
From balfanz@google.com  Thu Sep  1 17:15:12 2011
Return-Path: <balfanz@google.com>
X-Original-To: http-auth@ietfa.amsl.com
Delivered-To: http-auth@ietfa.amsl.com
Received: from localhost (localhost [127.0.0.1]) by ietfa.amsl.com (Postfix) with ESMTP id A552021F96E6 for <http-auth@ietfa.amsl.com>; Thu,  1 Sep 2011 17:15:12 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -105.976
X-Spam-Level: 
X-Spam-Status: No, score=-105.976 tagged_above=-999 required=5 tests=[BAYES_00=-2.599, FM_FORGED_GMAIL=0.622, HTML_MESSAGE=0.001, RCVD_IN_DNSWL_MED=-4, USER_IN_WHITELIST=-100]
Received: from mail.ietf.org ([12.22.58.30]) by localhost (ietfa.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id eHOH8fV9bhUW for <http-auth@ietfa.amsl.com>; Thu,  1 Sep 2011 17:15:10 -0700 (PDT)
Received: from smtp-out.google.com (smtp-out.google.com [216.239.44.51]) by ietfa.amsl.com (Postfix) with ESMTP id 9D6A221F96E4 for <http-auth@ietf.org>; Thu,  1 Sep 2011 17:15:10 -0700 (PDT)
Received: from hpaq1.eem.corp.google.com (hpaq1.eem.corp.google.com [172.25.149.1]) by smtp-out.google.com with ESMTP id p820GP51014617 for <http-auth@ietf.org>; Thu, 1 Sep 2011 17:16:25 -0700
DKIM-Signature: v=1; a=rsa-sha1; c=relaxed/relaxed; d=google.com; s=beta; t=1314922586; bh=yRkQKqLQT2uy+KIGJ12PW1JavL4=; h=MIME-Version:In-Reply-To:References:Date:Message-ID:Subject:From: To:Cc:Content-Type; b=j9K5E6BK4v5DK5SgkGge7qC/83tWPWnSVs6TOpfH4VBEUZS/TtwV39Ze4dNGmDp+a XAlYWi4EgXixq9sY7mauQ==
DomainKey-Signature: a=rsa-sha1; s=beta; d=google.com; c=nofws; q=dns; h=dkim-signature:mime-version:in-reply-to:references:date: message-id:subject:from:to:cc:content-type:x-system-of-record; b=oICe8qVke9vQHK0iVLCSg4RNhdHwfu5f/7kNRGasGj4YifNQSFwQAbzCvUoBpmqlT 894QHeRZhr2tylD/s7xMg==
Received: from yib19 (yib19.prod.google.com [10.243.65.83]) by hpaq1.eem.corp.google.com with ESMTP id p820GARM014084 (version=TLSv1/SSLv3 cipher=RC4-SHA bits=128 verify=NOT) for <http-auth@ietf.org>; Thu, 1 Sep 2011 17:16:24 -0700
Received: by yib19 with SMTP id 19so2571404yib.27 for <http-auth@ietf.org>; Thu, 01 Sep 2011 17:16:23 -0700 (PDT)
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=google.com; s=beta; h=mime-version:in-reply-to:references:date:message-id:subject:from:to :cc:content-type; bh=56N00pkc/PnlSoeWJ1eleb9aR1/XFVo2sIOQHyGX+T0=; b=UK/0XsaAQj9B6dCIll2SeegFcTGsoOfg/AfLaRsahZQBgTbZVNJBikEAixJJ4GHO+a bsx27Mau/uVydFIDxBzA==
Received: by 10.91.160.39 with SMTP id m39mr421655ago.17.1314922582745; Thu, 01 Sep 2011 17:16:22 -0700 (PDT)
MIME-Version: 1.0
Received: by 10.91.160.39 with SMTP id m39mr421643ago.17.1314922581751; Thu, 01 Sep 2011 17:16:21 -0700 (PDT)
Received: by 10.90.56.4 with HTTP; Thu, 1 Sep 2011 17:16:21 -0700 (PDT)
In-Reply-To: <CAL8DUN_p2i130VDox+SJcJn2SaafOE0JMuYi7Qb5r-2Hqy1gpw@mail.gmail.com>
References: <CADHfa2AMOeShxH_k5ZEB3DUVJAnOqvZmLMg5Yz8smtBDGkQsNg@mail.gmail.com> <4E5299F2.9050208@stpeter.im> <4E538469.8020506@cs.tcd.ie> <CADHfa2CYZ-8uODBKS6kHS-u_qGivUZFL3AF1D+SiptegLsGFQw@mail.gmail.com> <CAL8DUN_p2i130VDox+SJcJn2SaafOE0JMuYi7Qb5r-2Hqy1gpw@mail.gmail.com>
Date: Thu, 1 Sep 2011 17:16:21 -0700
Message-ID: <CADHfa2DJiQ1_4jCh=2YR9hn3Pq268TfzudNdVNeqU5aD1U-8pg@mail.gmail.com>
From: Dirk Balfanz <balfanz@google.com>
To: Yutaka OIWA <y.oiwa@aist.go.jp>
Content-Type: multipart/alternative; boundary=0016e6407cd665065604abea453b
X-System-Of-Record: true
Cc: draft-balfanz-tls-obc@tools.ietf.org, "http-auth@ietf.org" <http-auth@ietf.org>
Subject: Re: [http-auth] Fwd: [TLS] TLS-OBC proposal
X-BeenThere: http-auth@ietf.org
X-Mailman-Version: 2.1.12
Precedence: list
List-Id: HTTP authentication methods <http-auth.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/options/http-auth>, <mailto:http-auth-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/http-auth>
List-Post: <mailto:http-auth@ietf.org>
List-Help: <mailto:http-auth-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/http-auth>, <mailto:http-auth-request@ietf.org?subject=subscribe>
X-List-Received-Date: Fri, 02 Sep 2011 00:15:12 -0000

--0016e6407cd665065604abea453b
Content-Type: text/plain; charset=ISO-8859-1

On Wed, Aug 24, 2011 at 11:52 PM, Yutaka OIWA <y.oiwa@aist.go.jp> wrote:

> Hi Dirk,
>
> This does support for use of TLS, but does not support for doing
> authentication and authorization in TLS.
> Your point mentioned seems to be solved by the major solution which
> has already been widely deployed:
> using TLS for encryption and basic server authentication, and
> performing authentication in the application
> layer. What we need here more are strong authentication and channel binding
> (which I am proposing one of them).
>
> Doing authentication in the lower layer like TLS brings too much
> complexity and inconvenience to
> both applications and users.


Agreed. Which is why we don't actually attempt to authenticate the user at
the TLS layer. "Traditional" TLS ClientAuth attempts to do that (and so does
TLS/SRP, if my quick glance at the doc didn't deceive me), and I explain on
browserauth.net why I think that's broken:
http://www.browserauth.net/tls-client-authentication



>  Apart from TLS-OBC for a while, for
> example, TLS-SRP will work
> on Web only for very simple web application, but it cannot support any
> good user experiences
> for conventional large websites which usually need much control of how
> and when the users
> will be authenticated and authorized.
> TLS-OBC seems to be a bit different in this situation, but can you
> give me some more hint
> on the sales point of TLS-OBS, compared to using the combination of
> conventional TLS,
> some secure session management (e.g. secure cookie), HTTP or
> application-layer
> authentication (which the TLS-OBS is already relying on), and channel
> binding among those?
>

Channel-binding the cookie is exactly what TLS-OBC can support: see
http://www.browserauth.net/channel-bound-cookies That's what I imagine will
(continue to) be used to authenticate users - at the HTTP layer.


> And one another word, you've mentioned about use of TLS renegotiation
> for resetting the key,
> but please never rely on TLS renegotiation for any means
> other than re-establishing the session key independently with the
> upper-layer's
> requirements (e.g. solely controlled by the number of bits sent or the
> time since key generation).
> Controlling TLS's behavior from the application layer seems never be
> realistic to be implemented and
> deployed, especially on the Web server side.


Yes - agreed. Things like re-establishing the session key would totally be
up to the TLS layer - the layers on top of that don't control that. If we
channel-bind the cookie to the client's origin-bound cert like I propose,
then re-establishing the session doesn't break the cookie since the new
session will still be authenticated with the same client-side cert. So the
TLS layer can re-establish sessions without affecting the upper layers.

Dirk.

 This requirement is very strong
> especially on the Web system than others, because Web server (which has TLS
> capability) and Web applications are separate entities with loose
> coupling interfaces.
> This is not the case for other services like SMTP or IMAP: each mail
> server itself
> handles both TLS and the application logic in a single application, so
> they can e.g. kick in TLS-SRP using TLS renegotiation if desired.
>
> 2011/8/25 Dirk Balfanz <balfanz@google.com>:
> >
> >
> > On Tue, Aug 23, 2011 at 3:43 AM, Stephen Farrell <
> stephen.farrell@cs.tcd.ie>
> > wrote:
> >>
> >>
> >> On 08/22/2011 07:03 PM, Peter Saint-Andre wrote:
> >>>
> >>> Of interest.
> >>
> >> Indeed. This is quite like the idea I was trying
> >> to propose on this list before quebec, but does
> >> the negotiation inside TLS rather than as a HTTP
> >> authentication scheme.
> >>
> >> I *think* I'd still prefer the latter, but am
> >> wondering if the authors thought about that?
> >>
> >
> > I'll copy-and-paste a response I gave to a similar question on the TLS
> list
> > below::
> > I made a few attempts at doing this on HTTP layer - they didn't really
> work
> > out. One problem is security: if we look at integrity, for example, TLS
> > gives us that for free. To recreate that, say, at the HTTP layer, we'd
> have
> > to invent some form of signature on the HTTP request. So immediately the
> > question is: how much of the request do you sign? Do you sign the all
> > headers? The POST body? If you sign the whole POST body, then you can't
> > generate the signature until you've seen the whole POST body. What if the
> > POST body is a multi-gigabyte video? And if the signature goes into the
> HTTP
> > headers, which must come _before_ the POST body, this means you have read
> > the whole POST body for signing it, and then read it again to pump it out
> to
> > the network (assuming it's so big that it doesn't fit into whatever
> buffer
> > you have).
> >
> > Another problem was performance: Signing every single HTTP request with
> an
> > asymmetric key seems like overkill. So somehow, the server and client
> would
> > have to negotiate an HMAC key that would be used to sign HTTP requests.
> That
> > HMAC key would probably be visible to the attacker (which we assume sits
> in
> > the browser). To decrease the benefit that the attacker has from seeing
> the
> > HMAC key, we would re-negotiate a new HMAC key every so often, with the
> > client having to re-prove that it is in possession of its private key (we
> > assume that the attacker can't get to the private key, since that can sit
> in
> > a TPM). It's hard to imagine how to do that on the HTTP or application
> layer
> > without introducing a roundtrip between client and server that
> establishes
> > that new HMAC key. While that roundtrip is happening, you can't make
> other
> > calls to the server (well, you can make 3 others or so depending on how
> many
> > simultaneous HTTP connections your browser allows with a server, but the
> > point is that you're using one of your limited number of TCP channels for
> > this key renegotiation). In TLS, we have a similar problem - the session
> > master secret might be known to the attacker, so every now and then we'd
> > like to reset it by renegotiating the TLS session. In TLS, however, we
> can
> > interleave data and handshake messages, so renegotiating the session
> secret
> > doesn't mean that we're holding up a whole TCP connection.
> >
> > These (and a couple of other) nice features that TLS gives us for free we
> > would end up "reinventing" on the HTTP layer, up to the point where we
> would
> > have reinvented almost all of TLS.
> >
> >>
> >> One of the reasons it might be better to do this
> >> outside TLS is that I think then it'd be easier
> >> for a site to allow a user to bind >1 of these
> >> cert to the same account (you could provide the
> >> URL that allows that as a parameter maybe).
> >> Again though, maybe the authors thought that
> >> through, which I didn't;-)
> >
> > Just like today you can have multiple cookies "bound" to the same account
> > (e.g., one for the browser session you have open on your desktop, and one
> > for the browser session on your laptop), you can do the same thing with
> > TLS-OBC: you simply bind the cookie that you set on the desktop browser
> to
> > the desktop browser's cert, and the cookie that you set on the laptop
> > browser to the laptop browser's cert.
> > Does this make sense?
> > Dirk.
> >
> >>
> >> S.
> >>
> >> PS: Since the question relates to http auth, I've
> >> not cc'd the tls list.
> >>
> >>
> >>>
> >>> -------- Original Message --------
> >>> Subject:        [TLS] TLS-OBC proposal
> >>> Date:   Mon, 22 Aug 2011 10:53:05 -0700
> >>> From:   Dirk Balfanz<balfanz@google.com>
> >>> To:     tls@ietf.org
> >>>
> >>>
> >>>
> >>> Dear TLS-WG,
> >>>
> >>> I few weeks ago I presented a proposal for an "origin-bound
> >>> certificates" TLS extension at IETF 81. It's much easier to understand
> >>> what this extension is trying to accomplish when presented in a broader
> >>> context of web authentication (which I tried to give in Quebec), so I
> >>> put together a little web site that is supposed to do just that for
> >>> those who couldn't be at the IETF meeting: http://browserauth.net(This
> >>> took me a little while, hence the delay in sending out the I-D).
> >>>
> >>> Anyway, here is the
> >>> I-D: http://tools.ietf.org/html/draft-balfanz-tls-obc, which I submit
> >>> for your comments/feedback. (Again, remember to read
> >>> http://browserauth.net for some context.)
> >>>
> >>> All feedback is welcome, of course, but if you can't think of anything
> >>> to comment on, I have a couple particular questions:
> >>>
> >>> - in Section 2.1.1 we currently stipulate that the client include the
> >>> origin of the origin-bound cert as part of the OBC X.509 extension.
> When
> >>> we implemented this, we didn't really know what to do with that data on
> >>> the server side. If the client also uses TLS-SNI, then we could imagine
> >>> comparing the SNI and OBC origins in some manner on the server side,
> but
> >>> there isn't really a vehicle to signal an error back to the client
> >>> saying "your SNI and OBC origins don't match". Similarly, we could
> >>> perhaps at some higher layer compare the HTTP "Host" header with the
> OBC
> >>> origin, but then again - what do you do when they don't match? Respond
> >>> with a 400 at the HTTP layer? An alternative to the current language in
> >>> the I-D would be to simply mark the cert as origin-bound, but without
> >>> putting the origin itself into the cert. Any thoughts?
> >>>
> >>> - What do you think of the privacy-enhancing power of using per-origin
> >>> certs? The idea there is that different domains see different
> >>> "identities" (public keys) for the same client. Of course, if two
> >>> domains choose to collaborate, they can probably find a way to
> correlate
> >>> the two identities they see for the same client. This is similar to
> >>> cookies, where normally "identities" (cookies) for one domain are not
> >>> revealed to another domain unless, of course, they choose to
> >>> collaborate. How much privacy would we really lose if we just used one
> >>> certificate for all origins? It would certainly make the design and
> >>> implementation simpler. (I tend to favor per-origin certs, but I'm
> >>> curious as to what other people think.)
> >>>
> >>> Thanks for your consideration!
> >>>
> >>> Dirk.
> >>>
> >>>
> >>>
> >>>
> >>> _______________________________________________
> >>> http-auth mailing list
> >>> http-auth@ietf.org
> >>> https://www.ietf.org/mailman/listinfo/http-auth
> >
> >
> > _______________________________________________
> > http-auth mailing list
> > http-auth@ietf.org
> > https://www.ietf.org/mailman/listinfo/http-auth
> >
> >
>

--0016e6407cd665065604abea453b
Content-Type: text/html; charset=ISO-8859-1
Content-Transfer-Encoding: quoted-printable

<a href=3D"/"></a>On Wed, Aug 24, 2011 at 11:52 PM, Yutaka OIWA <span dir=
=3D"ltr">&lt;<a href=3D"mailto:y.oiwa@aist.go.jp">y.oiwa@aist.go.jp</a>&gt;=
</span> wrote:<br><div class=3D"gmail_quote"><blockquote class=3D"gmail_quo=
te" style=3D"margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex;=
">
Hi Dirk,<br>
<br>
This does support for use of TLS, but does not support for doing<br>
authentication and authorization in TLS.<br>
Your point mentioned seems to be solved by the major solution which<br>
has already been widely deployed:<br>
using TLS for encryption and basic server authentication, and<br>
performing authentication in the application<br>
layer. What we need here more are strong authentication and channel binding=
<br>
(which I am proposing one of them).<br>
<br>
Doing authentication in the lower layer like TLS brings too much<br>
complexity and inconvenience to<br>
both applications and users.</blockquote><div><br></div><div>Agreed. Which =
is why we don&#39;t actually attempt to authenticate the user at the TLS la=
yer. &quot;Traditional&quot; TLS ClientAuth attempts to do that (and so doe=
s TLS/SRP, if my quick glance at the doc didn&#39;t deceive me), and I expl=
ain on <a href=3D"http://browserauth.net">browserauth.net</a> why I think t=
hat&#39;s broken:=A0<a href=3D"http://www.browserauth.net/tls-client-authen=
tication">http://www.browserauth.net/tls-client-authentication</a></div>
<div><br></div><div>=A0</div><blockquote class=3D"gmail_quote" style=3D"mar=
gin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex;"> =A0Apart from=
 TLS-OBC for a while, for<br>
example, TLS-SRP will work<br>
on Web only for very simple web application, but it cannot support any<br>
good user experiences<br>
for conventional large websites which usually need much control of how<br>
and when the users<br>
will be authenticated and authorized.<br>
TLS-OBC seems to be a bit different in this situation, but can you<br>
give me some more hint<br>
on the sales point of TLS-OBS, compared to using the combination of<br>
conventional TLS,<br>
some secure session management (e.g. secure cookie), HTTP or application-la=
yer<br>
authentication (which the TLS-OBS is already relying on), and channel<br>
binding among those?<br></blockquote><div><br></div><div>Channel-binding th=
e cookie is exactly what TLS-OBC can support: see=A0<a href=3D"http://www.b=
rowserauth.net/channel-bound-cookies">http://www.browserauth.net/channel-bo=
und-cookies</a>=A0That&#39;s what I imagine will (continue to) be used to a=
uthenticate users - at the HTTP layer.=A0</div>
<div>=A0</div><blockquote class=3D"gmail_quote" style=3D"margin:0 0 0 .8ex;=
border-left:1px #ccc solid;padding-left:1ex;">And one another word, you&#39=
;ve mentioned about use of TLS renegotiation<br>
for resetting the key,<br>
but please never rely on TLS renegotiation for any means<br>
other than re-establishing the session key independently with the upper-lay=
er&#39;s<br>
requirements (e.g. solely controlled by the number of bits sent or the<br>
time since key generation).<br>
Controlling TLS&#39;s behavior from the application layer seems never be<br=
>
realistic to be implemented and<br>
deployed, especially on the Web server side.</blockquote><div><br></div><di=
v>Yes - agreed. Things like re-establishing the session key would totally b=
e up to the TLS layer - the layers on top of that don&#39;t control that. I=
f we channel-bind the cookie to the client&#39;s origin-bound cert like I p=
ropose, then re-establishing the session doesn&#39;t break the cookie since=
 the new session will still be authenticated with the same client-side cert=
. So the TLS layer can re-establish sessions without affecting the upper la=
yers. =A0</div>
<div>=A0</div><div>Dirk.</div><div><br></div><blockquote class=3D"gmail_quo=
te" style=3D"margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex;=
"> =A0This requirement is very strong<br>
especially on the Web system than others, because Web server (which has TLS=
<br>
capability) and Web applications are separate entities with loose<br>
coupling interfaces.<br>
This is not the case for other services like SMTP or IMAP: each mail<br>
server itself<br>
handles both TLS and the application logic in a single application, so<br>
they can e.g. kick in TLS-SRP using TLS renegotiation if desired.<br>
<br>
2011/8/25 Dirk Balfanz &lt;<a href=3D"mailto:balfanz@google.com">balfanz@go=
ogle.com</a>&gt;:<br>
<div class=3D"HOEnZb"><div></div><div class=3D"h5">&gt;<br>
&gt;<br>
&gt; On Tue, Aug 23, 2011 at 3:43 AM, Stephen Farrell &lt;<a href=3D"mailto=
:stephen.farrell@cs.tcd.ie">stephen.farrell@cs.tcd.ie</a>&gt;<br>
&gt; wrote:<br>
&gt;&gt;<br>
&gt;&gt;<br>
&gt;&gt; On 08/22/2011 07:03 PM, Peter Saint-Andre wrote:<br>
&gt;&gt;&gt;<br>
&gt;&gt;&gt; Of interest.<br>
&gt;&gt;<br>
&gt;&gt; Indeed. This is quite like the idea I was trying<br>
&gt;&gt; to propose on this list before quebec, but does<br>
&gt;&gt; the negotiation inside TLS rather than as a HTTP<br>
&gt;&gt; authentication scheme.<br>
&gt;&gt;<br>
&gt;&gt; I *think* I&#39;d still prefer the latter, but am<br>
&gt;&gt; wondering if the authors thought about that?<br>
&gt;&gt;<br>
&gt;<br>
&gt; I&#39;ll copy-and-paste a response I gave to a similar question on the=
 TLS list<br>
&gt; below::<br>
&gt; I made a few attempts at doing this on HTTP layer - they didn&#39;t re=
ally work<br>
&gt; out. One problem is security: if we look at integrity, for example, TL=
S<br>
&gt; gives us that for free. To recreate that, say, at the HTTP layer, we&#=
39;d have<br>
&gt; to invent some form of signature on the HTTP request. So immediately t=
he<br>
&gt; question is: how much of the request do you sign? Do you sign the all<=
br>
&gt; headers? The POST body? If you sign the whole POST body, then you can&=
#39;t<br>
&gt; generate the signature until you&#39;ve seen the whole POST body. What=
 if the<br>
&gt; POST body is a multi-gigabyte video? And if the signature goes into th=
e HTTP<br>
&gt; headers, which must come _before_ the POST body, this means you have r=
ead<br>
&gt; the whole POST body for signing it, and then read it again to pump it =
out to<br>
&gt; the network (assuming it&#39;s so big that it doesn&#39;t fit into wha=
tever buffer<br>
&gt; you have).<br>
&gt;<br>
&gt; Another problem was performance: Signing every single HTTP request wit=
h an<br>
&gt; asymmetric key seems like overkill. So somehow, the server and client =
would<br>
&gt; have to negotiate an HMAC key that would be used to sign HTTP requests=
. That<br>
&gt; HMAC key would probably be visible to the attacker (which we assume si=
ts in<br>
&gt; the browser). To decrease the benefit that the attacker has from seein=
g the<br>
&gt; HMAC key, we would re-negotiate a new HMAC key every so often, with th=
e<br>
&gt; client having to re-prove that it is in possession of its private key =
(we<br>
&gt; assume that the attacker can&#39;t get to the private key, since that =
can sit in<br>
&gt; a TPM). It&#39;s hard to imagine how to do that on the HTTP or applica=
tion layer<br>
&gt; without introducing a roundtrip between client and server that establi=
shes<br>
&gt; that new HMAC key. While that roundtrip is happening, you can&#39;t ma=
ke other<br>
&gt; calls to the server (well, you can make 3 others or so depending on ho=
w many<br>
&gt; simultaneous HTTP connections your browser allows with a server, but t=
he<br>
&gt; point is that you&#39;re using one of your limited number of TCP chann=
els for<br>
&gt; this key renegotiation). In TLS, we have a similar problem - the sessi=
on<br>
&gt; master secret might be known to the attacker, so every now and then we=
&#39;d<br>
&gt; like to reset it by renegotiating the TLS session. In TLS, however, we=
 can<br>
&gt; interleave data and handshake messages, so renegotiating the session s=
ecret<br>
&gt; doesn&#39;t mean that we&#39;re holding up a whole TCP connection.<br>
&gt;<br>
&gt; These (and a couple of other) nice features that TLS gives us for free=
 we<br>
&gt; would end up &quot;reinventing&quot; on the HTTP layer, up to the poin=
t where we would<br>
&gt; have reinvented almost all of TLS.<br>
&gt;<br>
&gt;&gt;<br>
&gt;&gt; One of the reasons it might be better to do this<br>
&gt;&gt; outside TLS is that I think then it&#39;d be easier<br>
&gt;&gt; for a site to allow a user to bind &gt;1 of these<br>
&gt;&gt; cert to the same account (you could provide the<br>
&gt;&gt; URL that allows that as a parameter maybe).<br>
&gt;&gt; Again though, maybe the authors thought that<br>
&gt;&gt; through, which I didn&#39;t;-)<br>
&gt;<br>
&gt; Just like today you can have multiple cookies &quot;bound&quot; to the=
 same account<br>
&gt; (e.g., one for the browser session you have open on your desktop, and =
one<br>
&gt; for the browser session on your laptop), you can do the same thing wit=
h<br>
&gt; TLS-OBC: you simply bind the cookie that you set on the desktop browse=
r to<br>
&gt; the desktop browser&#39;s cert, and the cookie that you set on the lap=
top<br>
&gt; browser to the laptop browser&#39;s cert.<br>
&gt; Does this make sense?<br>
&gt; Dirk.<br>
&gt;<br>
&gt;&gt;<br>
&gt;&gt; S.<br>
&gt;&gt;<br>
&gt;&gt; PS: Since the question relates to http auth, I&#39;ve<br>
&gt;&gt; not cc&#39;d the tls list.<br>
&gt;&gt;<br>
&gt;&gt;<br>
&gt;&gt;&gt;<br>
&gt;&gt;&gt; -------- Original Message --------<br>
&gt;&gt;&gt; Subject: =A0 =A0 =A0 =A0[TLS] TLS-OBC proposal<br>
&gt;&gt;&gt; Date: =A0 Mon, 22 Aug 2011 10:53:<a href=3D"tel:05%20-0700" va=
lue=3D"+4350700">05 -0700</a><br>
&gt;&gt;&gt; From: =A0 Dirk Balfanz&lt;<a href=3D"mailto:balfanz@google.com=
">balfanz@google.com</a>&gt;<br>
&gt;&gt;&gt; To: =A0 =A0 <a href=3D"mailto:tls@ietf.org">tls@ietf.org</a><b=
r>
&gt;&gt;&gt;<br>
&gt;&gt;&gt;<br>
&gt;&gt;&gt;<br>
&gt;&gt;&gt; Dear TLS-WG,<br>
&gt;&gt;&gt;<br>
&gt;&gt;&gt; I few weeks ago I presented a proposal for an &quot;origin-bou=
nd<br>
&gt;&gt;&gt; certificates&quot; TLS extension at IETF 81. It&#39;s much eas=
ier to understand<br>
&gt;&gt;&gt; what this extension is trying to accomplish when presented in =
a broader<br>
&gt;&gt;&gt; context of web authentication (which I tried to give in Quebec=
), so I<br>
&gt;&gt;&gt; put together a little web site that is supposed to do just tha=
t for<br>
&gt;&gt;&gt; those who couldn&#39;t be at the IETF meeting: <a href=3D"http=
://browserauth.net" target=3D"_blank">http://browserauth.net</a> (This<br>
&gt;&gt;&gt; took me a little while, hence the delay in sending out the I-D=
).<br>
&gt;&gt;&gt;<br>
&gt;&gt;&gt; Anyway, here is the<br>
&gt;&gt;&gt; I-D: <a href=3D"http://tools.ietf.org/html/draft-balfanz-tls-o=
bc" target=3D"_blank">http://tools.ietf.org/html/draft-balfanz-tls-obc</a>,=
 which I submit<br>
&gt;&gt;&gt; for your comments/feedback. (Again, remember to read<br>
&gt;&gt;&gt; <a href=3D"http://browserauth.net" target=3D"_blank">http://br=
owserauth.net</a> for some context.)<br>
&gt;&gt;&gt;<br>
&gt;&gt;&gt; All feedback is welcome, of course, but if you can&#39;t think=
 of anything<br>
&gt;&gt;&gt; to comment on, I have a couple particular questions:<br>
&gt;&gt;&gt;<br>
&gt;&gt;&gt; - in Section 2.1.1 we currently stipulate that the client incl=
ude the<br>
&gt;&gt;&gt; origin of the origin-bound cert as part of the OBC X.509 exten=
sion. When<br>
&gt;&gt;&gt; we implemented this, we didn&#39;t really know what to do with=
 that data on<br>
&gt;&gt;&gt; the server side. If the client also uses TLS-SNI, then we coul=
d imagine<br>
&gt;&gt;&gt; comparing the SNI and OBC origins in some manner on the server=
 side, but<br>
&gt;&gt;&gt; there isn&#39;t really a vehicle to signal an error back to th=
e client<br>
&gt;&gt;&gt; saying &quot;your SNI and OBC origins don&#39;t match&quot;. S=
imilarly, we could<br>
&gt;&gt;&gt; perhaps at some higher layer compare the HTTP &quot;Host&quot;=
 header with the OBC<br>
&gt;&gt;&gt; origin, but then again - what do you do when they don&#39;t ma=
tch? Respond<br>
&gt;&gt;&gt; with a 400 at the HTTP layer? An alternative to the current la=
nguage in<br>
&gt;&gt;&gt; the I-D would be to simply mark the cert as origin-bound, but =
without<br>
&gt;&gt;&gt; putting the origin itself into the cert. Any thoughts?<br>
&gt;&gt;&gt;<br>
&gt;&gt;&gt; - What do you think of the privacy-enhancing power of using pe=
r-origin<br>
&gt;&gt;&gt; certs? The idea there is that different domains see different<=
br>
&gt;&gt;&gt; &quot;identities&quot; (public keys) for the same client. Of c=
ourse, if two<br>
&gt;&gt;&gt; domains choose to collaborate, they can probably find a way to=
 correlate<br>
&gt;&gt;&gt; the two identities they see for the same client. This is simil=
ar to<br>
&gt;&gt;&gt; cookies, where normally &quot;identities&quot; (cookies) for o=
ne domain are not<br>
&gt;&gt;&gt; revealed to another domain unless, of course, they choose to<b=
r>
&gt;&gt;&gt; collaborate. How much privacy would we really lose if we just =
used one<br>
&gt;&gt;&gt; certificate for all origins? It would certainly make the desig=
n and<br>
&gt;&gt;&gt; implementation simpler. (I tend to favor per-origin certs, but=
 I&#39;m<br>
&gt;&gt;&gt; curious as to what other people think.)<br>
&gt;&gt;&gt;<br>
&gt;&gt;&gt; Thanks for your consideration!<br>
&gt;&gt;&gt;<br>
&gt;&gt;&gt; Dirk.<br>
&gt;&gt;&gt;<br>
&gt;&gt;&gt;<br>
&gt;&gt;&gt;<br>
&gt;&gt;&gt;<br>
&gt;&gt;&gt; _______________________________________________<br>
&gt;&gt;&gt; http-auth mailing list<br>
&gt;&gt;&gt; <a href=3D"mailto:http-auth@ietf.org">http-auth@ietf.org</a><b=
r>
&gt;&gt;&gt; <a href=3D"https://www.ietf.org/mailman/listinfo/http-auth" ta=
rget=3D"_blank">https://www.ietf.org/mailman/listinfo/http-auth</a><br>
&gt;<br>
&gt;<br>
&gt; _______________________________________________<br>
&gt; http-auth mailing list<br>
&gt; <a href=3D"mailto:http-auth@ietf.org">http-auth@ietf.org</a><br>
&gt; <a href=3D"https://www.ietf.org/mailman/listinfo/http-auth" target=3D"=
_blank">https://www.ietf.org/mailman/listinfo/http-auth</a><br>
&gt;<br>
&gt;<br>
</div></div></blockquote></div><br>

--0016e6407cd665065604abea453b--

From balfanz@google.com  Thu Sep  1 17:32:26 2011
Return-Path: <balfanz@google.com>
X-Original-To: http-auth@ietfa.amsl.com
Delivered-To: http-auth@ietfa.amsl.com
Received: from localhost (localhost [127.0.0.1]) by ietfa.amsl.com (Postfix) with ESMTP id 7E74621F929F for <http-auth@ietfa.amsl.com>; Thu,  1 Sep 2011 17:32:26 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -105.698
X-Spam-Level: 
X-Spam-Status: No, score=-105.698 tagged_above=-999 required=5 tests=[AWL=0.278, BAYES_00=-2.599, FM_FORGED_GMAIL=0.622, HTML_MESSAGE=0.001, RCVD_IN_DNSWL_MED=-4, USER_IN_WHITELIST=-100]
Received: from mail.ietf.org ([12.22.58.30]) by localhost (ietfa.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id Fw66GrHNCKrg for <http-auth@ietfa.amsl.com>; Thu,  1 Sep 2011 17:32:25 -0700 (PDT)
Received: from smtp-out.google.com (smtp-out.google.com [74.125.121.67]) by ietfa.amsl.com (Postfix) with ESMTP id 9BE6921F9234 for <http-auth@ietf.org>; Thu,  1 Sep 2011 17:32:24 -0700 (PDT)
Received: from wpaz33.hot.corp.google.com (wpaz33.hot.corp.google.com [172.24.198.97]) by smtp-out.google.com with ESMTP id p820XtpL013332 for <http-auth@ietf.org>; Thu, 1 Sep 2011 17:33:55 -0700
DKIM-Signature: v=1; a=rsa-sha1; c=relaxed/relaxed; d=google.com; s=beta; t=1314923636; bh=Wg07SncoV7qBdoXqHE4KbGHLvkA=; h=MIME-Version:In-Reply-To:References:Date:Message-ID:Subject:From: To:Cc:Content-Type; b=QP9RM+mo3uy0Gxvxgcomo9egamJvnNb8ZjCKPbafEIkDqKQaTyVVK0hVu+7GZqrNA f7uhrsRk/A2l47kRlrpWw==
DomainKey-Signature: a=rsa-sha1; s=beta; d=google.com; c=nofws; q=dns; h=dkim-signature:mime-version:in-reply-to:references:date: message-id:subject:from:to:cc:content-type:x-system-of-record; b=amIjYqcOuTI+Vg7A1tiE2k/y+0k0xNtZCHTCR53K4XzCxSYn5yLaq115o7abE933C i6n27zqHhyzvsMNt9KQtw==
Received: from yia27 (yia27.prod.google.com [10.243.65.27]) by wpaz33.hot.corp.google.com with ESMTP id p820XMwT027639 (version=TLSv1/SSLv3 cipher=RC4-SHA bits=128 verify=NOT) for <http-auth@ietf.org>; Thu, 1 Sep 2011 17:33:54 -0700
Received: by yia27 with SMTP id 27so2477382yia.16 for <http-auth@ietf.org>; Thu, 01 Sep 2011 17:33:54 -0700 (PDT)
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=google.com; s=beta; h=mime-version:in-reply-to:references:date:message-id:subject:from:to :cc:content-type; bh=72cupUArwxOgqeDsoCUtbXRLJ+sRhK4y7R6bJgZAU8s=; b=LNEZ6uzriCXcHPX/C/pIEWH15rjP2Hj+KGl5QZW77p7FRh3+3ysvRQNX8CSGLVcsOp Cy4vlTyvJZCqGbP8Kj6Q==
Received: by 10.91.47.27 with SMTP id z27mr99252agj.98.1314923633034; Thu, 01 Sep 2011 17:33:53 -0700 (PDT)
MIME-Version: 1.0
Received: by 10.91.47.27 with SMTP id z27mr99246agj.98.1314923632662; Thu, 01 Sep 2011 17:33:52 -0700 (PDT)
Received: by 10.90.56.4 with HTTP; Thu, 1 Sep 2011 17:33:52 -0700 (PDT)
In-Reply-To: <4E560CD6.9070100@cs.tcd.ie>
References: <CADHfa2AMOeShxH_k5ZEB3DUVJAnOqvZmLMg5Yz8smtBDGkQsNg@mail.gmail.com> <4E5299F2.9050208@stpeter.im> <4E538469.8020506@cs.tcd.ie> <CADHfa2CYZ-8uODBKS6kHS-u_qGivUZFL3AF1D+SiptegLsGFQw@mail.gmail.com> <4E560CD6.9070100@cs.tcd.ie>
Date: Thu, 1 Sep 2011 17:33:52 -0700
Message-ID: <CADHfa2DODBLLWBo67hpU-HUBpq1+cobFot4wow6gYxTZjJRbLg@mail.gmail.com>
From: Dirk Balfanz <balfanz@google.com>
To: Stephen Farrell <stephen.farrell@cs.tcd.ie>
Content-Type: multipart/alternative; boundary=001485f91e0e08557104abea846b
X-System-Of-Record: true
Cc: draft-balfanz-tls-obc@tools.ietf.org, "http-auth@ietf.org" <http-auth@ietf.org>
Subject: Re: [http-auth] Fwd: [TLS] TLS-OBC proposal
X-BeenThere: http-auth@ietf.org
X-Mailman-Version: 2.1.12
Precedence: list
List-Id: HTTP authentication methods <http-auth.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/options/http-auth>, <mailto:http-auth-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/http-auth>
List-Post: <mailto:http-auth@ietf.org>
List-Help: <mailto:http-auth-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/http-auth>, <mailto:http-auth-request@ietf.org?subject=subscribe>
X-List-Received-Date: Fri, 02 Sep 2011 00:32:26 -0000

--001485f91e0e08557104abea846b
Content-Type: text/plain; charset=ISO-8859-1

On Thu, Aug 25, 2011 at 1:50 AM, Stephen Farrell
<stephen.farrell@cs.tcd.ie>wrote:

>
>
> On 08/25/2011 07:08 AM, Dirk Balfanz wrote:
>
>> On Tue, Aug 23, 2011 at 3:43 AM, Stephen Farrell
>> <stephen.farrell@cs.tcd.ie>**wrote:
>>
>>
>>>
>>> On 08/22/2011 07:03 PM, Peter Saint-Andre wrote:
>>>
>>>  Of interest.
>>>>
>>>>
>>> Indeed. This is quite like the idea I was trying
>>> to propose on this list before quebec, but does
>>> the negotiation inside TLS rather than as a HTTP
>>> authentication scheme.
>>>
>>> I *think* I'd still prefer the latter, but am
>>> wondering if the authors thought about that?
>>>
>>>
>>>  I'll copy-and-paste a response I gave to a similar question on the TLS
>> list
>> below::
>>
>
> Yep, saw that there, but its the answer to a different question:-)
>
> What I was thinking about was having a HTTP authentication scheme
> that flagged that the server supported OBC (or equivalent) and
> then having the client (re)negotiate TLS with a new or existing
> OBC cert. So basically using TLS as-is in protocol terms, but
> flagging the OBC stuff in HTTP.


Oh I see.

How would that work in practice:

- during the first handshake, the server doesn't ask for a client cert (so
as to make sure the user-agent doesn't pop up useless UI)
- then, once the HTTP request comes in and the client says "I have an
origin-bound cert", the server instructs the TLS layer to renegotiate the
session
- this time, the server _does_ require a client cert, but somehow (how?)
makes it clear that the client should use an origin-bound cert, and not pop
up any UI to the user.

Did I get that right?

If so, then a few questions:

- doesn't that mean more round trips (immediate session renegotiation)?
- if a server _does_ want to use "traditional" TLS ClientAuth (maybe in an
enterprise setting where they have installed certs on employees' machines),
then presumably the server would send a certificate request during the
"first" handshake?
- when the TLS layer on the client side sees a certificate request, how does
it know whether to use a "traditional" cert (with UI), or an
origin-bound-cert?
- similarly, when the TLS layer on the server side receives a client
certificate, how does it know whether to employ "traditional" verification
techniques (looking for it to chain up to a corporate CA, for example), or
"OBC"-style techniques (allowing self-signed certs)? Sounds to me like the
TLS layers would have to instructed somehow by the upper layers to switch to
an "OBC" mode?

Generally speaking, what's the advantage of signalling support for OBCs on
the HTTP layer? Is it the fact that we seem be able to get away doing this
without actually defining any TLS extension?



>
>
>
>>  One of the reasons it might be better to do this
>>> outside TLS is that I think then it'd be easier
>>> for a site to allow a user to bind>1 of these
>>> cert to the same account (you could provide the
>>> URL that allows that as a parameter maybe).
>>> Again though, maybe the authors thought that
>>> through, which I didn't;-)
>>>
>>>
>> Just like today you can have multiple cookies "bound" to the same account
>> (e.g., one for the browser session you have open on your desktop, and one
>> for the browser session on your laptop), you can do the same thing with
>> TLS-OBC: you simply bind the cookie that you set on the desktop browser to
>> the desktop browser's cert, and the cookie that you set on the laptop
>> browser to the laptop browser's cert.
>>
>> Does this make sense?
>>
>
> Again I think you're answering a different question.
>
> The issue with any TLS-mutual auth scheme is that the user with >1
> device needs a way to bind those TLS client certs to one account.
> I think for any TLS mutual auth scheme to work we need a way
> to do that, including OBC or equivalent. I was asking if there's
> a way to do that with your scheme, and suggesting that maybe it'd
> be easier to do if the negotiation of OBC happened at the HTTP
> layer.
>

Mmh - that still sounds like you're asking what I though you were asking the
first time around. Maybe I'm still not quite understanding...

Anyway, here's another attempt at answering what (I think) you're asking:

- remember that the OBCs don't have user identity in them - they're
self-signed certs.
- so the two devices I have would each have their own OBC to talk to
www.example.com
- example.com, after authenticating me on the first device (in whichever
unspecified way), binds a cookie it sets on that device's browser to that
device's OBC.
- example.com, after authenticating me on the _second_ device (in whichever
unspecified way), binds a cookie it sets on that _seond_ device's browser to
that _second_ device's OBC.

So that's how the two certs would come to be bound to my account.

I'm having a feeling that I still didn't quite understand your question,
though...

Dirk.


>
> S.
>
>
>> Dirk.
>>
>>
>>
>>
>>> S.
>>>
>>> PS: Since the question relates to http auth, I've
>>> not cc'd the tls list.
>>>
>>>
>>>
>>>  -------- Original Message --------
>>>> Subject:        [TLS] TLS-OBC proposal
>>>> Date:   Mon, 22 Aug 2011 10:53:05 -0700
>>>> From:   Dirk Balfanz<balfanz@google.com>
>>>> To:     tls@ietf.org
>>>>
>>>>
>>>>
>>>> Dear TLS-WG,
>>>>
>>>> I few weeks ago I presented a proposal for an "origin-bound
>>>> certificates" TLS extension at IETF 81. It's much easier to understand
>>>> what this extension is trying to accomplish when presented in a broader
>>>> context of web authentication (which I tried to give in Quebec), so I
>>>> put together a little web site that is supposed to do just that for
>>>> those who couldn't be at the IETF meeting: http://browserauth.net (This
>>>> took me a little while, hence the delay in sending out the I-D).
>>>>
>>>> Anyway, here is the
>>>> I-D: http://tools.ietf.org/html/****draft-balfanz-tls-obc<http://tools.ietf.org/html/**draft-balfanz-tls-obc>
>>>> <http://**tools.ietf.org/html/draft-**balfanz-tls-obc<http://tools.ietf.org/html/draft-balfanz-tls-obc>
>>>> >,
>>>>
>>>> which I submit
>>>> for your comments/feedback. (Again, remember to read
>>>> http://browserauth.net for some context.)
>>>>
>>>> All feedback is welcome, of course, but if you can't think of anything
>>>> to comment on, I have a couple particular questions:
>>>>
>>>> - in Section 2.1.1 we currently stipulate that the client include the
>>>> origin of the origin-bound cert as part of the OBC X.509 extension. When
>>>> we implemented this, we didn't really know what to do with that data on
>>>> the server side. If the client also uses TLS-SNI, then we could imagine
>>>> comparing the SNI and OBC origins in some manner on the server side, but
>>>> there isn't really a vehicle to signal an error back to the client
>>>> saying "your SNI and OBC origins don't match". Similarly, we could
>>>> perhaps at some higher layer compare the HTTP "Host" header with the OBC
>>>> origin, but then again - what do you do when they don't match? Respond
>>>> with a 400 at the HTTP layer? An alternative to the current language in
>>>> the I-D would be to simply mark the cert as origin-bound, but without
>>>> putting the origin itself into the cert. Any thoughts?
>>>>
>>>> - What do you think of the privacy-enhancing power of using per-origin
>>>> certs? The idea there is that different domains see different
>>>> "identities" (public keys) for the same client. Of course, if two
>>>> domains choose to collaborate, they can probably find a way to correlate
>>>> the two identities they see for the same client. This is similar to
>>>> cookies, where normally "identities" (cookies) for one domain are not
>>>> revealed to another domain unless, of course, they choose to
>>>> collaborate. How much privacy would we really lose if we just used one
>>>> certificate for all origins? It would certainly make the design and
>>>> implementation simpler. (I tend to favor per-origin certs, but I'm
>>>> curious as to what other people think.)
>>>>
>>>> Thanks for your consideration!
>>>>
>>>> Dirk.
>>>>
>>>>
>>>>
>>>>
>>>> ______________________________****_________________
>>>> http-auth mailing list
>>>> http-auth@ietf.org
>>>> https://www.ietf.org/mailman/****listinfo/http-auth<https://www.ietf.org/mailman/**listinfo/http-auth>
>>>> <https://**www.ietf.org/mailman/listinfo/**http-auth<https://www.ietf.org/mailman/listinfo/http-auth>
>>>> >
>>>>
>>>>
>>>
>>

--001485f91e0e08557104abea846b
Content-Type: text/html; charset=ISO-8859-1
Content-Transfer-Encoding: quoted-printable

<br><br><div class=3D"gmail_quote">On Thu, Aug 25, 2011 at 1:50 AM, Stephen=
 Farrell <span dir=3D"ltr">&lt;<a href=3D"mailto:stephen.farrell@cs.tcd.ie"=
>stephen.farrell@cs.tcd.ie</a>&gt;</span> wrote:<br><blockquote class=3D"gm=
ail_quote" style=3D"margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-le=
ft:1ex;">
<div class=3D"im"><br>
<br>
On 08/25/2011 07:08 AM, Dirk Balfanz wrote:<br>
<blockquote class=3D"gmail_quote" style=3D"margin:0 0 0 .8ex;border-left:1p=
x #ccc solid;padding-left:1ex">
On Tue, Aug 23, 2011 at 3:43 AM, Stephen Farrell<br>
&lt;<a href=3D"mailto:stephen.farrell@cs.tcd.ie" target=3D"_blank">stephen.=
farrell@cs.tcd.ie</a>&gt;<u></u>wrote:<br>
<br>
<blockquote class=3D"gmail_quote" style=3D"margin:0 0 0 .8ex;border-left:1p=
x #ccc solid;padding-left:1ex">
<br>
<br>
On 08/22/2011 07:03 PM, Peter Saint-Andre wrote:<br>
<br>
<blockquote class=3D"gmail_quote" style=3D"margin:0 0 0 .8ex;border-left:1p=
x #ccc solid;padding-left:1ex">
Of interest.<br>
<br>
</blockquote>
<br>
Indeed. This is quite like the idea I was trying<br>
to propose on this list before quebec, but does<br>
the negotiation inside TLS rather than as a HTTP<br>
authentication scheme.<br>
<br>
I *think* I&#39;d still prefer the latter, but am<br>
wondering if the authors thought about that?<br>
<br>
<br>
</blockquote>
I&#39;ll copy-and-paste a response I gave to a similar question on the TLS =
list<br>
below::<br>
</blockquote>
<br></div>
Yep, saw that there, but its the answer to a different question:-)<br>
<br>
What I was thinking about was having a HTTP authentication scheme<br>
that flagged that the server supported OBC (or equivalent) and<br>
then having the client (re)negotiate TLS with a new or existing<br>
OBC cert. So basically using TLS as-is in protocol terms, but<br>
flagging the OBC stuff in HTTP.</blockquote><div><br></div><div>Oh I see.=
=A0</div><div><br></div><div>How would that work in practice:</div><div><br=
></div><div>- during the first handshake, the server doesn&#39;t ask for a =
client cert (so as to make sure the user-agent doesn&#39;t pop up useless U=
I)</div>
<div>- then, once the HTTP request comes in and the client says &quot;I hav=
e an origin-bound cert&quot;, the server instructs the TLS layer to renegot=
iate the session</div><div>- this time, the server _does_ require a client =
cert, but somehow (how?) makes it clear that the client should use an origi=
n-bound cert, and not pop up any UI to the user.</div>
<div><br></div><div>Did I get that right?=A0</div><div><br></div><div>If so=
, then a few questions:</div><div><br></div><div>- doesn&#39;t that mean mo=
re round trips (immediate session renegotiation)?</div><div>- if a server _=
does_ want to use &quot;traditional&quot; TLS ClientAuth (maybe in an enter=
prise setting where they have installed certs on employees&#39; machines), =
then presumably the server would send a certificate request during the &quo=
t;first&quot; handshake?</div>
<div>- when the TLS layer on the client side sees a certificate request, ho=
w does it know whether to use a &quot;traditional&quot; cert (with UI), or =
an origin-bound-cert?</div><div>- similarly, when the TLS layer on the serv=
er side receives a client certificate, how does it know whether to employ &=
quot;traditional&quot; verification techniques (looking for it to chain up =
to a corporate CA, for example), or &quot;OBC&quot;-style techniques (allow=
ing self-signed certs)? Sounds to me like the TLS layers would have to inst=
ructed somehow by the upper layers to switch to an &quot;OBC&quot; mode?</d=
iv>
<div><br></div><div>Generally speaking, what&#39;s the advantage of signall=
ing support for OBCs on the HTTP layer? Is it the fact that we seem be able=
 to get away doing this without actually defining any TLS extension?</div>
<div><br></div><div>=A0</div><blockquote class=3D"gmail_quote" style=3D"mar=
gin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex;"><div class=3D"=
im"><br>
<br>
<blockquote class=3D"gmail_quote" style=3D"margin:0 0 0 .8ex;border-left:1p=
x #ccc solid;padding-left:1ex">
<br>
<blockquote class=3D"gmail_quote" style=3D"margin:0 0 0 .8ex;border-left:1p=
x #ccc solid;padding-left:1ex">
One of the reasons it might be better to do this<br>
outside TLS is that I think then it&#39;d be easier<br>
for a site to allow a user to bind&gt;1 of these<br>
cert to the same account (you could provide the<br>
URL that allows that as a parameter maybe).<br>
Again though, maybe the authors thought that<br>
through, which I didn&#39;t;-)<br>
<br>
</blockquote>
<br>
Just like today you can have multiple cookies &quot;bound&quot; to the same=
 account<br>
(e.g., one for the browser session you have open on your desktop, and one<b=
r>
for the browser session on your laptop), you can do the same thing with<br>
TLS-OBC: you simply bind the cookie that you set on the desktop browser to<=
br>
the desktop browser&#39;s cert, and the cookie that you set on the laptop<b=
r>
browser to the laptop browser&#39;s cert.<br>
<br>
Does this make sense?<br>
</blockquote>
<br></div>
Again I think you&#39;re answering a different question.<br>
<br>
The issue with any TLS-mutual auth scheme is that the user with &gt;1<br>
device needs a way to bind those TLS client certs to one account.<br>
I think for any TLS mutual auth scheme to work we need a way<br>
to do that, including OBC or equivalent. I was asking if there&#39;s<br>
a way to do that with your scheme, and suggesting that maybe it&#39;d<br>
be easier to do if the negotiation of OBC happened at the HTTP<br>
layer.<br></blockquote><div><br></div><div>Mmh - that still sounds like you=
&#39;re asking what I though you were asking the first time around. Maybe I=
&#39;m still not quite understanding...</div><div><br></div><div>Anyway, he=
re&#39;s another attempt at answering what (I think) you&#39;re asking:</di=
v>
<div><br></div><div>- remember that the OBCs don&#39;t have user identity i=
n them - they&#39;re self-signed certs.</div><div>- so the two devices I ha=
ve would each have their own OBC to talk to <a href=3D"http://www.example.c=
om">www.example.com</a></div>
<div>- <a href=3D"http://example.com">example.com</a>, after authenticating=
 me on the first device (in whichever unspecified way), binds a cookie it s=
ets on that device&#39;s browser to that device&#39;s OBC.</div><div><div>
- <a href=3D"http://example.com">example.com</a>, after authenticating me o=
n the _second_ device (in whichever unspecified way), binds a cookie it set=
s on that _seond_ device&#39;s browser to that _second_ device&#39;s OBC.</=
div>
</div><div><br></div><div>So that&#39;s how the two certs would come to be =
bound to my account.=A0</div><div><br></div><div>I&#39;m having a feeling t=
hat I still didn&#39;t quite understand your question, though...=A0</div><d=
iv>
<br></div><div>Dirk.</div><div>=A0</div><blockquote class=3D"gmail_quote" s=
tyle=3D"margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex;">
<br>
S.<br>
<br>
<blockquote class=3D"gmail_quote" style=3D"margin:0 0 0 .8ex;border-left:1p=
x #ccc solid;padding-left:1ex">
<br>
Dirk.<br>
<br>
<br>
<br>
<blockquote class=3D"gmail_quote" style=3D"margin:0 0 0 .8ex;border-left:1p=
x #ccc solid;padding-left:1ex"><div class=3D"im">
<br>
S.<br>
<br>
PS: Since the question relates to http auth, I&#39;ve<br>
not cc&#39;d the tls list.<br>
<br>
<br>
<br>
</div><blockquote class=3D"gmail_quote" style=3D"margin:0 0 0 .8ex;border-l=
eft:1px #ccc solid;padding-left:1ex"><div class=3D"im">
-------- Original Message --------<br>
Subject: =A0 =A0 =A0 =A0[TLS] TLS-OBC proposal<br>
Date: =A0 Mon, 22 Aug 2011 10:53:<a href=3D"tel:05%20-0700" value=3D"+43507=
00" target=3D"_blank">05 -0700</a><br>
From: =A0 Dirk Balfanz&lt;<a href=3D"mailto:balfanz@google.com" target=3D"_=
blank">balfanz@google.com</a>&gt;<br>
To: =A0 =A0 <a href=3D"mailto:tls@ietf.org" target=3D"_blank">tls@ietf.org<=
/a><br>
<br>
<br>
<br>
Dear TLS-WG,<br>
<br>
I few weeks ago I presented a proposal for an &quot;origin-bound<br>
certificates&quot; TLS extension at IETF 81. It&#39;s much easier to unders=
tand<br>
what this extension is trying to accomplish when presented in a broader<br>
context of web authentication (which I tried to give in Quebec), so I<br>
put together a little web site that is supposed to do just that for<br>
those who couldn&#39;t be at the IETF meeting: <a href=3D"http://browseraut=
h.net" target=3D"_blank">http://browserauth.net</a> (This<br>
took me a little while, hence the delay in sending out the I-D).<br>
<br>
Anyway, here is the<br></div>
I-D: <a href=3D"http://tools.ietf.org/html/**draft-balfanz-tls-obc" target=
=3D"_blank">http://tools.ietf.org/html/**<u></u>draft-balfanz-tls-obc</a>&l=
t;<a href=3D"http://tools.ietf.org/html/draft-balfanz-tls-obc" target=3D"_b=
lank">http://<u></u>tools.ietf.org/html/draft-<u></u>balfanz-tls-obc</a>&gt=
;,<div>
<div></div><div class=3D"h5"><br>
which I submit<br>
for your comments/feedback. (Again, remember to read<br>
<a href=3D"http://browserauth.net" target=3D"_blank">http://browserauth.net=
</a> for some context.)<br>
<br>
All feedback is welcome, of course, but if you can&#39;t think of anything<=
br>
to comment on, I have a couple particular questions:<br>
<br>
- in Section 2.1.1 we currently stipulate that the client include the<br>
origin of the origin-bound cert as part of the OBC X.509 extension. When<br=
>
we implemented this, we didn&#39;t really know what to do with that data on=
<br>
the server side. If the client also uses TLS-SNI, then we could imagine<br>
comparing the SNI and OBC origins in some manner on the server side, but<br=
>
there isn&#39;t really a vehicle to signal an error back to the client<br>
saying &quot;your SNI and OBC origins don&#39;t match&quot;. Similarly, we =
could<br>
perhaps at some higher layer compare the HTTP &quot;Host&quot; header with =
the OBC<br>
origin, but then again - what do you do when they don&#39;t match? Respond<=
br>
with a 400 at the HTTP layer? An alternative to the current language in<br>
the I-D would be to simply mark the cert as origin-bound, but without<br>
putting the origin itself into the cert. Any thoughts?<br>
<br>
- What do you think of the privacy-enhancing power of using per-origin<br>
certs? The idea there is that different domains see different<br>
&quot;identities&quot; (public keys) for the same client. Of course, if two=
<br>
domains choose to collaborate, they can probably find a way to correlate<br=
>
the two identities they see for the same client. This is similar to<br>
cookies, where normally &quot;identities&quot; (cookies) for one domain are=
 not<br>
revealed to another domain unless, of course, they choose to<br>
collaborate. How much privacy would we really lose if we just used one<br>
certificate for all origins? It would certainly make the design and<br>
implementation simpler. (I tend to favor per-origin certs, but I&#39;m<br>
curious as to what other people think.)<br>
<br>
Thanks for your consideration!<br>
<br>
Dirk.<br>
<br>
<br>
<br>
<br>
______________________________<u></u>**_________________<br>
http-auth mailing list<br>
<a href=3D"mailto:http-auth@ietf.org" target=3D"_blank">http-auth@ietf.org<=
/a><br>
</div></div><a href=3D"https://www.ietf.org/mailman/**listinfo/http-auth" t=
arget=3D"_blank">https://www.ietf.org/mailman/*<u></u>*listinfo/http-auth</=
a>&lt;<a href=3D"https://www.ietf.org/mailman/listinfo/http-auth" target=3D=
"_blank">https://<u></u>www.ietf.org/mailman/listinfo/<u></u>http-auth</a>&=
gt;<br>

<br>
</blockquote>
<br>
</blockquote>
<br>
</blockquote>
</blockquote></div><br>

--001485f91e0e08557104abea846b--

From James.H.Manger@team.telstra.com  Fri Sep  2 05:40:28 2011
Return-Path: <James.H.Manger@team.telstra.com>
X-Original-To: http-auth@ietfa.amsl.com
Delivered-To: http-auth@ietfa.amsl.com
Received: from localhost (localhost [127.0.0.1]) by ietfa.amsl.com (Postfix) with ESMTP id 6A38C21F8E8D for <http-auth@ietfa.amsl.com>; Fri,  2 Sep 2011 05:40:28 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -3.545
X-Spam-Level: 
X-Spam-Status: No, score=-3.545 tagged_above=-999 required=5 tests=[AWL=-2.645, BAYES_00=-2.599, HELO_EQ_AU=0.377, HOST_EQ_AU=0.327, HTML_MESSAGE=0.001, RELAY_IS_203=0.994]
Received: from mail.ietf.org ([12.22.58.30]) by localhost (ietfa.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id ULLguHoR1H9B for <http-auth@ietfa.amsl.com>; Fri,  2 Sep 2011 05:40:27 -0700 (PDT)
Received: from ipxcno.tcif.telstra.com.au (ipxcno.tcif.telstra.com.au [203.35.82.208]) by ietfa.amsl.com (Postfix) with ESMTP id 4505621F8E8C for <http-auth@ietf.org>; Fri,  2 Sep 2011 05:40:25 -0700 (PDT)
X-IronPort-AV: E=Sophos;i="4.68,319,1312120800"; d="scan'208,217";a="44457273"
Received: from unknown (HELO ipcbni.tcif.telstra.com.au) ([10.97.216.204]) by ipocni.tcif.telstra.com.au with ESMTP; 02 Sep 2011 22:42:00 +1000
X-IronPort-AV: E=McAfee;i="5400,1158,6456"; a="36033543"
Received: from wsmsg3704.srv.dir.telstra.com ([172.49.40.197]) by ipcbni.tcif.telstra.com.au with ESMTP; 02 Sep 2011 22:42:00 +1000
Received: from WSMSG3153V.srv.dir.telstra.com ([172.49.40.159]) by WSMSG3704.srv.dir.telstra.com ([172.49.40.197]) with mapi; Fri, 2 Sep 2011 22:42:00 +1000
From: "Manger, James H" <James.H.Manger@team.telstra.com>
To: "http-auth@ietf.org" <http-auth@ietf.org>
Date: Fri, 2 Sep 2011 22:41:57 +1000
Thread-Topic: TLS-OBC: client software choosing session lifetime?
Thread-Index: AcxpNxnRimMdGTjGTF6mjZds23jpRgANjl6w
Message-ID: <255B9BB34FB7D647A506DC292726F6E1128DCF2BE3@WSMSG3153V.srv.dir.telstra.com>
Accept-Language: en-US, en-AU
Content-Language: en-US
X-MS-Has-Attach: 
X-MS-TNEF-Correlator: 
acceptlanguage: en-US, en-AU
Content-Type: multipart/alternative; boundary="_000_255B9BB34FB7D647A506DC292726F6E1128DCF2BE3WSMSG3153Vsrv_"
MIME-Version: 1.0
Subject: [http-auth] TLS-OBC: client software choosing session lifetime?
X-BeenThere: http-auth@ietf.org
X-Mailman-Version: 2.1.12
Precedence: list
List-Id: HTTP authentication methods <http-auth.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/options/http-auth>, <mailto:http-auth-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/http-auth>
List-Post: <mailto:http-auth@ietf.org>
List-Help: <mailto:http-auth-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/http-auth>, <mailto:http-auth-request@ietf.org?subject=subscribe>
X-List-Received-Date: Fri, 02 Sep 2011 12:40:28 -0000

--_000_255B9BB34FB7D647A506DC292726F6E1128DCF2BE3WSMSG3153Vsrv_
Content-Type: text/plain; charset="us-ascii"
Content-Transfer-Encoding: quoted-printable

The ideas for stronger web authentication at browserauth.net look really in=
teresting, such as Origin-Bound Certificates (TLS-OBC) [draft-balfanz-tls-o=
bc].

TLS-OBC effectively adds a (per-account per-device) identifier to TLS that =
persists across TLS sessions. The identifier is the public key in the anony=
mous self-signed origin-bound client certificate. An authenticated user-id =
(held in, say, a cookie) can be bound to this identifier.

It all works... until the origin-bound certificate is replaced.

I am wondering whether OBC's being replaced at a time of the client softwar=
e's choosing will cause problems (or at least annoyances). Today, services =
choose the lifetimes of the cookies they issue, and they can re-issue a rep=
lacement cookie if an existing one is about to expire. A service can certai=
nly choose the cookie lifetime so it doesn't expire in the next hour, for i=
nstance. An OBC-bound cookie, in contrast, cannot (securely) outlive the OB=
C. However an OBC may have been created a month ago and be due for replacem=
ent in 2 minutes. That could cause a seemingly-arbitrary glitch in the user=
 experience that the service is trying to offer.

Perhaps an OBC replacement will only occur when a TLS session cannot be res=
umed, which might generally imply a browser restart, which might be a big e=
nough break to warrant a new user login?
Perhaps the draft spec's recommended OBC lifetime of between 1 month and 1 =
year is long enough so that an OBC expiring at an inconvenient instant will=
 be rare enough to ignore?
Perhaps browsers can add heuristics to reduce OBC replacement occurring in =
the middle of something important (eg replace an OBC now if it expires in t=
he next 48 hours and it hasn't been used in the last hour)?


--
James Manger


--_000_255B9BB34FB7D647A506DC292726F6E1128DCF2BE3WSMSG3153Vsrv_
Content-Type: text/html; charset="us-ascii"
Content-Transfer-Encoding: quoted-printable

<html xmlns:v=3D"urn:schemas-microsoft-com:vml" xmlns:o=3D"urn:schemas-micr=
osoft-com:office:office" xmlns:w=3D"urn:schemas-microsoft-com:office:word" =
xmlns:m=3D"http://schemas.microsoft.com/office/2004/12/omml" xmlns=3D"http:=
//www.w3.org/TR/REC-html40"><head><META HTTP-EQUIV=3D"Content-Type" CONTENT=
=3D"text/html; charset=3Dus-ascii"><meta name=3DGenerator content=3D"Micros=
oft Word 12 (filtered medium)"><style><!--
/* Font Definitions */
@font-face
	{font-family:"Cambria Math";
	panose-1:2 4 5 3 5 4 6 3 2 4;}
@font-face
	{font-family:Calibri;
	panose-1:2 15 5 2 2 2 4 3 2 4;}
/* Style Definitions */
p.MsoNormal, li.MsoNormal, div.MsoNormal
	{margin:0cm;
	margin-bottom:.0001pt;
	font-size:11.0pt;
	font-family:"Calibri","sans-serif";}
a:link, span.MsoHyperlink
	{mso-style-priority:99;
	color:blue;
	text-decoration:underline;}
a:visited, span.MsoHyperlinkFollowed
	{mso-style-priority:99;
	color:purple;
	text-decoration:underline;}
span.EmailStyle17
	{mso-style-type:personal;
	font-family:"Calibri","sans-serif";
	color:windowtext;}
span.EmailStyle18
	{mso-style-type:personal-reply;
	font-family:"Calibri","sans-serif";
	color:#1F497D;}
.MsoChpDefault
	{mso-style-type:export-only;
	font-size:10.0pt;}
@page WordSection1
	{size:612.0pt 792.0pt;
	margin:72.0pt 72.0pt 72.0pt 72.0pt;}
div.WordSection1
	{page:WordSection1;}
--></style><!--[if gte mso 9]><xml>
<o:shapedefaults v:ext=3D"edit" spidmax=3D"1026" />
</xml><![endif]--><!--[if gte mso 9]><xml>
<o:shapelayout v:ext=3D"edit">
<o:idmap v:ext=3D"edit" data=3D"1" />
</o:shapelayout></xml><![endif]--></head><body lang=3DEN-AU link=3Dblue vli=
nk=3Dpurple><div class=3DWordSection1><p class=3DMsoNormal>The ideas for st=
ronger web authentication at browserauth.net look really interesting, such =
as Origin-Bound Certificates (TLS-OBC) [draft-balfanz-tls-obc].<o:p></o:p><=
/p><p class=3DMsoNormal><o:p>&nbsp;</o:p></p><p class=3DMsoNormal>TLS-OBC e=
ffectively adds a (per-account per-device) identifier to TLS that persists =
across TLS sessions. The identifier is the public key in the anonymous self=
-signed origin-bound client certificate. An authenticated user-id (held in,=
 say, a cookie) can be bound to this identifier.<o:p></o:p></p><p class=3DM=
soNormal><o:p>&nbsp;</o:p></p><p class=3DMsoNormal>It all works... until th=
e origin-bound certificate is replaced.<o:p></o:p></p><p class=3DMsoNormal>=
<o:p>&nbsp;</o:p></p><p class=3DMsoNormal>I am wondering whether OBC&#8217;=
s being replaced at a time of the client software&#8217;s choosing will cau=
se problems (or at least annoyances). Today, services choose the lifetimes =
of the cookies they issue, and they can re-issue a replacement cookie if an=
 existing one is about to expire. A service can certainly choose the cookie=
 lifetime so it doesn&#8217;t expire in the next hour, for instance. An OBC=
-bound cookie, in contrast, cannot (securely) outlive the OBC. However an O=
BC may have been created a month ago and be due for replacement in 2 minute=
s. That could cause a seemingly-arbitrary glitch in the user experience tha=
t the service is trying to offer.<o:p></o:p></p><p class=3DMsoNormal><o:p>&=
nbsp;</o:p></p><p class=3DMsoNormal>Perhaps an OBC replacement will only oc=
cur when a TLS session cannot be resumed, which might generally imply a bro=
wser restart, which might be a big enough break to warrant a new user login=
?<o:p></o:p></p><p class=3DMsoNormal>Perhaps the draft spec&#8217;s recomme=
nded OBC lifetime of between 1 month and 1 year is long enough so that an O=
BC expiring at an inconvenient instant will be rare enough to ignore?<o:p><=
/o:p></p><p class=3DMsoNormal>Perhaps browsers can add heuristics to reduce=
 OBC replacement occurring in the middle of something important (eg replace=
 an OBC now if it expires in the next 48 hours and it hasn&#8217;t been use=
d in the last hour)?<o:p></o:p></p><p class=3DMsoNormal><span style=3D'colo=
r:#1F497D'><o:p>&nbsp;</o:p></span></p><p class=3DMsoNormal><span style=3D'=
color:#1F497D'><o:p>&nbsp;</o:p></span></p><p class=3DMsoNormal>--<o:p></o:=
p></p><p class=3DMsoNormal>James Manger<o:p></o:p></p><p class=3DMsoNormal>=
<o:p>&nbsp;</o:p></p></div></body></html>=

--_000_255B9BB34FB7D647A506DC292726F6E1128DCF2BE3WSMSG3153Vsrv_--

From stephen.farrell@cs.tcd.ie  Wed Sep  7 06:30:39 2011
Return-Path: <stephen.farrell@cs.tcd.ie>
X-Original-To: http-auth@ietfa.amsl.com
Delivered-To: http-auth@ietfa.amsl.com
Received: from localhost (localhost [127.0.0.1]) by ietfa.amsl.com (Postfix) with ESMTP id 3F10621F8BB1 for <http-auth@ietfa.amsl.com>; Wed,  7 Sep 2011 06:30:39 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -106.599
X-Spam-Level: 
X-Spam-Status: No, score=-106.599 tagged_above=-999 required=5 tests=[BAYES_00=-2.599, RCVD_IN_DNSWL_MED=-4, USER_IN_WHITELIST=-100]
Received: from mail.ietf.org ([12.22.58.30]) by localhost (ietfa.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id 5koPcNZ5Ot0Y for <http-auth@ietfa.amsl.com>; Wed,  7 Sep 2011 06:30:35 -0700 (PDT)
Received: from scss.tcd.ie (hermes.cs.tcd.ie [134.226.32.56]) by ietfa.amsl.com (Postfix) with ESMTP id B6B7121F8C04 for <http-auth@ietf.org>; Wed,  7 Sep 2011 06:30:34 -0700 (PDT)
Received: from localhost (localhost [127.0.0.1]) by hermes.scss.tcd.ie (Postfix) with ESMTP id CF3D8171F4B; Wed,  7 Sep 2011 14:32:19 +0100 (IST)
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple; d=cs.tcd.ie; h= content-transfer-encoding:content-type:in-reply-to:references :subject:mime-version:user-agent:from:date:message-id:received :received:x-virus-scanned; s=cs; t=1315402335; bh=xdP/il4wabOgQC ZSxcUqOoBCoydL+1VlUH583js+ecE=; b=lOWq5ku4Xdl6R1hkvqd2TH8KOdYV23 u5JwIVcizPjDTdrCco9t0LdiknDPFMXazWtSPN6b4A8/s1MY6Aq1WBVD1TbvJR7U wnwwTyQVRdbYjPxFgQSR44YRLCcsXyssC03kXBk255/nZ3OY0eAnpXrMKL74DAWB BYkmcl3Pq4ArJBoAHErEG+jnRJq+ZHEZj9t223x0uyTa4jgOYr3FymEOiKe5pdwm xqEh9EYNyWPiHRt8oMMKjgM7U/ImDNfB1VWQ00OZ+f+RgqEjSp/w8/0nGNNOpDXA uyOK9O1b1yNzDuwW1XoyBYGUO++bAOBuP/n6vL8opDRLN1Khm2gO+G0w==
X-Virus-Scanned: Debian amavisd-new at scss.tcd.ie
Received: from scss.tcd.ie ([127.0.0.1]) by localhost (scss.tcd.ie [127.0.0.1]) (amavisd-new, port 10027) with ESMTP id rFun4FSOfY6i; Wed,  7 Sep 2011 14:32:15 +0100 (IST)
Received: from [10.87.48.8] (unknown [86.42.26.104]) by smtp.scss.tcd.ie (Postfix) with ESMTPSA id 1DC76171BFE; Wed,  7 Sep 2011 14:32:13 +0100 (IST)
Message-ID: <4E67725D.1070801@cs.tcd.ie>
Date: Wed, 07 Sep 2011 14:32:13 +0100
From: Stephen Farrell <stephen.farrell@cs.tcd.ie>
User-Agent: Mozilla/5.0 (X11; Linux i686 on x86_64; rv:6.0.1) Gecko/20110830 Thunderbird/6.0.1
MIME-Version: 1.0
To: Dirk Balfanz <balfanz@google.com>
References: <CADHfa2AMOeShxH_k5ZEB3DUVJAnOqvZmLMg5Yz8smtBDGkQsNg@mail.gmail.com> <4E5299F2.9050208@stpeter.im> <4E538469.8020506@cs.tcd.ie> <CADHfa2CYZ-8uODBKS6kHS-u_qGivUZFL3AF1D+SiptegLsGFQw@mail.gmail.com> <4E560CD6.9070100@cs.tcd.ie> <CADHfa2DODBLLWBo67hpU-HUBpq1+cobFot4wow6gYxTZjJRbLg@mail.gmail.com>
In-Reply-To: <CADHfa2DODBLLWBo67hpU-HUBpq1+cobFot4wow6gYxTZjJRbLg@mail.gmail.com>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit
Cc: draft-balfanz-tls-obc@tools.ietf.org, "http-auth@ietf.org" <http-auth@ietf.org>
Subject: Re: [http-auth] Fwd: [TLS] TLS-OBC proposal
X-BeenThere: http-auth@ietf.org
X-Mailman-Version: 2.1.12
Precedence: list
List-Id: HTTP authentication methods <http-auth.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/options/http-auth>, <mailto:http-auth-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/http-auth>
List-Post: <mailto:http-auth@ietf.org>
List-Help: <mailto:http-auth-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/http-auth>, <mailto:http-auth-request@ietf.org?subject=subscribe>
X-List-Received-Date: Wed, 07 Sep 2011 13:30:39 -0000

Hi Dirk,

Sorry for being slow responding. I will get back to it
though soon's I can.

S

On 09/02/2011 01:33 AM, Dirk Balfanz wrote:
> On Thu, Aug 25, 2011 at 1:50 AM, Stephen Farrell
> <stephen.farrell@cs.tcd.ie>wrote:
>
>>
>>
>> On 08/25/2011 07:08 AM, Dirk Balfanz wrote:
>>
>>> On Tue, Aug 23, 2011 at 3:43 AM, Stephen Farrell
>>> <stephen.farrell@cs.tcd.ie>**wrote:
>>>
>>>
>>>>
>>>> On 08/22/2011 07:03 PM, Peter Saint-Andre wrote:
>>>>
>>>>   Of interest.
>>>>>
>>>>>
>>>> Indeed. This is quite like the idea I was trying
>>>> to propose on this list before quebec, but does
>>>> the negotiation inside TLS rather than as a HTTP
>>>> authentication scheme.
>>>>
>>>> I *think* I'd still prefer the latter, but am
>>>> wondering if the authors thought about that?
>>>>
>>>>
>>>>   I'll copy-and-paste a response I gave to a similar question on the TLS
>>> list
>>> below::
>>>
>>
>> Yep, saw that there, but its the answer to a different question:-)
>>
>> What I was thinking about was having a HTTP authentication scheme
>> that flagged that the server supported OBC (or equivalent) and
>> then having the client (re)negotiate TLS with a new or existing
>> OBC cert. So basically using TLS as-is in protocol terms, but
>> flagging the OBC stuff in HTTP.
>
>
> Oh I see.
>
> How would that work in practice:
>
> - during the first handshake, the server doesn't ask for a client cert (so
> as to make sure the user-agent doesn't pop up useless UI)
> - then, once the HTTP request comes in and the client says "I have an
> origin-bound cert", the server instructs the TLS layer to renegotiate the
> session
> - this time, the server _does_ require a client cert, but somehow (how?)
> makes it clear that the client should use an origin-bound cert, and not pop
> up any UI to the user.
>
> Did I get that right?
>
> If so, then a few questions:
>
> - doesn't that mean more round trips (immediate session renegotiation)?
> - if a server _does_ want to use "traditional" TLS ClientAuth (maybe in an
> enterprise setting where they have installed certs on employees' machines),
> then presumably the server would send a certificate request during the
> "first" handshake?
> - when the TLS layer on the client side sees a certificate request, how does
> it know whether to use a "traditional" cert (with UI), or an
> origin-bound-cert?
> - similarly, when the TLS layer on the server side receives a client
> certificate, how does it know whether to employ "traditional" verification
> techniques (looking for it to chain up to a corporate CA, for example), or
> "OBC"-style techniques (allowing self-signed certs)? Sounds to me like the
> TLS layers would have to instructed somehow by the upper layers to switch to
> an "OBC" mode?
>
> Generally speaking, what's the advantage of signalling support for OBCs on
> the HTTP layer? Is it the fact that we seem be able to get away doing this
> without actually defining any TLS extension?
>
>
>
>>
>>
>>
>>>   One of the reasons it might be better to do this
>>>> outside TLS is that I think then it'd be easier
>>>> for a site to allow a user to bind>1 of these
>>>> cert to the same account (you could provide the
>>>> URL that allows that as a parameter maybe).
>>>> Again though, maybe the authors thought that
>>>> through, which I didn't;-)
>>>>
>>>>
>>> Just like today you can have multiple cookies "bound" to the same account
>>> (e.g., one for the browser session you have open on your desktop, and one
>>> for the browser session on your laptop), you can do the same thing with
>>> TLS-OBC: you simply bind the cookie that you set on the desktop browser to
>>> the desktop browser's cert, and the cookie that you set on the laptop
>>> browser to the laptop browser's cert.
>>>
>>> Does this make sense?
>>>
>>
>> Again I think you're answering a different question.
>>
>> The issue with any TLS-mutual auth scheme is that the user with>1
>> device needs a way to bind those TLS client certs to one account.
>> I think for any TLS mutual auth scheme to work we need a way
>> to do that, including OBC or equivalent. I was asking if there's
>> a way to do that with your scheme, and suggesting that maybe it'd
>> be easier to do if the negotiation of OBC happened at the HTTP
>> layer.
>>
>
> Mmh - that still sounds like you're asking what I though you were asking the
> first time around. Maybe I'm still not quite understanding...
>
> Anyway, here's another attempt at answering what (I think) you're asking:
>
> - remember that the OBCs don't have user identity in them - they're
> self-signed certs.
> - so the two devices I have would each have their own OBC to talk to
> www.example.com
> - example.com, after authenticating me on the first device (in whichever
> unspecified way), binds a cookie it sets on that device's browser to that
> device's OBC.
> - example.com, after authenticating me on the _second_ device (in whichever
> unspecified way), binds a cookie it sets on that _seond_ device's browser to
> that _second_ device's OBC.
>
> So that's how the two certs would come to be bound to my account.
>
> I'm having a feeling that I still didn't quite understand your question,
> though...
>
> Dirk.
>
>
>>
>> S.
>>
>>
>>> Dirk.
>>>
>>>
>>>
>>>
>>>> S.
>>>>
>>>> PS: Since the question relates to http auth, I've
>>>> not cc'd the tls list.
>>>>
>>>>
>>>>
>>>>   -------- Original Message --------
>>>>> Subject:        [TLS] TLS-OBC proposal
>>>>> Date:   Mon, 22 Aug 2011 10:53:05 -0700
>>>>> From:   Dirk Balfanz<balfanz@google.com>
>>>>> To:     tls@ietf.org
>>>>>
>>>>>
>>>>>
>>>>> Dear TLS-WG,
>>>>>
>>>>> I few weeks ago I presented a proposal for an "origin-bound
>>>>> certificates" TLS extension at IETF 81. It's much easier to understand
>>>>> what this extension is trying to accomplish when presented in a broader
>>>>> context of web authentication (which I tried to give in Quebec), so I
>>>>> put together a little web site that is supposed to do just that for
>>>>> those who couldn't be at the IETF meeting: http://browserauth.net (This
>>>>> took me a little while, hence the delay in sending out the I-D).
>>>>>
>>>>> Anyway, here is the
>>>>> I-D: http://tools.ietf.org/html/****draft-balfanz-tls-obc<http://tools.ietf.org/html/**draft-balfanz-tls-obc>
>>>>> <http://**tools.ietf.org/html/draft-**balfanz-tls-obc<http://tools.ietf.org/html/draft-balfanz-tls-obc>
>>>>>> ,
>>>>>
>>>>> which I submit
>>>>> for your comments/feedback. (Again, remember to read
>>>>> http://browserauth.net for some context.)
>>>>>
>>>>> All feedback is welcome, of course, but if you can't think of anything
>>>>> to comment on, I have a couple particular questions:
>>>>>
>>>>> - in Section 2.1.1 we currently stipulate that the client include the
>>>>> origin of the origin-bound cert as part of the OBC X.509 extension. When
>>>>> we implemented this, we didn't really know what to do with that data on
>>>>> the server side. If the client also uses TLS-SNI, then we could imagine
>>>>> comparing the SNI and OBC origins in some manner on the server side, but
>>>>> there isn't really a vehicle to signal an error back to the client
>>>>> saying "your SNI and OBC origins don't match". Similarly, we could
>>>>> perhaps at some higher layer compare the HTTP "Host" header with the OBC
>>>>> origin, but then again - what do you do when they don't match? Respond
>>>>> with a 400 at the HTTP layer? An alternative to the current language in
>>>>> the I-D would be to simply mark the cert as origin-bound, but without
>>>>> putting the origin itself into the cert. Any thoughts?
>>>>>
>>>>> - What do you think of the privacy-enhancing power of using per-origin
>>>>> certs? The idea there is that different domains see different
>>>>> "identities" (public keys) for the same client. Of course, if two
>>>>> domains choose to collaborate, they can probably find a way to correlate
>>>>> the two identities they see for the same client. This is similar to
>>>>> cookies, where normally "identities" (cookies) for one domain are not
>>>>> revealed to another domain unless, of course, they choose to
>>>>> collaborate. How much privacy would we really lose if we just used one
>>>>> certificate for all origins? It would certainly make the design and
>>>>> implementation simpler. (I tend to favor per-origin certs, but I'm
>>>>> curious as to what other people think.)
>>>>>
>>>>> Thanks for your consideration!
>>>>>
>>>>> Dirk.
>>>>>
>>>>>
>>>>>
>>>>>
>>>>> ______________________________****_________________
>>>>> http-auth mailing list
>>>>> http-auth@ietf.org
>>>>> https://www.ietf.org/mailman/****listinfo/http-auth<https://www.ietf.org/mailman/**listinfo/http-auth>
>>>>> <https://**www.ietf.org/mailman/listinfo/**http-auth<https://www.ietf.org/mailman/listinfo/http-auth>
>>>>>>
>>>>>
>>>>>
>>>>
>>>
>

From lef.mutualauth@gmail.com  Fri Sep 30 22:40:21 2011
Return-Path: <lef.mutualauth@gmail.com>
X-Original-To: http-auth@ietfa.amsl.com
Delivered-To: http-auth@ietfa.amsl.com
Received: from localhost (localhost [127.0.0.1]) by ietfa.amsl.com (Postfix) with ESMTP id 66AB921F8B3E; Fri, 30 Sep 2011 22:40:21 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -3.598
X-Spam-Level: 
X-Spam-Status: No, score=-3.598 tagged_above=-999 required=5 tests=[BAYES_00=-2.599, HTML_MESSAGE=0.001, RCVD_IN_DNSWL_LOW=-1]
Received: from mail.ietf.org ([12.22.58.30]) by localhost (ietfa.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id 6lZoHCF3oiJ0; Fri, 30 Sep 2011 22:40:20 -0700 (PDT)
Received: from mail-ww0-f44.google.com (mail-ww0-f44.google.com [74.125.82.44]) by ietfa.amsl.com (Postfix) with ESMTP id 11EA921F8B29; Fri, 30 Sep 2011 22:40:19 -0700 (PDT)
Received: by wwf22 with SMTP id 22so2512930wwf.13 for <multiple recipients>; Fri, 30 Sep 2011 22:43:15 -0700 (PDT)
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=gamma; h=mime-version:in-reply-to:references:date:message-id:subject:from:to :cc:content-type; bh=i7Bo8h5pZC9ukKVJUpfyqtdueo4nK4C7vKaqUMHkPH8=; b=fWaB+BEYLzObKMm5rK2UouwFVeDs9QNkRvHYhlRKvHfOTUx+hPgPQ2YmLhQtDTObjz QvAqdsPGdQotFR5VxdEWnhnJD+HwIt6qYc7RjgY31lb1vDGfwmPAKG4zqZPrBjJ1B/8Z 5lWxDSOjkamCgMNq9jrO2926Us6E3XOs9A36k=
MIME-Version: 1.0
Received: by 10.216.229.103 with SMTP id g81mr9037487weq.0.1317447795219; Fri, 30 Sep 2011 22:43:15 -0700 (PDT)
Received: by 10.216.37.146 with HTTP; Fri, 30 Sep 2011 22:43:15 -0700 (PDT)
In-Reply-To: <4E2CA293.70603@stpeter.im>
References: <5FA6AD59-7570-4A85-B6D1-3DC8E42688F1@mnot.net> <234F16BC-9875-474B-95B3-D61E8BE5A6E0@checkpoint.com> <CAK3OfOigCA1Jkv6qgc+kF-43Bavgxdv-twVs6au+B3qWWsbDvA@mail.gmail.com> <4E2CA293.70603@stpeter.im>
Date: Sat, 1 Oct 2011 14:43:15 +0900
Message-ID: <CAGipQFmN89+Q26AGsocnWbh1nzcH0xgJAC8oZOYVw-cn1L9mGw@mail.gmail.com>
From: "HAYASHI, Tatsuya" <lef.mutualauth@gmail.com>
To: Peter Saint-Andre <stpeter@stpeter.im>
Content-Type: multipart/alternative; boundary=0016e6469a2ed81bc904ae3637e9
Cc: "http-auth@ietf.org" <http-auth@ietf.org>, "apps-discuss@ietf.org Discuss" <apps-discuss@ietf.org>
Subject: Re: [http-auth] HTTP-Auth BoF in Quebec City Postponed
X-BeenThere: http-auth@ietf.org
X-Mailman-Version: 2.1.12
Precedence: list
List-Id: HTTP authentication methods <http-auth.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/options/http-auth>, <mailto:http-auth-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/http-auth>
List-Post: <mailto:http-auth@ietf.org>
List-Help: <mailto:http-auth-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/http-auth>, <mailto:http-auth-request@ietf.org?subject=subscribe>
X-List-Received-Date: Sat, 01 Oct 2011 05:40:21 -0000

--0016e6469a2ed81bc904ae3637e9
Content-Type: text/plain; charset=ISO-8859-1

Hi! Peter.

The cut-off date of BOF proposal requests in IETF Taipei is coming soon.
Taking account of the recent status of this list, I don't think we
have a BOF in Taipei. However, I want to improve the authentication in
the web too, so is there any intention to have a side meeting to
clarify the scope and the problem?

As a co-author of the problem statement draft by Yutaka Oiwa, I want
the draft enhanced by other guys familiar with authentication.
(We are updating a draft!)

Sincerely,

On Monday, July 25, 2011, Peter Saint-Andre <stpeter@stpeter.im> wrote:
> On 7/24/11 3:37 PM, Nico Williams wrote:
>> Sadly I cannot attend.  I planned my travel schedule for this week
>> around the original BoF, and so I depart Quebec in the afternoon.  I
>> look forward to the notes from any bar BoF :)
>
> A few of us got together just now in a side room to chat informally.
> We'll post further about some actions sometime soon. :)
>
> Peter
>
> --
> Peter Saint-Andre
> https://stpeter.im/
>
>
> _______________________________________________
> apps-discuss mailing list
> apps-discuss@ietf.org
> https://www.ietf.org/mailman/listinfo/apps-discuss
>

-- 
--
HAYASHI, Tatsuya
Lepidum Co. Ltd.

--0016e6469a2ed81bc904ae3637e9
Content-Type: text/html; charset=ISO-8859-1
Content-Transfer-Encoding: quoted-printable

<br>Hi! Peter.<br><br>The cut-off date of BOF proposal requests in IETF Tai=
pei is coming soon.<br>Taking account of the recent status of this list, I =
don&#39;t think we<br>have a BOF in Taipei. However, I want to improve the =
authentication in<br>
the web too, so is there any intention to have a side meeting to<br>clarify=
 the scope and the problem?<br><br>As a co-author of the problem statement =
draft by Yutaka Oiwa, I want<br>the draft enhanced by other guys familiar w=
ith authentication.<br>
(We are updating a draft!)<br><br>Sincerely,<br><br>On Monday, July 25, 201=
1, Peter Saint-Andre &lt;<a href=3D"mailto:stpeter@stpeter.im">stpeter@stpe=
ter.im</a>&gt; wrote:<br>&gt; On 7/24/11 3:37 PM, Nico Williams wrote:<br>
&gt;&gt; Sadly I cannot attend. =A0I planned my travel schedule for this we=
ek<br>&gt;&gt; around the original BoF, and so I depart Quebec in the after=
noon. =A0I<br>&gt;&gt; look forward to the notes from any bar BoF :)<br>&gt=
;<br>
&gt; A few of us got together just now in a side room to chat informally.<b=
r>&gt; We&#39;ll post further about some actions sometime soon. :)<br>&gt;<=
br>&gt; Peter<br>&gt;<br>&gt; --<br>&gt; Peter Saint-Andre<br>&gt; <a href=
=3D"https://stpeter.im/">https://stpeter.im/</a><br>
&gt;<br>&gt;<br>&gt; _______________________________________________<br>&gt=
; apps-discuss mailing list<br>&gt; <a href=3D"mailto:apps-discuss@ietf.org=
">apps-discuss@ietf.org</a><br>&gt; <a href=3D"https://www.ietf.org/mailman=
/listinfo/apps-discuss">https://www.ietf.org/mailman/listinfo/apps-discuss<=
/a><br>
&gt;<br><br>-- <br>--<br>HAYASHI, Tatsuya<br>Lepidum Co. Ltd.<br><br>

--0016e6469a2ed81bc904ae3637e9--
