IP-Addresses round-robin through DNS both don't work
with traditional Kerberos.  


-Martin

From klaas@cisco.com  Mon Aug 16 11:07:47 2010
Return-Path: <klaas@cisco.com>
X-Original-To: kitten@core3.amsl.com
Delivered-To: kitten@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id AAACA3A69FB for <kitten@core3.amsl.com>; Mon, 16 Aug 2010 11:07:47 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -10.3
X-Spam-Level: 
X-Spam-Status: No, score=-10.3 tagged_above=-999 required=5 tests=[AWL=0.299,  BAYES_00=-2.599, RCVD_IN_DNSWL_HI=-8]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id ic8dyHa+yFt7 for <kitten@core3.amsl.com>; Mon, 16 Aug 2010 11:07:40 -0700 (PDT)
Received: from rtp-iport-2.cisco.com (rtp-iport-2.cisco.com [64.102.122.149]) by core3.amsl.com (Postfix) with ESMTP id 5A6AD3A6A0C for <kitten@ietf.org>; Mon, 16 Aug 2010 11:07:40 -0700 (PDT)
Authentication-Results: rtp-iport-2.cisco.com; dkim=neutral (message not signed) header.i=none
X-IronPort-Anti-Spam-Filtered: true
X-IronPort-Anti-Spam-Result: AvsEANYbaUxAZnwN/2dsb2JhbACgRnGjLJtnhTsEiWKKOQ
X-IronPort-AV: E=Sophos;i="4.55,377,1278288000"; d="scan'208";a="148452377"
Received: from rtp-core-2.cisco.com ([64.102.124.13]) by rtp-iport-2.cisco.com with ESMTP; 16 Aug 2010 18:08:16 +0000
Received: from macmini.wierenga.net (ams-kwiereng-8715.cisco.com [10.55.220.246]) by rtp-core-2.cisco.com (8.13.8/8.14.3) with ESMTP id o7GI8FKM026314; Mon, 16 Aug 2010 18:08:15 GMT
Message-ID: <4C697E8F.5000106@cisco.com>
Date: Mon, 16 Aug 2010 20:08:15 +0200
From: Klaas Wierenga <klaas@cisco.com>
User-Agent: Mozilla/5.0 (Macintosh; U; Intel Mac OS X 10.6; en-GB; rv:1.9.2.8) Gecko/20100802 Thunderbird/3.1.2
MIME-Version: 1.0
To: Scott Cantor <cantor.2@osu.edu>
References: <3D3C75174CB95F42AD6BCC56E5555B4502E9C89F@FIESEXC015.nsn-intra.net> <20100816150915.GC14409@oracle.com> <006901cb3d57$39acdbf0$ad0693d0$@osu.edu> <20100816170613.GF14409@oracle.com> <000701cb3d67$84a012c0$8de03840$@osu.edu> <4C697736.1070005@cisco.com> <000901cb3d6a$b942edb0$2bc8c910$@osu.edu>
In-Reply-To: <000901cb3d6a$b942edb0$2bc8c910$@osu.edu>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit
Cc: kitten@ietf.org
Subject: Re: [kitten] Channel Bindings Discussion
X-BeenThere: kitten@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Common Authentication Technologies - Next Generation <kitten.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/kitten>
List-Post: <mailto:kitten@ietf.org>
List-Help: <mailto:kitten-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=subscribe>
X-List-Received-Date: Mon, 16 Aug 2010 18:07:47 -0000

On 8/16/10 7:44 PM, Scott Cantor wrote:
>> for the record, I support that, and it is the reason why I asked Scott
>> to write it down. I'd like to see an ECP mechanism get standardized and
>> deployed.
>
> Which I appreciate. ;-)
>
>> I believe OpenID has limited applicability, for low-value (for whatever
>> metric of value) transactions in the browser context, the SASL method
>> should not make this worse, but will also not make it any better.
>
> I don't want to get into the applicability discussion, because in my
> experience, "lightweight" reinvention does not historically stay in the box
> that its originators might think it should. I'm just looking purely at the
> technical side. To the extent that the technical side is squishy ("some
> people implement it this way, even though the spec doesn't really define
> that"), that's its own indictment.

sure, what I was trying to say is that for a certain use you need 
certain security properties. Usually when the scope creeps the technical 
side changes significantly (is LDAPv3 any more lightweight than DAP?). I 
find it for example telling that the OAuth folks are now talking about 
transporting SAML tokens. OpenID as it stands now doesn't satisfy the 
security constraints of the applications I care about, and even though I 
am not at all convinved that they will ever, I don't rule that out 
beforehand either. But we should not spend too much effort on making 1 
link very strong without working on the rest of the chain.

Klaas

From cantor.2@osu.edu  Mon Aug 16 11:16:37 2010
Return-Path: <cantor.2@osu.edu>
X-Original-To: kitten@core3.amsl.com
Delivered-To: kitten@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 545EE3A6A29 for <kitten@core3.amsl.com>; Mon, 16 Aug 2010 11:16:37 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -2.807
X-Spam-Level: 
X-Spam-Status: No, score=-2.807 tagged_above=-999 required=5 tests=[AWL=-0.208, BAYES_00=-2.599]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id fe3C1TS-ZifC for <kitten@core3.amsl.com>; Mon, 16 Aug 2010 11:16:35 -0700 (PDT)
Received: from defang19.it.ohio-state.edu (defang19.it.ohio-state.edu [128.146.216.133]) by core3.amsl.com (Postfix) with ESMTP id 5A9EB3A6A38 for <kitten@ietf.org>; Mon, 16 Aug 2010 11:16:35 -0700 (PDT)
Received: from defang6.it.ohio-state.edu (defang6.it.ohio-state.edu [128.146.216.86]) by defang19.it.ohio-state.edu (8.13.7/8.13.1) with ESMTP id o7GIHAt0020810; Mon, 16 Aug 2010 14:17:10 -0400
Received: from SNOWDOG (SNOWDOG.dyn.cio.osu.edu [164.107.161.86]) by defang6.it.ohio-state.edu (8.13.7/8.13.1) with ESMTP id o7GIH9iA027813; Mon, 16 Aug 2010 14:17:09 -0400
From: "Scott Cantor" <cantor.2@osu.edu>
To: "'Klaas Wierenga'" <klaas@cisco.com>
References: <3D3C75174CB95F42AD6BCC56E5555B4502E9C89F@FIESEXC015.nsn-intra.net> <20100816150915.GC14409@oracle.com> <006901cb3d57$39acdbf0$ad0693d0$@osu.edu> <20100816170613.GF14409@oracle.com> <000701cb3d67$84a012c0$8de03840$@osu.edu> <4C697736.1070005@cisco.com> <000901cb3d6a$b942edb0$2bc8c910$@osu.edu> <4C697E8F.5000106@cisco.com>
In-Reply-To: <4C697E8F.5000106@cisco.com>
Date: Mon, 16 Aug 2010 14:17:10 -0400
Organization: The Ohio State University
Message-ID: <001101cb3d6f$3e6ce820$bb46b860$@osu.edu>
MIME-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Content-Transfer-Encoding: 7bit
X-Mailer: Microsoft Outlook 14.0
Thread-Index: AQIms3cDeEFuGw/20EOge7duUn13bwKXmvTzAkLHIpABoz0xkAIxkV5qAbHmVXECL41rrgFDX6u6kb7WkkA=
Content-Language: en-us
X-CanIt-Geo: ip=128.146.216.86; country=US; region=OH; city=Columbus; latitude=39.9968; longitude=-82.9882; metrocode=535; areacode=614; http://maps.google.com/maps?q=39.9968,-82.9882&z=6
X-CanItPRO-Stream: outbound
X-Scanned-By: CanIt (www . roaringpenguin . com) on 128.146.216.133
Cc: kitten@ietf.org
Subject: Re: [kitten] Channel Bindings Discussion
X-BeenThere: kitten@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Common Authentication Technologies - Next Generation <kitten.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/kitten>
List-Post: <mailto:kitten@ietf.org>
List-Help: <mailto:kitten-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=subscribe>
X-List-Received-Date: Mon, 16 Aug 2010 18:16:37 -0000

> OpenID as it stands now doesn't satisfy the
> security constraints of the applications I care about, and even though I
> am not at all convinved that they will ever, I don't rule that out
> beforehand either. But we should not spend too much effort on making 1
> link very strong without working on the rest of the chain.

Yeah, the problem I was trying to highlight is that usually people say "we
don't need to fix that, because this isn't meant for applications that care
about that problem". Until inevitably people decide that because it's
simpler it must be better, so it needs to be shoehorned in, and oh well, we
still can't solve that problem, so let's make sure the policies get
rewritten to let us ignore it.

Maybe I've been stuck in too many disfunctional organizations, but policy
where I am has always been "what we have to allow so our chosen solution
will work".

-- Scott



From Josh.Howlett@ja.net  Mon Aug 16 12:01:48 2010
Return-Path: <Josh.Howlett@ja.net>
X-Original-To: kitten@core3.amsl.com
Delivered-To: kitten@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id BD4A73A6A77 for <kitten@core3.amsl.com>; Mon, 16 Aug 2010 12:01:48 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -100.461
X-Spam-Level: 
X-Spam-Status: No, score=-100.461 tagged_above=-999 required=5 tests=[AWL=-0.462, BAYES_50=0.001, USER_IN_WHITELIST=-100]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id cRzig6BYQmR8 for <kitten@core3.amsl.com>; Mon, 16 Aug 2010 12:01:47 -0700 (PDT)
Received: from har003676.ukerna.ac.uk (har003676.ukerna.ac.uk [194.82.140.75]) by core3.amsl.com (Postfix) with ESMTP id 6F0263A6A62 for <kitten@ietf.org>; Mon, 16 Aug 2010 12:01:35 -0700 (PDT)
Received: from har003676.ukerna.ac.uk (localhost.localdomain [127.0.0.1]) by localhost (Email Security Appliance) with SMTP id F3C7D4A6B69_C698B31B; Mon, 16 Aug 2010 19:02:09 +0000 (GMT)
Received: from EXC001.atlas.ukerna.ac.uk (exc001.atlas.ukerna.ac.uk [193.62.83.37]) by har003676.ukerna.ac.uk (Sophos Email Appliance) with ESMTP id DE2974A6B5E_C698B2FF; Mon, 16 Aug 2010 19:02:07 +0000 (GMT)
Received: from EXC001.atlas.ukerna.ac.uk ([193.62.83.37]) by EXC001.atlas.ukerna.ac.uk ([::1]) with mapi; Mon, 16 Aug 2010 20:02:16 +0100
From: Josh Howlett <Josh.Howlett@ja.net>
To: Klaas Wierenga <klaas@cisco.com>, Scott Cantor <cantor.2@osu.edu>
Thread-Topic: [kitten] Channel Bindings Discussion
Thread-Index: AQHLPVU85K7wbCw9B0SBCAD4HUfRB5LkIwQAgAAcOACAAAReAIAABDMAgAACN4CAAAaMgIAAG/AQ
Date: Mon, 16 Aug 2010 19:02:03 +0000
Message-ID: <40AC324AA6D81B4A8CF133BCFB7526E6055416@EXC001.atlas.ukerna.ac.uk>
References: <3D3C75174CB95F42AD6BCC56E5555B4502E9C89F@FIESEXC015.nsn-intra.net> <20100816150915.GC14409@oracle.com> <006901cb3d57$39acdbf0$ad0693d0$@osu.edu> <20100816170613.GF14409@oracle.com> <000701cb3d67$84a012c0$8de03840$@osu.edu>	<4C697736.1070005@cisco.com> <000901cb3d6a$b942edb0$2bc8c910$@osu.edu> <4C697E8F.5000106@cisco.com>
In-Reply-To: <4C697E8F.5000106@cisco.com>
Accept-Language: en-GB, en-US
Content-Language: en-US
X-MS-Has-Attach: 
X-MS-TNEF-Correlator: 
Content-Type: text/plain; charset="us-ascii"
Content-Transfer-Encoding: quoted-printable
MIME-Version: 1.0
Cc: "kitten@ietf.org" <kitten@ietf.org>, Josh Howlett <Josh.Howlett@ja.net>
Subject: Re: [kitten] Channel Bindings Discussion
X-BeenThere: kitten@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Common Authentication Technologies - Next Generation <kitten.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/kitten>
List-Post: <mailto:kitten@ietf.org>
List-Help: <mailto:kitten-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=subscribe>
X-List-Received-Date: Mon, 16 Aug 2010 19:01:49 -0000

Klaas wrote:
> I find it for example telling that the OAuth folks are now talking about
> transporting SAML tokens.

+1

Josh.

JANET(UK) is a trading name of The JNT Association, a company limited
by guarantee which is registered in England under No. 2881024=20
and whose Registered Office is at Lumen House, Library Avenue,
Harwell Science and Innovation Campus, Didcot, Oxfordshire. OX11 0SG


From klaas@cisco.com  Mon Aug 16 12:26:28 2010
Return-Path: <klaas@cisco.com>
X-Original-To: kitten@core3.amsl.com
Delivered-To: kitten@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 677E33A6A65 for <kitten@core3.amsl.com>; Mon, 16 Aug 2010 12:26:28 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -2.513
X-Spam-Level: 
X-Spam-Status: No, score=-2.513 tagged_above=-999 required=5 tests=[AWL=0.086,  BAYES_00=-2.599]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id M0AVUDNOfLCt for <kitten@core3.amsl.com>; Mon, 16 Aug 2010 12:26:27 -0700 (PDT)
Received: from teletubbie.het.net.je (teletubbie.het.net.je [IPv6:2001:610:508:110:192:87:110:29]) by core3.amsl.com (Postfix) with ESMTP id 305843A6A5C for <kitten@ietf.org>; Mon, 16 Aug 2010 12:26:27 -0700 (PDT)
Received: from 64-103-25-233.cisco.com ([64.103.25.233] helo=[10.55.220.245]) by teletubbie.het.net.je with esmtpsa (TLSv1:AES128-SHA:128) (Exim 4.71 (FreeBSD)) (envelope-from <klaas@cisco.com>) id 1Ol5KE-000Gsq-BO; Mon, 16 Aug 2010 21:26:26 +0200
References: <3D3C75174CB95F42AD6BCC56E5555B4502E9C89F@FIESEXC015.nsn-intra.net> <20100816150915.GC14409@oracle.com> <006901cb3d57$39acdbf0$ad0693d0$@osu.edu> <20100816170613.GF14409@oracle.com> <000701cb3d67$84a012c0$8de03840$@osu.edu> <4C697736.1070005@cisco.com> <000901cb3d6a$b942edb0$2bc8c910$@osu.edu> <4C697E8F.5000106@cisco.com> <001101cb3d6f$3e6ce820$bb46b860$@osu.edu>
Message-Id: <17C01F04-56D5-485B-B51D-1EDDE5ED9715@cisco.com>
From: Klaas Wierenga <klaas@cisco.com>
To: Scott Cantor <cantor.2@osu.edu>
In-Reply-To: <001101cb3d6f$3e6ce820$bb46b860$@osu.edu>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
X-Mailer: iPad Mail (7B405)
Mime-Version: 1.0 (iPad Mail 7B405)
Date: Mon, 16 Aug 2010 21:27:20 +0200
X-Antivirus: no malware found
Cc: "kitten@ietf.org" <kitten@ietf.org>
Subject: Re: [kitten] Channel Bindings Discussion
X-BeenThere: kitten@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Common Authentication Technologies - Next Generation <kitten.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/kitten>
List-Post: <mailto:kitten@ietf.org>
List-Help: <mailto:kitten-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=subscribe>
X-List-Received-Date: Mon, 16 Aug 2010 19:26:28 -0000

Sent from my iPad

On Aug 16, 2010, at 8:17 PM, "Scott Cantor" <cantor.2@osu.edu> wrote:

>> OpenID as it stands now doesn't satisfy the
>> security constraints of the applications I care about, and even though I
>> am not at all convinved that they will ever, I don't rule that out
>> beforehand either. But we should not spend too much effort on making 1
>> link very strong without working on the rest of the chain.
> 
> Yeah, the problem I was trying to highlight is that usually people say "we
> don't need to fix that, because this isn't meant for applications that care
> about that problem". Until inevitably people decide that because it's
> simpler it must be better, so it needs to be shoehorned in, and oh well, we
> still can't solve that problem, so let's make sure the policies get
> rewritten to let us ignore it.
> 
> Maybe I've been stuck in too many disfunctional organizations, but policy
> where I am has always been "what we have to allow so our chosen solution
> will work".
> 

No, that is a fair concern unfortunately


> -- Scott
> 
> 

From ghudson@mit.edu  Mon Aug 16 12:49:00 2010
Return-Path: <ghudson@mit.edu>
X-Original-To: kitten@core3.amsl.com
Delivered-To: kitten@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 005A83A6A2B for <kitten@core3.amsl.com>; Mon, 16 Aug 2010 12:49:00 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -4.199
X-Spam-Level: 
X-Spam-Status: No, score=-4.199 tagged_above=-999 required=5 tests=[AWL=-1.600, BAYES_00=-2.599]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id MKy7tb+i3ebl for <kitten@core3.amsl.com>; Mon, 16 Aug 2010 12:48:58 -0700 (PDT)
Received: from dmz-mailsec-scanner-7.mit.edu (DMZ-MAILSEC-SCANNER-7.MIT.EDU [18.7.68.36]) by core3.amsl.com (Postfix) with ESMTP id B45E03A6A47 for <kitten@ietf.org>; Mon, 16 Aug 2010 12:48:58 -0700 (PDT)
X-AuditID: 12074424-b7c35ae000000a07-c6-4c69963f0383
Received: from mailhub-auth-3.mit.edu (MAILHUB-AUTH-3.MIT.EDU [18.9.21.43]) by dmz-mailsec-scanner-7.mit.edu (Symantec Brightmail Gateway) with SMTP id E4.DF.02567.F36996C4; Mon, 16 Aug 2010 15:49:19 -0400 (EDT)
Received: from outgoing.mit.edu (OUTGOING-AUTH.MIT.EDU [18.7.22.103]) by mailhub-auth-3.mit.edu (8.13.8/8.9.2) with ESMTP id o7GJnYNw024547;  Mon, 16 Aug 2010 15:49:34 -0400
Received: from [10.0.0.102] (c-24-61-11-81.hsd1.ma.comcast.net [24.61.11.81]) (authenticated bits=0) (User authenticated as ghudson@ATHENA.MIT.EDU) by outgoing.mit.edu (8.13.6/8.12.4) with ESMTP id o7GJnWu7026832 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=NOT); Mon, 16 Aug 2010 15:49:33 -0400 (EDT)
From: Greg Hudson <ghudson@MIT.EDU>
To: "mrex@sap.com" <mrex@sap.com>
In-Reply-To: <201008161749.o7GHnpCc004587@fs4113.wdf.sap.corp>
References: <201008161749.o7GHnpCc004587@fs4113.wdf.sap.corp>
Content-Type: text/plain; charset="UTF-8"
Date: Mon, 16 Aug 2010 15:49:32 -0400
Message-ID: <1281988172.8066.992.camel@ray>
Mime-Version: 1.0
X-Mailer: Evolution 2.28.3 
Content-Transfer-Encoding: 7bit
X-Brightmail-Tracker: AAAAARWWQ4A=
Cc: "kitten@ietf.org" <kitten@ietf.org>
Subject: Re: [kitten] SAML GSS-API name type
X-BeenThere: kitten@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Common Authentication Technologies - Next Generation <kitten.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/kitten>
List-Post: <mailto:kitten@ietf.org>
List-Help: <mailto:kitten-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=subscribe>
X-List-Received-Date: Mon, 16 Aug 2010 19:49:00 -0000

On Mon, 2010-08-16 at 13:49 -0400, Martin Rex wrote:
> For those admins that would like to replace the host key on each re-install,
> a redistribution of the new key through a central known_hosts file to
> end-users machines would be another alternative (e.g. /etc/ssh/known_hosts).
> No admin is required to create the worst possible user experience.

Authentication of services certainly gets easier if you assume the site
admin has provided each client out of band with a flat file containing
all of the service public keys.  I don't get the impression that many
admins are fond of that requirement, though.

> I'm not currently aware of a browser that does this,
> although this feature could obviously cut down the number of
> leap-of-faiths of everyone's everyday usage considerably,
> and limit the vulnerability to very few inital leap-of-faith
> situations.

What you describe would make it easier for knowledgeable users to detect
wholesale forgery of SSL certificates--such as if one of the several
hundred CAs started issuing bogus certs.

I don't think average users would be able to interpret any information
that you'd be able to convey, and would generally interpret warn-on-
certificate-change to be meaningless and frequent noise, just like they
do SSH's warn-on-host-key-change.  (Frequent if for no other reason than
certificates tend to get renewed once a year, and users visit lots of
web sites.)

> Yes.  But Microsoft deployments have the other drawback that running
> multiple instances of a distributed service with Microsoft Kerberos
> authentication is a security problem (vulnerable to replay attacks).

This is a valid criticism of Kerberos, but not a fundamental problem of
service authentication.  Replay attacks are best thwarted by requiring
both sides of an exchange to contribute fresh key information, not
through replay caches.  So I don't see how this is relevant to the
larger question.

> You have much more freedom in the backend architecture of a
> distributed service when it doesn't matter for the client to which
> particular instance of a service it is talking, because they're all
> authenticating with the exact same identity.

I'm a bit lost.  I thought you were arguing against human-recognizable
service names (looking back at
http://www.ietf.org/mail-archive/web/kitten/current/msg01996.html where
you responded to Nico).  I didn't think this argument was about
identical vs. separate identities for multiple instances of a service.



From mrex@sap.com  Mon Aug 16 18:03:49 2010
Return-Path: <mrex@sap.com>
X-Original-To: kitten@core3.amsl.com
Delivered-To: kitten@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id BB7333A6858 for <kitten@core3.amsl.com>; Mon, 16 Aug 2010 18:03:49 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -9.519
X-Spam-Level: 
X-Spam-Status: No, score=-9.519 tagged_above=-999 required=5 tests=[AWL=0.130,  BAYES_00=-2.599, HELO_EQ_DE=0.35, J_CHICKENPOX_54=0.6, RCVD_IN_DNSWL_HI=-8]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id pgLAD0gJdaMA for <kitten@core3.amsl.com>; Mon, 16 Aug 2010 18:03:48 -0700 (PDT)
Received: from smtpde02.sap-ag.de (smtpde02.sap-ag.de [155.56.68.140]) by core3.amsl.com (Postfix) with ESMTP id 032AF3A65A5 for <kitten@ietf.org>; Mon, 16 Aug 2010 18:03:47 -0700 (PDT)
Received: from mail.sap.corp by smtpde02.sap-ag.de (26) with ESMTP id o7H14MCm024819 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=OK); Tue, 17 Aug 2010 03:04:22 +0200 (MEST)
From: Martin Rex <mrex@sap.com>
Message-Id: <201008170104.o7H14LbL029711@fs4113.wdf.sap.corp>
To: ghudson@MIT.EDU (Greg Hudson)
Date: Tue, 17 Aug 2010 03:04:21 +0200 (MEST)
In-Reply-To: <1281988172.8066.992.camel@ray> from "Greg Hudson" at Aug 16, 10 03:49:32 pm
MIME-Version: 1.0
Content-Type: text/plain; charset=ISO-8859-1
Content-Transfer-Encoding: 8bit
X-Scanner: Virus Scanner virwal07
X-SAP: out
Cc: kitten@ietf.org
Subject: Re: [kitten] SAML GSS-API name type
X-BeenThere: kitten@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
Reply-To: mrex@sap.com
List-Id: Common Authentication Technologies - Next Generation <kitten.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/kitten>
List-Post: <mailto:kitten@ietf.org>
List-Help: <mailto:kitten-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=subscribe>
X-List-Received-Date: Tue, 17 Aug 2010 01:03:49 -0000

Greg Hudson wrote:
> 
> On Mon, 2010-08-16 at 13:49 -0400, Martin Rex wrote:
> > For those admins that would like to replace the host key on each re-install,
> > a redistribution of the new key through a central known_hosts file to
> > end-users machines would be another alternative (e.g. /etc/ssh/known_hosts).
> > No admin is required to create the worst possible user experience.
> 
> Authentication of services certainly gets easier if you assume the site
> admin has provided each client out of band with a flat file containing
> all of the service public keys.  I don't get the impression that many
> admins are fond of that requirement, though.

There is nothing wrong with a preconfigured *secure* retrieval of
a mapping or new keys (no matter whether it is a push or pull/on-demand
update scheme).  It unlikey that admins are going to reinstall their
Kerberos KDC or Microsoft Windows Domain Controller on a monthly
basis without carrying over any users&keys.  Such activity usually affects
only application servers, and primarily for development and testing.

>
> > I'm not currently aware of a browser that does this,
> > although this feature could obviously cut down the number of
> > leap-of-faiths of everyone's everyday usage considerably,
> > and limit the vulnerability to very few inital leap-of-faith
> > situations.
> 
> What you describe would make it easier for knowledgeable users to detect
> wholesale forgery of SSL certificates--such as if one of the several
> hundred CAs started issuing bogus certs.
> 
> I don't think average users would be able to interpret any information
> that you'd be able to convey, and would generally interpret warn-on-
> certificate-change to be meaningless and frequent noise, just like they
> do SSH's warn-on-host-key-change.  (Frequent if for no other reason than
> certificates tend to get renewed once a year, and users visit lots of
> web sites.)

CA's and companies doing business on the web are much more sensitive
to negative user experiences caused by flawed administrative procedures
of their IT maintenance staff.

This is more about educating the IT admins to do cut down on negligence
rather than on educating end-users about doing less of the dangerous stuff.


Given how many Web-Sites require active content in Browsers to be enabled
these days and the neverending flood of critical bugs in the feature-bloated
WebBrowsers that are fixed every month, Security on the Internet
is a lost cause for Joe Average User anyway.

Compare the situation of 1996 with today,
  - back then you had no problem surfing the internet with
    active content disabled, today lots of sites are completely 
    unusable without active content enabled
  - the number of independent pre-trusted rootCA certs in browsers
    is *MUCH* larger
  - at least 10 critical bugs are fixed EVERY month that
    can be used for remote code execution with an active content
    enabled browser

Bottom line, if you honestly compare effective security when using
a Browser with HTTPS for electronic commerce on the internet,
it has become so much worse today that it brings tears to my eyes.


Only security geeks that go through the effort of using one or more
virtual sandboxes for connecting a Browser in a disposable installation
with enabled active content to the internet.  For them, a warn on key
change would be a security benefit, and would help them to more easily
detect attacks and to cut down on the number of distinct sandboxes
necessary to get sensitive stuff done over the internet.


> 
> > Yes.  But Microsoft deployments have the other drawback that running
> > multiple instances of a distributed service with Microsoft Kerberos
> > authentication is a security problem (vulnerable to replay attacks).
> 
> This is a valid criticism of Kerberos, but not a fundamental problem of
> service authentication.  Replay attacks are best thwarted by requiring
> both sides of an exchange to contribute fresh key information, not
> through replay caches.  So I don't see how this is relevant to the
> larger question.

That is not quite how message protection was defined for the original
rfc1964 Kerberos 5 GSS-API mechanism.  And for a very long time there
were authentication-only usage scenarios for Kerberos, where the
Kerberos authentication exchange was (ab)used as an OTP-style
authentication. That includes the original design of
rfc-4559 HTTP Negotiate before the addition of TLS channel bindings
(are the channel bindings available on WindowsXP as well, which
 is still at > 50% of the Microsoft Windows Market share and still
 sold on some new Netbooks even today).


> 
> > You have much more freedom in the backend architecture of a
> > distributed service when it doesn't matter for the client to which
> > particular instance of a service it is talking, because they're all
> > authenticating with the exact same identity.
> 
> I'm a bit lost.  I thought you were arguing against human-recognizable
> service names (looking back at
> http://www.ietf.org/mail-archive/web/kitten/current/msg01996.html where
> you responded to Nico).  I didn't think this argument was about
> identical vs. separate identities for multiple instances of a service.

That is a misunderstanding.  I am against insecure transformations of
human-recognizable service names into real server identites.  And
even more so against repeated insecure transformations for every connet.

When there is no secure transformation is available, an insecure
translation involving a counscious leap-of-faith or out-of-band
authentication by the end user ought to be performed, and the users
decision ought to be memorized, just like SSH has been doing it 
the last 15 years.


In traditional SSH, the identities of peers are raw public keys.
These are admittedly hard to memorize and hard to recognize.
Even the public key fingerprints are hard to memorize and hard to
recognize.  Which is why *secure* translation facilities from a
server name into the servers public key, such as $HOME/.ssh/known_hosts
and /etc/ssh/known_hosts are sensible and improve usability.


Replacing the existing, fairly secure authentication of peers with
a scheme based on X.509 certs, a zoo of ~40 completely unrelated
rootCA certs from largely unknown commercial CAs around the globe
and doing an pretty weak matching of the network address to small
parts of the server's certificate does not even come close to
authentication.  (Which might be one reason why rfc-2818 "HTTP over TLS"
does not use the word authentication at all).



If you would permit root/adminstrative access over the internet
to one of your own machines that stores sensitive and valuable
information, how would you enable that access if you were using
public key credentials?

- Would you create an X.509v3 cert and get it signed under one
  of the ~40 distinct rootCAs preconfigured in common browsers and
  then configure your host to match only on parts of the Subject DName
  and allow arbitrary certification pathes to any of those ~40 CAs

- Or would you rather configure only your public key (naked or
  embedded in an X.509 cert) for access, a key that you created all
  by yourself pretty much what traditional SSH has been successfully
  doing for both, servers and clients, since it was invented.

Configuring the public key or cert directly has the big advantage that it
cuts a huge amount of questionable code, a very complex technology (PKIX)
and the entire faith in ~40 liability-free trustanchors and
an unknown amounts of intermedaites from the risk management equation
for remote administrative access to my host.


-Martin

From Nicolas.Williams@oracle.com  Thu Aug 19 09:57:29 2010
Return-Path: <Nicolas.Williams@oracle.com>
X-Original-To: kitten@core3.amsl.com
Delivered-To: kitten@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 5196F3A696F for <kitten@core3.amsl.com>; Thu, 19 Aug 2010 09:57:29 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -6.544
X-Spam-Level: 
X-Spam-Status: No, score=-6.544 tagged_above=-999 required=5 tests=[AWL=0.054,  BAYES_00=-2.599, RCVD_IN_DNSWL_MED=-4, UNPARSEABLE_RELAY=0.001]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id nxIzIUKUOIqX for <kitten@core3.amsl.com>; Thu, 19 Aug 2010 09:57:28 -0700 (PDT)
Received: from rcsinet10.oracle.com (rcsinet10.oracle.com [148.87.113.121]) by core3.amsl.com (Postfix) with ESMTP id 6EBB33A68CB for <kitten@ietf.org>; Thu, 19 Aug 2010 09:57:28 -0700 (PDT)
Received: from acsinet15.oracle.com (acsinet15.oracle.com [141.146.126.227]) by rcsinet10.oracle.com (Switch-3.4.2/Switch-3.4.2) with ESMTP id o7JGw0Ag026405 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=OK); Thu, 19 Aug 2010 16:58:02 GMT
Received: from acsmt355.oracle.com (acsmt355.oracle.com [141.146.40.155]) by acsinet15.oracle.com (Switch-3.4.2/Switch-3.4.1) with ESMTP id o7J2oYGQ006893; Thu, 19 Aug 2010 16:57:59 GMT
Received: from abhmt019.oracle.com by acsmt354.oracle.com with ESMTP id 512651721282237003; Thu, 19 Aug 2010 09:56:43 -0700
Received: from oracle.com (/129.153.128.104) by default (Oracle Beehive Gateway v4.0) with ESMTP ; Thu, 19 Aug 2010 09:56:41 -0700
Date: Thu, 19 Aug 2010 11:56:36 -0500
From: Nicolas Williams <Nicolas.Williams@oracle.com>
To: Scott Cantor <cantor.2@osu.edu>
Message-ID: <20100819165636.GM14409@oracle.com>
References: <3D3C75174CB95F42AD6BCC56E5555B4502E9C89F@FIESEXC015.nsn-intra.net> <20100816150915.GC14409@oracle.com> <006901cb3d57$39acdbf0$ad0693d0$@osu.edu> <20100816170613.GF14409@oracle.com> <000701cb3d67$84a012c0$8de03840$@osu.edu>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <000701cb3d67$84a012c0$8de03840$@osu.edu>
User-Agent: Mutt/1.5.20 (2010-03-02)
Cc: kitten@ietf.org
Subject: Re: [kitten] Channel Bindings Discussion
X-BeenThere: kitten@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Common Authentication Technologies - Next Generation <kitten.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/kitten>
List-Post: <mailto:kitten@ietf.org>
List-Help: <mailto:kitten-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=subscribe>
X-List-Received-Date: Thu, 19 Aug 2010 16:57:29 -0000

On Mon, Aug 16, 2010 at 01:21:52PM -0400, Scott Cantor wrote:
> I'm not asking about what's theoretically possible in such SSO protocols,
> but what the OpenID protocol is defined to be. Believe it or not, the lack
> of such a step is part of its raison d'etre.

Given the figure and explanation given by Hannes I am willing to believe
that OpenID can provide mutual authentication and channel binding in a
manner not unlike the one you've sketched for SAML ECP.  The result
might involve IdPs doing leap-of-faith learning of servers' certs, and
so might be weak, but likely still usable.

Nico
-- 

From wmills@yahoo-inc.com  Thu Aug 19 10:00:43 2010
Return-Path: <wmills@yahoo-inc.com>
X-Original-To: kitten@core3.amsl.com
Delivered-To: kitten@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 34ECD3A6A69 for <kitten@core3.amsl.com>; Thu, 19 Aug 2010 10:00:43 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -17.389
X-Spam-Level: 
X-Spam-Status: No, score=-17.389 tagged_above=-999 required=5 tests=[AWL=-0.125, BAYES_00=-2.599, IP_NOT_FRIENDLY=0.334, NO_RDNS_DOTCOM_HELO=0.001, USER_IN_DEF_WHITELIST=-15]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id ZhgO2tmKmAAN for <kitten@core3.amsl.com>; Thu, 19 Aug 2010 10:00:41 -0700 (PDT)
Received: from mrout1-b.corp.re1.yahoo.com (mrout1-b.corp.re1.yahoo.com [69.147.107.20]) by core3.amsl.com (Postfix) with ESMTP id B9C433A6A55 for <kitten@ietf.org>; Thu, 19 Aug 2010 10:00:38 -0700 (PDT)
Received: from SP2-EX07CAS02.ds.corp.yahoo.com (sp2-ex07cas02.corp.sp2.yahoo.com [98.137.59.38]) by mrout1-b.corp.re1.yahoo.com (8.13.8/8.13.8/y.out) with ESMTP id o7JH0mrU078332 for <kitten@ietf.org>; Thu, 19 Aug 2010 10:00:48 -0700 (PDT)
DomainKey-Signature: a=rsa-sha1; s=serpent; d=yahoo-inc.com; c=nofws; q=dns; h=received:from:to:date:subject:thread-topic:thread-index: message-id:accept-language:content-language:x-ms-has-attach: x-ms-tnef-correlator:acceptlanguage:content-type: content-transfer-encoding:mime-version; b=w7/VeaELNVUg00V0Eb6JVKaqd86ky75onHtVuiREoctExvUNhoKN2x71tQUDJwE1
Received: from SP2-EX07VS06.ds.corp.yahoo.com ([98.137.59.24]) by SP2-EX07CAS02.ds.corp.yahoo.com ([98.137.59.38]) with mapi; Thu, 19 Aug 2010 10:00:48 -0700
From: William Mills <wmills@yahoo-inc.com>
To: "kitten@ietf.org" <kitten@ietf.org>
Date: Thu, 19 Aug 2010 10:00:46 -0700
Thread-Topic: Re: Concerns about oauth sasl mechanism
Thread-Index: Acs/wBD9nhQhYR4YQzGNxviwt8+CMQ==
Message-ID: <FFDFD7371D517847AD71FBB08F9A31562E7CD85E@SP2-EX07VS06.ds.corp.yahoo.com>
Accept-Language: en-US
Content-Language: en-US
X-MS-Has-Attach: 
X-MS-TNEF-Correlator: 
acceptlanguage: en-US
Content-Type: text/plain; charset="us-ascii"
Content-Transfer-Encoding: quoted-printable
MIME-Version: 1.0
Subject: Re: [kitten] Concerns about oauth sasl mechanism
X-BeenThere: kitten@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Common Authentication Technologies - Next Generation <kitten.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/kitten>
List-Post: <mailto:kitten@ietf.org>
List-Help: <mailto:kitten-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=subscribe>
X-List-Received-Date: Thu, 19 Aug 2010 17:00:43 -0000

Hi,

I've included the initial post in the thread below, and this is in response=
 to reading all of the thread I see in the archive.  I'll add the disclaime=
r that I'm new to the writing of RFCs and I'm likely to make some novice mi=
stakes.  That said....

The initial dracft of the OAuth mechanism spec did not include a security l=
ayer because I was both punting to see if anyone objected, and leaning on t=
he OAuth design to provide options for cleartext communication that are sti=
ll secure (this I realise is a point of conversation still).

My goal is pretty simple, I want to get people to stop storing their passwo=
rd in devices (phones, pdas, etc) and giving their passwords to folks like =
moble phone providers, and specifically I'm looking at the mail (IMAP/POP/S=
MTP) space.  A side effect of this is that passwords stop flowing in the cl=
ear to mail services that choose not to use TLS.

In the space I'm focussing on transport security is TLS if it's implemented=
.  Channel binding to TLS (I am in the camp that thinks TLS is not perfect =
but it's good enough, even in light of the fact that almost no one checks r=
evocations) seems reasonable, BUT I think there are a bunch of places out t=
here where an OAuth mechanims in SASL is still a win that will not have TLS=
 in place, and I think that means I don't want to require TLS.  I can be ar=
gued off this point perhaps, and driving people to use TLS would be a good =
thing too.  This focus is not necessailry the right solution for the genera=
l case.  Is channel binding the right strategy here?

A security layer is certainly a possibility, but I did not know where to st=
art on making that part of the initial draft and I wanted to get something =
out there to get working on the things that are more well defined for me.

Another question is: OAuth itself (in it's current form) allows for either =
a bearer token (password substitute which may expire) or signed requests, a=
nd relys on the implementor to choose the crrect security model for their a=
pplication.  My presumption was that a SASL mechanism should not change the=
 protocol being tunneled too much.  What's the right strategy for a success=
ful and implementable spec here?

Thanks,

-bill

________________________________

*	To: kitten at ietf.org <mailto:kitten@DOMAIN.HIDDEN>=20
*	Subject: Concerns about oauth sasl mechanism
*	From: Sam Hartman <hartmans-ietf at mit.edu <mailto:hartmans-ietf@DOMAIN.=
HIDDEN> >
*	Date: Wed, 11 Aug 2010 14:55:46 -0400
*	Delivered-to: kitten at core3.amsl.com <mailto:kitten@DOMAIN.HIDDEN>=20
*	List-archive: <http://www.ietf.org/mail-archive/web/kitten>
*	List-help: <mailto:kitten-request@ietf.org?subject=3Dhelp>
*	List-id: Common Authentication Technologies - Next Generation <kitten.iet=
f.org>
*	List-post: <mailto:kitten@ietf.org>
*	List-subscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:k=
itten-request@ietf.org?subject=3Dsubscribe>
*	List-unsubscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto=
:kitten-request@ietf.org?subject=3Dunsubscribe>
*	User-agent: Gnus/5.110009 (No Gnus v0.9) Emacs/22.3 (gnu/linux)

________________________________

I'll admit that I'm a bit concerned about the never of SASL/GSS
mechanisms we're creating that have no better security than cram-md5 and
may have somewhat less.

I don't think we need any more SASL mechanisms that provide neither
security layers nor channel binding.  I haven't followed the OAuth
standards in a while. If they currently have something that can be used
for channel binding or if they gain such a facility, I think
standardizing such a mechanism here would be great.  Otherwise, I don't
think it's a good use of this WG's time.

Obviously, nothing stops proponents of such a mechanism of registering
it or even getting an non-IETF RFC.  I guess nothing even stops them
from trying to recharter OAuth, although I would bring up the same
concern in that case as well.

I understand that a similar argument can be made about the SAML and
OpenID mechanism. First, I have some hope for the SAML mechanism.  I'll
admit that I'm quite conflicted on the OpenID mechanism and that I can't
entirely justify why I didn't raise this objection there. Roughly, I
think that's very close to the line and this is over the line.

--Sam

________________________________


*	Follow-Ups:=20

	*	RE: Concerns about oauth sasl mechanism <http://www.ietf.org/mail-archiv=
e/web/kitten/current/msg01935.html> =20

		*	From: Tschofenig, Hannes (NSN - FI/Espoo)

		*	Re: Concerns about oauth sasl mechanism <http://www.ietf.org/mail-archi=
ve/web/kitten/current/msg01920.html> =20

		*	From: Peter Saint-Andre

*	Prev by Date: Feedback solicited for Privacy and Identity Management Term=
inology <http://www.ietf.org/mail-archive/web/kitten/current/msg01918.html>=
 =20
*	Next by Date: Re: Concerns about oauth sasl mechanism <http://www.ietf.or=
g/mail-archive/web/kitten/current/msg01920.html> =20
*	Previous by thread: Feedback solicited for Privacy and Identity Managemen=
t Terminology <http://www.ietf.org/mail-archive/web/kitten/current/msg01918=
.html> =20
*	Next by thread: Re: Concerns about oauth sasl mechanism <http://www.ietf.=
org/mail-archive/web/kitten/current/msg01920.html> =20
*	Index(es):=20

	*	Date <http://www.ietf.org/mail-archive/web/kitten/current/maillist.html#=
01919>=20
	*	Thread <http://www.ietf.org/mail-archive/web/kitten/current/threads.html=
#01919>=20


Note: Messages sent to this list are the opinions of the senders and do not=
 imply endorsement by the IETF.=20


From Nicolas.Williams@oracle.com  Thu Aug 19 10:33:57 2010
Return-Path: <Nicolas.Williams@oracle.com>
X-Original-To: kitten@core3.amsl.com
Delivered-To: kitten@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 5BDC93A6967 for <kitten@core3.amsl.com>; Thu, 19 Aug 2010 10:33:57 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -6.547
X-Spam-Level: 
X-Spam-Status: No, score=-6.547 tagged_above=-999 required=5 tests=[AWL=0.051,  BAYES_00=-2.599, RCVD_IN_DNSWL_MED=-4, UNPARSEABLE_RELAY=0.001]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id kibznYQqbm6e for <kitten@core3.amsl.com>; Thu, 19 Aug 2010 10:33:55 -0700 (PDT)
Received: from rcsinet10.oracle.com (rcsinet10.oracle.com [148.87.113.121]) by core3.amsl.com (Postfix) with ESMTP id C1AC73A68CB for <kitten@ietf.org>; Thu, 19 Aug 2010 10:33:54 -0700 (PDT)
Received: from rcsinet13.oracle.com (rcsinet13.oracle.com [148.87.113.125]) by rcsinet10.oracle.com (Switch-3.4.2/Switch-3.4.2) with ESMTP id o7JHYREa026606 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=OK); Thu, 19 Aug 2010 17:34:29 GMT
Received: from acsmt353.oracle.com (acsmt353.oracle.com [141.146.40.153]) by rcsinet13.oracle.com (Switch-3.4.2/Switch-3.4.1) with ESMTP id o7JGqlBY019865; Thu, 19 Aug 2010 17:34:26 GMT
Received: from abhmt018.oracle.com by acsmt355.oracle.com with ESMTP id 532641101282239208; Thu, 19 Aug 2010 10:33:28 -0700
Received: from oracle.com (/129.153.128.104) by default (Oracle Beehive Gateway v4.0) with ESMTP ; Thu, 19 Aug 2010 10:33:28 -0700
Date: Thu, 19 Aug 2010 12:33:23 -0500
From: Nicolas Williams <Nicolas.Williams@oracle.com>
To: Scott Cantor <cantor.2@osu.edu>
Message-ID: <20100819173323.GS14409@oracle.com>
References: <3D3C75174CB95F42AD6BCC56E5555B4502E9C89F@FIESEXC015.nsn-intra.net> <20100816150915.GC14409@oracle.com> <006901cb3d57$39acdbf0$ad0693d0$@osu.edu> <20100816170613.GF14409@oracle.com> <000701cb3d67$84a012c0$8de03840$@osu.edu> <20100819165636.GM14409@oracle.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <20100819165636.GM14409@oracle.com>
User-Agent: Mutt/1.5.20 (2010-03-02)
Cc: kitten@ietf.org
Subject: Re: [kitten] Channel Bindings Discussion
X-BeenThere: kitten@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Common Authentication Technologies - Next Generation <kitten.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/kitten>
List-Post: <mailto:kitten@ietf.org>
List-Help: <mailto:kitten-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=subscribe>
X-List-Received-Date: Thu, 19 Aug 2010 17:33:57 -0000

On Thu, Aug 19, 2010 at 11:56:36AM -0500, Nicolas Williams wrote:
> On Mon, Aug 16, 2010 at 01:21:52PM -0400, Scott Cantor wrote:
> > I'm not asking about what's theoretically possible in such SSO protocols,
> > but what the OpenID protocol is defined to be. Believe it or not, the lack
> > of such a step is part of its raison d'etre.
> 
> Given the figure and explanation given by Hannes I am willing to believe
> that OpenID can provide mutual authentication and channel binding in a

s/can/could eventually/

> manner not unlike the one you've sketched for SAML ECP.  The result
> might involve IdPs doing leap-of-faith learning of servers' certs, and
> so might be weak, but likely still usable.



From Nicolas.Williams@oracle.com  Thu Aug 19 10:46:30 2010
Return-Path: <Nicolas.Williams@oracle.com>
X-Original-To: kitten@core3.amsl.com
Delivered-To: kitten@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 7DDC13A6A49 for <kitten@core3.amsl.com>; Thu, 19 Aug 2010 10:46:30 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -6.547
X-Spam-Level: 
X-Spam-Status: No, score=-6.547 tagged_above=-999 required=5 tests=[AWL=0.051,  BAYES_00=-2.599, RCVD_IN_DNSWL_MED=-4, UNPARSEABLE_RELAY=0.001]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id rnN+Nqc9KtNa for <kitten@core3.amsl.com>; Thu, 19 Aug 2010 10:46:19 -0700 (PDT)
Received: from rcsinet10.oracle.com (rcsinet10.oracle.com [148.87.113.121]) by core3.amsl.com (Postfix) with ESMTP id 87E013A6A38 for <kitten@ietf.org>; Thu, 19 Aug 2010 10:46:18 -0700 (PDT)
Received: from rcsinet15.oracle.com (rcsinet15.oracle.com [148.87.113.117]) by rcsinet10.oracle.com (Switch-3.4.2/Switch-3.4.2) with ESMTP id o7JHkpm1028143 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=OK); Thu, 19 Aug 2010 17:46:53 GMT
Received: from acsmt354.oracle.com (acsmt354.oracle.com [141.146.40.154]) by rcsinet15.oracle.com (Switch-3.4.2/Switch-3.4.1) with ESMTP id o7JHkpYU010057; Thu, 19 Aug 2010 17:46:51 GMT
Received: from abhmt008.oracle.com by acsmt354.oracle.com with ESMTP id 532674001282239905; Thu, 19 Aug 2010 10:45:05 -0700
Received: from oracle.com (/129.153.128.104) by default (Oracle Beehive Gateway v4.0) with ESMTP ; Thu, 19 Aug 2010 10:45:04 -0700
Date: Thu, 19 Aug 2010 12:44:59 -0500
From: Nicolas Williams <Nicolas.Williams@oracle.com>
To: William Mills <wmills@yahoo-inc.com>
Message-ID: <20100819174459.GT14409@oracle.com>
References: <FFDFD7371D517847AD71FBB08F9A31562E7CD85E@SP2-EX07VS06.ds.corp.yahoo.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <FFDFD7371D517847AD71FBB08F9A31562E7CD85E@SP2-EX07VS06.ds.corp.yahoo.com>
User-Agent: Mutt/1.5.20 (2010-03-02)
Cc: "kitten@ietf.org" <kitten@ietf.org>
Subject: Re: [kitten] Concerns about oauth sasl mechanism
X-BeenThere: kitten@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Common Authentication Technologies - Next Generation <kitten.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/kitten>
List-Post: <mailto:kitten@ietf.org>
List-Help: <mailto:kitten-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=subscribe>
X-List-Received-Date: Thu, 19 Aug 2010 17:46:30 -0000

On Thu, Aug 19, 2010 at 10:00:46AM -0700, William Mills wrote:
> I've included the initial post in the thread below, and this is in
> response to reading all of the thread I see in the archive.  I'll add
> the disclaimer that I'm new to the writing of RFCs and I'm likely to
> make some novice mistakes.  That said....
> 
> The initial dracft of the OAuth mechanism spec did not include a
> security layer because I was both punting to see if anyone objected,
> and leaning on the OAuth design to provide options for cleartext
> communication that are still secure (this I realise is a point of
> conversation still).

What do you mean by "options for cleartext communication that are still
secure"?  Integrity protection without confidentiality protection?

There's nothing wrong with an option for integrity protection w/o
confidentiality protection.  But note that we still call that a
"security layer" (in SASL terminology).

> My goal is pretty simple, I want to get people to stop storing their
> password in devices (phones, pdas, etc) and giving their passwords to
> folks like moble phone providers, and specifically I'm looking at the
> mail (IMAP/POP/SMTP) space.  A side effect of this is that passwords
> stop flowing in the clear to mail services that choose not to use TLS.
> 
> In the space I'm focussing on transport security is TLS if it's
> implemented.  Channel binding to TLS (I am in the camp that thinks TLS
> is not perfect but it's good enough, even in light of the fact that
> almost no one checks revocations) seems reasonable, BUT I think there
> are a bunch of places out there where an OAuth mechanims in SASL is
> still a win that will not have TLS in place, and I think that means I
> don't want to require TLS.  I can be argued off this point perhaps,
> and driving people to use TLS would be a good thing too.  This focus
> is not necessailry the right solution for the general case.  Is
> channel binding the right strategy here?

Channel binding by definition requires a secure channel.  The lack of a
secure channel implies that channel binding cannot be done.

You can have the option to not have a secure channel.  But when there is
one, channel binding is a good thing to do.

> A security layer is certainly a possibility, but I did not know where
> to start on making that part of the initial draft and I wanted to get
> something out there to get working on the things that are more well
> defined for me.

The SASL community has effectively decided that security layers are
obsolete, that instead of security layers applications should do channel
binding to TLS.

(This happened in the context of the SCRAM mechanism.  SCRAM does not
provide security layers.  It provides channel binding instead.  There
was a lengthy thread about this that concluded with consensus that
security layers were undesirable.)

> Another question is: OAuth itself (in it's current form) allows for
> either a bearer token (password substitute which may expire) or signed
> requests, and relys on the implementor to choose the crrect security
> model for their application.  My presumption was that a SASL mechanism
> should not change the protocol being tunneled too much.  What's the
> right strategy for a successful and implementable spec here?

I'd need to know more to answer this.  But my gut feeling is that you'd
want to base the SASL mechanism on the signed request/reply profile.

Does OAuth provide for mutual authentication?  I.e., can the server's
name (and service name) be authenticated to the client by the user's
OpenID provider?

Nico
-- 

From wmills@yahoo-inc.com  Thu Aug 19 11:19:46 2010
Return-Path: <wmills@yahoo-inc.com>
X-Original-To: kitten@core3.amsl.com
Delivered-To: kitten@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id CA80E3A6953 for <kitten@core3.amsl.com>; Thu, 19 Aug 2010 11:19:46 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -17.531
X-Spam-Level: 
X-Spam-Status: No, score=-17.531 tagged_above=-999 required=5 tests=[AWL=0.067, BAYES_00=-2.599, NO_RDNS_DOTCOM_HELO=0.001, USER_IN_DEF_WHITELIST=-15]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id 9BP4f0I4lk45 for <kitten@core3.amsl.com>; Thu, 19 Aug 2010 11:19:45 -0700 (PDT)
Received: from mrout2.yahoo.com (mrout2.yahoo.com [216.145.54.172]) by core3.amsl.com (Postfix) with ESMTP id AA09A3A68B5 for <kitten@ietf.org>; Thu, 19 Aug 2010 11:19:45 -0700 (PDT)
Received: from SP2-EX07CAS05.ds.corp.yahoo.com (sp2-ex07cas05.corp.sp2.yahoo.com [98.137.59.39]) by mrout2.yahoo.com (8.13.8/8.13.8/y.out) with ESMTP id o7JIJtaq015593;  Thu, 19 Aug 2010 11:19:55 -0700 (PDT)
DomainKey-Signature: a=rsa-sha1; s=serpent; d=yahoo-inc.com; c=nofws; q=dns; h=received:from:to:cc:date:subject:thread-topic: thread-index:message-id:references:in-reply-to:accept-language: content-language:x-ms-has-attach:x-ms-tnef-correlator:acceptlanguage: content-type:content-transfer-encoding:mime-version; b=auwCgdZDQM5ep+cgS+3BvHaLlEWvOi/tIDMpaw/5rmL0UxvQQaJvg40UExLhjbgN
Received: from SP2-EX07VS06.ds.corp.yahoo.com ([98.137.59.24]) by SP2-EX07CAS05.ds.corp.yahoo.com ([98.137.59.39]) with mapi; Thu, 19 Aug 2010 13:19:55 -0500
From: William Mills <wmills@yahoo-inc.com>
To: Nicolas Williams <Nicolas.Williams@oracle.com>
Date: Thu, 19 Aug 2010 13:19:53 -0500
Thread-Topic: [kitten] Concerns about oauth sasl mechanism
Thread-Index: Acs/xo03Ewh//dJVQuWMZdoME6leSQAATaow
Message-ID: <FFDFD7371D517847AD71FBB08F9A31562E7CD8B4@SP2-EX07VS06.ds.corp.yahoo.com>
References: <FFDFD7371D517847AD71FBB08F9A31562E7CD85E@SP2-EX07VS06.ds.corp.yahoo.com> <20100819174459.GT14409@oracle.com>
In-Reply-To: <20100819174459.GT14409@oracle.com>
Accept-Language: en-US
Content-Language: en-US
X-MS-Has-Attach: 
X-MS-TNEF-Correlator: 
acceptlanguage: en-US
Content-Type: text/plain; charset="us-ascii"
Content-Transfer-Encoding: quoted-printable
MIME-Version: 1.0
Cc: "kitten@ietf.org" <kitten@ietf.org>
Subject: Re: [kitten] Concerns about oauth sasl mechanism
X-BeenThere: kitten@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Common Authentication Technologies - Next Generation <kitten.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/kitten>
List-Post: <mailto:kitten@ietf.org>
List-Help: <mailto:kitten-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=subscribe>
X-List-Received-Date: Thu, 19 Aug 2010 18:19:46 -0000

=20

> -----Original Message-----
> From: Nicolas Williams [mailto:Nicolas.Williams@oracle.com]=20
> Sent: Thursday, August 19, 2010 10:45 AM
> To: William Mills
> Cc: kitten@ietf.org
> Subject: Re: [kitten] Concerns about oauth sasl mechanism
>=20
> On Thu, Aug 19, 2010 at 10:00:46AM -0700, William Mills wrote:
> > I've included the initial post in the thread below, and this is in=20
> > response to reading all of the thread I see in the archive.=20
>  I'll add=20
> > the disclaimer that I'm new to the writing of RFCs and I'm=20
> likely to=20
> > make some novice mistakes.  That said....
> >=20
> > The initial dracft of the OAuth mechanism spec did not include a=20
> > security layer because I was both punting to see if anyone=20
> objected,=20
> > and leaning on the OAuth design to provide options for cleartext=20
> > communication that are still secure (this I realise is a point of=20
> > conversation still).
>=20
> What do you mean by "options for cleartext communication that=20
> are still secure"?  Integrity protection without=20
> confidentiality protection?

Yes, the tokens are cryptographically signed, or the signed exchange is sig=
ned with a shared secret.

>=20
> There's nothing wrong with an option for integrity protection=20
> w/o confidentiality protection.  But note that we still call=20
> that a "security layer" (in SASL terminology).
>=20

So in the initial draft we're leaning on the integrity protection in OAuth =
rather than spellign it out in the SASL mechanism.  Does that work, or do w=
e need to pull something in to the mechanism spec?

> > My goal is pretty simple, I want to get people to stop=20
> storing their=20
> > password in devices (phones, pdas, etc) and giving their=20
> passwords to=20
> > folks like moble phone providers, and specifically I'm=20
> looking at the=20
> > mail (IMAP/POP/SMTP) space.  A side effect of this is that=20
> passwords=20
> > stop flowing in the clear to mail services that choose not=20
> to use TLS.
> >=20
> > In the space I'm focussing on transport security is TLS if it's=20
> > implemented.  Channel binding to TLS (I am in the camp that=20
> thinks TLS=20
> > is not perfect but it's good enough, even in light of the fact that=20
> > almost no one checks revocations) seems reasonable, BUT I=20
> think there=20
> > are a bunch of places out there where an OAuth mechanims in SASL is=20
> > still a win that will not have TLS in place, and I think=20
> that means I=20
> > don't want to require TLS.  I can be argued off this point perhaps,=20
> > and driving people to use TLS would be a good thing too. =20
> This focus=20
> > is not necessailry the right solution for the general case.  Is=20
> > channel binding the right strategy here?
>=20
> Channel binding by definition requires a secure channel.  The=20
> lack of a secure channel implies that channel binding cannot be done.
>=20
> You can have the option to not have a secure channel.  But=20
> when there is one, channel binding is a good thing to do.

We can certainly add that.  Is there a spec we should use as an example to =
draw from? =20

>=20
> > A security layer is certainly a possibility, but I did not=20
> know where=20
> > to start on making that part of the initial draft and I=20
> wanted to get=20
> > something out there to get working on the things that are more well=20
> > defined for me.
>=20
> The SASL community has effectively decided that security=20
> layers are obsolete, that instead of security layers=20
> applications should do channel binding to TLS.
>=20
> (This happened in the context of the SCRAM mechanism.  SCRAM=20
> does not provide security layers.  It provides channel=20
> binding instead.  There was a lengthy thread about this that=20
> concluded with consensus that security layers were undesirable.)
>=20
> > Another question is: OAuth itself (in it's current form) allows for=20
> > either a bearer token (password substitute which may=20
> expire) or signed=20
> > requests, and relys on the implementor to choose the crrect=20
> security=20
> > model for their application.  My presumption was that a=20
> SASL mechanism=20
> > should not change the protocol being tunneled too much.  What's the=20
> > right strategy for a successful and implementable spec here?
>=20
> I'd need to know more to answer this.  But my gut feeling is=20
> that you'd want to base the SASL mechanism on the signed=20
> request/reply profile.
>=20
> Does OAuth provide for mutual authentication?  I.e., can the=20
> server's name (and service name) be authenticated to the=20
> client by the user's OpenID provider?

No, there is no provision currently for the client to authenticate the serv=
er (protected resource is the OAuth term).  The current direction is to hav=
e endpoint discovery (via WebFinger for example), but that's not sufficient=
 to call it authentication.  It's a really interesting problem, and my init=
ial thought is that OpenID my be used to verify the server's identity, but =
that starts to get into a weird hybrid of both OpenID and OAuth.

>=20
> Nico
> --=20
> =

From Nicolas.Williams@oracle.com  Thu Aug 19 11:38:58 2010
Return-Path: <Nicolas.Williams@oracle.com>
X-Original-To: kitten@core3.amsl.com
Delivered-To: kitten@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 9E5D93A6844 for <kitten@core3.amsl.com>; Thu, 19 Aug 2010 11:38:58 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -6.549
X-Spam-Level: 
X-Spam-Status: No, score=-6.549 tagged_above=-999 required=5 tests=[AWL=0.049,  BAYES_00=-2.599, RCVD_IN_DNSWL_MED=-4, UNPARSEABLE_RELAY=0.001]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id iXZbYju-eeeZ for <kitten@core3.amsl.com>; Thu, 19 Aug 2010 11:38:57 -0700 (PDT)
Received: from rcsinet10.oracle.com (rcsinet10.oracle.com [148.87.113.121]) by core3.amsl.com (Postfix) with ESMTP id A00C73A6782 for <kitten@ietf.org>; Thu, 19 Aug 2010 11:38:57 -0700 (PDT)
Received: from rcsinet13.oracle.com (rcsinet13.oracle.com [148.87.113.125]) by rcsinet10.oracle.com (Switch-3.4.2/Switch-3.4.2) with ESMTP id o7JIdRmw030759 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=OK); Thu, 19 Aug 2010 18:39:28 GMT
Received: from acsmt355.oracle.com (acsmt355.oracle.com [141.146.40.155]) by rcsinet13.oracle.com (Switch-3.4.2/Switch-3.4.1) with ESMTP id o7JEokl0008948; Thu, 19 Aug 2010 18:39:26 GMT
Received: from abhmt020.oracle.com by acsmt355.oracle.com with ESMTP id 532823551282243135; Thu, 19 Aug 2010 11:38:55 -0700
Received: from oracle.com (/129.153.128.104) by default (Oracle Beehive Gateway v4.0) with ESMTP ; Thu, 19 Aug 2010 11:38:54 -0700
Date: Thu, 19 Aug 2010 13:38:49 -0500
From: Nicolas Williams <Nicolas.Williams@oracle.com>
To: William Mills <wmills@yahoo-inc.com>
Message-ID: <20100819183848.GX14409@oracle.com>
References: <FFDFD7371D517847AD71FBB08F9A31562E7CD85E@SP2-EX07VS06.ds.corp.yahoo.com> <20100819174459.GT14409@oracle.com> <FFDFD7371D517847AD71FBB08F9A31562E7CD8B4@SP2-EX07VS06.ds.corp.yahoo.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <FFDFD7371D517847AD71FBB08F9A31562E7CD8B4@SP2-EX07VS06.ds.corp.yahoo.com>
User-Agent: Mutt/1.5.20 (2010-03-02)
Cc: "kitten@ietf.org" <kitten@ietf.org>
Subject: Re: [kitten] Concerns about oauth sasl mechanism
X-BeenThere: kitten@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Common Authentication Technologies - Next Generation <kitten.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/kitten>
List-Post: <mailto:kitten@ietf.org>
List-Help: <mailto:kitten-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=subscribe>
X-List-Received-Date: Thu, 19 Aug 2010 18:38:58 -0000

On Thu, Aug 19, 2010 at 01:19:53PM -0500, William Mills wrote:
> > What do you mean by "options for cleartext communication that 
> > are still secure"?  Integrity protection without 
> > confidentiality protection?
> 
> Yes, the tokens are cryptographically signed, or the signed exchange
> is signed with a shared secret.

OK.  That is still referred to as a "security layer".

> So in the initial draft we're leaning on the integrity protection in
> OAuth rather than spellign it out in the SASL mechanism.  Does that
> work, or do we need to pull something in to the mechanism spec?

The messages themselves might be defined outside the SASL mechanism
itself; the SASL mechanism specification can make normative references
to other documents.

> > You can have the option to not have a secure channel.  But 
> > when there is one, channel binding is a good thing to do.
> 
> We can certainly add that.  Is there a spec we should use as an
> example to draw from?  

SCRAM [RFC5802] is probably a good example.

> > I'd need to know more to answer this.  But my gut feeling is 
> > that you'd want to base the SASL mechanism on the signed 
> > request/reply profile.
> > 
> > Does OAuth provide for mutual authentication?  I.e., can the 
> > server's name (and service name) be authenticated to the 
> > client by the user's OpenID provider?
> 
> No, there is no provision currently for the client to authenticate the
> server (protected resource is the OAuth term).  The current direction
> is to have endpoint discovery (via WebFinger for example), but that's
> not sufficient to call it authentication.  It's a really interesting
> problem, and my initial thought is that OpenID my be used to verify
> the server's identity, but that starts to get into a weird hybrid of
> both OpenID and OAuth.

You will need to have host naming and authentication in your mechanism.
This is practically unavoidable given the way that GSS-API and SASL are
used in Internet protocols.  A combination of mechanisms is certainly
not out of the question.

Nico
-- 

From mrex@sap.com  Thu Aug 19 12:43:41 2010
Return-Path: <mrex@sap.com>
X-Original-To: kitten@core3.amsl.com
Delivered-To: kitten@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id D3CF43A680D for <kitten@core3.amsl.com>; Thu, 19 Aug 2010 12:43:41 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -9.68
X-Spam-Level: 
X-Spam-Status: No, score=-9.68 tagged_above=-999 required=5 tests=[AWL=0.254,  BAYES_00=-2.599, HELO_EQ_DE=0.35, RCVD_IN_DNSWL_HI=-8, SARE_MILLIONSOF=0.315]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id 9jvNoG0kswwA for <kitten@core3.amsl.com>; Thu, 19 Aug 2010 12:43:41 -0700 (PDT)
Received: from smtpde02.sap-ag.de (smtpde02.sap-ag.de [155.56.68.140]) by core3.amsl.com (Postfix) with ESMTP id B8D413A659C for <kitten@ietf.org>; Thu, 19 Aug 2010 12:43:40 -0700 (PDT)
Received: from mail.sap.corp by smtpde02.sap-ag.de (26) with ESMTP id o7JJiEX7005102 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=OK); Thu, 19 Aug 2010 21:44:14 +0200 (MEST)
From: Martin Rex <mrex@sap.com>
Message-Id: <201008191944.o7JJiETU025823@fs4113.wdf.sap.corp>
To: wmills@yahoo-inc.com (William Mills)
Date: Thu, 19 Aug 2010 21:44:14 +0200 (MEST)
In-Reply-To: <FFDFD7371D517847AD71FBB08F9A31562E7CD85E@SP2-EX07VS06.ds.corp.yahoo.com> from "William Mills" at Aug 19, 10 10:00:46 am
MIME-Version: 1.0
Content-Type: text/plain; charset=ISO-8859-1
Content-Transfer-Encoding: 8bit
X-Scanner: Virus Scanner virwal07
X-SAP: out
Cc: kitten@ietf.org
Subject: Re: [kitten] Concerns about oauth sasl mechanism
X-BeenThere: kitten@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
Reply-To: mrex@sap.com
List-Id: Common Authentication Technologies - Next Generation <kitten.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/kitten>
List-Post: <mailto:kitten@ietf.org>
List-Help: <mailto:kitten-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=subscribe>
X-List-Received-Date: Thu, 19 Aug 2010 19:43:41 -0000

William Mills wrote:
> 
> My goal is pretty simple, I want to get people to stop storing their
> password in devices (phones, pdas, etc) and giving their passwords
> to folks like moble phone providers, and specifically I'm looking
> at the mail (IMAP/POP/SMTP) space.  A side effect of this is that
> passwords stop flowing in the clear to mail services that choose
> not to use TLS.

You must be kidding.  Sharing your passwords for a personal service
with some other organization than that which provides that particular
service is unconditionally a bad idea.

If there are currently providers that do not offer TLS-protected
logons, they must be explicitly UNsupported.  Let market pressure
force them to offer password-based authentication over encrypted
communication links.

Why would you want to put potentially millions of user credentials
in jeopardy if there's at most a few hundred service providers which
should be exterminated anyway if they show so little little respect
for the security and privacy of their own users.


-Martin

From simon@josefsson.org  Tue Aug 31 08:51:29 2010
Return-Path: <simon@josefsson.org>
X-Original-To: kitten@core3.amsl.com
Delivered-To: kitten@core3.amsl.com
Received: from localhost (localhost [127.0.0.1]) by core3.amsl.com (Postfix) with ESMTP id 480373A6881; Tue, 31 Aug 2010 08:51:29 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -102.689
X-Spam-Level: 
X-Spam-Status: No, score=-102.689 tagged_above=-999 required=5 tests=[AWL=-0.090, BAYES_00=-2.599, USER_IN_WHITELIST=-100]
Received: from mail.ietf.org ([64.170.98.32]) by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id z6YRvkUPfMNk; Tue, 31 Aug 2010 08:51:28 -0700 (PDT)
Received: from yxa-v.extundo.com (yxa-v.extundo.com [83.241.177.39]) by core3.amsl.com (Postfix) with ESMTP id BDACA3A6804; Tue, 31 Aug 2010 08:51:26 -0700 (PDT)
Received: from mocca (c80-216-27-64.bredband.comhem.se [80.216.27.64]) (authenticated bits=0) by yxa-v.extundo.com (8.14.3/8.14.3/Debian-5+lenny1) with ESMTP id o7VFpncU032443 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES128-SHA bits=128 verify=NOT); Tue, 31 Aug 2010 17:51:52 +0200
From: Simon Josefsson <simon@josefsson.org>
To: IETF Secretariat <ietf-ipr@ietf.org>
References: <20100831151412.DD8713A688D@core3.amsl.com>
OpenPGP: id=B565716F; url=http://josefsson.org/key.txt
X-Hashcash: 1:22:100831:ietf-ipr@ietf.org::+dRdgv+mpUFaWeOH:EexU
X-Hashcash: 1:22:100831:tlyu@mit.edu::RWFU2eUbzp4PuRxJ:Lg7P
X-Hashcash: 1:22:100831:kitten@ietf.org::PqY9BFioR8GePSEI:dN94
X-Hashcash: 1:22:100831:shawn.emery@oracle.com::M4CsfYY0ODdA3HFv:LvjJ
X-Hashcash: 1:22:100831:hartmans-ietf@mit.edu::ujnww/MozLdJgkxg:S13L
X-Hashcash: 1:22:100831:ipr-announce@ietf.org::6CVcWlMRosEhGmyB:doDB
Date: Tue, 31 Aug 2010 17:51:47 +0200
In-Reply-To: <20100831151412.DD8713A688D@core3.amsl.com> (IETF Secretariat's message of "Tue, 31 Aug 2010 08:14:12 -0700 (PDT)")
Message-ID: <87eiden5kc.fsf@mocca.josefsson.org>
User-Agent: Gnus/5.110011 (No Gnus v0.11) Emacs/23.2 (gnu/linux)
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
X-Virus-Scanned: clamav-milter 0.96.1 at yxa-v
X-Virus-Status: Clean
Cc: kitten@ietf.org, hartmans-ietf@mit.edu, ipr-announce@ietf.org
Subject: Re: [kitten] IPR Disclosure: CNRS's Statement about IPR related to RFC 4768, RFC 2069, RFC 4169, and draft-ietf-httpbis-p7-auth-11
X-BeenThere: kitten@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Common Authentication Technologies - Next Generation <kitten.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/kitten>
List-Post: <mailto:kitten@ietf.org>
List-Help: <mailto:kitten-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=subscribe>
X-List-Received-Date: Tue, 31 Aug 2010 15:51:29 -0000

IETF Secretariat <ietf-ipr@ietf.org> writes:

> Dear Sam Hartman:
>
> An IPR disclosure that pertains to your RFC entitled "Desired Enhancements to
> Generic Security Services Application Program Interface (GSS-API) Version 3
> Naming" (RFC4768) was submitted to the IETF Secretariat on 2010-08-31 and has
> been posted on the "IETF Page of Intellectual Property Rights Disclosures"
> (https://datatracker.ietf.org/ipr/1398/). The title of the IPR disclosure is
> "CNRS's Statement about IPR related to RFC 4768, RFC 2069, RFC 4169, and
> draft-ietf-httpbis-p7-auth-11."

The licensing information refers to an "attached pdf file", but I do not
see any way to download it.  Can a link to it be added to the web page?

/Simon

From wwwrun@core3.amsl.com  Tue Aug 31 08:14:13 2010
Return-Path: <wwwrun@core3.amsl.com>
X-Original-To: kitten@ietf.org
Delivered-To: kitten@core3.amsl.com
Received: by core3.amsl.com (Postfix, from userid 30) id DD8713A688D; Tue, 31 Aug 2010 08:14:12 -0700 (PDT)
To: hartmans-ietf@mit.edu
From: IETF Secretariat <ietf-ipr@ietf.org>
Message-Id: <20100831151412.DD8713A688D@core3.amsl.com>
Date: Tue, 31 Aug 2010 08:14:12 -0700 (PDT)
X-Mailman-Approved-At: Tue, 31 Aug 2010 09:07:43 -0700
Cc: tim.polk@nist.gov, housley@vigilsec.com, kitten@ietf.org, turners@ieca.com, ipr-announce@ietf.org
Subject: [kitten] IPR Disclosure: CNRS's Statement about IPR related to RFC 4768, RFC 2069, RFC 4169, and draft-ietf-httpbis-p7-auth-11
X-BeenThere: kitten@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: Common Authentication Technologies - Next Generation <kitten.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/kitten>
List-Post: <mailto:kitten@ietf.org>
List-Help: <mailto:kitten-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/kitten>, <mailto:kitten-request@ietf.org?subject=subscribe>
X-List-Received-Date: Tue, 31 Aug 2010 15:14:13 -0000

Dear Sam Hartman:

An IPR disclosure that pertains to your RFC entitled "Desired Enhancements to
Generic Security Services Application Program Interface (GSS-API) Version 3
Naming" (RFC4768) was submitted to the IETF Secretariat on 2010-08-31 and has
been posted on the "IETF Page of Intellectual Property Rights Disclosures"
(https://datatracker.ietf.org/ipr/1398/). The title of the IPR disclosure is
"CNRS's Statement about IPR related to RFC 4768, RFC 2069, RFC 4169, and
draft-ietf-httpbis-p7-auth-11."

The IETF Secretariat


