
From elf-pavlik@hackers4peace.net  Fri Sep  3 06:55:20 2021
Return-Path: <elf-pavlik@hackers4peace.net>
X-Original-To: txauth@ietfa.amsl.com
Delivered-To: txauth@ietfa.amsl.com
Received: from localhost (localhost [127.0.0.1]) by ietfa.amsl.com (Postfix) with ESMTP id 907023A1F48 for <txauth@ietfa.amsl.com>; Fri,  3 Sep 2021 06:55:20 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -1.701
X-Spam-Level: 
X-Spam-Status: No, score=-1.701 tagged_above=-999 required=5 tests=[BAYES_00=-1.9, DKIM_INVALID=0.1, DKIM_SIGNED=0.1, RCVD_IN_MSPIKE_H2=-0.001, SPF_PASS=-0.001, URIBL_BLOCKED=0.001] autolearn=no autolearn_force=no
Authentication-Results: ietfa.amsl.com (amavisd-new); dkim=neutral reason="invalid (public key: not available)" header.d=hackers4peace.net
Received: from mail.ietf.org ([4.31.198.44]) by localhost (ietfa.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id 32xVy-S-Wt0X for <txauth@ietfa.amsl.com>; Fri,  3 Sep 2021 06:55:14 -0700 (PDT)
Received: from email.ecobytes.net (email.ecobytes.net [176.9.199.32]) (using TLSv1.2 with cipher ADH-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by ietfa.amsl.com (Postfix) with ESMTPS id 984153A1F4B for <txauth@ietf.org>; Fri,  3 Sep 2021 06:55:13 -0700 (PDT)
Received: from localhost (localhost [127.0.0.1]) by email.ecobytes.net (Postfix) with ESMTP id 471BF620CE for <txauth@ietf.org>; Fri,  3 Sep 2021 13:55:08 +0000 (UTC)
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=hackers4peace.net; s=modoboa; t=1630677308; bh=DUUubME+EHNUwZqTgViJyez3Xpof9ayxroYbo0yE4ic=; h=Date:From:To:Subject:From; b=l7g2Bz3e+ZMTA1vXzT6Yw24neqiEZvLtIX3clH9W6UWt2AXm3IpVzcK7LSnq08MBD Ug3EmJxGQQOrvj/rMpZIURyP6s3YTNSIjzyoqTKpEmwdO9SD7JrQOBxU3KKipZ4Qoe EG8V9eQu+jbSsW92HB8eBzagEp2y57xCQG694vEMMyGvTfPjYkyFTvZVlJzOENR21y K28szz2pqLGjO2vUjWxeNi1p0r+f/h6IhENCL0ymZqHQLDRkShNDfmaXKRf1BInnoP ciWTi/tlGKEivpz8Pu0bC7ON+/jIJnllOWUSCutKXa/tA1uBeP+NyEYSSlcbFalzGX vyMcszaZtX0qQ==
X-Virus-Scanned: Debian amavisd-new at email.ecobytes.net
Received: from email.ecobytes.net ([127.0.0.1]) by localhost (email.ecobytes.net [127.0.0.1]) (amavisd-new, port 10026) with ESMTP id mkgenr_RZtPs for <txauth@ietf.org>; Fri,  3 Sep 2021 13:55:06 +0000 (UTC)
Received: from webmail.ecobytes.net (geogale.ecobytes.net [88.99.144.132]) (using TLSv1.3 with cipher TLS_AES_256_GCM_SHA384 (256/256 bits) key-exchange ECDHE (P-256) server-signature RSA-PSS (2048 bits) server-digest SHA256) (No client certificate requested) by email.ecobytes.net (Postfix) with ESMTPSA for <txauth@ietf.org>; Fri,  3 Sep 2021 13:55:06 +0000 (UTC)
MIME-Version: 1.0
Date: Fri, 03 Sep 2021 08:55:06 -0500
From: elf-pavlik@hackers4peace.net
To: txauth@ietf.org
User-Agent: Roundcube Webmail/1.4.11
Message-ID: <10c8c1d93688494588a9d7032a0aa37d@hackers4peace.net>
X-Sender: elf-pavlik@hackers4peace.net
Content-Type: text/plain; charset=US-ASCII; format=flowed
Content-Transfer-Encoding: 7bit
Archived-At: <https://mailarchive.ietf.org/arch/msg/txauth/nhV4f59Xf5EaZ8c-YjkWWN6oCr0>
Subject: [GNAP] Distinct role used by End-user to give permissions to client instances.
X-BeenThere: txauth@ietf.org
X-Mailman-Version: 2.1.29
Precedence: list
List-Id: GNAP <txauth.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/options/txauth>, <mailto:txauth-request@ietf.org?subject=unsubscribe>
List-Archive: <https://mailarchive.ietf.org/arch/browse/txauth/>
List-Post: <mailto:txauth@ietf.org>
List-Help: <mailto:txauth-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/txauth>, <mailto:txauth-request@ietf.org?subject=subscribe>
X-List-Received-Date: Fri, 03 Sep 2021 13:57:11 -0000

Hello,

I'm participating in AuthN and AuthZ related work of 
https://solidproject.org/

I will intentionally keep this email short and provide further details 
based on responses.
Given that I'll try to focus on one problem we are trying to address and 
we can start zooming out from there as needed.

In our scenarios different End-user and Resource Owner is very common 
scenario.
We are working on few proposals for handing User Authorization, in that 
case ACME organization (Resource Owner) could authorize Alice (End-user) 
to access specific data.
In the same way ACME authorizes Bob (another End-user) to access some 
specific data. Data that Alice and Bob can access have certain overlap.

Now both Alice and Bob should be able to independently permission / 
authorize various applications (Client instances) they wan to access 
that data owned by ACME.
In fact both Alice and Bob have broader peer networks, including other 
Organizations and Individuals. We are also defining common permission 
model,
which would allow expressing application access to span across data 
owned by any number of owners.

In the scenario above, we consider assumption that given Resource Owner 
owns all the data on given Resource Server.
In that case Authorization Server which issues Access Tokens stays 
associated with RS, so also with specific Resource Owner.
Since each End-user needs some party which would take responsibility for 
them to give consent and permission apps (Client instances) they are 
using.
We look at another role let's just call it EUXS (End-user's X Server), 
similar to OIDC Provider. It stays associated with End-user in similar 
way to rel="http://openid.net/specs/connect/1.0/issuer" links to OP.

 From here it would be really great to see how GNAP would help handing 
following:

End-user's X Server, in a similar way to OIDC Provider, issues at token 
(let's call it a permissions token) to Client instance representing 
permissions applicable to a specific RS.
Once again very often that RS is controlled by Resource Owner other than 
End-user.
Client instance uses that 'permissions' token with AS to get Access 
Token which finally can be used when making requests to RS.
Here I make assumption that AS understands that permissions token and 
later either makes it available to RS or translates it to something that 
RS can enforce.
Oh, also RS would hint their AS with as_uri in WWW-Authenticate, since 
client doesn't have that prior knowledge.

As I mentioned User Authorization is handled separately so AS & RS have 
prior knowledge of what access Resource Owner granted to End-user.
In this email I mostly focus on End-user having another party associated 
with them (EUXS for the sake of this discussion) where they
can further delegate that access their received to apps (Client 
instances) which they want to use to exercise access to data.

I hope description above is sufficient to roughly illustrate the 
scenario. I can fill in more details as needed.

I very much appreciate any hints on how GNAP can handle described 
scenario in elegant way.

Best regards,
elf Pavlik


From nobody Fri Sep  3 12:50:50 2021
Return-Path: <agropper@gmail.com>
X-Original-To: txauth@ietfa.amsl.com
Delivered-To: txauth@ietfa.amsl.com
Received: from localhost (localhost [127.0.0.1]) by ietfa.amsl.com (Postfix) with ESMTP id A7EBA3A2B57 for <txauth@ietfa.amsl.com>; Fri,  3 Sep 2021 12:50:48 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -1.645
X-Spam-Level: 
X-Spam-Status: No, score=-1.645 tagged_above=-999 required=5 tests=[BAYES_00=-1.9, FREEMAIL_FORGED_FROMDOMAIN=0.249, FREEMAIL_FROM=0.001, HEADER_FROM_DIFFERENT_DOMAINS=0.001, HTML_MESSAGE=0.001, RCVD_IN_MSPIKE_H3=0.001, RCVD_IN_MSPIKE_WL=0.001, SPF_HELO_NONE=0.001, SPF_PASS=-0.001, URIBL_BLOCKED=0.001] autolearn=no autolearn_force=no
Received: from mail.ietf.org ([4.31.198.44]) by localhost (ietfa.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id Q8n5eV91Ko8G for <txauth@ietfa.amsl.com>; Fri,  3 Sep 2021 12:50:43 -0700 (PDT)
Received: from mail-qt1-f182.google.com (mail-qt1-f182.google.com [209.85.160.182]) (using TLSv1.2 with cipher ECDHE-RSA-AES128-GCM-SHA256 (128/128 bits)) (No client certificate requested) by ietfa.amsl.com (Postfix) with ESMTPS id 9C29B3A2B56 for <txauth@ietf.org>; Fri,  3 Sep 2021 12:50:43 -0700 (PDT)
Received: by mail-qt1-f182.google.com with SMTP id s15so172350qta.10 for <txauth@ietf.org>; Fri, 03 Sep 2021 12:50:43 -0700 (PDT)
X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20161025; h=x-gm-message-state:mime-version:references:in-reply-to:from:date :message-id:subject:to:cc; bh=9gY3GZDTyiEvevwJeo75sEbahuqEtQolaIiy8RtLOkg=; b=cU2vaNIzO9uq4AukkhAWYBfzlgMaEZWN4bYwE3cYWS37XO0wdTnw3k4wG4drVYKHIP murqAVsREkPB5pS4uOyz7vc86sAdELRqU2gHQTVR5lkoJd/irhSn1QhGc+ihY1y12Kw8 O5JicbJGsFsuK3hL9XlbJBL2Kmf12n45dDhvYSJcl6HNE7O5+7F0X9ILyF5fVQJ2hO8l dS7pl4GfOOPfmbxsDXAEknFQ2mwdR2LHrB0YGhrRMUve7L2d/M+jbMfo3GC77V+MmzaM LAP4L0KTNe3mVLL1Zani/lbWvMnfOkH/3ebe2Fl/jNrIqER5B1ADnuHE0/dtWsj5nRXU 0/CA==
X-Gm-Message-State: AOAM530ERJaDqaL+YrI+TCM0ldVedrTjSexRX+bEWrcEN/7mJZgoLekz 73qSOx/3Wh6cajFDUC1v9z4JBodQrSwJnyRIQFj20/ypXRbEKg==
X-Google-Smtp-Source: ABdhPJyfpZZcHalWjN5ZTZNTtwHTF0RC7jww+26Qb0vL/gQdLKUqS7qH1mSeH/P28kx0cj1OERmjTiS7Sf6Gr7R6wE4=
X-Received: by 2002:ac8:5693:: with SMTP id h19mr653335qta.84.1630698642457; Fri, 03 Sep 2021 12:50:42 -0700 (PDT)
MIME-Version: 1.0
References: <10c8c1d93688494588a9d7032a0aa37d@hackers4peace.net>
In-Reply-To: <10c8c1d93688494588a9d7032a0aa37d@hackers4peace.net>
From: Adrian Gropper <agropper@healthurl.com>
Date: Fri, 3 Sep 2021 15:50:31 -0400
Message-ID: <CANYRo8j1XMSE9XWuLb8y0rB1Bpd47hNoa314qTXKNHHTGx0t-w@mail.gmail.com>
To: elf-pavlik@hackers4peace.net
Cc: txauth@ietf.org
Content-Type: multipart/alternative; boundary="00000000000051ca3e05cb1c9de4"
Archived-At: <https://mailarchive.ietf.org/arch/msg/txauth/l6hNCndCwKEx_DrMs9ZZTpxOaMc>
Subject: Re: [GNAP] Distinct role used by End-user to give permissions to client instances.
X-BeenThere: txauth@ietf.org
X-Mailman-Version: 2.1.29
Precedence: list
List-Id: GNAP <txauth.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/options/txauth>, <mailto:txauth-request@ietf.org?subject=unsubscribe>
List-Archive: <https://mailarchive.ietf.org/arch/browse/txauth/>
List-Post: <mailto:txauth@ietf.org>
List-Help: <mailto:txauth-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/txauth>, <mailto:txauth-request@ietf.org?subject=subscribe>
X-List-Received-Date: Fri, 03 Sep 2021 19:50:49 -0000

--00000000000051ca3e05cb1c9de4
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable

An interesting use-case. One way to discuss the protocol issues is from a
transparency perspective.

If the RS / RO is presumed to be operating as a (regulated or informal)
fiduciary with no conflict of interest for its own profit, then both Alice
and Bob should have total transparency on what is actually shared with the
other. That transparency and potentially semi-autonomous behavior would be
managed by the respective EUXS agents.

If the RS / RO is truly acting as an RO and owes relatively little to Alice
and Bob, then the protocol would be easier because the EUXS would have less
information to work with. In the limit, Alice=E2=80=99s EUXS would not even=
 be
notified of Bob=E2=80=99s access to the shared RS.

Notification as part of an authorization protocol is, in my experience with
UMA 2, an option. Even when the RO is the subject and separate from the RS,
the RS may override the scope of an access token. This could be due to
security (e.g. DoS attack) or jurisdictional constraints (cross-border
clients). In these cases, transparency demands that the RS notify the RO or
the AS of the exception.

- Adrian

On Fri, Sep 3, 2021 at 9:57 AM <elf-pavlik@hackers4peace.net> wrote:

> Hello,
>
> I'm participating in AuthN and AuthZ related work of
> https://solidproject.org/
>
> I will intentionally keep this email short and provide further details
> based on responses.
> Given that I'll try to focus on one problem we are trying to address and
> we can start zooming out from there as needed.
>
> In our scenarios different End-user and Resource Owner is very common
> scenario.
> We are working on few proposals for handing User Authorization, in that
> case ACME organization (Resource Owner) could authorize Alice (End-user)
> to access specific data.
> In the same way ACME authorizes Bob (another End-user) to access some
> specific data. Data that Alice and Bob can access have certain overlap.
>
> Now both Alice and Bob should be able to independently permission /
> authorize various applications (Client instances) they wan to access
> that data owned by ACME.
> In fact both Alice and Bob have broader peer networks, including other
> Organizations and Individuals. We are also defining common permission
> model,
> which would allow expressing application access to span across data
> owned by any number of owners.
>
> In the scenario above, we consider assumption that given Resource Owner
> owns all the data on given Resource Server.
> In that case Authorization Server which issues Access Tokens stays
> associated with RS, so also with specific Resource Owner.
> Since each End-user needs some party which would take responsibility for
> them to give consent and permission apps (Client instances) they are
> using.
> We look at another role let's just call it EUXS (End-user's X Server),
> similar to OIDC Provider. It stays associated with End-user in similar
> way to rel=3D"http://openid.net/specs/connect/1.0/issuer" links to OP.
>
>  From here it would be really great to see how GNAP would help handing
> following:
>
> End-user's X Server, in a similar way to OIDC Provider, issues at token
> (let's call it a permissions token) to Client instance representing
> permissions applicable to a specific RS.
> Once again very often that RS is controlled by Resource Owner other than
> End-user.
> Client instance uses that 'permissions' token with AS to get Access
> Token which finally can be used when making requests to RS.
> Here I make assumption that AS understands that permissions token and
> later either makes it available to RS or translates it to something that
> RS can enforce.
> Oh, also RS would hint their AS with as_uri in WWW-Authenticate, since
> client doesn't have that prior knowledge.
>
> As I mentioned User Authorization is handled separately so AS & RS have
> prior knowledge of what access Resource Owner granted to End-user.
> In this email I mostly focus on End-user having another party associated
> with them (EUXS for the sake of this discussion) where they
> can further delegate that access their received to apps (Client
> instances) which they want to use to exercise access to data.
>
> I hope description above is sufficient to roughly illustrate the
> scenario. I can fill in more details as needed.
>
> I very much appreciate any hints on how GNAP can handle described
> scenario in elegant way.
>
> Best regards,
> elf Pavlik
>
> --
> TXAuth mailing list
> TXAuth@ietf.org
> https://www.ietf.org/mailman/listinfo/txauth
>

--00000000000051ca3e05cb1c9de4
Content-Type: text/html; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable

<div dir=3D"auto">An interesting use-case. One way to discuss the protocol =
issues is from a transparency perspective.</div><div dir=3D"auto"><br></div=
><div dir=3D"auto">If the RS / RO is presumed to be operating as a (regulat=
ed or informal) fiduciary with no conflict=C2=A0of interest for its own pro=
fit, then both Alice and Bob should have total transparency on what is actu=
ally shared with the other. That transparency and potentially semi-autonomo=
us behavior would be managed by the respective EUXS agents.=C2=A0</div><div=
 dir=3D"auto"><br></div><div dir=3D"auto">If the RS / RO is truly acting as=
 an RO and owes relatively little to Alice and Bob, then the protocol would=
 be easier because the EUXS would have less information to work with. In th=
e limit, Alice=E2=80=99s EUXS would not even be notified of Bob=E2=80=99s a=
ccess to the shared RS.=C2=A0</div><div dir=3D"auto"><br></div><div dir=3D"=
auto">Notification as part of an authorization protocol is, in my experienc=
e with UMA 2, an option. Even when the RO is the subject and separate from =
the RS, the RS may override the scope of an access token. This could be due=
 to security (e.g. DoS attack) or jurisdictional constraints (cross-border =
clients). In these cases, transparency demands that the RS notify the RO or=
 the AS of the exception.=C2=A0</div><div dir=3D"auto"><br></div><div dir=
=3D"auto">- Adrian</div><div><br><div class=3D"gmail_quote"><div dir=3D"ltr=
" class=3D"gmail_attr">On Fri, Sep 3, 2021 at 9:57 AM &lt;<a href=3D"mailto=
:elf-pavlik@hackers4peace.net">elf-pavlik@hackers4peace.net</a>&gt; wrote:<=
br></div><blockquote class=3D"gmail_quote" style=3D"margin:0px 0px 0px 0.8e=
x;border-left-width:1px;border-left-style:solid;padding-left:1ex;border-lef=
t-color:rgb(204,204,204)">Hello,<br>
<br>
I&#39;m participating in AuthN and AuthZ related work of <br>
<a href=3D"https://solidproject.org/" rel=3D"noreferrer" target=3D"_blank">=
https://solidproject.org/</a><br>
<br>
I will intentionally keep this email short and provide further details <br>
based on responses.<br>
Given that I&#39;ll try to focus on one problem we are trying to address an=
d <br>
we can start zooming out from there as needed.<br>
<br>
In our scenarios different End-user and Resource Owner is very common <br>
scenario.<br>
We are working on few proposals for handing User Authorization, in that <br=
>
case ACME organization (Resource Owner) could authorize Alice (End-user) <b=
r>
to access specific data.<br>
In the same way ACME authorizes Bob (another End-user) to access some <br>
specific data. Data that Alice and Bob can access have certain overlap.<br>
<br>
Now both Alice and Bob should be able to independently permission / <br>
authorize various applications (Client instances) they wan to access <br>
that data owned by ACME.<br>
In fact both Alice and Bob have broader peer networks, including other <br>
Organizations and Individuals. We are also defining common permission <br>
model,<br>
which would allow expressing application access to span across data <br>
owned by any number of owners.<br>
<br>
In the scenario above, we consider assumption that given Resource Owner <br=
>
owns all the data on given Resource Server.<br>
In that case Authorization Server which issues Access Tokens stays <br>
associated with RS, so also with specific Resource Owner.<br>
Since each End-user needs some party which would take responsibility for <b=
r>
them to give consent and permission apps (Client instances) they are <br>
using.<br>
We look at another role let&#39;s just call it EUXS (End-user&#39;s X Serve=
r), <br>
similar to OIDC Provider. It stays associated with End-user in similar <br>
way to rel=3D&quot;<a href=3D"http://openid.net/specs/connect/1.0/issuer" r=
el=3D"noreferrer" target=3D"_blank">http://openid.net/specs/connect/1.0/iss=
uer</a>&quot; links to OP.<br>
<br>
=C2=A0From here it would be really great to see how GNAP would help handing=
 <br>
following:<br>
<br>
End-user&#39;s X Server, in a similar way to OIDC Provider, issues at token=
 <br>
(let&#39;s call it a permissions token) to Client instance representing <br=
>
permissions applicable to a specific RS.<br>
Once again very often that RS is controlled by Resource Owner other than <b=
r>
End-user.<br>
Client instance uses that &#39;permissions&#39; token with AS to get Access=
 <br>
Token which finally can be used when making requests to RS.<br>
Here I make assumption that AS understands that permissions token and <br>
later either makes it available to RS or translates it to something that <b=
r>
RS can enforce.<br>
Oh, also RS would hint their AS with as_uri in WWW-Authenticate, since <br>
client doesn&#39;t have that prior knowledge.<br>
<br>
As I mentioned User Authorization is handled separately so AS &amp; RS have=
 <br>
prior knowledge of what access Resource Owner granted to End-user.<br>
In this email I mostly focus on End-user having another party associated <b=
r>
with them (EUXS for the sake of this discussion) where they<br>
can further delegate that access their received to apps (Client <br>
instances) which they want to use to exercise access to data.<br>
<br>
I hope description above is sufficient to roughly illustrate the <br>
scenario. I can fill in more details as needed.<br>
<br>
I very much appreciate any hints on how GNAP can handle described <br>
scenario in elegant way.<br>
<br>
Best regards,<br>
elf Pavlik<br>
<br>
-- <br>
TXAuth mailing list<br>
<a href=3D"mailto:TXAuth@ietf.org" target=3D"_blank">TXAuth@ietf.org</a><br=
>
<a href=3D"https://www.ietf.org/mailman/listinfo/txauth" rel=3D"noreferrer"=
 target=3D"_blank">https://www.ietf.org/mailman/listinfo/txauth</a><br>
</blockquote></div></div>

--00000000000051ca3e05cb1c9de4--


From nobody Sat Sep  4 00:35:25 2021
Return-Path: <fabien.imbault@gmail.com>
X-Original-To: txauth@ietfa.amsl.com
Delivered-To: txauth@ietfa.amsl.com
Received: from localhost (localhost [127.0.0.1]) by ietfa.amsl.com (Postfix) with ESMTP id 4DB023A1EB0 for <txauth@ietfa.amsl.com>; Sat,  4 Sep 2021 00:35:23 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -2.097
X-Spam-Level: 
X-Spam-Status: No, score=-2.097 tagged_above=-999 required=5 tests=[BAYES_00=-1.9, DKIM_SIGNED=0.1, DKIM_VALID=-0.1, DKIM_VALID_AU=-0.1, DKIM_VALID_EF=-0.1, FREEMAIL_FROM=0.001, HTML_MESSAGE=0.001, SPF_HELO_NONE=0.001, SPF_PASS=-0.001, URIBL_BLOCKED=0.001] autolearn=ham autolearn_force=no
Authentication-Results: ietfa.amsl.com (amavisd-new); dkim=pass (2048-bit key) header.d=gmail.com
Received: from mail.ietf.org ([4.31.198.44]) by localhost (ietfa.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id KbAMATQL0hmU for <txauth@ietfa.amsl.com>; Sat,  4 Sep 2021 00:35:18 -0700 (PDT)
Received: from mail-io1-xd34.google.com (mail-io1-xd34.google.com [IPv6:2607:f8b0:4864:20::d34]) (using TLSv1.2 with cipher ECDHE-RSA-AES128-GCM-SHA256 (128/128 bits)) (No client certificate requested) by ietfa.amsl.com (Postfix) with ESMTPS id 237D83A1EAB for <txauth@ietf.org>; Sat,  4 Sep 2021 00:35:18 -0700 (PDT)
Received: by mail-io1-xd34.google.com with SMTP id m11so1447379ioo.6 for <txauth@ietf.org>; Sat, 04 Sep 2021 00:35:18 -0700 (PDT)
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=20210112;  h=mime-version:references:in-reply-to:from:date:message-id:subject:to :cc; bh=pmS1go0WnpP439kgcncMA/71AtbysBGfCqgsvsA6p2w=; b=Rep6CSYYKab2PTzjQqYB4uKOtRBoUgY0bAVD5VTr4iI17zaC0Uu6+rWSxWHuVVPm1d 4kIJnmw0KlTaz5L22JcY1yNYlH0cr2dA0fq3qg0yQHd5tssPjkRNXSNELOQDUFwhfrSM u2Zin0USOfpnH/2rP2St/FLiZLvh/oFGyejucCdt7NJxtRNLWV1tya5zSWu8AP9IaHdz R+mHHZVnoDx1dEPY9vdbONVvks65wX78OfJkCttESr5j/+RFXnkHsmA5kzyqG82iWMoM HXx1cQUbZOkK1nhsFTkYqPgpyCaqFZw5H4eglTqWfotPyynZhcxENdWOkSYIGN4LG6hP GNTA==
X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20161025; h=x-gm-message-state:mime-version:references:in-reply-to:from:date :message-id:subject:to:cc; bh=pmS1go0WnpP439kgcncMA/71AtbysBGfCqgsvsA6p2w=; b=byYST4i43xQB+Vp/bCBRDB1dbBuTFLLxVpxDLVeR7bY1XMCOqRtu2o3kFGWKYTwOxv SkD3dXkzzGf/BJRJQ0QlHg3ybWUYdrdvnNTbtP+esQqwMc0tKk/7BfnL3PHYWF14DgFa 9o+WnqrDjj7zADZwOrM36h8s+RB64cVGr8cIjeRNRPQIwZ+4VI+gDGDg/ERLl3/sKBlc 38d3LHINUWfHPINX4EReaBw/CJF1JLJTbErk+jXyDk2d312CA7vRAwSV6wM+8jmozUmU MiS6eZEh2VIqnMiNc6Iai9ijbKVH1HthmOxjsBhjVhwKls0ERvYPkK0Ijbvkj5xtGCmF FyQg==
X-Gm-Message-State: AOAM533uaQ5aZdVbrTc/kwik0AUd4ycOmTBXFPtziimSBtsfA3IaPV3C UU6ilsmPI2C/gpo2p/pFptWL/1WWD14Vh9xkD1bw8poX97o=
X-Google-Smtp-Source: ABdhPJyRphGJh/XG/XkG8vIEoaYXNjhEDm+oPssJ0noCPy5Bx3hC/Z5tcK2y6R0brqk0f8NP+xCwGFoP0vC6pmEToxY=
X-Received: by 2002:a5d:9da4:: with SMTP id ay36mr2196031iob.153.1630740916513;  Sat, 04 Sep 2021 00:35:16 -0700 (PDT)
MIME-Version: 1.0
References: <10c8c1d93688494588a9d7032a0aa37d@hackers4peace.net> <CANYRo8j1XMSE9XWuLb8y0rB1Bpd47hNoa314qTXKNHHTGx0t-w@mail.gmail.com>
In-Reply-To: <CANYRo8j1XMSE9XWuLb8y0rB1Bpd47hNoa314qTXKNHHTGx0t-w@mail.gmail.com>
From: Fabien Imbault <fabien.imbault@gmail.com>
Date: Sat, 4 Sep 2021 09:35:05 +0200
Message-ID: <CAM8feuRLZ6fmmpAU9C+SWZkMbsQhK+OpgWOqmd7ASXFDFG=Zrw@mail.gmail.com>
To: Adrian Gropper <agropper@healthurl.com>
Cc: elf-pavlik@hackers4peace.net, GNAP Mailing List <txauth@ietf.org>
Content-Type: multipart/alternative; boundary="0000000000000cb28105cb267522"
Archived-At: <https://mailarchive.ietf.org/arch/msg/txauth/_aA1bJCTK9W97CpsqN2JssFzthM>
Subject: Re: [GNAP] Distinct role used by End-user to give permissions to client instances.
X-BeenThere: txauth@ietf.org
X-Mailman-Version: 2.1.29
Precedence: list
List-Id: GNAP <txauth.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/options/txauth>, <mailto:txauth-request@ietf.org?subject=unsubscribe>
List-Archive: <https://mailarchive.ietf.org/arch/browse/txauth/>
List-Post: <mailto:txauth@ietf.org>
List-Help: <mailto:txauth-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/txauth>, <mailto:txauth-request@ietf.org?subject=subscribe>
X-List-Received-Date: Sat, 04 Sep 2021 07:35:23 -0000

--0000000000000cb28105cb267522
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable

Hi Elf,

Indeed that's interesting. I'm not entirely sure I get all the ins and outs
of what you're trying to do, but I'll give a first answer based on my
limited understanding.

In GNAP section 4 (of the core spec) theorically allows the case when the
end-user and the RO are different. The details of how that would work in
still quite thin (and the rest of section 4 considers only the case when
they're the same).

Seems to me that your use case is also relatively close to having to manage
several ASs, with one that's associated to the RS (a case entirely made
possible in GNAP) and one handling permissions tokens (to take your
wording). Similar scenarii have been discussed in the past. Currently we
assume in GNAP core that there's only one AS involved, but it's entirely
possible that an extension would really get into the weeds of how more
complex cases would work. That's kind of the spirit behind the RS part of
the spec, and the entire discussion on capabilities. Please see
https://github.com/ietf-wg-gnap/gnap-resource-servers/issues/25 as a
pointer.

Now to answer your question in detail, I guess we'd need to map the flows
and see how that goes.

Cheers
Fabien



Le ven. 3 sept. 2021 =C3=A0 21:50, Adrian Gropper <agropper@healthurl.com> =
a
=C3=A9crit :

> An interesting use-case. One way to discuss the protocol issues is from a
> transparency perspective.
>
> If the RS / RO is presumed to be operating as a (regulated or informal)
> fiduciary with no conflict of interest for its own profit, then both Alic=
e
> and Bob should have total transparency on what is actually shared with th=
e
> other. That transparency and potentially semi-autonomous behavior would b=
e
> managed by the respective EUXS agents.
>
> If the RS / RO is truly acting as an RO and owes relatively little to
> Alice and Bob, then the protocol would be easier because the EUXS would
> have less information to work with. In the limit, Alice=E2=80=99s EUXS wo=
uld not
> even be notified of Bob=E2=80=99s access to the shared RS.
>
> Notification as part of an authorization protocol is, in my experience
> with UMA 2, an option. Even when the RO is the subject and separate from
> the RS, the RS may override the scope of an access token. This could be d=
ue
> to security (e.g. DoS attack) or jurisdictional constraints (cross-border
> clients). In these cases, transparency demands that the RS notify the RO =
or
> the AS of the exception.
>
> - Adrian
>
> On Fri, Sep 3, 2021 at 9:57 AM <elf-pavlik@hackers4peace.net> wrote:
>
>> Hello,
>>
>> I'm participating in AuthN and AuthZ related work of
>> https://solidproject.org/
>>
>> I will intentionally keep this email short and provide further details
>> based on responses.
>> Given that I'll try to focus on one problem we are trying to address and
>> we can start zooming out from there as needed.
>>
>> In our scenarios different End-user and Resource Owner is very common
>> scenario.
>> We are working on few proposals for handing User Authorization, in that
>> case ACME organization (Resource Owner) could authorize Alice (End-user)
>> to access specific data.
>> In the same way ACME authorizes Bob (another End-user) to access some
>> specific data. Data that Alice and Bob can access have certain overlap.
>>
>> Now both Alice and Bob should be able to independently permission /
>> authorize various applications (Client instances) they wan to access
>> that data owned by ACME.
>> In fact both Alice and Bob have broader peer networks, including other
>> Organizations and Individuals. We are also defining common permission
>> model,
>> which would allow expressing application access to span across data
>> owned by any number of owners.
>>
>> In the scenario above, we consider assumption that given Resource Owner
>> owns all the data on given Resource Server.
>> In that case Authorization Server which issues Access Tokens stays
>> associated with RS, so also with specific Resource Owner.
>> Since each End-user needs some party which would take responsibility for
>> them to give consent and permission apps (Client instances) they are
>> using.
>> We look at another role let's just call it EUXS (End-user's X Server),
>> similar to OIDC Provider. It stays associated with End-user in similar
>> way to rel=3D"http://openid.net/specs/connect/1.0/issuer" links to OP.
>>
>>  From here it would be really great to see how GNAP would help handing
>> following:
>>
>> End-user's X Server, in a similar way to OIDC Provider, issues at token
>> (let's call it a permissions token) to Client instance representing
>> permissions applicable to a specific RS.
>> Once again very often that RS is controlled by Resource Owner other than
>> End-user.
>> Client instance uses that 'permissions' token with AS to get Access
>> Token which finally can be used when making requests to RS.
>> Here I make assumption that AS understands that permissions token and
>> later either makes it available to RS or translates it to something that
>> RS can enforce.
>> Oh, also RS would hint their AS with as_uri in WWW-Authenticate, since
>> client doesn't have that prior knowledge.
>>
>> As I mentioned User Authorization is handled separately so AS & RS have
>> prior knowledge of what access Resource Owner granted to End-user.
>> In this email I mostly focus on End-user having another party associated
>> with them (EUXS for the sake of this discussion) where they
>> can further delegate that access their received to apps (Client
>> instances) which they want to use to exercise access to data.
>>
>> I hope description above is sufficient to roughly illustrate the
>> scenario. I can fill in more details as needed.
>>
>> I very much appreciate any hints on how GNAP can handle described
>> scenario in elegant way.
>>
>> Best regards,
>> elf Pavlik
>>
>> --
>> TXAuth mailing list
>> TXAuth@ietf.org
>> https://www.ietf.org/mailman/listinfo/txauth
>>
> --
> TXAuth mailing list
> TXAuth@ietf.org
> https://www.ietf.org/mailman/listinfo/txauth
>

--0000000000000cb28105cb267522
Content-Type: text/html; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable

<div dir=3D"auto"><div>Hi Elf,=C2=A0</div><div dir=3D"auto"><br></div><div =
dir=3D"auto">Indeed that&#39;s interesting. I&#39;m not entirely sure I get=
 all the ins and outs of what you&#39;re trying to do, but I&#39;ll give a =
first answer based on my limited understanding.=C2=A0</div><div dir=3D"auto=
"><br></div><div dir=3D"auto">In GNAP section 4 (of the core spec) theorica=
lly allows the case when the end-user and the RO are different. The details=
 of how that would work in still quite thin (and the rest of section 4 cons=
iders only the case when they&#39;re the same).=C2=A0</div><div dir=3D"auto=
"><br></div><div dir=3D"auto">Seems to me that your use case is also relati=
vely close to having to manage several ASs, with one that&#39;s associated =
to the RS (a case entirely made possible in GNAP) and one handling permissi=
ons tokens (to take your wording). Similar scenarii have been discussed in =
the past. Currently we assume in GNAP core that there&#39;s only one AS inv=
olved, but it&#39;s entirely possible that an extension would really get in=
to the weeds of how more complex cases would work. That&#39;s kind of the s=
pirit behind the RS part of the spec, and the entire discussion on capabili=
ties. Please see=C2=A0<a href=3D"https://github.com/ietf-wg-gnap/gnap-resou=
rce-servers/issues/25">https://github.com/ietf-wg-gnap/gnap-resource-server=
s/issues/25</a> as a pointer.=C2=A0</div><div dir=3D"auto"><br></div><div d=
ir=3D"auto">Now to answer your question in detail, I guess we&#39;d need to=
 map the flows and see how that goes.</div><div dir=3D"auto"><br></div><div=
 dir=3D"auto">Cheers</div><div dir=3D"auto">Fabien</div><div dir=3D"auto"><=
br></div><div dir=3D"auto"><br><br><div class=3D"gmail_quote" dir=3D"auto">=
<div dir=3D"ltr" class=3D"gmail_attr">Le ven. 3 sept. 2021 =C3=A0 21:50, Ad=
rian Gropper &lt;<a href=3D"mailto:agropper@healthurl.com" target=3D"_blank=
" rel=3D"noreferrer">agropper@healthurl.com</a>&gt; a =C3=A9crit=C2=A0:<br>=
</div><blockquote class=3D"gmail_quote" style=3D"margin:0 0 0 .8ex;border-l=
eft:1px #ccc solid;padding-left:1ex"><div dir=3D"auto">An interesting use-c=
ase. One way to discuss the protocol issues is from a transparency perspect=
ive.</div><div dir=3D"auto"><br></div><div dir=3D"auto">If the RS / RO is p=
resumed to be operating as a (regulated or informal) fiduciary with no conf=
lict=C2=A0of interest for its own profit, then both Alice and Bob should ha=
ve total transparency on what is actually shared with the other. That trans=
parency and potentially semi-autonomous behavior would be managed by the re=
spective EUXS agents.=C2=A0</div><div dir=3D"auto"><br></div><div dir=3D"au=
to">If the RS / RO is truly acting as an RO and owes relatively little to A=
lice and Bob, then the protocol would be easier because the EUXS would have=
 less information to work with. In the limit, Alice=E2=80=99s EUXS would no=
t even be notified of Bob=E2=80=99s access to the shared RS.=C2=A0</div><di=
v dir=3D"auto"><br></div><div dir=3D"auto">Notification as part of an autho=
rization protocol is, in my experience with UMA 2, an option. Even when the=
 RO is the subject and separate from the RS, the RS may override the scope =
of an access token. This could be due to security (e.g. DoS attack) or juri=
sdictional constraints (cross-border clients). In these cases, transparency=
 demands that the RS notify the RO or the AS of the exception.=C2=A0</div><=
div dir=3D"auto"><br></div><div dir=3D"auto">- Adrian</div><div><br><div cl=
ass=3D"gmail_quote"><div dir=3D"ltr" class=3D"gmail_attr">On Fri, Sep 3, 20=
21 at 9:57 AM &lt;<a href=3D"mailto:elf-pavlik@hackers4peace.net" rel=3D"no=
referrer noreferrer" target=3D"_blank">elf-pavlik@hackers4peace.net</a>&gt;=
 wrote:<br></div><blockquote class=3D"gmail_quote" style=3D"margin:0px 0px =
0px 0.8ex;border-left-width:1px;border-left-style:solid;padding-left:1ex;bo=
rder-left-color:rgb(204,204,204)">Hello,<br>
<br>
I&#39;m participating in AuthN and AuthZ related work of <br>
<a href=3D"https://solidproject.org/" rel=3D"noreferrer noreferrer noreferr=
er" target=3D"_blank">https://solidproject.org/</a><br>
<br>
I will intentionally keep this email short and provide further details <br>
based on responses.<br>
Given that I&#39;ll try to focus on one problem we are trying to address an=
d <br>
we can start zooming out from there as needed.<br>
<br>
In our scenarios different End-user and Resource Owner is very common <br>
scenario.<br>
We are working on few proposals for handing User Authorization, in that <br=
>
case ACME organization (Resource Owner) could authorize Alice (End-user) <b=
r>
to access specific data.<br>
In the same way ACME authorizes Bob (another End-user) to access some <br>
specific data. Data that Alice and Bob can access have certain overlap.<br>
<br>
Now both Alice and Bob should be able to independently permission / <br>
authorize various applications (Client instances) they wan to access <br>
that data owned by ACME.<br>
In fact both Alice and Bob have broader peer networks, including other <br>
Organizations and Individuals. We are also defining common permission <br>
model,<br>
which would allow expressing application access to span across data <br>
owned by any number of owners.<br>
<br>
In the scenario above, we consider assumption that given Resource Owner <br=
>
owns all the data on given Resource Server.<br>
In that case Authorization Server which issues Access Tokens stays <br>
associated with RS, so also with specific Resource Owner.<br>
Since each End-user needs some party which would take responsibility for <b=
r>
them to give consent and permission apps (Client instances) they are <br>
using.<br>
We look at another role let&#39;s just call it EUXS (End-user&#39;s X Serve=
r), <br>
similar to OIDC Provider. It stays associated with End-user in similar <br>
way to rel=3D&quot;<a href=3D"http://openid.net/specs/connect/1.0/issuer" r=
el=3D"noreferrer noreferrer noreferrer" target=3D"_blank">http://openid.net=
/specs/connect/1.0/issuer</a>&quot; links to OP.<br>
<br>
=C2=A0From here it would be really great to see how GNAP would help handing=
 <br>
following:<br>
<br>
End-user&#39;s X Server, in a similar way to OIDC Provider, issues at token=
 <br>
(let&#39;s call it a permissions token) to Client instance representing <br=
>
permissions applicable to a specific RS.<br>
Once again very often that RS is controlled by Resource Owner other than <b=
r>
End-user.<br>
Client instance uses that &#39;permissions&#39; token with AS to get Access=
 <br>
Token which finally can be used when making requests to RS.<br>
Here I make assumption that AS understands that permissions token and <br>
later either makes it available to RS or translates it to something that <b=
r>
RS can enforce.<br>
Oh, also RS would hint their AS with as_uri in WWW-Authenticate, since <br>
client doesn&#39;t have that prior knowledge.<br>
<br>
As I mentioned User Authorization is handled separately so AS &amp; RS have=
 <br>
prior knowledge of what access Resource Owner granted to End-user.<br>
In this email I mostly focus on End-user having another party associated <b=
r>
with them (EUXS for the sake of this discussion) where they<br>
can further delegate that access their received to apps (Client <br>
instances) which they want to use to exercise access to data.<br>
<br>
I hope description above is sufficient to roughly illustrate the <br>
scenario. I can fill in more details as needed.<br>
<br>
I very much appreciate any hints on how GNAP can handle described <br>
scenario in elegant way.<br>
<br>
Best regards,<br>
elf Pavlik<br>
<br>
-- <br>
TXAuth mailing list<br>
<a href=3D"mailto:TXAuth@ietf.org" rel=3D"noreferrer noreferrer" target=3D"=
_blank">TXAuth@ietf.org</a><br>
<a href=3D"https://www.ietf.org/mailman/listinfo/txauth" rel=3D"noreferrer =
noreferrer noreferrer" target=3D"_blank">https://www.ietf.org/mailman/listi=
nfo/txauth</a><br>
</blockquote></div></div>
-- <br>
TXAuth mailing list<br>
<a href=3D"mailto:TXAuth@ietf.org" rel=3D"noreferrer noreferrer" target=3D"=
_blank">TXAuth@ietf.org</a><br>
<a href=3D"https://www.ietf.org/mailman/listinfo/txauth" rel=3D"noreferrer =
noreferrer noreferrer" target=3D"_blank">https://www.ietf.org/mailman/listi=
nfo/txauth</a><br>
</blockquote></div></div></div>

--0000000000000cb28105cb267522--


From nobody Sat Sep  4 07:43:31 2021
Return-Path: <elf-pavlik@hackers4peace.net>
X-Original-To: txauth@ietfa.amsl.com
Delivered-To: txauth@ietfa.amsl.com
Received: from localhost (localhost [127.0.0.1]) by ietfa.amsl.com (Postfix) with ESMTP id 5D6693A18F3 for <txauth@ietfa.amsl.com>; Sat,  4 Sep 2021 07:43:29 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -1.7
X-Spam-Level: 
X-Spam-Status: No, score=-1.7 tagged_above=-999 required=5 tests=[BAYES_00=-1.9, DKIM_INVALID=0.1, DKIM_SIGNED=0.1, SPF_PASS=-0.001, URIBL_BLOCKED=0.001] autolearn=no autolearn_force=no
Authentication-Results: ietfa.amsl.com (amavisd-new); dkim=neutral reason="invalid (public key: not available)" header.d=hackers4peace.net
Received: from mail.ietf.org ([4.31.198.44]) by localhost (ietfa.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id 7ju11CmfV2HI for <txauth@ietfa.amsl.com>; Sat,  4 Sep 2021 07:43:24 -0700 (PDT)
Received: from email.ecobytes.net (email.ecobytes.net [IPv6:2a01:4f8:10a:28c1::32]) (using TLSv1.2 with cipher ADH-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by ietfa.amsl.com (Postfix) with ESMTPS id 031753A18F1 for <txauth@ietf.org>; Sat,  4 Sep 2021 07:43:22 -0700 (PDT)
Received: from localhost (localhost [127.0.0.1]) by email.ecobytes.net (Postfix) with ESMTP id 35EED6B5D8 for <txauth@ietf.org>; Sat,  4 Sep 2021 14:43:14 +0000 (UTC)
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=hackers4peace.net; s=modoboa; t=1630766594; bh=2LlNPS+abv5Cat4deMHCRPyngZNSg7oGM20GXKO60Zc=; h=Date:From:To:Subject:In-Reply-To:References:From; b=VAPwAub297KdlkjHOk7mSW3gRfPF4YsKss4qvewBvbb6QZ5vR7vS2UsqXcNiiExqY WdFzRNPszAaSwKbxWf2ywaTpbbg6PSAKidmIU/iQfqpkkDgA/E/EZEcD/MfrjRK8HA QL48hGH/3qabaecYPrvreo78nBXs0rXxvLd9HmF15umLvxERvc52cbGDx8oUk2KiLe 3mqY3LptEmXjsT7HZHssfgupeWIXsqq32a02arukH68sS+6xd0BJSD1Bku73RBvhZR wrZx5w5AZGv2OgWPd4gkAhPwMjknPuOD4c4eOys1gBUw+yLNUhXMhlFFe7jkCpYoCS 3qorxVXq85Q0A==
X-Virus-Scanned: Debian amavisd-new at email.ecobytes.net
Received: from email.ecobytes.net ([127.0.0.1]) by localhost (email.ecobytes.net [127.0.0.1]) (amavisd-new, port 10026) with ESMTP id GF6JwMCltGJ7 for <txauth@ietf.org>; Sat,  4 Sep 2021 14:43:12 +0000 (UTC)
Received: from webmail.ecobytes.net (geogale.ecobytes.net [88.99.144.132]) (using TLSv1.3 with cipher TLS_AES_256_GCM_SHA384 (256/256 bits) key-exchange ECDHE (P-256) server-signature RSA-PSS (2048 bits) server-digest SHA256) (No client certificate requested) by email.ecobytes.net (Postfix) with ESMTPSA for <txauth@ietf.org>; Sat,  4 Sep 2021 14:43:12 +0000 (UTC)
MIME-Version: 1.0
Date: Sat, 04 Sep 2021 09:43:12 -0500
From: elf-pavlik@hackers4peace.net
To: txauth@ietf.org
In-Reply-To: <CAM8feuRLZ6fmmpAU9C+SWZkMbsQhK+OpgWOqmd7ASXFDFG=Zrw@mail.gmail.com>
References: <10c8c1d93688494588a9d7032a0aa37d@hackers4peace.net> <CANYRo8j1XMSE9XWuLb8y0rB1Bpd47hNoa314qTXKNHHTGx0t-w@mail.gmail.com> <CAM8feuRLZ6fmmpAU9C+SWZkMbsQhK+OpgWOqmd7ASXFDFG=Zrw@mail.gmail.com>
User-Agent: Roundcube Webmail/1.4.11
Message-ID: <996bafb5b7b9cc95639ddd8ad3f16728@hackers4peace.net>
X-Sender: elf-pavlik@hackers4peace.net
Content-Type: text/plain; charset=UTF-8; format=flowed
Content-Transfer-Encoding: 8bit
Archived-At: <https://mailarchive.ietf.org/arch/msg/txauth/xwUIf1JVSgSt3FGC8Qof5qq2SqM>
Subject: Re: [GNAP] Distinct role used by End-user to give permissions to client instances.
X-BeenThere: txauth@ietf.org
X-Mailman-Version: 2.1.29
Precedence: list
List-Id: GNAP <txauth.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/options/txauth>, <mailto:txauth-request@ietf.org?subject=unsubscribe>
List-Archive: <https://mailarchive.ietf.org/arch/browse/txauth/>
List-Post: <mailto:txauth@ietf.org>
List-Help: <mailto:txauth-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/txauth>, <mailto:txauth-request@ietf.org?subject=subscribe>
X-List-Received-Date: Sat, 04 Sep 2021 14:43:30 -0000

Hi Adrian, Fabien,

Thank you for your responsens!

> If the RS / RO is truly acting as an RO and owes relatively little
> to Alice and Bob, then the protocol would be easier because the EUXS
> would have less information to work with. In the limit, Alice’s
> EUXS would not even be notified of Bob’s access to the shared RS.

I think I'm discussing this case. Here both End-users Alice and Bob
each one has EUXS they choose to use. Similar to how each Alice and Bob
can choose their OIDC Provider.
When ACME (RO) sets access policy, they grant access to End-users like 
Alice (EU) and Bob (EU)
and let them permission clients independently using their EUXSs.
We work on notification system so that Alice's EUXS would get notified
when ACME shared access with Alice. When ACME shares access with Bob
Alice's EUXS has NO knowledge of that happening, only Bob's EUXS
will get notified.

>> Notification as part of an authorization protocol is, in my
>> experience with UMA 2, an option.

Yes, as I mentioned when RO gives access to some EU, that EUXS of that 
EU gets
notified. To clarify, similar as OIDC Provider, EUXS can be discovered 
for any EU.


> In GNAP section 4 (of the core spec) theorically allows the case when
> the end-user and the RO are different. The details of how that would
> work in still quite thin (and the rest of section 4 considers only the
> case when they're the same).

In the scenarios we work with, different EU and RO is a very common 
case.
Here I'm trying to focus on most common scenario where RO sets access 
policies
only based on identities of EUs (In our case IRIs are used to identify 
EUs).
Later EUs, independently from RO, can permission clients they are using.
In the end RS and AS associated with it, the one that RS hints in
WWW-Authenticate using as_uri, need to take into account policy set by 
RO
which determines access for given EU, as well as permissions given by EU 
to
the Client. Mentioned


> Seems to me that your use case is also relatively close to having to
> manage several ASs, with one that's associated to the RS (a case
> entirely made possible in GNAP) and one handling permissions tokens
> (to take your wording). Similar scenarii have been discussed in the
> past. Currently we assume in GNAP core that there's only one AS
> involved, but it's entirely possible that an extension would really
> get into the weeds of how more complex cases would work. That's kind
> of the spirit behind the RS part of the spec, and the entire
> discussion on capabilities. Please see
> https://github.com/ietf-wg-gnap/gnap-resource-servers/issues/25 as a
> pointer.

I think we would always have specific AS associated with RS. Again RS 
advertises it
via WWW-Authenticate as_uri . It issues Access Token which client is 
using with that RS.

I probably haven't mentioned that in our scenario End-user is using 
client to seamlessly
access resources in any number of Resource Servers, where each RS can be 
controlled by different RO.
Client would keep track of which Access Token should be used with which 
RS.
For that reason we also don't expect client to redirect EU to any of RS 
associated ASs.
EU would only give consent to the client at their EUXS.

Suggested EUXS would be something closer to OIDC Provider. It stays 
associated
with EU, similar to rel="http://openid.net/specs/connect/1.0/issuer".
EUXS keeps track of all the data that any number of ROs shared with the 
EU. Based on that it presents
consent screen to EU where they can decide which of that data they want 
to access using given client.
As I mentioned, EU shouldn't interact with any of RS associated ASs.
Client should be able to get 'permissions token' from EUXS for each 
distinct RS
(each RS can be in control of different RO). Later Client would exchange 
that 'permissions token'
with AS associated with RS that this token was created for. This should 
happen without End-user
interaction.

Probably it can be useful to discuss it in terms of delegation chains.

1. ACME (RO) delegates some access to Alice (EU)
2. Alice (EU) further delegates some of the access, which she received 
from ACME (RO), to a Projectron (Client)

More advanced case, which I intentionally tried not to start with would 
allow longer chains

1. ACME (RO) delegates some access to Omni (EU-ish) - Omni another 
organization
2. Omni (EU-ish) further delegates some access their received from ACME 
(RO) to Alice (EU)
3. Alice (EU) further delegates some access, which she received from 
Omni (EU-ish),
    which they received from ACME (RO), to a Projectron (Client)

Having just one RO associated with any RS probably makes both scenarios 
simpler.

Again I think only RS associated AS should issue Access Token that it's 
used with RS.
EUXS would issue 'permission tokens', similar to OIDC Provider issuing 
ID Tokens.
I imagine those 'permission tokens' later being exchanged with RS 
associated AS
without End-user interaction.

BTW I think this comment in mentioned issue provides a useful hint.

> If an AS receives and processes an access token, it is acting as an RS.
> There are two transactions at the same time and one entity plays 
> different roles in each.
> If an RS requests a token in response to an incoming request, that RS 
> is now a client instance.



> Now to answer your question in detail, I guess we'd need to map the
> flows and see how that goes.

I have some sequence diagrams which I can adjust to illustrate discussed 
scenario better.
I will also try to clarify there which role any given party plays in 
each interaction.
I'll write a followup message once I have them ready.

Thank you for all your feedback!
elf Pavlik


From nobody Fri Sep 10 13:47:03 2021
Return-Path: <jricher@mit.edu>
X-Original-To: txauth@ietfa.amsl.com
Delivered-To: txauth@ietfa.amsl.com
Received: from localhost (localhost [127.0.0.1]) by ietfa.amsl.com (Postfix) with ESMTP id CADCD3A1A99 for <txauth@ietfa.amsl.com>; Fri, 10 Sep 2021 13:47:01 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -1.498
X-Spam-Level: 
X-Spam-Status: No, score=-1.498 tagged_above=-999 required=5 tests=[BAYES_00=-1.9, KHOP_HELO_FCRDNS=0.399, SPF_HELO_NONE=0.001, SPF_NONE=0.001, URIBL_BLOCKED=0.001] autolearn=no autolearn_force=no
Received: from mail.ietf.org ([4.31.198.44]) by localhost (ietfa.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id fMGZR_SSY9kt for <txauth@ietfa.amsl.com>; Fri, 10 Sep 2021 13:46:57 -0700 (PDT)
Received: from outgoing.mit.edu (outgoing-auth-1.mit.edu [18.9.28.11]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by ietfa.amsl.com (Postfix) with ESMTPS id 02E063A1A98 for <txauth@ietf.org>; Fri, 10 Sep 2021 13:46:56 -0700 (PDT)
Received: from [192.168.1.28] (static-71-174-62-56.bstnma.fios.verizon.net [71.174.62.56]) (authenticated bits=0) (User authenticated as jricher@ATHENA.MIT.EDU) by outgoing.mit.edu (8.14.7/8.12.4) with ESMTP id 18AKkj6R005111 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-GCM-SHA384 bits=256 verify=NOT); Fri, 10 Sep 2021 16:46:46 -0400
Content-Type: text/plain; charset=utf-8
Mime-Version: 1.0 (Mac OS X Mail 13.4 \(3608.120.23.2.7\))
From: Justin Richer <jricher@mit.edu>
In-Reply-To: <996bafb5b7b9cc95639ddd8ad3f16728@hackers4peace.net>
Date: Fri, 10 Sep 2021 16:46:45 -0400
Cc: txauth@ietf.org
Content-Transfer-Encoding: quoted-printable
Message-Id: <77DFE057-4DD2-4786-9090-3F47AE70881E@mit.edu>
References: <10c8c1d93688494588a9d7032a0aa37d@hackers4peace.net> <CANYRo8j1XMSE9XWuLb8y0rB1Bpd47hNoa314qTXKNHHTGx0t-w@mail.gmail.com> <CAM8feuRLZ6fmmpAU9C+SWZkMbsQhK+OpgWOqmd7ASXFDFG=Zrw@mail.gmail.com> <996bafb5b7b9cc95639ddd8ad3f16728@hackers4peace.net>
To: elf-pavlik@hackers4peace.net
X-Mailer: Apple Mail (2.3608.120.23.2.7)
Archived-At: <https://mailarchive.ietf.org/arch/msg/txauth/U8MjWc5d3T-4z6-zvByQl6WKQ6w>
Subject: Re: [GNAP] Distinct role used by End-user to give permissions to client instances.
X-BeenThere: txauth@ietf.org
X-Mailman-Version: 2.1.29
Precedence: list
List-Id: GNAP <txauth.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/options/txauth>, <mailto:txauth-request@ietf.org?subject=unsubscribe>
List-Archive: <https://mailarchive.ietf.org/arch/browse/txauth/>
List-Post: <mailto:txauth@ietf.org>
List-Help: <mailto:txauth-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/txauth>, <mailto:txauth-request@ietf.org?subject=subscribe>
X-List-Received-Date: Fri, 10 Sep 2021 20:47:02 -0000

Hi elf, thanks so much for the detailed writeup. I=E2=80=99d like to add =
some of my thoughts to this discussion, and hopefully I=E2=80=99m =
following your use case correctly. Please correct me if I=E2=80=99m off =
at all!

I believe the scenario you=E2=80=99re describing is in line with the =
more-general descriptions around =E2=80=9Cuser interaction and =
consent=E2=80=9D that are described in section 4 of GNAP=E2=80=99s =
current draft. While I agree that some sequence diagrams would be =
helpful, this is what I=E2=80=99m picturing in your scenario, mapped to =
GNAP=E2=80=99s messages and terms:


Phase 1: Client Instance makes a request to EUXS for an assertion =
representing the current user. The nature and format of this assertion =
are out of scope for GNAP itself, but would be in the purview of =
Solid=E2=80=99s profile of the technologies =E2=80=94 for now we=E2=80=99l=
l call this a Solid Assertion Token (SAT) within this description. A JWT =
would be a reasonable carrier for such an assertion. What GNAP offers is =
a standard way for the client instance to request this assertion and =
have it delivered directly. We=E2=80=99re assuming that the end user =
interacts with or otherwise associates with the EUXS so that the EUXS =
knows who the end user is. Or possibly, the EUXS knows that the client =
in question has already been authorized and can issue the SAT directly. =
GNAP allows for both scenarios to come out of the same request message, =
allowing for a clean fallback in different situations.

Phase 2: Client instance talks to the AS that controls access to the RS. =
The client instance presents the SAT as part of its =E2=80=9Cuser=E2=80=9D=
 information field in the request. The AS validates the SAT based on =
Solid=E2=80=99s rules. The AS contacts the RO to get permission for the =
client instance on the end user=E2=80=99s behalf. How that happens is up =
to the AS, but the SAT would likely have some kind of RO identifier =
within it: namely representing what=E2=80=99s being requested, not who =
the SAT was issued to. Or it could be based on the RS and permissions =
the client instance is asking for =E2=80=94 so the client instance and =
end user wouldn=E2=80=99t necessarily know who the RO was at all, just =
that they want access to the RS and it=E2=80=99s controlled by =E2=80=A6 =
someone. Once the RO approves it, the AS issues the access token to the =
client instance.

Phase 3: Client instance calls the RS with the access token from AS. It =
may or may not contain any information from the SAT, since the only =
thing the RS ultimately needs to know is if the token is good for the =
request being made. Solid can further profile the access token itself =
(either its format or its introspected data) to convey specific =
information. GNAP remains agnostic on this decision, but provisions for =
standardizing these elements are in the GNAP-RS document.


An alternative process inverts some of the phases:


Phase 1 (alt): Client instance talks to the AS to get access and says it =
can interact with the end user through the EUXS and get an assertion =
(SAT) as a result. The AS passes whatever information it needs to =
complete this back to the client instance. The end user interacts at =
EUXS to get the SAT generated and handed back to the client instance.

Phase 2 (alt): Client instance presents the SAT back to the AS, which =
then goes on to talk to the RO, wherever they are.

Phase 3 continues as above.


Do either of these processes align with what you had in mind, or am I =
missing some important subtleties to this use case? We=E2=80=99ve =
discussed both of these patterns here in the WG before and =
accommodations for them are built into the GNAP core protocol: both what =
it defines and what it allows but keeps out of scope of the core =
protocol definitions.=20

 =E2=80=94 Justin

> On Sep 4, 2021, at 10:43 AM, elf-pavlik@hackers4peace.net wrote:
>=20
> Hi Adrian, Fabien,
>=20
> Thank you for your responsens!
>=20
>> If the RS / RO is truly acting as an RO and owes relatively little
>> to Alice and Bob, then the protocol would be easier because the EUXS
>> would have less information to work with. In the limit, Alice=E2=80=99s=

>> EUXS would not even be notified of Bob=E2=80=99s access to the shared =
RS.
>=20
> I think I'm discussing this case. Here both End-users Alice and Bob
> each one has EUXS they choose to use. Similar to how each Alice and =
Bob
> can choose their OIDC Provider.
> When ACME (RO) sets access policy, they grant access to End-users like =
Alice (EU) and Bob (EU)
> and let them permission clients independently using their EUXSs.
> We work on notification system so that Alice's EUXS would get notified
> when ACME shared access with Alice. When ACME shares access with Bob
> Alice's EUXS has NO knowledge of that happening, only Bob's EUXS
> will get notified.
>=20
>>> Notification as part of an authorization protocol is, in my
>>> experience with UMA 2, an option.
>=20
> Yes, as I mentioned when RO gives access to some EU, that EUXS of that =
EU gets
> notified. To clarify, similar as OIDC Provider, EUXS can be discovered =
for any EU.
>=20
>=20
>> In GNAP section 4 (of the core spec) theorically allows the case when
>> the end-user and the RO are different. The details of how that would
>> work in still quite thin (and the rest of section 4 considers only =
the
>> case when they're the same).
>=20
> In the scenarios we work with, different EU and RO is a very common =
case.
> Here I'm trying to focus on most common scenario where RO sets access =
policies
> only based on identities of EUs (In our case IRIs are used to identify =
EUs).
> Later EUs, independently from RO, can permission clients they are =
using.
> In the end RS and AS associated with it, the one that RS hints in
> WWW-Authenticate using as_uri, need to take into account policy set by =
RO
> which determines access for given EU, as well as permissions given by =
EU to
> the Client. Mentioned
>=20
>=20
>> Seems to me that your use case is also relatively close to having to
>> manage several ASs, with one that's associated to the RS (a case
>> entirely made possible in GNAP) and one handling permissions tokens
>> (to take your wording). Similar scenarii have been discussed in the
>> past. Currently we assume in GNAP core that there's only one AS
>> involved, but it's entirely possible that an extension would really
>> get into the weeds of how more complex cases would work. That's kind
>> of the spirit behind the RS part of the spec, and the entire
>> discussion on capabilities. Please see
>> https://github.com/ietf-wg-gnap/gnap-resource-servers/issues/25 as a
>> pointer.
>=20
> I think we would always have specific AS associated with RS. Again RS =
advertises it
> via WWW-Authenticate as_uri . It issues Access Token which client is =
using with that RS.
>=20
> I probably haven't mentioned that in our scenario End-user is using =
client to seamlessly
> access resources in any number of Resource Servers, where each RS can =
be controlled by different RO.
> Client would keep track of which Access Token should be used with =
which RS.
> For that reason we also don't expect client to redirect EU to any of =
RS associated ASs.
> EU would only give consent to the client at their EUXS.
>=20
> Suggested EUXS would be something closer to OIDC Provider. It stays =
associated
> with EU, similar to rel=3D"http://openid.net/specs/connect/1.0/issuer".
> EUXS keeps track of all the data that any number of ROs shared with =
the EU. Based on that it presents
> consent screen to EU where they can decide which of that data they =
want to access using given client.
> As I mentioned, EU shouldn't interact with any of RS associated ASs.
> Client should be able to get 'permissions token' from EUXS for each =
distinct RS
> (each RS can be in control of different RO). Later Client would =
exchange that 'permissions token'
> with AS associated with RS that this token was created for. This =
should happen without End-user
> interaction.
>=20
> Probably it can be useful to discuss it in terms of delegation chains.
>=20
> 1. ACME (RO) delegates some access to Alice (EU)
> 2. Alice (EU) further delegates some of the access, which she received =
from ACME (RO), to a Projectron (Client)
>=20
> More advanced case, which I intentionally tried not to start with =
would allow longer chains
>=20
> 1. ACME (RO) delegates some access to Omni (EU-ish) - Omni another =
organization
> 2. Omni (EU-ish) further delegates some access their received from =
ACME (RO) to Alice (EU)
> 3. Alice (EU) further delegates some access, which she received from =
Omni (EU-ish),
>   which they received from ACME (RO), to a Projectron (Client)
>=20
> Having just one RO associated with any RS probably makes both =
scenarios simpler.
>=20
> Again I think only RS associated AS should issue Access Token that =
it's used with RS.
> EUXS would issue 'permission tokens', similar to OIDC Provider issuing =
ID Tokens.
> I imagine those 'permission tokens' later being exchanged with RS =
associated AS
> without End-user interaction.
>=20
> BTW I think this comment in mentioned issue provides a useful hint.
>=20
>> If an AS receives and processes an access token, it is acting as an =
RS.
>> There are two transactions at the same time and one entity plays =
different roles in each.
>> If an RS requests a token in response to an incoming request, that RS =
is now a client instance.
>=20
>=20
>=20
>> Now to answer your question in detail, I guess we'd need to map the
>> flows and see how that goes.
>=20
> I have some sequence diagrams which I can adjust to illustrate =
discussed scenario better.
> I will also try to clarify there which role any given party plays in =
each interaction.
> I'll write a followup message once I have them ready.
>=20
> Thank you for all your feedback!
> elf Pavlik
>=20
> --=20
> TXAuth mailing list
> TXAuth@ietf.org
> https://www.ietf.org/mailman/listinfo/txauth


From nobody Wed Sep 15 20:04:36 2021
Return-Path: <elf-pavlik@hackers4peace.net>
X-Original-To: txauth@ietfa.amsl.com
Delivered-To: txauth@ietfa.amsl.com
Received: from localhost (localhost [127.0.0.1]) by ietfa.amsl.com (Postfix) with ESMTP id 2E8773A0BEF for <txauth@ietfa.amsl.com>; Wed, 15 Sep 2021 20:04:35 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -1.7
X-Spam-Level: 
X-Spam-Status: No, score=-1.7 tagged_above=-999 required=5 tests=[BAYES_00=-1.9, DKIM_INVALID=0.1, DKIM_SIGNED=0.1, SPF_PASS=-0.001, URIBL_BLOCKED=0.001] autolearn=no autolearn_force=no
Authentication-Results: ietfa.amsl.com (amavisd-new); dkim=neutral reason="invalid (public key: not available)" header.d=hackers4peace.net
Received: from mail.ietf.org ([4.31.198.44]) by localhost (ietfa.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id OlhFnwPhAgcZ for <txauth@ietfa.amsl.com>; Wed, 15 Sep 2021 20:04:30 -0700 (PDT)
Received: from email.ecobytes.net (email.ecobytes.net [IPv6:2a01:4f8:10a:28c1::32]) (using TLSv1.2 with cipher ADH-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by ietfa.amsl.com (Postfix) with ESMTPS id 5057D3A0BE9 for <txauth@ietf.org>; Wed, 15 Sep 2021 20:04:28 -0700 (PDT)
Received: from localhost (localhost [127.0.0.1]) by email.ecobytes.net (Postfix) with ESMTP id 4FADC8E474 for <txauth@ietf.org>; Thu, 16 Sep 2021 03:04:22 +0000 (UTC)
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=hackers4peace.net; s=modoboa; t=1631761462; bh=PEq33pOQ1wDEE/+CZMvAhT14yoZdKkZ1syYxp8vGuBU=; h=Date:From:To:Subject:In-Reply-To:References:From; b=EPvc3I4kJedz26Rad1QOrpQBrmHfZhqOHoqQxOIgfM7ClwkrQXgqFXusXwyxkoBLl SF0gd2EqGAI9spV6z7Oa72KGuQzQc1N2EfoQ8yph7w4w44fRykGE/s7yNxN+5H2gUl 7ll0xE89yZRS9ZTITJjN+bwAbIJM6debaV+x9gYIvmnBFDrbs8PJ7E/bHxsYMPsrDf RO6dkI+JKoL4GKJ70Vp/AUVRpmp8fw8PpPnj6c0c5bA7xo3EddMMUESHNUjOkflQzp mUIiNzZvDdkkkeXXdssh0zkJm2VLAcZIygSAEDgYcsBfYxHHpt0PuDb1HSI9I0Ya9x sMjKcgIiERFhA==
X-Virus-Scanned: Debian amavisd-new at email.ecobytes.net
Received: from email.ecobytes.net ([127.0.0.1]) by localhost (email.ecobytes.net [127.0.0.1]) (amavisd-new, port 10026) with ESMTP id OYXPAuZ353hN for <txauth@ietf.org>; Thu, 16 Sep 2021 03:04:20 +0000 (UTC)
Received: from webmail.ecobytes.net (geogale.ecobytes.net [88.99.144.132]) (using TLSv1.3 with cipher TLS_AES_256_GCM_SHA384 (256/256 bits) key-exchange ECDHE (P-256) server-signature RSA-PSS (2048 bits) server-digest SHA256) (No client certificate requested) by email.ecobytes.net (Postfix) with ESMTPSA for <txauth@ietf.org>; Thu, 16 Sep 2021 03:04:20 +0000 (UTC)
MIME-Version: 1.0
Date: Wed, 15 Sep 2021 22:04:19 -0500
From: elf-pavlik@hackers4peace.net
To: txauth@ietf.org
In-Reply-To: <77DFE057-4DD2-4786-9090-3F47AE70881E@mit.edu>
References: <10c8c1d93688494588a9d7032a0aa37d@hackers4peace.net> <CANYRo8j1XMSE9XWuLb8y0rB1Bpd47hNoa314qTXKNHHTGx0t-w@mail.gmail.com> <CAM8feuRLZ6fmmpAU9C+SWZkMbsQhK+OpgWOqmd7ASXFDFG=Zrw@mail.gmail.com> <996bafb5b7b9cc95639ddd8ad3f16728@hackers4peace.net> <77DFE057-4DD2-4786-9090-3F47AE70881E@mit.edu>
User-Agent: Roundcube Webmail/1.4.11
Message-ID: <840de5cc7c7bdf64198592e21edcbd92@hackers4peace.net>
X-Sender: elf-pavlik@hackers4peace.net
Content-Type: text/plain; charset=UTF-8; format=flowed
Content-Transfer-Encoding: 8bit
Archived-At: <https://mailarchive.ietf.org/arch/msg/txauth/ykStc2DlkU63bi6wmoHBczGM0No>
Subject: Re: [GNAP] Distinct role used by End-user to give permissions to client instances.
X-BeenThere: txauth@ietf.org
X-Mailman-Version: 2.1.29
Precedence: list
List-Id: GNAP <txauth.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/options/txauth>, <mailto:txauth-request@ietf.org?subject=unsubscribe>
List-Archive: <https://mailarchive.ietf.org/arch/browse/txauth/>
List-Post: <mailto:txauth@ietf.org>
List-Help: <mailto:txauth-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/txauth>, <mailto:txauth-request@ietf.org?subject=subscribe>
X-List-Received-Date: Thu, 16 Sep 2021 03:04:35 -0000

Hi Justin,

Thank you for your in-depth reply!
I'll try to respond in-line.

> Phase 1: Client Instance makes a request to EUXS for an assertion
> representing the current user. The nature and format of this assertion
> are out of scope for GNAP itself, but would be in the purview of
> Solid’s profile of the technologies — for now we’ll call this a Solid
> Assertion Token (SAT) within this description. A JWT would be a
> reasonable carrier for such an assertion. What GNAP offers is a
> standard way for the client instance to request this assertion and
> have it delivered directly. We’re assuming that the end user interacts
> with or otherwise associates with the EUXS so that the EUXS knows who
> the end user is. Or possibly, the EUXS knows that the client in
> question has already been authorized and can issue the SAT directly.
> GNAP allows for both scenarios to come out of the same request
> message, allowing for a clean fallback in different situations.

I haven't mentioned it (trying to start as simple as possible), but we 
consider
possibility of having separate IdP and EUXS.
The IdP would stay in line with OIDC and issue assertion representing 
the current end-user.
Trying to follow your naming we could call it Solid Identity Assertion 
Token (SIAT)
The EUXS would stay responsible for issuing an assertion representing
'permissions' - what access end-user delegates to the given client 
instance.
We could call this one Solid Permissions Assertion Token (SPAT)

While we evaluate defining separate IdP and EUXS, we also want to
support combined IdP+EUXS which could combine user interaction
into a single flow.

So far I think both IdP and EUXS act as AS. Until now also the
End-user acts as RO who controls issurance of both assertion SIAT and 
SPAT.

> 
> Phase 2: Client instance talks to the AS that controls access to the
> RS. The client instance presents the SAT as part of its “user”
> information field in the request. The AS validates the SAT based on
> Solid’s rules. The AS contacts the RO to get permission for the client
> instance on the end user’s behalf. How that happens is up to the AS,
> but the SAT would likely have some kind of RO identifier within it:
> namely representing what’s being requested, not who the SAT was issued
> to. Or it could be based on the RS and permissions the client instance
> is asking for — so the client instance and end user wouldn’t
> necessarily know who the RO was at all, just that they want access to
> the RS and it’s controlled by … someone. Once the RO approves it, the
> AS issues the access token to the client instance.

The SPAT would indeed contain information about what client can access
on a given RS. Actually EUXS would most likely need to issue one SPAT
per RS that End-user wants to access using the client.

Currently we look at approache where RS has only one RO which has
full control over all the Protected Resources on that RS.

At the sime time End-user will be using the client to access resources
across any number of Resource Servers, where each RS can be in control
of different RO.

Having one SPAT per RS should preven any undesired leak of information
across those different Resource Servers.

What I need to emphasis here, RO who controls given RS, doesn't really
get involved in authorizaing client instances (unless they also act as
End-user of that client). RO grants access directly to End-user,
than End-user can authorize/permission clients they use via EUXS,
independently of RO associated with RS.

Here also EUXS is responsible for authenticating the client,
AS associated with RS relies on client identity from SPAT.

When it comes to flow where End-user wants to request access to
resources on RS (RO != End-user). We consider their EUXS would
handle it for them. Once the End-user has access, they can delegate
it to any clients they want to use to excercise it.

> 
> Phase 3: Client instance calls the RS with the access token from AS.
> It may or may not contain any information from the SAT, since the only
> thing the RS ultimately needs to know is if the token is good for the
> request being made. Solid can further profile the access token itself
> (either its format or its introspected data) to convey specific
> information. GNAP remains agnostic on this decision, but provisions
> for standardizing these elements are in the GNAP-RS document.

Yes, I think this part seems straight forward.

> 
> 
> An alternative process inverts some of the phases:
> 
> 
> Phase 1 (alt): Client instance talks to the AS to get access and says
> it can interact with the end user through the EUXS and get an
> assertion (SAT) as a result. The AS passes whatever information it
> needs to complete this back to the client instance. The end user
> interacts at EUXS to get the SAT generated and handed back to the
> client instance.

I need to think further about this flow. Currently we focus mostly
on flow where client starts 'following their nose' from End-user
identiy, discovers their IdP and EUXS (this information is public).
Later EUXS provides an entry point which has index of all the data
across all ther RS that End-user wants to access using given client.
Given above client interacts with EUXS before interacting with any
RS and disovering AS that controlls access to it.


> 
> Do either of these processes align with what you had in mind, or am I
> missing some important subtleties to this use case? We’ve discussed
> both of these patterns here in the WG before and accommodations for
> them are built into the GNAP core protocol: both what it defines and
> what it allows but keeps out of scope of the core protocol
> definitions.

Yes I think it's pretty close. I hope I was able to highlight
additional nuances above.

Thank you again for taking your time to patiently disect
scenarios I'm working with.

Best regards,
elf Pavlik


From nobody Sun Sep 19 01:36:17 2021
Return-Path: <do_not_reply@mnot.net>
X-Original-To: txauth@ietfa.amsl.com
Delivered-To: txauth@ietfa.amsl.com
Received: from localhost (localhost [127.0.0.1]) by ietfa.amsl.com (Postfix) with ESMTP id 677823A2240 for <txauth@ietfa.amsl.com>; Sun, 19 Sep 2021 01:36:14 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -2.1
X-Spam-Level: 
X-Spam-Status: No, score=-2.1 tagged_above=-999 required=5 tests=[BAYES_00=-1.9, DKIM_SIGNED=0.1, DKIM_VALID=-0.1, DKIM_VALID_AU=-0.1, DKIM_VALID_EF=-0.1, HTML_MESSAGE=0.001, RCVD_IN_MSPIKE_H2=-0.001, SPF_PASS=-0.001, URIBL_BLOCKED=0.001] autolearn=ham autolearn_force=no
Authentication-Results: ietfa.amsl.com (amavisd-new); dkim=pass (2048-bit key) header.d=mnot.net header.b=Uz5aPXji; dkim=pass (2048-bit key) header.d=messagingengine.com header.b=sXZ28hX8
Received: from mail.ietf.org ([4.31.198.44]) by localhost (ietfa.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id hyK_s9nXpOwC for <txauth@ietfa.amsl.com>; Sun, 19 Sep 2021 01:36:09 -0700 (PDT)
Received: from wout2-smtp.messagingengine.com (wout2-smtp.messagingengine.com [64.147.123.25]) (using TLSv1.2 with cipher ADH-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by ietfa.amsl.com (Postfix) with ESMTPS id A24383A223F for <txauth@ietf.org>; Sun, 19 Sep 2021 01:36:09 -0700 (PDT)
Received: from compute1.internal (compute1.nyi.internal [10.202.2.41]) by mailout.west.internal (Postfix) with ESMTP id 0942B320046F for <txauth@ietf.org>; Sun, 19 Sep 2021 03:39:01 -0400 (EDT)
Received: from mailfrontend2 ([10.202.2.163]) by compute1.internal (MEProxy); Sun, 19 Sep 2021 03:39:02 -0400
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=mnot.net; h= content-type:mime-version:from:to:subject; s=fm3; bh=7XVqtjTG7pO cN5Kn1yIPY/ISC6P3X2PHhtvWpTauJP0=; b=Uz5aPXjiNph0f8UkVKpdtUaGbx4 vmJkpf/d4aQ0PP/wDa8p2szwNDuy44SZfKFxELcLJnbmcWxgxRA6YyXUADwM0v21 er4AG38xtJpNNhH0foZz0LXtfBKYPqz5pJexc30faciay45zLMblNL6+qU4tKnId QJ+HUSTWqTtk8rDeE0vdgrhzNI//mpVMnq/HjBW2/YHXmh2CUfDyj9PBSwLZrwot d6kEH0MKUCobPAmtQlDMfpxGpGZ7zBmyAlZijTUCzqCEkechCbM03AVggr1tVhY8 R7T1dVG7pRtjDMZJvfm3lMvLl6RmB+qnrlhI3u1jKL9ldpYx+EpdfU2eFgQ==
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d= messagingengine.com; h=content-type:from:mime-version:subject:to :x-me-proxy:x-me-proxy:x-me-sender:x-me-sender:x-sasl-enc; s= fm3; bh=7XVqtjTG7pOcN5Kn1yIPY/ISC6P3X2PHhtvWpTauJP0=; b=sXZ28hX8 6AWe6xzklTFT4aEeuohWAvI7psmpIDkwLthlSnh487n4NvUeBS589HSCGiFrtOWb mGZevuuONRxn1aJMpPn2R5+GB8iB3PzySVKC067LyL2Ssp4k8rEAqmfTWQDqI3Sa wZIQOgggniQCdufQaJM27xx6UkgEdoo9QAyir6E5/jdoA8fhTFcnmB6Qs83FOdH+ bb3Sq8sOrv1353DAH6tQHgLOAAjCr5vrcXW7pRt6wVPs1XUVeYhnsXrC29cB5L+W eTJQwqM6fU6y3918KkWvcrNx5GsK8saCj2eeONGcB+mlF+Wmy5kiyMUG1G6+480w 9pfV5vBXfCUvxg==
X-ME-Sender: <xms:FelGYfa5BRPu84xgWAiWJ6esGC4YDOs-OblVIMwqcpCuOg0QdsF4xw> <xme:FelGYeZNQUmgcBbG0G9yF6w3-Hoo-1zGgZyfUzYjl4aDWGyZO3MPUrbfIeAnURH4x EKW5i84VPdrPGq7dg>
X-ME-Received: <xmr:FelGYR8nMxjaTxUytFyL3ANhjm2Oni-LlZHwbcaGfv10xqCZe36kP_lDQEDMuzCD4pRX04X08g72qvr9kDNkxF4C7fTSyYSf-Mah0b-pHwXlt4pNFdvtOh9Uy5h9_AqYr6QXww>
X-ME-Proxy-Cause: gggruggvucftvghtrhhoucdtuddrgedvtddrudehledguddulecutefuodetggdotefrod ftvfcurfhrohhfihhlvgemucfhrghsthforghilhdpqfgfvfdpuffrtefokffrpgfnqfgh necuuegrihhlohhuthemuceftddtnecupfhoucgurghtvgcufhhivghlugculdegledmne cujfgurheptggghffvufesrgdttdertddtjeenucfhrhhomheptfgvphhoshhithhorhih ucettghtihhvihhthicuufhumhhmrghrhicuuehothcuoeguohgpnhhothgprhgvphhlhi esmhhnohhtrdhnvghtqeenucggtffrrghtthgvrhhnpeekfedvudetjedvfeekheeiveeu gfefhfetteevgeffkefffeetffdvleehudeiteenucffohhmrghinhepghhithhhuhgsrd gtohhmnecuvehluhhsthgvrhfuihiivgeptdenucfrrghrrghmpehmrghilhhfrhhomhep ughopghnohhtpghrvghplhihsehmnhhothdrnhgvth
X-ME-Proxy: <xmx:FelGYVruZHrHD5qoWPvpFJJytOazObZAjMTfDK26EOelI5nFKQ7tfQ> <xmx:FelGYaocWs7_Z3J4byg9miVYFem6yQMyed0HQIArjkTwGf8Bzh8glA> <xmx:FelGYbSMvSttJMQgd0VIDx0SXr1B-LWxNgyDLnCCri1nk0WPFQvyQw> <xmx:FelGYR1VyAgYYUo9xaM68GBYf93WQ4sGuWdGMmaCBTbzV5BENZUX3w>
Received: by mail.messagingengine.com (Postfix) with ESMTPA for <txauth@ietf.org>; Sun, 19 Sep 2021 03:39:01 -0400 (EDT)
Content-Type: multipart/alternative; boundary="===============7651401569174353880=="
MIME-Version: 1.0
From: Repository Activity Summary Bot <do_not_reply@mnot.net>
To: txauth@ietf.org
Message-Id: <20210919083609.A24383A223F@ietfa.amsl.com>
Date: Sun, 19 Sep 2021 01:36:09 -0700 (PDT)
Archived-At: <https://mailarchive.ietf.org/arch/msg/txauth/r3EHG27naeTxroRnMDhedDEt3hM>
Subject: [GNAP] Weekly github digest (GNAP Weekly GitHub Activity Summary)
X-BeenThere: txauth@ietf.org
X-Mailman-Version: 2.1.29
Precedence: list
List-Id: GNAP <txauth.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/options/txauth>, <mailto:txauth-request@ietf.org?subject=unsubscribe>
List-Archive: <https://mailarchive.ietf.org/arch/browse/txauth/>
List-Post: <mailto:txauth@ietf.org>
List-Help: <mailto:txauth-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/txauth>, <mailto:txauth-request@ietf.org?subject=subscribe>
X-List-Received-Date: Sun, 19 Sep 2021 08:36:15 -0000

--===============7651401569174353880==
MIME-Version: 1.0
Content-Transfer-Encoding: quoted-printable
Content-Type: text/plain; charset="utf-8"; format="flowed"




Events without label "editorial"

Issues
------
* ietf-wg-gnap/core-protocol (+0/-0/=F0=9F=92=AC7)
  1 issues received 7 new comments:
  - #291 Requesting RO Information or Requesting Subject Information? (7 by=
 elf-pavlik, fimbault, jricher)
    https://github.com/ietf-wg-gnap/gnap-core-protocol/issues/291=20




Repositories tracked by this digest:
-----------------------------------
* https://github.com/ietf-wg-gnap/core-protocol
* https://github.com/ietf-wg-gnap/gnap-resource-servers

--===============7651401569174353880==
Content-Type: text/html; charset="utf-8"
MIME-Version: 1.0
Content-Transfer-Encoding: quoted-printable

<!doctype html>
<html lang=3D"en">
<head>
<meta charset=3D"utf-8">
<title>Weekly github digest (GNAP Weekly GitHub Activity Summary)</title>
<style>
body { font-family: Gotham, "Helvetica Neue", Helvetica, Arial, sans-serif;=
 font-size: 14px; }
h2 { margin-top: 3em; color: #A52A2A; font-style: italic; font-weight: norm=
al; }
h3 { margin-bottom:0; margin-top: 2em; font-size: 1.2em; }
h1+h2 { margin-top: 1em; }
a { color: #bb6219; text-decoration: none; }
li { margin-bottom: .35em; }
.repos { margin-bottom: 0; margin-top:0; line-height: 1.2; }
.new { color: red; }
.label { display: inline;
	padding: .2em .6em .3em;
	font-size: 75%;
	font-weight: 700;
	line-height: 1;
	color: #fff;
	text-align: center;
	white-space: nowrap;
	vertical-align: baseline;
	border-radius: .25em;
}
</style>
</head>

<body>
<h1>Sunday September 19, 2021</h1>

<p>Events without label "editorial"</p>

<h2>Issues</h2>

<h3>ietf-wg-gnap/core-protocol (+0/-0/=F0=9F=92=AC7)</h3>

  <p>1 issues received 7 new comments:</p>
  <ul>
  <li>#291 <a href=3D"https://github.com/ietf-wg-gnap/gnap-core-protocol/is=
sues/291">Requesting RO Information or Requesting Subject Information?</a> =
(7 by elf-pavlik, fimbault, jricher) </li>
  </ul>





<h2>Repositories tracked by this digest:</h2>
<ul class=3D"repos">
  <li><a href=3D"https://github.com/ietf-wg-gnap/core-protocol">https://git=
hub.com/ietf-wg-gnap/core-protocol</a></li>
  <li><a href=3D"https://github.com/ietf-wg-gnap/gnap-resource-servers">htt=
ps://github.com/ietf-wg-gnap/gnap-resource-servers</a></li>
  </ul>
</body>
</html>

--===============7651401569174353880==--


From nobody Wed Sep 22 10:11:25 2021
Return-Path: <jricher@mit.edu>
X-Original-To: txauth@ietfa.amsl.com
Delivered-To: txauth@ietfa.amsl.com
Received: from localhost (localhost [127.0.0.1]) by ietfa.amsl.com (Postfix) with ESMTP id 31C213A29BB for <txauth@ietfa.amsl.com>; Wed, 22 Sep 2021 10:11:20 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -1.498
X-Spam-Level: 
X-Spam-Status: No, score=-1.498 tagged_above=-999 required=5 tests=[BAYES_00=-1.9, HTML_MESSAGE=0.001, KHOP_HELO_FCRDNS=0.399, SPF_HELO_NONE=0.001, SPF_NONE=0.001] autolearn=no autolearn_force=no
Received: from mail.ietf.org ([4.31.198.44]) by localhost (ietfa.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id Mvmf_9yqpf00 for <txauth@ietfa.amsl.com>; Wed, 22 Sep 2021 10:11:18 -0700 (PDT)
Received: from outgoing.mit.edu (outgoing-auth-1.mit.edu [18.9.28.11]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by ietfa.amsl.com (Postfix) with ESMTPS id D4B2C3A29AB for <txauth@ietf.org>; Wed, 22 Sep 2021 10:11:15 -0700 (PDT)
Received: from [192.168.1.28] (static-71-174-62-56.bstnma.fios.verizon.net [71.174.62.56]) (authenticated bits=0) (User authenticated as jricher@ATHENA.MIT.EDU) by outgoing.mit.edu (8.14.7/8.12.4) with ESMTP id 18MHBCuX021908 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-GCM-SHA384 bits=256 verify=NOT) for <txauth@ietf.org>; Wed, 22 Sep 2021 13:11:13 -0400
From: Justin Richer <jricher@mit.edu>
Content-Type: multipart/alternative; boundary="Apple-Mail=_F628CB5E-F25F-4C5B-BFE0-0232A4F640AC"
Mime-Version: 1.0 (Mac OS X Mail 13.4 \(3608.120.23.2.7\))
Message-Id: <32F934B0-1E11-49A6-BF40-94742475BC86@mit.edu>
Date: Wed, 22 Sep 2021 13:11:12 -0400
To: GNAP Mailing List <txauth@ietf.org>
X-Mailer: Apple Mail (2.3608.120.23.2.7)
Archived-At: <https://mailarchive.ietf.org/arch/msg/txauth/CHCe3vpE1lpT88hhYIDxCYxxUPs>
Subject: [GNAP] Draft Updates for Interim Meeting
X-BeenThere: txauth@ietf.org
X-Mailman-Version: 2.1.29
Precedence: list
List-Id: GNAP <txauth.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/options/txauth>, <mailto:txauth-request@ietf.org?subject=unsubscribe>
List-Archive: <https://mailarchive.ietf.org/arch/browse/txauth/>
List-Post: <mailto:txauth@ietf.org>
List-Help: <mailto:txauth-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/txauth>, <mailto:txauth-request@ietf.org?subject=subscribe>
X-List-Received-Date: Wed, 22 Sep 2021 17:11:20 -0000

--Apple-Mail=_F628CB5E-F25F-4C5B-BFE0-0232A4F640AC
Content-Transfer-Encoding: quoted-printable
Content-Type: text/plain;
	charset=utf-8

Hi everyone,

The editors met today and we are targeting the following four PRs for =
inclusion into the draft ahead of the next Interim on October 5th:

Security Considerations
https://github.com/ietf-wg-gnap/gnap-core-protocol/pull/304 =
<https://github.com/ietf-wg-gnap/gnap-core-protocol/pull/304>

Trust Relationships
https://github.com/ietf-wg-gnap/gnap-core-protocol/pull/306 =
<https://github.com/ietf-wg-gnap/gnap-core-protocol/pull/306>

Privacy Considerations
https://github.com/ietf-wg-gnap/gnap-core-protocol/pull/307 =
<https://github.com/ietf-wg-gnap/gnap-core-protocol/pull/307>

Discussing these sections will be the major topic of the upcoming =
interim meeting. None of these are meant to be comprehensive and final, =
but all of them provide significant improvement over the placeholders in =
the text to date.=20

In addition, one technical change is being finalized that we expect to =
be included:

Use opaque identifier instead of user handle
https://github.com/ietf-wg-gnap/gnap-core-protocol/pull/305 =
<https://github.com/ietf-wg-gnap/gnap-core-protocol/pull/305>


The editors plan to include all of these in a draft to be published this =
Friday as -07, to give people time to read the draft before the interim =
meeting.

 =E2=80=94 Justin=

--Apple-Mail=_F628CB5E-F25F-4C5B-BFE0-0232A4F640AC
Content-Transfer-Encoding: quoted-printable
Content-Type: text/html;
	charset=utf-8

<html><head><meta http-equiv=3D"Content-Type" content=3D"text/html; =
charset=3Dutf-8"></head><body style=3D"word-wrap: break-word; =
-webkit-nbsp-mode: space; line-break: after-white-space;" class=3D"">Hi =
everyone,<div class=3D""><br class=3D""></div><div class=3D"">The =
editors met today and we are targeting the following four PRs for =
inclusion into the draft ahead of the next Interim on October =
5th:</div><div class=3D""><br class=3D""></div><div class=3D"">Security =
Considerations</div><div class=3D""><a =
href=3D"https://github.com/ietf-wg-gnap/gnap-core-protocol/pull/304" =
class=3D"">https://github.com/ietf-wg-gnap/gnap-core-protocol/pull/304</a>=
</div><div class=3D""><br class=3D""></div><div class=3D"">Trust =
Relationships</div><div class=3D""><a =
href=3D"https://github.com/ietf-wg-gnap/gnap-core-protocol/pull/306" =
class=3D"">https://github.com/ietf-wg-gnap/gnap-core-protocol/pull/306</a>=
</div><div class=3D""><br class=3D""></div><div class=3D"">Privacy =
Considerations</div><div class=3D""><a =
href=3D"https://github.com/ietf-wg-gnap/gnap-core-protocol/pull/307" =
class=3D"">https://github.com/ietf-wg-gnap/gnap-core-protocol/pull/307</a>=
</div><div class=3D""><br class=3D""></div><div class=3D"">Discussing =
these sections will be the major topic of the upcoming interim meeting. =
None of these are meant to be comprehensive and final, but all of them =
provide significant improvement over the placeholders in the text to =
date.&nbsp;</div><div class=3D""><br class=3D""></div><div class=3D"">In =
addition, one technical change is being finalized that we expect to be =
included:</div><div class=3D""><br class=3D""></div><div class=3D"">Use =
opaque identifier instead of user handle</div><div class=3D""><a =
href=3D"https://github.com/ietf-wg-gnap/gnap-core-protocol/pull/305" =
class=3D"">https://github.com/ietf-wg-gnap/gnap-core-protocol/pull/305</a>=
</div><div class=3D""><br class=3D""></div><div class=3D""><br =
class=3D""></div><div class=3D"">The editors plan to include all of =
these in a draft to be published this Friday as -07, to give people time =
to read the draft before the interim meeting.</div><div class=3D""><br =
class=3D""></div><div class=3D"">&nbsp;=E2=80=94 =
Justin</div></body></html>=

--Apple-Mail=_F628CB5E-F25F-4C5B-BFE0-0232A4F640AC--


From nobody Fri Sep 24 11:10:51 2021
Return-Path: <internet-drafts@ietf.org>
X-Original-To: txauth@ietf.org
Delivered-To: txauth@ietfa.amsl.com
Received: from ietfa.amsl.com (localhost [IPv6:::1]) by ietfa.amsl.com (Postfix) with ESMTP id B98D03A0A83; Fri, 24 Sep 2021 11:09:54 -0700 (PDT)
MIME-Version: 1.0
Content-Type: text/plain; charset="utf-8"
Content-Transfer-Encoding: 7bit
From: internet-drafts@ietf.org
To: <i-d-announce@ietf.org>
Cc: txauth@ietf.org
X-Test-IDTracker: no
X-IETF-IDTracker: 7.38.0
Auto-Submitted: auto-generated
Precedence: bulk
Reply-To: txauth@ietf.org
Message-ID: <163250699468.15335.5613363144392423196@ietfa.amsl.com>
Date: Fri, 24 Sep 2021 11:09:54 -0700
Archived-At: <https://mailarchive.ietf.org/arch/msg/txauth/VD6SLXKHuNyNFCL6zOFKT0tv0kU>
Subject: [GNAP] I-D Action: draft-ietf-gnap-core-protocol-07.txt
X-BeenThere: txauth@ietf.org
X-Mailman-Version: 2.1.29
List-Id: GNAP <txauth.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/options/txauth>, <mailto:txauth-request@ietf.org?subject=unsubscribe>
List-Archive: <https://mailarchive.ietf.org/arch/browse/txauth/>
List-Post: <mailto:txauth@ietf.org>
List-Help: <mailto:txauth-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/txauth>, <mailto:txauth-request@ietf.org?subject=subscribe>
X-List-Received-Date: Fri, 24 Sep 2021 18:09:55 -0000

A New Internet-Draft is available from the on-line Internet-Drafts directories.
This draft is a work item of the Grant Negotiation and Authorization Protocol WG of the IETF.

        Title           : Grant Negotiation and Authorization Protocol
        Authors         : Justin Richer
                          Aaron Parecki
                          Fabien Imbault
	Filename        : draft-ietf-gnap-core-protocol-07.txt
	Pages           : 151
	Date            : 2021-09-24

Abstract:
   GNAP defines a mechanism for delegating authorization to a piece of
   software, and conveying that delegation to the software.  This
   delegation can include access to a set of APIs as well as information
   passed directly to the software.


The IETF datatracker status page for this draft is:
https://datatracker.ietf.org/doc/draft-ietf-gnap-core-protocol/

There is also an HTML version available at:
https://www.ietf.org/archive/id/draft-ietf-gnap-core-protocol-07.html

A diff from the previous version is available at:
https://www.ietf.org/rfcdiff?url2=draft-ietf-gnap-core-protocol-07


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



From nobody Sun Sep 26 00:56:32 2021
Return-Path: <do_not_reply@mnot.net>
X-Original-To: txauth@ietfa.amsl.com
Delivered-To: txauth@ietfa.amsl.com
Received: from localhost (localhost [127.0.0.1]) by ietfa.amsl.com (Postfix) with ESMTP id EEEA73A15D5 for <txauth@ietfa.amsl.com>; Sun, 26 Sep 2021 00:55:56 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -2.1
X-Spam-Level: 
X-Spam-Status: No, score=-2.1 tagged_above=-999 required=5 tests=[BAYES_00=-1.9, DKIM_SIGNED=0.1, DKIM_VALID=-0.1, DKIM_VALID_AU=-0.1, DKIM_VALID_EF=-0.1, HTML_MESSAGE=0.001, RCVD_IN_MSPIKE_H2=-0.001, SPF_PASS=-0.001, URIBL_BLOCKED=0.001] autolearn=ham autolearn_force=no
Authentication-Results: ietfa.amsl.com (amavisd-new); dkim=pass (2048-bit key) header.d=mnot.net header.b=ZBoSymmM; dkim=pass (2048-bit key) header.d=messagingengine.com header.b=Lx9Iz4Av
Received: from mail.ietf.org ([4.31.198.44]) by localhost (ietfa.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id lBx7ZDv8gr45 for <txauth@ietfa.amsl.com>; Sun, 26 Sep 2021 00:55:51 -0700 (PDT)
Received: from out3-smtp.messagingengine.com (out3-smtp.messagingengine.com [66.111.4.27]) (using TLSv1.2 with cipher ADH-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by ietfa.amsl.com (Postfix) with ESMTPS id BACA33A15D6 for <txauth@ietf.org>; Sun, 26 Sep 2021 00:55:51 -0700 (PDT)
Received: from compute2.internal (compute2.nyi.internal [10.202.2.42]) by mailout.nyi.internal (Postfix) with ESMTP id 924895C0114 for <txauth@ietf.org>; Sun, 26 Sep 2021 03:38:47 -0400 (EDT)
Received: from mailfrontend1 ([10.202.2.162]) by compute2.internal (MEProxy); Sun, 26 Sep 2021 03:38:47 -0400
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=mnot.net; h= content-type:mime-version:from:to:subject; s=fm3; bh=8ckrSSZwtS3 cFtIALxnQsOSZbjDpL7s4W1iQof+06D0=; b=ZBoSymmMC54ym9vOJa0VhJV8j/R Lnb/eqORdcrq/djfOsJjJ30SZEJ0PwJ1sONIteuYmK65g6rAjiuLz2JNzw+DL1xX HBiVzQ8vVFf2T0DyTQHCBaMs/lNvlNaiGKtPztKdRH3rnC54IiKhEA0C+yEJUizR RFGsMhbMok6mP2Fso1sF3bwEJvv/4z3V5c/EjHnyFCxpNistpyy4A8kxT6msd9kP biPM2psf6tSkyK6ZG5FpAQK4CeGExfL06f9xPNjQ8aGP8Rj6po9qo7LPiTzoeKy+ njtKY61dPoUAWVxje/2O68aYGiPRLvkoG0h0vHUiHA0WHjckkk+3/Mcjflg==
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d= messagingengine.com; h=content-type:from:mime-version:subject:to :x-me-proxy:x-me-proxy:x-me-sender:x-me-sender:x-sasl-enc; s= fm3; bh=8ckrSSZwtS3cFtIALxnQsOSZbjDpL7s4W1iQof+06D0=; b=Lx9Iz4Av V7qV6acLPNdcDBIBYH7p9XoJrPQc175P/TFPTkjJwr2xZukRuQ8gbMtsDFdnjVQl QjZ78cCewcOAqcvlzxE3pq6w2TYZNemqDjXt3PqNDnk6MRa4MPYB90eB6YXKhRri Q5gFjHYYsJTX+A98b+cb5Y5oofVkuDk+7i4xawlq8aVfRhiLStdV7hwtvc1Kdq3Y eW5yFDCH+mXsGnmsl5oCa4+seJjtP/Md6L2vf/YDyugirKYknRE/Yh8SUkzIAt2j LfLSfK04v9JTM1dwW3ROaTBkM2GFz48ZxvNmJrRWeglmhu55yolKDTCH8Y8cWGeY YCQSMjcqjTYvDw==
X-ME-Sender: <xms:hyNQYSAvpLRTAByN04gYtIflurFSf-wA5sHNTAL42YkLIthS5G8afA> <xme:hyNQYcjstOW58n5yIQA0rDn0i1GrauxhE1PF_X4AwSQRP-4jRZ9qBFj7NADy92PU3 TQ7GUDxHtfKFbw9MA>
X-ME-Received: <xmr:hyNQYVnZhAog058oFPL8hm4Pf-FdxA5KEeAnZ7qQ4WQZ2mMBjo_lIgt5DW7Gv3sDn0nLWBayygPGYIQDCH97-Hs3JSxKsW05BYnOTUUUxms2H_DDuhL1nozO3PP6bQzYl5XOqQ>
X-ME-Proxy-Cause: gggruggvucftvghtrhhoucdtuddrgedvtddrudejhedgudeiucetufdoteggodetrfdotf fvucfrrhhofhhilhgvmecuhfgrshhtofgrihhlpdfqfgfvpdfurfetoffkrfgpnffqhgen uceurghilhhouhhtmecufedttdenucfpohcuuggrthgvuchfihgvlhguucdlgeelmdenuc fjughrpegtggfhvffusegrtddtredttdejnecuhfhrohhmpeftvghpohhsihhtohhrhicu tegtthhivhhithihucfuuhhmmhgrrhihuceuohhtuceoughopghnohhtpghrvghplhihse hmnhhothdrnhgvtheqnecuggftrfgrthhtvghrnhepkeefvdduteejvdefkeehieevuefg fefhteetveegffekffefteffvdelheduieetnecuffhomhgrihhnpehgihhthhhusgdrtg homhenucevlhhushhtvghrufhiiigvpedvnecurfgrrhgrmhepmhgrihhlfhhrohhmpegu ohgpnhhothgprhgvphhlhiesmhhnohhtrdhnvght
X-ME-Proxy: <xmx:hyNQYQztxaZSaamURyDNZj-0wtUd69Hpup1NRLbvauZcrHj_OcNfHQ> <xmx:hyNQYXR6kgBiMd3XYvfygaNFs-KpH2rmiPHV2toB-G6rKTcWYSKoIA> <xmx:hyNQYbb3ONQfELPg6Zqig5Ve_-jAtOhvMfX1lKBZSGPebqOHqMG03g> <xmx:hyNQYSfifojcu3IC_WqPhXp87QXjy1gcDrquL4lNyoJav9kNmWno9A>
Received: by mail.messagingengine.com (Postfix) with ESMTPA for <txauth@ietf.org>; Sun, 26 Sep 2021 03:38:47 -0400 (EDT)
Content-Type: multipart/alternative; boundary="===============5407714161980052886=="
MIME-Version: 1.0
From: Repository Activity Summary Bot <do_not_reply@mnot.net>
To: txauth@ietf.org
Message-Id: <20210926075551.BACA33A15D6@ietfa.amsl.com>
Date: Sun, 26 Sep 2021 00:55:51 -0700 (PDT)
Archived-At: <https://mailarchive.ietf.org/arch/msg/txauth/6OFK1LxEcbCaCfjnOxmxQjJNf1g>
Subject: [GNAP] Weekly github digest (GNAP Weekly GitHub Activity Summary)
X-BeenThere: txauth@ietf.org
X-Mailman-Version: 2.1.29
Precedence: list
List-Id: GNAP <txauth.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/options/txauth>, <mailto:txauth-request@ietf.org?subject=unsubscribe>
List-Archive: <https://mailarchive.ietf.org/arch/browse/txauth/>
List-Post: <mailto:txauth@ietf.org>
List-Help: <mailto:txauth-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/txauth>, <mailto:txauth-request@ietf.org?subject=subscribe>
X-List-Received-Date: Sun, 26 Sep 2021 07:55:57 -0000

--===============5407714161980052886==
MIME-Version: 1.0
Content-Transfer-Encoding: quoted-printable
Content-Type: text/plain; charset="utf-8"; format="flowed"




Events without label "editorial"

Issues
------
* ietf-wg-gnap/core-protocol (+0/-3/=F0=9F=92=AC6)
  1 issues received 6 new comments:
  - #291 Requesting RO Information or Requesting Subject Information? (6 by=
 agropper, fimbault, jricher)
    https://github.com/ietf-wg-gnap/gnap-core-protocol/issues/291=20

  3 issues closed:
  - User reference as an assertion https://github.com/ietf-wg-gnap/gnap-cor=
e-protocol/issues/51=20
  - Use opaque identifer instead of user handle https://github.com/ietf-wg-=
gnap/gnap-core-protocol/issues/302=20
  - Requesting User Information https://github.com/ietf-wg-gnap/gnap-core-p=
rotocol/issues/197=20



Pull requests
-------------
* ietf-wg-gnap/core-protocol (+5/-6/=F0=9F=92=AC6)
  5 pull requests submitted:
  - Fix label of arrow (by pq2)
    https://github.com/ietf-wg-gnap/gnap-core-protocol/pull/311=20
  - Change apostrophes to backticks (by pq2)
    https://github.com/ietf-wg-gnap/gnap-core-protocol/pull/310=20
  - Clarified use of user identifiers by reference (by jricher)
    https://github.com/ietf-wg-gnap/gnap-core-protocol/pull/309=20
  - Consistent naming of 'resource reference' (by jmandel)
    https://github.com/ietf-wg-gnap/gnap-core-protocol/pull/308=20
  - initial draft of privacy considerations section (by aaronpk)
    https://github.com/ietf-wg-gnap/gnap-core-protocol/pull/307=20

  5 pull requests received 6 new comments:
  - #311 Fix label of arrow (1 by netlify)
    https://github.com/ietf-wg-gnap/gnap-core-protocol/pull/311=20
  - #310 Change apostrophes to backticks (1 by netlify)
    https://github.com/ietf-wg-gnap/gnap-core-protocol/pull/310=20
  - #309 Clarified use of user identifiers by reference (1 by netlify)
    https://github.com/ietf-wg-gnap/gnap-core-protocol/pull/309=20
  - #308 Consistent naming of 'resource reference' (2 by jricher, netlify)
    https://github.com/ietf-wg-gnap/gnap-core-protocol/pull/308 [Editorial]=
=20
  - #307 initial draft of privacy considerations section (1 by netlify)
    https://github.com/ietf-wg-gnap/gnap-core-protocol/pull/307=20

  6 pull requests merged:
  - Clarified use of user identifiers by reference
    https://github.com/ietf-wg-gnap/gnap-core-protocol/pull/309 [Pending Me=
rge]=20
  - Use opaque identifier instead of user handle
    https://github.com/ietf-wg-gnap/gnap-core-protocol/pull/305 [Pending Me=
rge]=20
  - initial draft of trust relationships
    https://github.com/ietf-wg-gnap/gnap-core-protocol/pull/306 [Pending Me=
rge]=20
  - Consistent naming of 'resource reference'
    https://github.com/ietf-wg-gnap/gnap-core-protocol/pull/308 [Editorial]=
=20
  - initial draft of privacy considerations section
    https://github.com/ietf-wg-gnap/gnap-core-protocol/pull/307 [Pending Me=
rge]=20
  - Security Considerations
    https://github.com/ietf-wg-gnap/gnap-core-protocol/pull/304 [Pending Me=
rge]=20


Repositories tracked by this digest:
-----------------------------------
* https://github.com/ietf-wg-gnap/core-protocol
* https://github.com/ietf-wg-gnap/gnap-resource-servers

--===============5407714161980052886==
Content-Type: text/html; charset="utf-8"
MIME-Version: 1.0
Content-Transfer-Encoding: quoted-printable

<!doctype html>
<html lang=3D"en">
<head>
<meta charset=3D"utf-8">
<title>Weekly github digest (GNAP Weekly GitHub Activity Summary)</title>
<style>
body { font-family: Gotham, "Helvetica Neue", Helvetica, Arial, sans-serif;=
 font-size: 14px; }
h2 { margin-top: 3em; color: #A52A2A; font-style: italic; font-weight: norm=
al; }
h3 { margin-bottom:0; margin-top: 2em; font-size: 1.2em; }
h1+h2 { margin-top: 1em; }
a { color: #bb6219; text-decoration: none; }
li { margin-bottom: .35em; }
.repos { margin-bottom: 0; margin-top:0; line-height: 1.2; }
.new { color: red; }
.label { display: inline;
	padding: .2em .6em .3em;
	font-size: 75%;
	font-weight: 700;
	line-height: 1;
	color: #fff;
	text-align: center;
	white-space: nowrap;
	vertical-align: baseline;
	border-radius: .25em;
}
</style>
</head>

<body>
<h1>Sunday September 26, 2021</h1>

<p>Events without label "editorial"</p>

<h2>Issues</h2>

<h3>ietf-wg-gnap/core-protocol (+0/-3/=F0=9F=92=AC6)</h3>

  <p>1 issues received 6 new comments:</p>
  <ul>
  <li>#291 <a href=3D"https://github.com/ietf-wg-gnap/gnap-core-protocol/is=
sues/291">Requesting RO Information or Requesting Subject Information?</a> =
(6 by agropper, fimbault, jricher) </li>
  </ul>

  <p>3 issues closed:</p>
  <ul>
  <li>#51 <a href=3D"https://github.com/ietf-wg-gnap/gnap-core-protocol/iss=
ues/51">User reference as an assertion</a> </li>
 =20
  <li>#302 <a href=3D"https://github.com/ietf-wg-gnap/gnap-core-protocol/is=
sues/302">Use opaque identifer instead of user handle</a> </li>
 =20
  <li>#197 <a href=3D"https://github.com/ietf-wg-gnap/gnap-core-protocol/is=
sues/197">Requesting User Information</a> </li>
  </ul>



<h2>Pull requests</h2>
<h3>ietf-wg-gnap/core-protocol (+5/-6/=F0=9F=92=AC6)</h3>
  <p class=3D"new">5 pull requests submitted:</p>
  <ul>
  <li>#311 <a href=3D"https://github.com/ietf-wg-gnap/gnap-core-protocol/pu=
ll/311">Fix label of arrow</a> (by pq2) </li>
 =20
  <li>#310 <a href=3D"https://github.com/ietf-wg-gnap/gnap-core-protocol/pu=
ll/310">Change apostrophes to backticks</a> (by pq2) </li>
 =20
  <li>#309 <a href=3D"https://github.com/ietf-wg-gnap/gnap-core-protocol/pu=
ll/309">Clarified use of user identifiers by reference</a> (by jricher) </l=
i>
 =20
  <li>#308 <a href=3D"https://github.com/ietf-wg-gnap/gnap-core-protocol/pu=
ll/308">Consistent naming of &#x27;resource reference&#x27;</a> (by jmandel=
) </li>
 =20
  <li>#307 <a href=3D"https://github.com/ietf-wg-gnap/gnap-core-protocol/pu=
ll/307">initial draft of privacy considerations section</a> (by aaronpk) </=
li>
  </ul>

  <p>5 pull requests received 6 new comments:</p>
  <ul>
  <li>#311 <a href=3D"https://github.com/ietf-wg-gnap/gnap-core-protocol/pu=
ll/311">Fix label of arrow</a> (1 by netlify) </li>
 =20
  <li>#310 <a href=3D"https://github.com/ietf-wg-gnap/gnap-core-protocol/pu=
ll/310">Change apostrophes to backticks</a> (1 by netlify) </li>
 =20
  <li>#309 <a href=3D"https://github.com/ietf-wg-gnap/gnap-core-protocol/pu=
ll/309">Clarified use of user identifiers by reference</a> (1 by netlify) <=
/li>
 =20
  <li>#308 <a href=3D"https://github.com/ietf-wg-gnap/gnap-core-protocol/pu=
ll/308">Consistent naming of &#x27;resource reference&#x27;</a> (2 by jrich=
er, netlify) <span class=3D"label" style=3D"background-color: #bfd4f2; colo=
r: #000000">Editorial</span> </li>
 =20
  <li>#307 <a href=3D"https://github.com/ietf-wg-gnap/gnap-core-protocol/pu=
ll/307">initial draft of privacy considerations section</a> (1 by netlify) =
</li>
  </ul>

  <p>6 pull requests merged:</p>
  <ul>
  <li>#309 <a href=3D"https://github.com/ietf-wg-gnap/gnap-core-protocol/pu=
ll/309">Clarified use of user identifiers by reference</a> <span class=3D"l=
abel" style=3D"background-color: #a6f490; color: #">Pending Merge</span> </=
li>
 =20
  <li>#305 <a href=3D"https://github.com/ietf-wg-gnap/gnap-core-protocol/pu=
ll/305">Use opaque identifier instead of user handle</a> <span class=3D"lab=
el" style=3D"background-color: #a6f490; color: #">Pending Merge</span> </li>
 =20
  <li>#306 <a href=3D"https://github.com/ietf-wg-gnap/gnap-core-protocol/pu=
ll/306">initial draft of trust relationships</a> <span class=3D"label" styl=
e=3D"background-color: #a6f490; color: #">Pending Merge</span> </li>
 =20
  <li>#308 <a href=3D"https://github.com/ietf-wg-gnap/gnap-core-protocol/pu=
ll/308">Consistent naming of &#x27;resource reference&#x27;</a> <span class=
=3D"label" style=3D"background-color: #bfd4f2; color: #">Editorial</span> <=
/li>
 =20
  <li>#307 <a href=3D"https://github.com/ietf-wg-gnap/gnap-core-protocol/pu=
ll/307">initial draft of privacy considerations section</a> <span class=3D"=
label" style=3D"background-color: #a6f490; color: #">Pending Merge</span> <=
/li>
 =20
  <li>#304 <a href=3D"https://github.com/ietf-wg-gnap/gnap-core-protocol/pu=
ll/304">Security Considerations</a> <span class=3D"label" style=3D"backgrou=
nd-color: #a6f490; color: #">Pending Merge</span> </li>
  </ul>


<h2>Repositories tracked by this digest:</h2>
<ul class=3D"repos">
  <li><a href=3D"https://github.com/ietf-wg-gnap/core-protocol">https://git=
hub.com/ietf-wg-gnap/core-protocol</a></li>
  <li><a href=3D"https://github.com/ietf-wg-gnap/gnap-resource-servers">htt=
ps://github.com/ietf-wg-gnap/gnap-resource-servers</a></li>
  </ul>
</body>
</html>

--===============5407714161980052886==--


From nobody Mon Sep 27 10:56:42 2021
Return-Path: <jmandel@gmail.com>
X-Original-To: txauth@ietfa.amsl.com
Delivered-To: txauth@ietfa.amsl.com
Received: from localhost (localhost [127.0.0.1]) by ietfa.amsl.com (Postfix) with ESMTP id 1602C3A1010 for <txauth@ietfa.amsl.com>; Mon, 27 Sep 2021 10:56:40 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -2.098
X-Spam-Level: 
X-Spam-Status: No, score=-2.098 tagged_above=-999 required=5 tests=[BAYES_00=-1.9, DKIM_SIGNED=0.1, DKIM_VALID=-0.1, DKIM_VALID_AU=-0.1, DKIM_VALID_EF=-0.1, FREEMAIL_FROM=0.001, HTML_MESSAGE=0.001, SPF_HELO_NONE=0.001, SPF_PASS=-0.001] autolearn=ham autolearn_force=no
Authentication-Results: ietfa.amsl.com (amavisd-new); dkim=pass (2048-bit key) header.d=gmail.com
Received: from mail.ietf.org ([4.31.198.44]) by localhost (ietfa.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id FpWc_XQ4F4Mz for <txauth@ietfa.amsl.com>; Mon, 27 Sep 2021 10:56:36 -0700 (PDT)
Received: from mail-yb1-xb2a.google.com (mail-yb1-xb2a.google.com [IPv6:2607:f8b0:4864:20::b2a]) (using TLSv1.2 with cipher ECDHE-RSA-AES128-GCM-SHA256 (128/128 bits)) (No client certificate requested) by ietfa.amsl.com (Postfix) with ESMTPS id 6C67B3A100F for <txauth@ietf.org>; Mon, 27 Sep 2021 10:56:36 -0700 (PDT)
Received: by mail-yb1-xb2a.google.com with SMTP id i84so25835386ybc.12 for <txauth@ietf.org>; Mon, 27 Sep 2021 10:56:36 -0700 (PDT)
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=20210112;  h=mime-version:from:date:message-id:subject:to; bh=GFY/pfgbKvdZiMxXcOoYhn9S+DihUicMI+e+7/FIrWY=; b=hSkLyfhgUNvpO0Q83WOlIGsp+suwlo37GcQirSfo5CBBXDSNeuVPM6B9JhiBlI6cPT pjxMekz4qoDJs6FGyhOelKc8Zsoqk/VLj046dcTKhvavacY11wtmlI9NZlm2qxfFR4RC GVHCRIuM2mBYtiZcDpeggrBF70fanNEUtdQrxNBt2tT2JhgXdjhoSduUdvkhe8LDUm0z H1zAjvkr7YmPdBXA/qb3Cja9RHxfbC4ELw/QJOJ74Zef4Gkj9kUX9n2MZcfxbIejdE1p 7evH/d4Od6g/WiVmxwQH1Wby1Rd1Eo5ePAo1F+rNhfS82/QrHadEQXUX5pIcX75J97Wl 3x1g==
X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20210112; h=x-gm-message-state:mime-version:from:date:message-id:subject:to; bh=GFY/pfgbKvdZiMxXcOoYhn9S+DihUicMI+e+7/FIrWY=; b=WebYBiPavQOY9Mj+aUkVaBHP3bw0taSu9a2W7LoF6qZtEKi4YmuZvN8rLt3sw22zgE SJc+OS2MXVNGk0GsvYQ6kMRm+9fmIVSLVs5rjFmOSl88VnSlyT12ekEQOYmIHwu7ruiJ rYcc0wMMRct0k6VGwVMtmNUWVMsl48x+EbAh1ZMVC5ZtRU0/dLzYgL7MS3D0Irskv/ne BVVcYP4qbjX9wK2nMGlBIydrx75E0gLSKRKbtGNTA+GncWnIWw+xpnRTiKeU4W6I4CcB I1dxh9TmBvuOV8/prqFS/aly6VDzoS6BQxuiOKVgjG+QSk4B9269yPosUGMpH16sFULy PpDQ==
X-Gm-Message-State: AOAM531z3XQ/zInY75ryPnrKlo++t0t+i3e4IwZpGTVmi3ppgLuc25Za SeQlRK0oGfH1aMABhWBaoqVP5879SInF5ASSkKXUu7hxFU4=
X-Google-Smtp-Source: ABdhPJxmMEBsa3IcSLapOnMtkjcFTY3rl5G8qbig5xIMwFH50UigyngZU89HsfFlI2fPpCL1uRpJDZX136ZapQ1/CIQ=
X-Received: by 2002:a25:1246:: with SMTP id 67mr1178612ybs.461.1632765394947;  Mon, 27 Sep 2021 10:56:34 -0700 (PDT)
MIME-Version: 1.0
From: Josh Mandel <jmandel@gmail.com>
Date: Mon, 27 Sep 2021 12:56:18 -0500
Message-ID: <CANSMLKHx4xKrR+c-KzdpxF0==c_=wqPexQJTVNTEXVmU5Usm9g@mail.gmail.com>
To: txauth@ietf.org
Content-Type: multipart/alternative; boundary="0000000000005e1c1e05ccfdd1a6"
Archived-At: <https://mailarchive.ietf.org/arch/msg/txauth/yyEorsXEy7ZR8v2lWx79kjIQ_u0>
Subject: [GNAP] Accounting for Resource Owner in RS::AS interactions
X-BeenThere: txauth@ietf.org
X-Mailman-Version: 2.1.29
Precedence: list
List-Id: GNAP <txauth.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/options/txauth>, <mailto:txauth-request@ietf.org?subject=unsubscribe>
List-Archive: <https://mailarchive.ietf.org/arch/browse/txauth/>
List-Post: <mailto:txauth@ietf.org>
List-Help: <mailto:txauth-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/txauth>, <mailto:txauth-request@ietf.org?subject=subscribe>
X-List-Received-Date: Mon, 27 Sep 2021 17:56:40 -0000

--0000000000005e1c1e05ccfdd1a6
Content-Type: text/plain; charset="UTF-8"

Hi All,

I'm interested in exploring how GNAP can/does/might account for the
Resource Owner associated with specific protected resources, in scenarios
where a single AS is trusted by multiple independent RS's (specifically in
scenarios where these RS's don't know or trust one another). Justin
suggestd I might reach out here.

My sense is that there are security considerations related to these
questions, and perhaps some best practices or even areas where the core
spec's data models could be enhanced to support these scenarios. I'm not
proposing anything in particular at this stage; just trying to get the lay
of the land.

See assumptions and scenarios below.

-Josh

--

*Assumptions. *I'll focus on a very simple API from the RS perspective: a
standardized file access API where resource servers would like to make
access control decisions based on access tokens whose RARs look something
like:

{
  "type": "file-access",
  "actions": ["read"],
  "location": ["{{URL of one protected file}}"]
}

---

*Scenario 1. *Consider the case where a single RS manages files owned by
multiple end users. The RS uses a single AS for all files (but the AS also
serves other untrusted RS's). Each RO is allowed to set authorization
policies for their own files. In this scenario, how does the AS learn which
RO is responsible for setting policies for each file, so it can present
Alice with a list of current policies, allow her to review/tweak access,
etc? (If an AS is confused on this point, there's a risk that Bob can set
policies for Alice's files.) It's possible that the Resource Registration
API could be extended to model this kind of ownership, if both servers have
some interoperable way to communicate user identities. I'm not sure if
there are existing patterns to recommend here? It's also very important for
the RS to know, whenever it receives an access token, that the RS is an
intended audience for that token. The RAR "locations" aren't enough to
accomplish this, unless the AS has a reliable scheme in place to prevent
one RS from crafting RARs that refer to resource locations within another
RS.  And as the spec says, the AS introspection endpoint needs to ensure
that access tokens are "appropriate for presentation at the identified RS".
But that's especially challenging when a single access token can include
multiple RARs, potentially intended for different RS's (which appears to be
allowed by the base spec). And for structured token formats where
introspection isn't required, the same questions hold; is there existing
work on this?

*Scenario 2. *Now consider the case where 10 organizations each manage a
File Sharing RS. Each RS hosts files for multiple ROs, and the ROs get to
choose which AS should be trusted for their own files. In this context, I
suppose the workflow that introduces the RS to the AS can be mediated by
the RO, so that the RO is logged into both systems when making the
introduction, and the RS can essentially register a new client instance
with the AS for each RO who decides to establish a connection. This pattern
has the advantage that there's no confusion about *who* controls a given
resource -- but perhaps there are other, better ways to accomplish this.

--0000000000005e1c1e05ccfdd1a6
Content-Type: text/html; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable

<div dir=3D"ltr"><div dir=3D"auto">Hi All,</div><div dir=3D"auto"><br></div=
><div dir=3D"auto">I&#39;m interested in exploring how GNAP can/does/might =
account for the Resource Owner associated with specific protected resources=
, in scenarios where a single AS is trusted by multiple independent RS&#39;=
s (specifically in scenarios where these RS&#39;s don&#39;t know or trust o=
ne another). Justin suggestd I might reach out here.<br></div><div dir=3D"a=
uto"><br></div><div dir=3D"auto"><div>My sense is that there are security c=
onsiderations related to these questions, and perhaps some best practices o=
r even areas where the core spec&#39;s data models could be enhanced to sup=
port these scenarios. I&#39;m not proposing anything in particular at this =
stage; just trying to get the lay of the land.</div><div><br></div><div>See=
 assumptions and scenarios below.</div><div><br></div><div>-Josh</div><div>=
<br></div><div>--</div></div><div dir=3D"auto"><div dir=3D"auto"><br></div>=
<div dir=3D"auto"><b>Assumptions.=C2=A0</b>I&#39;ll focus on a very simple =
API from the RS perspective: a standardized file access API where resource =
servers would like to make access control decisions based on access tokens =
whose RARs look something like:</div><div dir=3D"auto"><br></div><div><font=
 size=3D"1" face=3D"monospace">{</font></div><div><font size=3D"1" face=3D"=
monospace">=C2=A0 &quot;type&quot;: &quot;file-access&quot;,</font></div><d=
iv><font size=3D"1" face=3D"monospace">=C2=A0 &quot;actions&quot;: [&quot;r=
ead&quot;],</font></div><div><font size=3D"1" face=3D"monospace">=C2=A0 &qu=
ot;location&quot;: [&quot;{{URL of one protected file}}&quot;]</font></div>=
<div><font size=3D"1" face=3D"monospace">}</font></div><div dir=3D"auto"><d=
iv dir=3D"auto"><br></div><div dir=3D"auto">---</div><div dir=3D"auto"><br>=
</div><div dir=3D"auto"><b>Scenario 1. </b>Consider the case where a=C2=A0s=
ingle RS manages files owned by multiple end users. The RS uses a single AS=
 for all files (but the AS also serves other untrusted RS&#39;s). Each RO i=
s allowed to set authorization policies for their own files. In this scenar=
io, how does the AS learn which RO is responsible for setting policies for =
each file,=C2=A0so it can present Alice with a list of current policies,=C2=
=A0allow her to review/tweak access, etc? (If an AS is confused on this poi=
nt, there&#39;s a risk that Bob can set policies for Alice&#39;s files.) It=
&#39;s possible that the Resource Registration API could be extended to mod=
el this kind of ownership, if both servers have some interoperable way to c=
ommunicate user identities. I&#39;m not sure if there are existing patterns=
 to recommend here? It&#39;s also very important for the RS to know, whenev=
er it receives an access token, that the RS is an intended audience for tha=
t token. The RAR &quot;locations&quot; aren&#39;t enough to accomplish this=
, unless the AS has a reliable scheme in place to prevent one RS from craft=
ing RARs that refer to resource locations within another RS.=C2=A0 And as t=
he spec says, the AS introspection endpoint needs to=C2=A0ensure that=C2=A0=
access tokens are &quot;appropriate for presentation at the identified RS&q=
uot;. But=C2=A0that&#39;s especially challenging when a single access token=
 can include multiple RARs, potentially intended for different RS&#39;s (wh=
ich appears to be allowed by the base spec). And for structured token forma=
ts where introspection isn&#39;t required, the same questions hold; is ther=
e existing work on this?<br></div><div dir=3D"auto"><br></div><div><b>Scena=
rio 2.=C2=A0</b>Now consider the case where 10 organizations each manage a =
File Sharing RS. Each RS hosts files for multiple ROs, and the ROs get to c=
hoose which AS should be trusted for their own files. In this context, I su=
ppose the workflow that introduces the RS to the AS can be mediated by the =
RO, so that the RO is logged into both systems when making the introduction=
, and the RS can essentially register a new client instance with the AS for=
 each RO who decides to establish a connection. This pattern has the advant=
age that there&#39;s no confusion about *who* controls a given resource -- =
but perhaps there are other, better ways to accomplish this.</div><div dir=
=3D"auto"><div dir=3D"auto"><br></div><div dir=3D"auto"><br></div></div></d=
iv></div>
</div>

--0000000000005e1c1e05ccfdd1a6--


From nobody Mon Sep 27 11:40:55 2021
Return-Path: <agropper@gmail.com>
X-Original-To: txauth@ietfa.amsl.com
Delivered-To: txauth@ietfa.amsl.com
Received: from localhost (localhost [127.0.0.1]) by ietfa.amsl.com (Postfix) with ESMTP id 295933A1417 for <txauth@ietfa.amsl.com>; Mon, 27 Sep 2021 11:40:53 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -1.648
X-Spam-Level: 
X-Spam-Status: No, score=-1.648 tagged_above=-999 required=5 tests=[BAYES_00=-1.9, FREEMAIL_FORGED_FROMDOMAIN=0.249, FREEMAIL_FROM=0.001, HEADER_FROM_DIFFERENT_DOMAINS=0.001, HTML_MESSAGE=0.001, RCVD_IN_MSPIKE_H2=-0.001, SPF_HELO_NONE=0.001, SPF_PASS=-0.001, URIBL_BLOCKED=0.001] autolearn=no autolearn_force=no
Received: from mail.ietf.org ([4.31.198.44]) by localhost (ietfa.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id Y2OpnV0k-e0a for <txauth@ietfa.amsl.com>; Mon, 27 Sep 2021 11:40:49 -0700 (PDT)
Received: from mail-qt1-f176.google.com (mail-qt1-f176.google.com [209.85.160.176]) (using TLSv1.2 with cipher ECDHE-RSA-AES128-GCM-SHA256 (128/128 bits)) (No client certificate requested) by ietfa.amsl.com (Postfix) with ESMTPS id E55FC3A1409 for <txauth@ietf.org>; Mon, 27 Sep 2021 11:40:48 -0700 (PDT)
Received: by mail-qt1-f176.google.com with SMTP id j13so17570957qtq.6 for <txauth@ietf.org>; Mon, 27 Sep 2021 11:40:48 -0700 (PDT)
X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20210112; h=x-gm-message-state:mime-version:references:in-reply-to:from:date :message-id:subject:to:cc; bh=JLxtM/wvszpZemoaY3h3f8fZ2ExMrRTjgT0TrEsmaUM=; b=3sw0NOXonB/oY8Wu+f6Q9YJAG1QeQ2oEv8Kq8bnm8Y6FK3SncvGbGltcL8KuGaFoZx VoJENEQmY18cSkz4xRFvBar3pZNJybhYsBgJMMHp94pykDnel2zwOklHs5VkfH+tsWzB 116c5E4+rSVxgTEi/xqPRyNAjvQZRyjCENDTg3Is6btdWUYHyI1cVJmXA1reHTmTm6d5 rPc9cfoKI5+malcad6RZx731PCIpLtkIN6Sasto6ufRPPK2oNheFplnA68sEDEhT0H33 F4BoMrJrTENVUVwcrjs/s8KNeedN7090/Z4idUDftVi4GrLEdkafISbDwczGJHHQG0nv qQhw==
X-Gm-Message-State: AOAM532reRvlTCRdaqlVgcUFgtnbLW4Yc39hWnB2LWKVrJwLXGQPf6/C en9rXLUJ0DoIScl4UEOftWifzOE8XMNdDC7BqaI=
X-Google-Smtp-Source: ABdhPJy6wCwYyfVSWzcaZSbtribAomwVmGL07LUpt/VaO8RZZzmA1FpkFsqzNfkFvRU2wJbpy05IWTkKeFZIp+3tDvI=
X-Received: by 2002:ac8:720f:: with SMTP id a15mr1443457qtp.84.1632768047782;  Mon, 27 Sep 2021 11:40:47 -0700 (PDT)
MIME-Version: 1.0
References: <CANSMLKHx4xKrR+c-KzdpxF0==c_=wqPexQJTVNTEXVmU5Usm9g@mail.gmail.com>
In-Reply-To: <CANSMLKHx4xKrR+c-KzdpxF0==c_=wqPexQJTVNTEXVmU5Usm9g@mail.gmail.com>
From: Adrian Gropper <agropper@healthurl.com>
Date: Mon, 27 Sep 2021 14:40:34 -0400
Message-ID: <CANYRo8hB+73Z0jBewrWiFNXHaeEGPoUem1to__rGCmXODGMHPw@mail.gmail.com>
To: Josh Mandel <jmandel@gmail.com>
Cc: GNAP Mailing List <txauth@ietf.org>
Content-Type: multipart/alternative; boundary="0000000000007d1ee905ccfe6f47"
Archived-At: <https://mailarchive.ietf.org/arch/msg/txauth/JyMNLiZJV_v3VMePDVGevzEO_Hg>
Subject: Re: [GNAP] Accounting for Resource Owner in RS::AS interactions
X-BeenThere: txauth@ietf.org
X-Mailman-Version: 2.1.29
Precedence: list
List-Id: GNAP <txauth.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/options/txauth>, <mailto:txauth-request@ietf.org?subject=unsubscribe>
List-Archive: <https://mailarchive.ietf.org/arch/browse/txauth/>
List-Post: <mailto:txauth@ietf.org>
List-Help: <mailto:txauth-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/txauth>, <mailto:txauth-request@ietf.org?subject=subscribe>
X-List-Received-Date: Mon, 27 Sep 2021 18:40:53 -0000

--0000000000007d1ee905ccfe6f47
Content-Type: text/plain; charset="UTF-8"

This was being discussed in this thread
https://mailarchive.ietf.org/arch/msg/txauth/nhV4f59Xf5EaZ8c-YjkWWN6oCr0/

In my opinion, Scenario 2 or RS-issued capabilities that allow the RO to
effectively choose their AS is the only solution.

Scenario 1, even if it could be supported by GNAP, does not deal with the
data minimization privacy issue. As an RO, I don't want to share my
policies with a multi-tenant RS. In UMA, we spent almost a decade trying to
deal with "cascading ASs" and similar dreams of RS-restricted agency for
the convenience of the RS. If GNAP is to be a delegation protocol, then
restricting the RO's delegation options will always be a problem because
such restrictions are a matter of law, not protocol.

In my opinion, GNAP would be much easier to understand if we defined the RO
as the entity that controls policy at the GNAP AS. An AS in the same trust
domain as an RS is just a federation of RSs and we're back to OAuth.

- Adrian

On Mon, Sep 27, 2021 at 1:56 PM Josh Mandel <jmandel@gmail.com> wrote:

> Hi All,
>
> I'm interested in exploring how GNAP can/does/might account for the
> Resource Owner associated with specific protected resources, in scenarios
> where a single AS is trusted by multiple independent RS's (specifically in
> scenarios where these RS's don't know or trust one another). Justin
> suggestd I might reach out here.
>
> My sense is that there are security considerations related to these
> questions, and perhaps some best practices or even areas where the core
> spec's data models could be enhanced to support these scenarios. I'm not
> proposing anything in particular at this stage; just trying to get the lay
> of the land.
>
> See assumptions and scenarios below.
>
> -Josh
>
> --
>
> *Assumptions. *I'll focus on a very simple API from the RS perspective: a
> standardized file access API where resource servers would like to make
> access control decisions based on access tokens whose RARs look something
> like:
>
> {
>   "type": "file-access",
>   "actions": ["read"],
>   "location": ["{{URL of one protected file}}"]
> }
>
> ---
>
> *Scenario 1. *Consider the case where a single RS manages files owned by
> multiple end users. The RS uses a single AS for all files (but the AS also
> serves other untrusted RS's). Each RO is allowed to set authorization
> policies for their own files. In this scenario, how does the AS learn which
> RO is responsible for setting policies for each file, so it can present
> Alice with a list of current policies, allow her to review/tweak access,
> etc? (If an AS is confused on this point, there's a risk that Bob can set
> policies for Alice's files.) It's possible that the Resource Registration
> API could be extended to model this kind of ownership, if both servers have
> some interoperable way to communicate user identities. I'm not sure if
> there are existing patterns to recommend here? It's also very important for
> the RS to know, whenever it receives an access token, that the RS is an
> intended audience for that token. The RAR "locations" aren't enough to
> accomplish this, unless the AS has a reliable scheme in place to prevent
> one RS from crafting RARs that refer to resource locations within another
> RS.  And as the spec says, the AS introspection endpoint needs to ensure
> that access tokens are "appropriate for presentation at the identified RS".
> But that's especially challenging when a single access token can include
> multiple RARs, potentially intended for different RS's (which appears to be
> allowed by the base spec). And for structured token formats where
> introspection isn't required, the same questions hold; is there existing
> work on this?
>
> *Scenario 2. *Now consider the case where 10 organizations each manage a
> File Sharing RS. Each RS hosts files for multiple ROs, and the ROs get to
> choose which AS should be trusted for their own files. In this context, I
> suppose the workflow that introduces the RS to the AS can be mediated by
> the RO, so that the RO is logged into both systems when making the
> introduction, and the RS can essentially register a new client instance
> with the AS for each RO who decides to establish a connection. This pattern
> has the advantage that there's no confusion about *who* controls a given
> resource -- but perhaps there are other, better ways to accomplish this.
>
>
> --
> TXAuth mailing list
> TXAuth@ietf.org
> https://www.ietf.org/mailman/listinfo/txauth
>

--0000000000007d1ee905ccfe6f47
Content-Type: text/html; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable

<div dir=3D"ltr"><div>This was being discussed in this thread <a href=3D"ht=
tps://mailarchive.ietf.org/arch/msg/txauth/nhV4f59Xf5EaZ8c-YjkWWN6oCr0/">ht=
tps://mailarchive.ietf.org/arch/msg/txauth/nhV4f59Xf5EaZ8c-YjkWWN6oCr0/</a>=
</div><div><br></div><div>In my opinion, Scenario 2 or RS-issued capabiliti=
es that allow the RO to effectively choose their AS is the only solution. <=
br></div><div><br></div><div>Scenario 1, even if it could be supported by G=
NAP, does not deal with the data minimization privacy issue. As an RO, I do=
n&#39;t want to share my policies with a multi-tenant RS. In UMA, we spent =
almost a decade trying to deal with &quot;cascading ASs&quot; and similar d=
reams of RS-restricted agency for the convenience of the RS. If GNAP is to =
be a delegation protocol, then restricting the RO&#39;s delegation options =
will always be a problem because such restrictions are a matter of law, not=
 protocol. <br></div><div><br></div><div>In my opinion, GNAP would be much =
easier to understand if we defined the RO as the entity that controls polic=
y at the GNAP AS. An AS in the same trust domain as an RS is just a federat=
ion of RSs and we&#39;re back to OAuth. <br></div><div><br></div><div>- Adr=
ian<br></div></div><br><div class=3D"gmail_quote"><div dir=3D"ltr" class=3D=
"gmail_attr">On Mon, Sep 27, 2021 at 1:56 PM Josh Mandel &lt;<a href=3D"mai=
lto:jmandel@gmail.com">jmandel@gmail.com</a>&gt; wrote:<br></div><blockquot=
e class=3D"gmail_quote" style=3D"margin:0px 0px 0px 0.8ex;border-left:1px s=
olid rgb(204,204,204);padding-left:1ex"><div dir=3D"ltr"><div dir=3D"auto">=
Hi All,</div><div dir=3D"auto"><br></div><div dir=3D"auto">I&#39;m interest=
ed in exploring how GNAP can/does/might account for the Resource Owner asso=
ciated with specific protected resources, in scenarios where a single AS is=
 trusted by multiple independent RS&#39;s (specifically in scenarios where =
these RS&#39;s don&#39;t know or trust one another). Justin suggestd I migh=
t reach out here.<br></div><div dir=3D"auto"><br></div><div dir=3D"auto"><d=
iv>My sense is that there are security considerations related to these ques=
tions, and perhaps some best practices or even areas where the core spec&#3=
9;s data models could be enhanced to support these scenarios. I&#39;m not p=
roposing anything in particular at this stage; just trying to get the lay o=
f the land.</div><div><br></div><div>See assumptions and scenarios below.</=
div><div><br></div><div>-Josh</div><div><br></div><div>--</div></div><div d=
ir=3D"auto"><div dir=3D"auto"><br></div><div dir=3D"auto"><b>Assumptions.=
=C2=A0</b>I&#39;ll focus on a very simple API from the RS perspective: a st=
andardized file access API where resource servers would like to make access=
 control decisions based on access tokens whose RARs look something like:</=
div><div dir=3D"auto"><br></div><div><font size=3D"1" face=3D"monospace">{<=
/font></div><div><font size=3D"1" face=3D"monospace">=C2=A0 &quot;type&quot=
;: &quot;file-access&quot;,</font></div><div><font size=3D"1" face=3D"monos=
pace">=C2=A0 &quot;actions&quot;: [&quot;read&quot;],</font></div><div><fon=
t size=3D"1" face=3D"monospace">=C2=A0 &quot;location&quot;: [&quot;{{URL o=
f one protected file}}&quot;]</font></div><div><font size=3D"1" face=3D"mon=
ospace">}</font></div><div dir=3D"auto"><div dir=3D"auto"><br></div><div di=
r=3D"auto">---</div><div dir=3D"auto"><br></div><div dir=3D"auto"><b>Scenar=
io 1. </b>Consider the case where a=C2=A0single RS manages files owned by m=
ultiple end users. The RS uses a single AS for all files (but the AS also s=
erves other untrusted RS&#39;s). Each RO is allowed to set authorization po=
licies for their own files. In this scenario, how does the AS learn which R=
O is responsible for setting policies for each file,=C2=A0so it can present=
 Alice with a list of current policies,=C2=A0allow her to review/tweak acce=
ss, etc? (If an AS is confused on this point, there&#39;s a risk that Bob c=
an set policies for Alice&#39;s files.) It&#39;s possible that the Resource=
 Registration API could be extended to model this kind of ownership, if bot=
h servers have some interoperable way to communicate user identities. I&#39=
;m not sure if there are existing patterns to recommend here? It&#39;s also=
 very important for the RS to know, whenever it receives an access token, t=
hat the RS is an intended audience for that token. The RAR &quot;locations&=
quot; aren&#39;t enough to accomplish this, unless the AS has a reliable sc=
heme in place to prevent one RS from crafting RARs that refer to resource l=
ocations within another RS.=C2=A0 And as the spec says, the AS introspectio=
n endpoint needs to=C2=A0ensure that=C2=A0access tokens are &quot;appropria=
te for presentation at the identified RS&quot;. But=C2=A0that&#39;s especia=
lly challenging when a single access token can include multiple RARs, poten=
tially intended for different RS&#39;s (which appears to be allowed by the =
base spec). And for structured token formats where introspection isn&#39;t =
required, the same questions hold; is there existing work on this?<br></div=
><div dir=3D"auto"><br></div><div><b>Scenario 2.=C2=A0</b>Now consider the =
case where 10 organizations each manage a File Sharing RS. Each RS hosts fi=
les for multiple ROs, and the ROs get to choose which AS should be trusted =
for their own files. In this context, I suppose the workflow that introduce=
s the RS to the AS can be mediated by the RO, so that the RO is logged into=
 both systems when making the introduction, and the RS can essentially regi=
ster a new client instance with the AS for each RO who decides to establish=
 a connection. This pattern has the advantage that there&#39;s no confusion=
 about *who* controls a given resource -- but perhaps there are other, bett=
er ways to accomplish this.</div><div dir=3D"auto"><div dir=3D"auto"><br></=
div><div dir=3D"auto"><br></div></div></div></div>
</div>
-- <br>
TXAuth mailing list<br>
<a href=3D"mailto:TXAuth@ietf.org" target=3D"_blank">TXAuth@ietf.org</a><br=
>
<a href=3D"https://www.ietf.org/mailman/listinfo/txauth" rel=3D"noreferrer"=
 target=3D"_blank">https://www.ietf.org/mailman/listinfo/txauth</a><br>
</blockquote></div>

--0000000000007d1ee905ccfe6f47--


From nobody Tue Sep 28 11:57:38 2021
Return-Path: <jricher@mit.edu>
X-Original-To: txauth@ietfa.amsl.com
Delivered-To: txauth@ietfa.amsl.com
Received: from localhost (localhost [127.0.0.1]) by ietfa.amsl.com (Postfix) with ESMTP id 449F73A0B08 for <txauth@ietfa.amsl.com>; Tue, 28 Sep 2021 11:57:36 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -1.497
X-Spam-Level: 
X-Spam-Status: No, score=-1.497 tagged_above=-999 required=5 tests=[BAYES_00=-1.9, HTML_MESSAGE=0.001, KHOP_HELO_FCRDNS=0.399, SPF_HELO_NONE=0.001, SPF_NONE=0.001, URIBL_BLOCKED=0.001] autolearn=no autolearn_force=no
Received: from mail.ietf.org ([4.31.198.44]) by localhost (ietfa.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id 27YtHxr6GsjY for <txauth@ietfa.amsl.com>; Tue, 28 Sep 2021 11:57:32 -0700 (PDT)
Received: from outgoing.mit.edu (outgoing-auth-1.mit.edu [18.9.28.11]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by ietfa.amsl.com (Postfix) with ESMTPS id BFD943A0B05 for <txauth@ietf.org>; Tue, 28 Sep 2021 11:57:31 -0700 (PDT)
Received: from [192.168.1.28] (static-71-174-62-56.bstnma.fios.verizon.net [71.174.62.56]) (authenticated bits=0) (User authenticated as jricher@ATHENA.MIT.EDU) by outgoing.mit.edu (8.14.7/8.12.4) with ESMTP id 18SIvSx7000521 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-GCM-SHA384 bits=256 verify=NOT); Tue, 28 Sep 2021 14:57:28 -0400
From: Justin Richer <jricher@mit.edu>
Message-Id: <5AE1FBFF-C475-4B05-BB94-5C97B3B5CA22@mit.edu>
Content-Type: multipart/alternative; boundary="Apple-Mail=_CE480443-F779-464D-848A-A47A809405A8"
Mime-Version: 1.0 (Mac OS X Mail 13.4 \(3608.120.23.2.7\))
Date: Tue, 28 Sep 2021 14:57:28 -0400
In-Reply-To: <CANSMLKHx4xKrR+c-KzdpxF0==c_=wqPexQJTVNTEXVmU5Usm9g@mail.gmail.com>
Cc: txauth@ietf.org
To: Josh Mandel <jmandel@gmail.com>
References: <CANSMLKHx4xKrR+c-KzdpxF0==c_=wqPexQJTVNTEXVmU5Usm9g@mail.gmail.com>
X-Mailer: Apple Mail (2.3608.120.23.2.7)
Archived-At: <https://mailarchive.ietf.org/arch/msg/txauth/uhxH5KFM5lQefTydxrmKZ9glkZ8>
Subject: Re: [GNAP] Accounting for Resource Owner in RS::AS interactions
X-BeenThere: txauth@ietf.org
X-Mailman-Version: 2.1.29
Precedence: list
List-Id: GNAP <txauth.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/options/txauth>, <mailto:txauth-request@ietf.org?subject=unsubscribe>
List-Archive: <https://mailarchive.ietf.org/arch/browse/txauth/>
List-Post: <mailto:txauth@ietf.org>
List-Help: <mailto:txauth-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/txauth>, <mailto:txauth-request@ietf.org?subject=subscribe>
X-List-Received-Date: Tue, 28 Sep 2021 18:57:37 -0000

--Apple-Mail=_CE480443-F779-464D-848A-A47A809405A8
Content-Transfer-Encoding: quoted-printable
Content-Type: text/plain;
	charset=utf-8

Hi Josh, thanks for bringing up this use case. An AS protecting several =
different unrelated RS=E2=80=99s is absolutely in scope, as this is a =
common deployment for OAuth 2 today. If I understand everything you=E2=80=99=
re asking, currently GNAP would allow everything here, but it doesn=E2=80=99=
t specify the details of every part of it. I=E2=80=99m going to reply to =
your scenarios inline below. Apologies in advance for the long email, =
there=E2=80=99s a lot to unpack here!

> On Sep 27, 2021, at 1:56 PM, Josh Mandel <jmandel@gmail.com> wrote:
>=20
> Hi All,
>=20
> I'm interested in exploring how GNAP can/does/might account for the =
Resource Owner associated with specific protected resources, in =
scenarios where a single AS is trusted by multiple independent RS's =
(specifically in scenarios where these RS's don't know or trust one =
another). Justin suggestd I might reach out here.
>=20
> My sense is that there are security considerations related to these =
questions, and perhaps some best practices or even areas where the core =
spec's data models could be enhanced to support these scenarios. I'm not =
proposing anything in particular at this stage; just trying to get the =
lay of the land.
>=20
> See assumptions and scenarios below.
>=20
> -Josh
>=20
> --
>=20
> Assumptions. I'll focus on a very simple API from the RS perspective: =
a standardized file access API where resource servers would like to make =
access control decisions based on access tokens whose RARs look =
something like:
>=20
> {
>   "type": "file-access",
>   "actions": ["read"],
>   "location": ["{{URL of one protected file}}"]
> }
>=20
> ---
>=20
> Scenario 1. Consider the case where a single RS manages files owned by =
multiple end users. The RS uses a single AS for all files (but the AS =
also serves other untrusted RS's). Each RO is allowed to set =
authorization policies for their own files. In this scenario, how does =
the AS learn which RO is responsible for setting policies for each file, =
so it can present Alice with a list of current policies, allow her to =
review/tweak access, etc?

Setting and updating policies are outside the scope of GNAP core, =
including any interfaces for doing so.

The policies are controlled at the AS and enforced by the RS. It=E2=80=99s=
 really up to the individual deployments how and where to make that =
split: an AS could issue tokens that are simply tied to the specific =
data, the result of processing the security context of the delegation. =
Or the AS could send over all the context information to the RS to let =
it make its own decisions entirely, while the AS only has an idea of =
something existing and not who owns or controls it. Or something in =
between. And all of these are functionally the same whether you=E2=80=99re=
 introspecting or using structured tokens, so I=E2=80=99m not going to =
get into the differences here.

Doing this statically would be pretty easy, since the decisions can just =
be set in code and executed. Doing it dynamically feels a lot harder and =
more interesting. There are a lot of pieces that need to be fulfilled to =
even get to this point, and I=E2=80=99ll try to add some thoughts on =
each:=20

- How does the AS know who Alice is? In OAuth, it=E2=80=99s assumed that =
Alice can =E2=80=9Clog in=E2=80=9D to the AS using her account. In GNAP =
we also allow that but further let Alice have other ways of interacting =
with the AS, and doesn=E2=80=99t assume that Alice has an =E2=80=9Caccount=
=E2=80=9D at the AS in a traditional sense. Even so, it=E2=80=99s likely =
in this scenario that the AS will have some set of identifiers that it =
has bound to Alice, which can be associated with that policy.

- How does the RS know about the AS? GNAP does offer some discovery for =
the technical connections =
(https://www.ietf.org/archive/id/draft-ietf-gnap-resource-servers-01.html#=
name-rs-facing-as-discovery), but more important is the question of how =
the RS decides to trust the AS to protect any given resource. In your =
scenario this sounds like a configuration time decision, but it might =
not always be.


> (If an AS is confused on this point, there's a risk that Bob can set =
policies for Alice's files.) It's possible that the Resource =
Registration API could be extended to model this kind of ownership, if =
both servers have some interoperable way to communicate user identities. =
I'm not sure if there are existing patterns to recommend here?

- How does the AS know about what the RS is protecting? GNAP allows the =
RS to dynamically introduce a new resource set to the AS and get back a =
reference for it =
(https://www.ietf.org/archive/id/draft-ietf-gnap-resource-servers-01.html#=
name-registering-a-resource-set =
<https://www.ietf.org/archive/id/draft-ietf-gnap-resource-servers-01.html#=
name-registering-a-resource-set>), and like you, I could see this being =
expanded with an ability to say =E2=80=9Cthis RO / set of ROs controls =
this resource set=E2=80=9D as part of that registration step. As long as =
the identifiers aligned, it would let the AS do some enforcement at =
token issuance time. But that really isn=E2=80=99t meant to set =
policies, it=E2=80=99s meant to allow an RS to register a set of rights =
with a simple identifier (like a scope with a random value). Perhaps a =
different API is actually needed for policy management? And with that in =
mind, it would only be needed if we wanted the AS to make sure Bob is =
never prompted for one of Alice=E2=80=99s files at the AS, ever. If you =
don=E2=80=99t want that information to leak to the AS, you can allow Bob =
to get an ineffective token for Alice=E2=80=99s file resource, and count =
on the RS to say =E2=80=9Chey you aren=E2=80=99t Alice, go away=E2=80=9D. =
Again, this goes back to where you draw the enforcement lines, which =
often has more to do with the nature of the API you=E2=80=99re =
protecting than anything.

- How does the RS know how to associate a specific resource with =
=E2=80=9CAlice=E2=80=9D, and not =E2=80=9CBob"? This requires the RS to =
know who =E2=80=9CAlice=E2=80=9D is and know that a given file is hers =
to control. Doing this dynamically feels really difficult

- How do the RS and AS agree that =E2=80=9CAlice=E2=80=9D is the same =
person with the same identifiers? It=E2=80=99s obvious that simply =
trusting a global identifier here won=E2=80=99t solve it on its own =E2=80=
=94 confusion remains trivial, it would just take an RS to apply Bob=E2=80=
=99s global identifier to Alice=E2=80=99s data, or for Bob to get =
assigned Alice=E2=80=99s identifier by a bad AS. But GNAP does have a =
whole notion of =E2=80=9Csubject identifiers=E2=80=9D that are scoped to =
the AS=E2=80=99s trust domain that can be used here: if the RS refers to =
Alice by an identifier known to the AS, and it trusts the AS to enforce =
that binding for Alice, then it will work.


> It's also very important for the RS to know, whenever it receives an =
access token, that the RS is an intended audience for that token. The =
RAR "locations" aren't enough to accomplish this, unless the AS has a =
reliable scheme in place to prevent one RS from crafting RARs that refer =
to resource locations within another RS.  And as the spec says, the AS =
introspection endpoint needs to ensure that access tokens are =
"appropriate for presentation at the identified RS". But that's =
especially challenging when a single access token can include multiple =
RARs, potentially intended for different RS's (which appears to be =
allowed by the base spec). And for structured token formats where =
introspection isn't required, the same questions hold; is there existing =
work on this?

The =E2=80=98locations=E2=80=99 field on its own isn=E2=80=99t enough =
for this, but it can be a place to hold identifiers for each RS. These =
need not be URLs to the APIs in question, though in practice I=E2=80=99d =
expect them to be. It then also begs the question of how the RS =
identifiers came to be, and if they=E2=80=99re then required for this =
type of processing.=20

>=20
> Scenario 2. Now consider the case where 10 organizations each manage a =
File Sharing RS. Each RS hosts files for multiple ROs, and the ROs get =
to choose which AS should be trusted for their own files. In this =
context, I suppose the workflow that introduces the RS to the AS can be =
mediated by the RO, so that the RO is logged into both systems when =
making the introduction, and the RS can essentially register a new =
client instance with the AS for each RO who decides to establish a =
connection. This pattern has the advantage that there's no confusion =
about *who* controls a given resource -- but perhaps there are other, =
better ways to accomplish this.

That was what UMA tried to do with resource set registration, in a way. =
It was awkward from a technology perspective, and it made RS owners =
nervous in a lot of ways to have a new AS introduced. I appreciate the =
ideology, but to this day I don=E2=80=99t believe it=E2=80=99s realistic =
to implement or deploy.

I think a better pattern is to allow introduction of =E2=80=9CAlice=E2=80=9D=
 at the AS level instead of at the RS level. In other words, Alice =
doesn=E2=80=99t bring her own =E2=80=9CAS=E2=80=9D, but she could bring =
her own identity and policy to the AS that protects the RS. That way the =
RS-AS relationship can be more tightly aligned, but the user =
capabilities are greatly expanded. It changes the fit of the components =
from what we might be used to, but I strongly believe there=E2=80=99s a =
lot of promise in this pattern. And in fact, we see a similar pattern in =
OAuth 2 today with API Gateways that can be configured to allow =
federated logins, for example.

Finally, I think there=E2=80=99s an even more interesting scenario: when =
Alice says that Bob should be able to access her stuff. Who does she =
tell and who needs to care? There are a lot of ways to slice this, but =
this jumps to my mind immediately:

 - Alice has a resource A at the RS, which is protected by the AS
 - The AS has a policy that says Alice can authorize client instances to =
access resource A
 - The AS also has an interface for Alice to set additional policies
 - Alice says to the AS that a user identified as Bob can also access A
 - When either Alice or Bob come to the AS, the AS can figure out from =
this policy to create an access token that can access A
 - The RS doesn=E2=80=99t need to know about Alice or Bob at all, only =
whether the token is good for accessing A

 =E2=80=94 Justin



>=20
>=20
> --=20
> TXAuth mailing list
> TXAuth@ietf.org
> https://www.ietf.org/mailman/listinfo/txauth


--Apple-Mail=_CE480443-F779-464D-848A-A47A809405A8
Content-Transfer-Encoding: quoted-printable
Content-Type: text/html;
	charset=utf-8

<html><head><meta http-equiv=3D"Content-Type" content=3D"text/html; =
charset=3Dutf-8"></head><body style=3D"word-wrap: break-word; =
-webkit-nbsp-mode: space; line-break: after-white-space;" class=3D"">Hi =
Josh, thanks for bringing up this use case. An AS protecting several =
different unrelated RS=E2=80=99s is absolutely in scope, as this is a =
common deployment for OAuth 2 today. If I understand everything you=E2=80=99=
re asking, currently GNAP would allow everything here, but it doesn=E2=80=99=
t specify the details of every part of it. I=E2=80=99m going to reply to =
your scenarios inline below. Apologies in advance for the long email, =
there=E2=80=99s a lot to unpack here!<div class=3D""><div><br =
class=3D""><blockquote type=3D"cite" class=3D""><div class=3D"">On Sep =
27, 2021, at 1:56 PM, Josh Mandel &lt;<a href=3D"mailto:jmandel@gmail.com"=
 class=3D"">jmandel@gmail.com</a>&gt; wrote:</div><br =
class=3D"Apple-interchange-newline"><div class=3D""><meta =
http-equiv=3D"Content-Type" content=3D"text/html; charset=3Dutf-8" =
class=3D""><div dir=3D"ltr" class=3D""><div dir=3D"auto" class=3D"">Hi =
All,</div><div dir=3D"auto" class=3D""><br class=3D""></div><div =
dir=3D"auto" class=3D"">I'm interested in exploring how GNAP =
can/does/might account for the Resource Owner associated with specific =
protected resources, in scenarios where a single AS is trusted by =
multiple independent RS's (specifically in scenarios where these RS's =
don't know or trust one another). Justin suggestd I might reach out =
here.<br class=3D""></div><div dir=3D"auto" class=3D""><br =
class=3D""></div><div dir=3D"auto" class=3D""><div class=3D"">My sense =
is that there are security considerations related to these questions, =
and perhaps some best practices or even areas where the core spec's data =
models could be enhanced to support these scenarios. I'm not proposing =
anything in particular at this stage; just trying to get the lay of the =
land.</div><div class=3D""><br class=3D""></div><div class=3D"">See =
assumptions and scenarios below.</div><div class=3D""><br =
class=3D""></div><div class=3D"">-Josh</div><div class=3D""><br =
class=3D""></div><div class=3D"">--</div></div><div dir=3D"auto" =
class=3D""><div dir=3D"auto" class=3D""><br class=3D""></div><div =
dir=3D"auto" class=3D""><b class=3D"">Assumptions.&nbsp;</b>I'll focus =
on a very simple API from the RS perspective: a standardized file access =
API where resource servers would like to make access control decisions =
based on access tokens whose RARs look something like:</div><div =
dir=3D"auto" class=3D""><br class=3D""></div><div class=3D""><font =
size=3D"1" face=3D"monospace" class=3D"">{</font></div><div =
class=3D""><font size=3D"1" face=3D"monospace" class=3D"">&nbsp; "type": =
"file-access",</font></div><div class=3D""><font size=3D"1" =
face=3D"monospace" class=3D"">&nbsp; "actions": =
["read"],</font></div><div class=3D""><font size=3D"1" face=3D"monospace" =
class=3D"">&nbsp; "location": ["{{URL of one protected =
file}}"]</font></div><div class=3D""><font size=3D"1" face=3D"monospace" =
class=3D"">}</font></div><div dir=3D"auto" class=3D""><div dir=3D"auto" =
class=3D""><br class=3D""></div><div dir=3D"auto" class=3D"">---</div><div=
 dir=3D"auto" class=3D""><br class=3D""></div><div dir=3D"auto" =
class=3D""><b class=3D"">Scenario 1. </b>Consider the case where =
a&nbsp;single RS manages files owned by multiple end users. The RS uses =
a single AS for all files (but the AS also serves other untrusted RS's). =
Each RO is allowed to set authorization policies for their own files. In =
this scenario, how does the AS learn which RO is responsible for setting =
policies for each file,&nbsp;so it can present Alice with a list of =
current policies,&nbsp;allow her to review/tweak access, =
etc?</div></div></div></div></div></blockquote><div><br =
class=3D""></div><div>Setting and updating policies are outside the =
scope of GNAP core, including any interfaces for doing so.</div><div><br =
class=3D""></div><div>The policies are controlled at the AS and enforced =
by the RS. It=E2=80=99s really up to the individual deployments how and =
where to make that split: an AS could issue tokens that are simply tied =
to the specific data, the result of processing the security context of =
the delegation. Or the AS could send over all the context information to =
the RS to let it make its own decisions entirely, while the AS only has =
an idea of something existing and not who owns or controls it. Or =
something in between. And all of these are functionally the same whether =
you=E2=80=99re introspecting or using structured tokens, so I=E2=80=99m =
not going to get into the differences here.</div><div><br =
class=3D""></div><div>Doing this statically would be pretty easy, since =
the decisions can just be set in code and executed. Doing it dynamically =
feels a lot harder and more interesting. There are a lot of pieces that =
need to be fulfilled to even get to this point, and I=E2=80=99ll try to =
add some thoughts on each:&nbsp;</div><div><br class=3D""></div><div>- =
How does the AS know who Alice is? In OAuth, it=E2=80=99s assumed that =
Alice can =E2=80=9Clog in=E2=80=9D to the AS using her account. In GNAP =
we also allow that but further let Alice have other ways of interacting =
with the AS, and doesn=E2=80=99t assume that Alice has an =E2=80=9Caccount=
=E2=80=9D at the AS in a traditional sense. Even so, it=E2=80=99s likely =
in this scenario that the AS will have some set of identifiers that it =
has bound to Alice, which can be associated with that =
policy.</div><div><br class=3D""></div><div>- How does the RS know about =
the AS? GNAP does offer some discovery for the technical connections (<a =
href=3D"https://www.ietf.org/archive/id/draft-ietf-gnap-resource-servers-0=
1.html#name-rs-facing-as-discovery" =
class=3D"">https://www.ietf.org/archive/id/draft-ietf-gnap-resource-server=
s-01.html#name-rs-facing-as-discovery</a>), but more important is the =
question of how the RS decides to trust the AS to protect any given =
resource. In your scenario this sounds like a configuration time =
decision, but it might not always be.</div><div><br =
class=3D""></div><div><br class=3D""></div><blockquote type=3D"cite" =
class=3D""><div class=3D""><div dir=3D"ltr" class=3D""><div dir=3D"auto" =
class=3D""><div dir=3D"auto" class=3D""><div dir=3D"auto" class=3D""> =
(If an AS is confused on this point, there's a risk that Bob can set =
policies for Alice's files.) It's possible that the Resource =
Registration API could be extended to model this kind of ownership, if =
both servers have some interoperable way to communicate user identities. =
I'm not sure if there are existing patterns to recommend here? =
</div></div></div></div></div></blockquote><div><br =
class=3D""></div><div><div>- How does the AS know about what the RS is =
protecting? GNAP allows the RS to dynamically introduce a new resource =
set to the AS and get back a reference for it (<a =
href=3D"https://www.ietf.org/archive/id/draft-ietf-gnap-resource-servers-0=
1.html#name-registering-a-resource-set" =
class=3D"">https://www.ietf.org/archive/id/draft-ietf-gnap-resource-server=
s-01.html#name-registering-a-resource-set</a>), and like you, I could =
see this being expanded with an ability to say =E2=80=9Cthis RO / set of =
ROs controls this resource set=E2=80=9D as part of that registration =
step. As long as the identifiers aligned, it would let the AS do some =
enforcement at token issuance time. But that really isn=E2=80=99t meant =
to set policies, it=E2=80=99s meant to allow an RS to register a set of =
rights with a simple identifier (like a scope with a random value). =
Perhaps a different API is actually needed for policy management? And =
with that in mind, it would only be needed if we wanted the AS to make =
sure Bob is never prompted for one of Alice=E2=80=99s files at the AS, =
ever. If you don=E2=80=99t want that information to leak to the AS, you =
can allow Bob to get an ineffective token for Alice=E2=80=99s file =
resource, and count on the RS to say =E2=80=9Chey you aren=E2=80=99t =
Alice, go away=E2=80=9D. Again, this goes back to where you draw the =
enforcement lines, which often has more to do with the nature of the API =
you=E2=80=99re protecting than anything.</div><div class=3D""><br =
class=3D""></div></div><div><div>- How does the RS know how to associate =
a specific resource with =E2=80=9CAlice=E2=80=9D, and not =E2=80=9CBob"? =
This requires the RS to know who =E2=80=9CAlice=E2=80=9D is and know =
that a given file is hers to control. Doing this dynamically feels =
really difficult</div><div><br class=3D""></div><div>- How do the RS and =
AS agree that =E2=80=9CAlice=E2=80=9D is the same person with the same =
identifiers? It=E2=80=99s obvious that simply trusting a global =
identifier here won=E2=80=99t solve it on its own =E2=80=94 confusion =
remains trivial, it would just take an RS to apply Bob=E2=80=99s global =
identifier to Alice=E2=80=99s data, or for Bob to get assigned Alice=E2=80=
=99s identifier by a bad AS. But GNAP does have a whole notion of =
=E2=80=9Csubject identifiers=E2=80=9D that are scoped to the AS=E2=80=99s =
trust domain that can be used here: if the RS refers to Alice by an =
identifier known to the AS, and it trusts the AS to enforce that binding =
for Alice, then it will work.</div><div class=3D""><br =
class=3D""></div></div><br class=3D""><blockquote type=3D"cite" =
class=3D""><div class=3D""><div dir=3D"ltr" class=3D""><div dir=3D"auto" =
class=3D""><div dir=3D"auto" class=3D""><div dir=3D"auto" class=3D"">It's =
also very important for the RS to know, whenever it receives an access =
token, that the RS is an intended audience for that token. The RAR =
"locations" aren't enough to accomplish this, unless the AS has a =
reliable scheme in place to prevent one RS from crafting RARs that refer =
to resource locations within another RS.&nbsp; And as the spec says, the =
AS introspection endpoint needs to&nbsp;ensure that&nbsp;access tokens =
are "appropriate for presentation at the identified RS". But&nbsp;that's =
especially challenging when a single access token can include multiple =
RARs, potentially intended for different RS's (which appears to be =
allowed by the base spec). And for structured token formats where =
introspection isn't required, the same questions hold; is there existing =
work on this?<br =
class=3D""></div></div></div></div></div></blockquote><div><br =
class=3D""></div><div>The =E2=80=98locations=E2=80=99 field on its own =
isn=E2=80=99t enough for this, but it can be a place to hold identifiers =
for each RS. These need not be URLs to the APIs in question, though in =
practice I=E2=80=99d expect them to be. It then also begs the question =
of how the RS identifiers came to be, and if they=E2=80=99re then =
required for this type of processing.&nbsp;</div><br =
class=3D""><blockquote type=3D"cite" class=3D""><div class=3D""><div =
dir=3D"ltr" class=3D""><div dir=3D"auto" class=3D""><div dir=3D"auto" =
class=3D""><div dir=3D"auto" class=3D""><br class=3D""></div><div =
class=3D""><b class=3D"">Scenario 2.&nbsp;</b>Now consider the case =
where 10 organizations each manage a File Sharing RS. Each RS hosts =
files for multiple ROs, and the ROs get to choose which AS should be =
trusted for their own files. In this context, I suppose the workflow =
that introduces the RS to the AS can be mediated by the RO, so that the =
RO is logged into both systems when making the introduction, and the RS =
can essentially register a new client instance with the AS for each RO =
who decides to establish a connection. This pattern has the advantage =
that there's no confusion about *who* controls a given resource -- but =
perhaps there are other, better ways to accomplish =
this.</div></div></div></div></div></blockquote><div><br =
class=3D""></div><div>That was what UMA tried to do with resource set =
registration, in a way. It was awkward from a technology perspective, =
and it made RS owners nervous in a lot of ways to have a new AS =
introduced. I appreciate the ideology, but to this day I don=E2=80=99t =
believe it=E2=80=99s realistic to implement or deploy.</div><div><br =
class=3D""></div><div>I think a better pattern is to allow introduction =
of =E2=80=9CAlice=E2=80=9D at the AS level instead of at the RS level. =
In other words, Alice doesn=E2=80=99t bring her own =E2=80=9CAS=E2=80=9D, =
but she could bring her own identity and policy to the AS that protects =
the RS. That way the RS-AS relationship can be more tightly aligned, but =
the user capabilities are greatly expanded. It changes the fit of the =
components from what we might be used to, but I strongly believe =
there=E2=80=99s a lot of promise in this pattern. And in fact, we see a =
similar pattern in OAuth 2 today with API Gateways that can be =
configured to allow federated logins, for example.</div><div><br =
class=3D""></div><div>Finally, I think there=E2=80=99s an even more =
interesting scenario: when Alice says that Bob should be able to access =
her stuff. Who does she tell and who needs to care? There are a lot of =
ways to slice this, but this jumps to my mind immediately:</div><div><br =
class=3D""></div><div>&nbsp;- Alice has a resource A at the RS, which is =
protected by the AS</div><div>&nbsp;- The AS has a policy that says =
Alice can authorize client instances to access resource =
A</div><div>&nbsp;- The AS also has an interface for Alice to set =
additional policies</div><div>&nbsp;- Alice says to the AS that a user =
identified as Bob can also access A</div><div>&nbsp;- When either Alice =
or Bob come to the AS, the AS can figure out from this policy to create =
an access token that can access A</div><div>&nbsp;- The RS doesn=E2=80=99t=
 need to know about Alice or Bob at all, only whether the token is good =
for accessing A</div><div><br class=3D""></div><div>&nbsp;=E2=80=94 =
Justin</div><div><br class=3D""></div><div><br class=3D""></div><br =
class=3D""><blockquote type=3D"cite" class=3D""><div class=3D""><div =
dir=3D"ltr" class=3D""><div dir=3D"auto" class=3D""><div dir=3D"auto" =
class=3D""><div dir=3D"auto" class=3D""><div dir=3D"auto" class=3D""><br =
class=3D""></div><div dir=3D"auto" class=3D""><br =
class=3D""></div></div></div></div>
</div>
-- <br class=3D"">TXAuth mailing list<br class=3D""><a =
href=3D"mailto:TXAuth@ietf.org" class=3D"">TXAuth@ietf.org</a><br =
class=3D"">https://www.ietf.org/mailman/listinfo/txauth<br =
class=3D""></div></blockquote></div><br class=3D""></div></body></html>=

--Apple-Mail=_CE480443-F779-464D-848A-A47A809405A8--


From nobody Tue Sep 28 12:20:44 2021
Return-Path: <jmandel@gmail.com>
X-Original-To: txauth@ietfa.amsl.com
Delivered-To: txauth@ietfa.amsl.com
Received: from localhost (localhost [127.0.0.1]) by ietfa.amsl.com (Postfix) with ESMTP id C25E13A0C33 for <txauth@ietfa.amsl.com>; Tue, 28 Sep 2021 12:20:42 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -2.097
X-Spam-Level: 
X-Spam-Status: No, score=-2.097 tagged_above=-999 required=5 tests=[BAYES_00=-1.9, DKIM_SIGNED=0.1, DKIM_VALID=-0.1, DKIM_VALID_AU=-0.1, DKIM_VALID_EF=-0.1, FREEMAIL_FROM=0.001, HTML_MESSAGE=0.001, SPF_HELO_NONE=0.001, SPF_PASS=-0.001, URIBL_BLOCKED=0.001] autolearn=ham autolearn_force=no
Authentication-Results: ietfa.amsl.com (amavisd-new); dkim=pass (2048-bit key) header.d=gmail.com
Received: from mail.ietf.org ([4.31.198.44]) by localhost (ietfa.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id nE0qEB0LVZNo for <txauth@ietfa.amsl.com>; Tue, 28 Sep 2021 12:20:38 -0700 (PDT)
Received: from mail-yb1-xb2d.google.com (mail-yb1-xb2d.google.com [IPv6:2607:f8b0:4864:20::b2d]) (using TLSv1.2 with cipher ECDHE-RSA-AES128-GCM-SHA256 (128/128 bits)) (No client certificate requested) by ietfa.amsl.com (Postfix) with ESMTPS id 2182D3A0C40 for <txauth@ietf.org>; Tue, 28 Sep 2021 12:20:38 -0700 (PDT)
Received: by mail-yb1-xb2d.google.com with SMTP id z5so41309ybj.2 for <txauth@ietf.org>; Tue, 28 Sep 2021 12:20:38 -0700 (PDT)
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=20210112;  h=mime-version:references:in-reply-to:from:date:message-id:subject:to :cc; bh=iH4nwCho3BJa9zCbmUeeWRnhG57nySyAkcahW3540kI=; b=BwXh8kLvsq+prC88j/RVm1UKNFwgOA1agdNEGm5R2WrGu3JMqTBVWuiUd7t6MYYTw7 DCgQDRhJgoNKvjjSlxjR/kUb80ZHYhmkqB6SXtBvY5sgDtRgq9tNWi0YSCnfq3EzoNVX ekn0j21c/XcsbJaVHcYkySvZH+YPVT+AYaoDFcAlI6I+a32GPwT/MsW2XzTuDVxcW4oe 2TPTLyWzNejAZsQCIugkw+WnTYhgeV7GBcpS2kSd+rBb4DEj9uKZol+0MzcS2fGW7vVT e92T0LqNfM/RLjRR4VhwJa91bFuYj/7hHRSF+86IuX6wZmGykNDm2wRdU3rbAYx3vm3F xxkQ==
X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20210112; h=x-gm-message-state:mime-version:references:in-reply-to:from:date :message-id:subject:to:cc; bh=iH4nwCho3BJa9zCbmUeeWRnhG57nySyAkcahW3540kI=; b=4reW5ozV/5gl2/lCmtH56d9GcHXNfvGYDU725XClbSZVOJtqrxGYkf0oviI7VrL6+H BZBC6n6YogsQSlcCm8g4eX15IIgPxIWcDwOav9qQmnVcZFC8OJ+2cwn2OosOz55Pq5o/ yTRHEFFy7rZvM/f7PNZlWXf0TJUiiqALBngKcxpVZ0n1KVNITkolG///ofiSz8/g+K4x 6PCqMgqnifcTrVDi1aRzUeNCQwc/gT9js5d/ZWfpSPqm0FCWxE4LnSSvK8GRWK2vfoLD nE++ymeDwhH9LbW4gTH0Q697neyP+CbKlPmtzZMuiWYI8OIcdWyXVINYPZGcj9txFB4G R/dA==
X-Gm-Message-State: AOAM531SOsGh6A2FFmhA5Z70U6jetmnRN1j6w/y2zZT51evsHuUoCvCZ gqgFrTn01Ao1C8Gl2oXNeOGKTn5b3fRvKJ7grh66M16YV8M=
X-Google-Smtp-Source: ABdhPJy4ElkT3oUbz6pgXx9ApZt+TZAomn8gDFwiJl+qHSWaqGdLmWO3znzxnqmmBVHe0yXpA5Ko+/N0iyNnWpLKwfc=
X-Received: by 2002:a25:1545:: with SMTP id 66mr8116211ybv.521.1632856836919;  Tue, 28 Sep 2021 12:20:36 -0700 (PDT)
MIME-Version: 1.0
References: <CANSMLKHx4xKrR+c-KzdpxF0==c_=wqPexQJTVNTEXVmU5Usm9g@mail.gmail.com> <5AE1FBFF-C475-4B05-BB94-5C97B3B5CA22@mit.edu>
In-Reply-To: <5AE1FBFF-C475-4B05-BB94-5C97B3B5CA22@mit.edu>
From: Josh Mandel <jmandel@gmail.com>
Date: Tue, 28 Sep 2021 14:20:19 -0500
Message-ID: <CANSMLKFR=9QH+ZCCVj3ePDwA9TnVrMALan3W=icg+pz0+fV4rA@mail.gmail.com>
To: Justin Richer <jricher@mit.edu>
Cc: txauth@ietf.org
Content-Type: multipart/alternative; boundary="000000000000bbd7d705cd131b6d"
Archived-At: <https://mailarchive.ietf.org/arch/msg/txauth/n9jrO1GTkmkUZIAtJp5S06mxBuo>
Subject: Re: [GNAP] Accounting for Resource Owner in RS::AS interactions
X-BeenThere: txauth@ietf.org
X-Mailman-Version: 2.1.29
Precedence: list
List-Id: GNAP <txauth.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/options/txauth>, <mailto:txauth-request@ietf.org?subject=unsubscribe>
List-Archive: <https://mailarchive.ietf.org/arch/browse/txauth/>
List-Post: <mailto:txauth@ietf.org>
List-Help: <mailto:txauth-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/txauth>, <mailto:txauth-request@ietf.org?subject=subscribe>
X-List-Received-Date: Tue, 28 Sep 2021 19:20:43 -0000

--000000000000bbd7d705cd131b6d
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable

Thanks Justin -- this whole response is extremely helpful. I particularly
appreciate your very clear declaration of scope and intention:

*> That was what UMA tried to do with resource set registration, in a way.
It was awkward from a technology perspective, and it made RS owners nervous
in a lot of ways to have a new AS introduced. I appreciate the ideology,
but to this day I don=E2=80=99t believe it=E2=80=99s realistic to implement=
 or deploy.*

I agree entirely with this assessment, and I agree that some tighter degree
of AS::RS coupling is needed to support real-world system design.
Understanding that this UMA-like level of RS::AS flexibility is *not* the
design intent for GNAP alleviates several of my concerns in one fell swoop.
I should mention that for readers familiar with UMA (or at least... for me)
this was not clear from reading GNAP; I came in assuming GNAP intended to
cover this space.

On Tue, Sep 28, 2021 at 1:57 PM Justin Richer <jricher@mit.edu> wrote:

> Hi Josh, thanks for bringing up this use case. An AS protecting several
> different unrelated RS=E2=80=99s is absolutely in scope, as this is a com=
mon
> deployment for OAuth 2 today. If I understand everything you=E2=80=99re a=
sking,
> currently GNAP would allow everything here, but it doesn=E2=80=99t specif=
y the
> details of every part of it. I=E2=80=99m going to reply to your scenarios=
 inline
> below. Apologies in advance for the long email, there=E2=80=99s a lot to =
unpack
> here!
>
> On Sep 27, 2021, at 1:56 PM, Josh Mandel <jmandel@gmail.com> wrote:
>
> Hi All,
>
> I'm interested in exploring how GNAP can/does/might account for the
> Resource Owner associated with specific protected resources, in scenarios
> where a single AS is trusted by multiple independent RS's (specifically i=
n
> scenarios where these RS's don't know or trust one another). Justin
> suggestd I might reach out here.
>
> My sense is that there are security considerations related to these
> questions, and perhaps some best practices or even areas where the core
> spec's data models could be enhanced to support these scenarios. I'm not
> proposing anything in particular at this stage; just trying to get the la=
y
> of the land.
>
> See assumptions and scenarios below.
>
> -Josh
>
> --
>
> *Assumptions. *I'll focus on a very simple API from the RS perspective: a
> standardized file access API where resource servers would like to make
> access control decisions based on access tokens whose RARs look something
> like:
>
> {
>   "type": "file-access",
>   "actions": ["read"],
>   "location": ["{{URL of one protected file}}"]
> }
>
> ---
>
> *Scenario 1. *Consider the case where a single RS manages files owned by
> multiple end users. The RS uses a single AS for all files (but the AS als=
o
> serves other untrusted RS's). Each RO is allowed to set authorization
> policies for their own files. In this scenario, how does the AS learn whi=
ch
> RO is responsible for setting policies for each file, so it can present
> Alice with a list of current policies, allow her to review/tweak access,
> etc?
>
>
> Setting and updating policies are outside the scope of GNAP core,
> including any interfaces for doing so.
>
> The policies are controlled at the AS and enforced by the RS. It=E2=80=99=
s really
> up to the individual deployments how and where to make that split: an AS
> could issue tokens that are simply tied to the specific data, the result =
of
> processing the security context of the delegation. Or the AS could send
> over all the context information to the RS to let it make its own decisio=
ns
> entirely, while the AS only has an idea of something existing and not who
> owns or controls it. Or something in between. And all of these are
> functionally the same whether you=E2=80=99re introspecting or using struc=
tured
> tokens, so I=E2=80=99m not going to get into the differences here.
>
> Doing this statically would be pretty easy, since the decisions can just
> be set in code and executed. Doing it dynamically feels a lot harder and
> more interesting. There are a lot of pieces that need to be fulfilled to
> even get to this point, and I=E2=80=99ll try to add some thoughts on each=
:
>
> - How does the AS know who Alice is? In OAuth, it=E2=80=99s assumed that =
Alice can
> =E2=80=9Clog in=E2=80=9D to the AS using her account. In GNAP we also all=
ow that but
> further let Alice have other ways of interacting with the AS, and doesn=
=E2=80=99t
> assume that Alice has an =E2=80=9Caccount=E2=80=9D at the AS in a traditi=
onal sense. Even
> so, it=E2=80=99s likely in this scenario that the AS will have some set o=
f
> identifiers that it has bound to Alice, which can be associated with that
> policy.
>
> - How does the RS know about the AS? GNAP does offer some discovery for
> the technical connections (
> https://www.ietf.org/archive/id/draft-ietf-gnap-resource-servers-01.html#=
name-rs-facing-as-discovery),
> but more important is the question of how the RS decides to trust the AS =
to
> protect any given resource. In your scenario this sounds like a
> configuration time decision, but it might not always be.
>
>
> (If an AS is confused on this point, there's a risk that Bob can set
> policies for Alice's files.) It's possible that the Resource Registration
> API could be extended to model this kind of ownership, if both servers ha=
ve
> some interoperable way to communicate user identities. I'm not sure if
> there are existing patterns to recommend here?
>
>
> - How does the AS know about what the RS is protecting? GNAP allows the R=
S
> to dynamically introduce a new resource set to the AS and get back a
> reference for it (
> https://www.ietf.org/archive/id/draft-ietf-gnap-resource-servers-01.html#=
name-registering-a-resource-set),
> and like you, I could see this being expanded with an ability to say =E2=
=80=9Cthis
> RO / set of ROs controls this resource set=E2=80=9D as part of that regis=
tration
> step. As long as the identifiers aligned, it would let the AS do some
> enforcement at token issuance time. But that really isn=E2=80=99t meant t=
o set
> policies, it=E2=80=99s meant to allow an RS to register a set of rights w=
ith a
> simple identifier (like a scope with a random value). Perhaps a different
> API is actually needed for policy management? And with that in mind, it
> would only be needed if we wanted the AS to make sure Bob is never prompt=
ed
> for one of Alice=E2=80=99s files at the AS, ever. If you don=E2=80=99t wa=
nt that
> information to leak to the AS, you can allow Bob to get an ineffective
> token for Alice=E2=80=99s file resource, and count on the RS to say =E2=
=80=9Chey you aren=E2=80=99t
> Alice, go away=E2=80=9D. Again, this goes back to where you draw the enfo=
rcement
> lines, which often has more to do with the nature of the API you=E2=80=99=
re
> protecting than anything.
>
> - How does the RS know how to associate a specific resource with =E2=80=
=9CAlice=E2=80=9D,
> and not =E2=80=9CBob"? This requires the RS to know who =E2=80=9CAlice=E2=
=80=9D is and know that a
> given file is hers to control. Doing this dynamically feels really diffic=
ult
>
> - How do the RS and AS agree that =E2=80=9CAlice=E2=80=9D is the same per=
son with the same
> identifiers? It=E2=80=99s obvious that simply trusting a global identifie=
r here
> won=E2=80=99t solve it on its own =E2=80=94 confusion remains trivial, it=
 would just take
> an RS to apply Bob=E2=80=99s global identifier to Alice=E2=80=99s data, o=
r for Bob to get
> assigned Alice=E2=80=99s identifier by a bad AS. But GNAP does have a who=
le notion
> of =E2=80=9Csubject identifiers=E2=80=9D that are scoped to the AS=E2=80=
=99s trust domain that can
> be used here: if the RS refers to Alice by an identifier known to the AS,
> and it trusts the AS to enforce that binding for Alice, then it will work=
.
>
>
> It's also very important for the RS to know, whenever it receives an
> access token, that the RS is an intended audience for that token. The RAR
> "locations" aren't enough to accomplish this, unless the AS has a reliabl=
e
> scheme in place to prevent one RS from crafting RARs that refer to resour=
ce
> locations within another RS.  And as the spec says, the AS introspection
> endpoint needs to ensure that access tokens are "appropriate for
> presentation at the identified RS". But that's especially challenging whe=
n
> a single access token can include multiple RARs, potentially intended for
> different RS's (which appears to be allowed by the base spec). And for
> structured token formats where introspection isn't required, the same
> questions hold; is there existing work on this?
>
>
> The =E2=80=98locations=E2=80=99 field on its own isn=E2=80=99t enough for=
 this, but it can be a
> place to hold identifiers for each RS. These need not be URLs to the APIs
> in question, though in practice I=E2=80=99d expect them to be. It then al=
so begs
> the question of how the RS identifiers came to be, and if they=E2=80=99re=
 then
> required for this type of processing.
>
>
> *Scenario 2. *Now consider the case where 10 organizations each manage a
> File Sharing RS. Each RS hosts files for multiple ROs, and the ROs get to
> choose which AS should be trusted for their own files. In this context, I
> suppose the workflow that introduces the RS to the AS can be mediated by
> the RO, so that the RO is logged into both systems when making the
> introduction, and the RS can essentially register a new client instance
> with the AS for each RO who decides to establish a connection. This patte=
rn
> has the advantage that there's no confusion about *who* controls a given
> resource -- but perhaps there are other, better ways to accomplish this.
>
>
> That was what UMA tried to do with resource set registration, in a way. I=
t
> was awkward from a technology perspective, and it made RS owners nervous =
in
> a lot of ways to have a new AS introduced. I appreciate the ideology, but
> to this day I don=E2=80=99t believe it=E2=80=99s realistic to implement o=
r deploy.
>
> I think a better pattern is to allow introduction of =E2=80=9CAlice=E2=80=
=9D at the AS
> level instead of at the RS level. In other words, Alice doesn=E2=80=99t b=
ring her
> own =E2=80=9CAS=E2=80=9D, but she could bring her own identity and policy=
 to the AS that
> protects the RS. That way the RS-AS relationship can be more tightly
> aligned, but the user capabilities are greatly expanded. It changes the f=
it
> of the components from what we might be used to, but I strongly believe
> there=E2=80=99s a lot of promise in this pattern. And in fact, we see a s=
imilar
> pattern in OAuth 2 today with API Gateways that can be configured to allo=
w
> federated logins, for example.
>
> Finally, I think there=E2=80=99s an even more interesting scenario: when =
Alice
> says that Bob should be able to access her stuff. Who does she tell and w=
ho
> needs to care? There are a lot of ways to slice this, but this jumps to m=
y
> mind immediately:
>
>  - Alice has a resource A at the RS, which is protected by the AS
>  - The AS has a policy that says Alice can authorize client instances to
> access resource A
>  - The AS also has an interface for Alice to set additional policies
>  - Alice says to the AS that a user identified as Bob can also access A
>  - When either Alice or Bob come to the AS, the AS can figure out from
> this policy to create an access token that can access A
>  - The RS doesn=E2=80=99t need to know about Alice or Bob at all, only wh=
ether the
> token is good for accessing A
>
>  =E2=80=94 Justin
>
>
>
>
>
> --
> TXAuth mailing list
> TXAuth@ietf.org
> https://www.ietf.org/mailman/listinfo/txauth
>
>
>

--000000000000bbd7d705cd131b6d
Content-Type: text/html; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable

<div dir=3D"ltr">Thanks Justin -- this whole response is extremely helpful.=
 I particularly appreciate=C2=A0your very clear declaration of scope and in=
tention:<div><br></div><div><i>&gt; That was what UMA tried to do with reso=
urce set registration, in a way. It was awkward from a technology perspecti=
ve, and it made RS owners nervous in a lot of ways to have a new AS introdu=
ced. I appreciate the ideology, but to this day I don=E2=80=99t believe it=
=E2=80=99s realistic to implement or deploy.</i></div><div><br></div><div>I=
 agree entirely with this assessment, and I agree that some tighter degree =
of AS::RS coupling is needed to support real-world system design. Understan=
ding that this UMA-like level of RS::AS flexibility is *not* the design int=
ent for GNAP alleviates several of my concerns in one fell swoop. I should =
mention that for readers familiar with UMA (or at least... for me) this was=
 not clear from reading GNAP; I came in assuming GNAP intended to cover thi=
s space.</div></div><br><div class=3D"gmail_quote"><div dir=3D"ltr" class=
=3D"gmail_attr">On Tue, Sep 28, 2021 at 1:57 PM Justin Richer &lt;<a href=
=3D"mailto:jricher@mit.edu">jricher@mit.edu</a>&gt; wrote:<br></div><blockq=
uote class=3D"gmail_quote" style=3D"margin:0px 0px 0px 0.8ex;border-left:1p=
x solid rgb(204,204,204);padding-left:1ex"><div style=3D"overflow-wrap: bre=
ak-word;">Hi Josh, thanks for bringing up this use case. An AS protecting s=
everal different unrelated RS=E2=80=99s is absolutely in scope, as this is =
a common deployment for OAuth 2 today. If I understand everything you=E2=80=
=99re asking, currently GNAP would allow everything here, but it doesn=E2=
=80=99t specify the details of every part of it. I=E2=80=99m going to reply=
 to your scenarios inline below. Apologies in advance for the long email, t=
here=E2=80=99s a lot to unpack here!<div><div><br><blockquote type=3D"cite"=
><div>On Sep 27, 2021, at 1:56 PM, Josh Mandel &lt;<a href=3D"mailto:jmande=
l@gmail.com" target=3D"_blank">jmandel@gmail.com</a>&gt; wrote:</div><br><d=
iv><div dir=3D"ltr"><div dir=3D"auto">Hi All,</div><div dir=3D"auto"><br></=
div><div dir=3D"auto">I&#39;m interested in exploring how GNAP can/does/mig=
ht account for the Resource Owner associated with specific protected resour=
ces, in scenarios where a single AS is trusted by multiple independent RS&#=
39;s (specifically in scenarios where these RS&#39;s don&#39;t know or trus=
t one another). Justin suggestd I might reach out here.<br></div><div dir=
=3D"auto"><br></div><div dir=3D"auto"><div>My sense is that there are secur=
ity considerations related to these questions, and perhaps some best practi=
ces or even areas where the core spec&#39;s data models could be enhanced t=
o support these scenarios. I&#39;m not proposing anything in particular at =
this stage; just trying to get the lay of the land.</div><div><br></div><di=
v>See assumptions and scenarios below.</div><div><br></div><div>-Josh</div>=
<div><br></div><div>--</div></div><div dir=3D"auto"><div dir=3D"auto"><br><=
/div><div dir=3D"auto"><b>Assumptions.=C2=A0</b>I&#39;ll focus on a very si=
mple API from the RS perspective: a standardized file access API where reso=
urce servers would like to make access control decisions based on access to=
kens whose RARs look something like:</div><div dir=3D"auto"><br></div><div>=
<font size=3D"1" face=3D"monospace">{</font></div><div><font size=3D"1" fac=
e=3D"monospace">=C2=A0 &quot;type&quot;: &quot;file-access&quot;,</font></d=
iv><div><font size=3D"1" face=3D"monospace">=C2=A0 &quot;actions&quot;: [&q=
uot;read&quot;],</font></div><div><font size=3D"1" face=3D"monospace">=C2=
=A0 &quot;location&quot;: [&quot;{{URL of one protected file}}&quot;]</font=
></div><div><font size=3D"1" face=3D"monospace">}</font></div><div dir=3D"a=
uto"><div dir=3D"auto"><br></div><div dir=3D"auto">---</div><div dir=3D"aut=
o"><br></div><div dir=3D"auto"><b>Scenario 1. </b>Consider the case where a=
=C2=A0single RS manages files owned by multiple end users. The RS uses a si=
ngle AS for all files (but the AS also serves other untrusted RS&#39;s). Ea=
ch RO is allowed to set authorization policies for their own files. In this=
 scenario, how does the AS learn which RO is responsible for setting polici=
es for each file,=C2=A0so it can present Alice with a list of current polic=
ies,=C2=A0allow her to review/tweak access, etc?</div></div></div></div></d=
iv></blockquote><div><br></div><div>Setting and updating policies are outsi=
de the scope of GNAP core, including any interfaces for doing so.</div><div=
><br></div><div>The policies are controlled at the AS and enforced by the R=
S. It=E2=80=99s really up to the individual deployments how and where to ma=
ke that split: an AS could issue tokens that are simply tied to the specifi=
c data, the result of processing the security context of the delegation. Or=
 the AS could send over all the context information to the RS to let it mak=
e its own decisions entirely, while the AS only has an idea of something ex=
isting and not who owns or controls it. Or something in between. And all of=
 these are functionally the same whether you=E2=80=99re introspecting or us=
ing structured tokens, so I=E2=80=99m not going to get into the differences=
 here.</div><div><br></div><div>Doing this statically would be pretty easy,=
 since the decisions can just be set in code and executed. Doing it dynamic=
ally feels a lot harder and more interesting. There are a lot of pieces tha=
t need to be fulfilled to even get to this point, and I=E2=80=99ll try to a=
dd some thoughts on each:=C2=A0</div><div><br></div><div>- How does the AS =
know who Alice is? In OAuth, it=E2=80=99s assumed that Alice can =E2=80=9Cl=
og in=E2=80=9D to the AS using her account. In GNAP we also allow that but =
further let Alice have other ways of interacting with the AS, and doesn=E2=
=80=99t assume that Alice has an =E2=80=9Caccount=E2=80=9D at the AS in a t=
raditional sense. Even so, it=E2=80=99s likely in this scenario that the AS=
 will have some set of identifiers that it has bound to Alice, which can be=
 associated with that policy.</div><div><br></div><div>- How does the RS kn=
ow about the AS? GNAP does offer some discovery for the technical connectio=
ns (<a href=3D"https://www.ietf.org/archive/id/draft-ietf-gnap-resource-ser=
vers-01.html#name-rs-facing-as-discovery" target=3D"_blank">https://www.iet=
f.org/archive/id/draft-ietf-gnap-resource-servers-01.html#name-rs-facing-as=
-discovery</a>), but more important is the question of how the RS decides t=
o trust the AS to protect any given resource. In your scenario this sounds =
like a configuration time decision, but it might not always be.</div><div><=
br></div><div><br></div><blockquote type=3D"cite"><div><div dir=3D"ltr"><di=
v dir=3D"auto"><div dir=3D"auto"><div dir=3D"auto"> (If an AS is confused o=
n this point, there&#39;s a risk that Bob can set policies for Alice&#39;s =
files.) It&#39;s possible that the Resource Registration API could be exten=
ded to model this kind of ownership, if both servers have some interoperabl=
e way to communicate user identities. I&#39;m not sure if there are existin=
g patterns to recommend here? </div></div></div></div></div></blockquote><d=
iv><br></div><div><div>- How does the AS know about what the RS is protecti=
ng? GNAP allows the RS to dynamically introduce a new resource set to the A=
S and get back a reference for it (<a href=3D"https://www.ietf.org/archive/=
id/draft-ietf-gnap-resource-servers-01.html#name-registering-a-resource-set=
" target=3D"_blank">https://www.ietf.org/archive/id/draft-ietf-gnap-resourc=
e-servers-01.html#name-registering-a-resource-set</a>), and like you, I cou=
ld see this being expanded with an ability to say =E2=80=9Cthis RO / set of=
 ROs controls this resource set=E2=80=9D as part of that registration step.=
 As long as the identifiers aligned, it would let the AS do some enforcemen=
t at token issuance time. But that really isn=E2=80=99t meant to set polici=
es, it=E2=80=99s meant to allow an RS to register a set of rights with a si=
mple identifier (like a scope with a random value). Perhaps a different API=
 is actually needed for policy management? And with that in mind, it would =
only be needed if we wanted the AS to make sure Bob is never prompted for o=
ne of Alice=E2=80=99s files at the AS, ever. If you don=E2=80=99t want that=
 information to leak to the AS, you can allow Bob to get an ineffective tok=
en for Alice=E2=80=99s file resource, and count on the RS to say =E2=80=9Ch=
ey you aren=E2=80=99t Alice, go away=E2=80=9D. Again, this goes back to whe=
re you draw the enforcement lines, which often has more to do with the natu=
re of the API you=E2=80=99re protecting than anything.</div><div><br></div>=
</div><div><div>- How does the RS know how to associate a specific resource=
 with =E2=80=9CAlice=E2=80=9D, and not =E2=80=9CBob&quot;? This requires th=
e RS to know who =E2=80=9CAlice=E2=80=9D is and know that a given file is h=
ers to control. Doing this dynamically feels really difficult</div><div><br=
></div><div>- How do the RS and AS agree that =E2=80=9CAlice=E2=80=9D is th=
e same person with the same identifiers? It=E2=80=99s obvious that simply t=
rusting a global identifier here won=E2=80=99t solve it on its own =E2=80=
=94 confusion remains trivial, it would just take an RS to apply Bob=E2=80=
=99s global identifier to Alice=E2=80=99s data, or for Bob to get assigned =
Alice=E2=80=99s identifier by a bad AS. But GNAP does have a whole notion o=
f =E2=80=9Csubject identifiers=E2=80=9D that are scoped to the AS=E2=80=99s=
 trust domain that can be used here: if the RS refers to Alice by an identi=
fier known to the AS, and it trusts the AS to enforce that binding for Alic=
e, then it will work.</div><div><br></div></div><br><blockquote type=3D"cit=
e"><div><div dir=3D"ltr"><div dir=3D"auto"><div dir=3D"auto"><div dir=3D"au=
to">It&#39;s also very important for the RS to know, whenever it receives a=
n access token, that the RS is an intended audience for that token. The RAR=
 &quot;locations&quot; aren&#39;t enough to accomplish this, unless the AS =
has a reliable scheme in place to prevent one RS from crafting RARs that re=
fer to resource locations within another RS.=C2=A0 And as the spec says, th=
e AS introspection endpoint needs to=C2=A0ensure that=C2=A0access tokens ar=
e &quot;appropriate for presentation at the identified RS&quot;. But=C2=A0t=
hat&#39;s especially challenging when a single access token can include mul=
tiple RARs, potentially intended for different RS&#39;s (which appears to b=
e allowed by the base spec). And for structured token formats where introsp=
ection isn&#39;t required, the same questions hold; is there existing work =
on this?<br></div></div></div></div></div></blockquote><div><br></div><div>=
The =E2=80=98locations=E2=80=99 field on its own isn=E2=80=99t enough for t=
his, but it can be a place to hold identifiers for each RS. These need not =
be URLs to the APIs in question, though in practice I=E2=80=99d expect them=
 to be. It then also begs the question of how the RS identifiers came to be=
, and if they=E2=80=99re then required for this type of processing.=C2=A0</=
div><br><blockquote type=3D"cite"><div><div dir=3D"ltr"><div dir=3D"auto"><=
div dir=3D"auto"><div dir=3D"auto"><br></div><div><b>Scenario 2.=C2=A0</b>N=
ow consider the case where 10 organizations each manage a File Sharing RS. =
Each RS hosts files for multiple ROs, and the ROs get to choose which AS sh=
ould be trusted for their own files. In this context, I suppose the workflo=
w that introduces the RS to the AS can be mediated by the RO, so that the R=
O is logged into both systems when making the introduction, and the RS can =
essentially register a new client instance with the AS for each RO who deci=
des to establish a connection. This pattern has the advantage that there&#3=
9;s no confusion about *who* controls a given resource -- but perhaps there=
 are other, better ways to accomplish this.</div></div></div></div></div></=
blockquote><div><br></div><div>That was what UMA tried to do with resource =
set registration, in a way. It was awkward from a technology perspective, a=
nd it made RS owners nervous in a lot of ways to have a new AS introduced. =
I appreciate the ideology, but to this day I don=E2=80=99t believe it=E2=80=
=99s realistic to implement or deploy.</div><div><br></div><div>I think a b=
etter pattern is to allow introduction of =E2=80=9CAlice=E2=80=9D at the AS=
 level instead of at the RS level. In other words, Alice doesn=E2=80=99t br=
ing her own =E2=80=9CAS=E2=80=9D, but she could bring her own identity and =
policy to the AS that protects the RS. That way the RS-AS relationship can =
be more tightly aligned, but the user capabilities are greatly expanded. It=
 changes the fit of the components from what we might be used to, but I str=
ongly believe there=E2=80=99s a lot of promise in this pattern. And in fact=
, we see a similar pattern in OAuth 2 today with API Gateways that can be c=
onfigured to allow federated logins, for example.</div><div><br></div><div>=
Finally, I think there=E2=80=99s an even more interesting scenario: when Al=
ice says that Bob should be able to access her stuff. Who does she tell and=
 who needs to care? There are a lot of ways to slice this, but this jumps t=
o my mind immediately:</div><div><br></div><div>=C2=A0- Alice has a resourc=
e A at the RS, which is protected by the AS</div><div>=C2=A0- The AS has a =
policy that says Alice can authorize client instances to access resource A<=
/div><div>=C2=A0- The AS also has an interface for Alice to set additional =
policies</div><div>=C2=A0- Alice says to the AS that a user identified as B=
ob can also access A</div><div>=C2=A0- When either Alice or Bob come to the=
 AS, the AS can figure out from this policy to create an access token that =
can access A</div><div>=C2=A0- The RS doesn=E2=80=99t need to know about Al=
ice or Bob at all, only whether the token is good for accessing A</div><div=
><br></div><div>=C2=A0=E2=80=94 Justin</div><div><br></div><div><br></div><=
br><blockquote type=3D"cite"><div><div dir=3D"ltr"><div dir=3D"auto"><div d=
ir=3D"auto"><div dir=3D"auto"><div dir=3D"auto"><br></div><div dir=3D"auto"=
><br></div></div></div></div>
</div>
-- <br>TXAuth mailing list<br><a href=3D"mailto:TXAuth@ietf.org" target=3D"=
_blank">TXAuth@ietf.org</a><br><a href=3D"https://www.ietf.org/mailman/list=
info/txauth" target=3D"_blank">https://www.ietf.org/mailman/listinfo/txauth=
</a><br></div></blockquote></div><br></div></div></blockquote></div>

--000000000000bbd7d705cd131b6d--


From nobody Tue Sep 28 12:35:31 2021
Return-Path: <jricher@mit.edu>
X-Original-To: txauth@ietfa.amsl.com
Delivered-To: txauth@ietfa.amsl.com
Received: from localhost (localhost [127.0.0.1]) by ietfa.amsl.com (Postfix) with ESMTP id D6A8C3A0D3B for <txauth@ietfa.amsl.com>; Tue, 28 Sep 2021 12:35:28 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -1.497
X-Spam-Level: 
X-Spam-Status: No, score=-1.497 tagged_above=-999 required=5 tests=[BAYES_00=-1.9, HTML_MESSAGE=0.001, KHOP_HELO_FCRDNS=0.399, SPF_HELO_NONE=0.001, SPF_NONE=0.001, URIBL_BLOCKED=0.001] autolearn=no autolearn_force=no
Received: from mail.ietf.org ([4.31.198.44]) by localhost (ietfa.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id 1fAVpStWRjIj for <txauth@ietfa.amsl.com>; Tue, 28 Sep 2021 12:35:24 -0700 (PDT)
Received: from outgoing.mit.edu (outgoing-auth-1.mit.edu [18.9.28.11]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by ietfa.amsl.com (Postfix) with ESMTPS id DA0A23A0D29 for <txauth@ietf.org>; Tue, 28 Sep 2021 12:35:23 -0700 (PDT)
Received: from [192.168.1.28] (static-71-174-62-56.bstnma.fios.verizon.net [71.174.62.56]) (authenticated bits=0) (User authenticated as jricher@ATHENA.MIT.EDU) by outgoing.mit.edu (8.14.7/8.12.4) with ESMTP id 18SJZIYF017995 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-GCM-SHA384 bits=256 verify=NOT); Tue, 28 Sep 2021 15:35:19 -0400
From: Justin Richer <jricher@mit.edu>
Message-Id: <7D238C14-3FEB-4166-91CD-D54086CFEFA1@mit.edu>
Content-Type: multipart/alternative; boundary="Apple-Mail=_E68346AB-D2A0-4279-9633-140650342DFB"
Mime-Version: 1.0 (Mac OS X Mail 13.4 \(3608.120.23.2.7\))
Date: Tue, 28 Sep 2021 15:35:18 -0400
In-Reply-To: <CANSMLKFR=9QH+ZCCVj3ePDwA9TnVrMALan3W=icg+pz0+fV4rA@mail.gmail.com>
Cc: txauth@ietf.org
To: Josh Mandel <jmandel@gmail.com>
References: <CANSMLKHx4xKrR+c-KzdpxF0==c_=wqPexQJTVNTEXVmU5Usm9g@mail.gmail.com> <5AE1FBFF-C475-4B05-BB94-5C97B3B5CA22@mit.edu> <CANSMLKFR=9QH+ZCCVj3ePDwA9TnVrMALan3W=icg+pz0+fV4rA@mail.gmail.com>
X-Mailer: Apple Mail (2.3608.120.23.2.7)
Archived-At: <https://mailarchive.ietf.org/arch/msg/txauth/ESSXi_nXsYS662NumqCXJhPpoWw>
Subject: Re: [GNAP] Accounting for Resource Owner in RS::AS interactions
X-BeenThere: txauth@ietf.org
X-Mailman-Version: 2.1.29
Precedence: list
List-Id: GNAP <txauth.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/options/txauth>, <mailto:txauth-request@ietf.org?subject=unsubscribe>
List-Archive: <https://mailarchive.ietf.org/arch/browse/txauth/>
List-Post: <mailto:txauth@ietf.org>
List-Help: <mailto:txauth-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/txauth>, <mailto:txauth-request@ietf.org?subject=subscribe>
X-List-Received-Date: Tue, 28 Sep 2021 19:35:29 -0000

--Apple-Mail=_E68346AB-D2A0-4279-9633-140650342DFB
Content-Transfer-Encoding: quoted-printable
Content-Type: text/plain;
	charset=utf-8

On Sep 28, 2021, at 3:20 PM, Josh Mandel <jmandel@gmail.com> wrote:
>=20
> Thanks Justin -- this whole response is extremely helpful. I =
particularly appreciate your very clear declaration of scope and =
intention:
>=20
> > That was what UMA tried to do with resource set registration, in a =
way. It was awkward from a technology perspective, and it made RS owners =
nervous in a lot of ways to have a new AS introduced. I appreciate the =
ideology, but to this day I don=E2=80=99t believe it=E2=80=99s realistic =
to implement or deploy.
>=20
> I agree entirely with this assessment, and I agree that some tighter =
degree of AS::RS coupling is needed to support real-world system design. =
Understanding that this UMA-like level of RS::AS flexibility is *not* =
the design intent for GNAP alleviates several of my concerns in one fell =
swoop.

To be clear, this is my own take on the goals of GNAP vs. experience =
with UMA, and the stance I=E2=80=99ve been advocating for here for some =
time. I think GNAP could allow that, in that GNAP doesn=E2=80=99t =
:require: a tight coupling per se, but it also doesn=E2=80=99t assume a =
completely loose coupling or force the protocol through those hoops for =
all other use cases. UMA tried to do the latter, but just about every =
UMA deployment I saw didn=E2=80=99t make use of that loose coupling, but =
you still had to go through the bits of the protocol that were built =
just in case that happened.=20

It=E2=80=99s a goal of GNAP to learn from the good and bad of previous =
protocols, including UMA, and this is one area that I think we can =
improve.

> I should mention that for readers familiar with UMA (or at least... =
for me) this was not clear from reading GNAP; I came in assuming GNAP =
intended to cover this space.
>=20

I think what might be confusing here is that GNAP does intend to cover =
the use case space of having a user delegate to another user, and =
externalizing a lot of the user rights aspects of the ore delegation =
process too. But instead of doing it through =E2=80=9Ceveryone brings =
their own AS=E2=80=9D type of things, GNAP is trying to do it through =
=E2=80=9Cthe AS can now be flexible enough to get a lot of different =
information=E2=80=9D. There were some slides on this a while back:

=
https://datatracker.ietf.org/meeting/110/materials/slides-110-gnap-core-pr=
otocol-slides-00 =
<https://datatracker.ietf.org/meeting/110/materials/slides-110-gnap-core-p=
rotocol-slides-00>

Starting at slide 35 you=E2=80=99ll see diagrams and points of some of =
the alternatives. I know the meeting is recorded but I=E2=80=9Dm not =
sure where that lives if you want to review it.

 - Justin


> On Tue, Sep 28, 2021 at 1:57 PM Justin Richer <jricher@mit.edu =
<mailto:jricher@mit.edu>> wrote:
> Hi Josh, thanks for bringing up this use case. An AS protecting =
several different unrelated RS=E2=80=99s is absolutely in scope, as this =
is a common deployment for OAuth 2 today. If I understand everything =
you=E2=80=99re asking, currently GNAP would allow everything here, but =
it doesn=E2=80=99t specify the details of every part of it. I=E2=80=99m =
going to reply to your scenarios inline below. Apologies in advance for =
the long email, there=E2=80=99s a lot to unpack here!
>=20
>> On Sep 27, 2021, at 1:56 PM, Josh Mandel <jmandel@gmail.com =
<mailto:jmandel@gmail.com>> wrote:
>>=20
>> Hi All,
>>=20
>> I'm interested in exploring how GNAP can/does/might account for the =
Resource Owner associated with specific protected resources, in =
scenarios where a single AS is trusted by multiple independent RS's =
(specifically in scenarios where these RS's don't know or trust one =
another). Justin suggestd I might reach out here.
>>=20
>> My sense is that there are security considerations related to these =
questions, and perhaps some best practices or even areas where the core =
spec's data models could be enhanced to support these scenarios. I'm not =
proposing anything in particular at this stage; just trying to get the =
lay of the land.
>>=20
>> See assumptions and scenarios below.
>>=20
>> -Josh
>>=20
>> --
>>=20
>> Assumptions. I'll focus on a very simple API from the RS perspective: =
a standardized file access API where resource servers would like to make =
access control decisions based on access tokens whose RARs look =
something like:
>>=20
>> {
>>   "type": "file-access",
>>   "actions": ["read"],
>>   "location": ["{{URL of one protected file}}"]
>> }
>>=20
>> ---
>>=20
>> Scenario 1. Consider the case where a single RS manages files owned =
by multiple end users. The RS uses a single AS for all files (but the AS =
also serves other untrusted RS's). Each RO is allowed to set =
authorization policies for their own files. In this scenario, how does =
the AS learn which RO is responsible for setting policies for each file, =
so it can present Alice with a list of current policies, allow her to =
review/tweak access, etc?
>=20
> Setting and updating policies are outside the scope of GNAP core, =
including any interfaces for doing so.
>=20
> The policies are controlled at the AS and enforced by the RS. It=E2=80=99=
s really up to the individual deployments how and where to make that =
split: an AS could issue tokens that are simply tied to the specific =
data, the result of processing the security context of the delegation. =
Or the AS could send over all the context information to the RS to let =
it make its own decisions entirely, while the AS only has an idea of =
something existing and not who owns or controls it. Or something in =
between. And all of these are functionally the same whether you=E2=80=99re=
 introspecting or using structured tokens, so I=E2=80=99m not going to =
get into the differences here.
>=20
> Doing this statically would be pretty easy, since the decisions can =
just be set in code and executed. Doing it dynamically feels a lot =
harder and more interesting. There are a lot of pieces that need to be =
fulfilled to even get to this point, and I=E2=80=99ll try to add some =
thoughts on each:=20
>=20
> - How does the AS know who Alice is? In OAuth, it=E2=80=99s assumed =
that Alice can =E2=80=9Clog in=E2=80=9D to the AS using her account. In =
GNAP we also allow that but further let Alice have other ways of =
interacting with the AS, and doesn=E2=80=99t assume that Alice has an =
=E2=80=9Caccount=E2=80=9D at the AS in a traditional sense. Even so, =
it=E2=80=99s likely in this scenario that the AS will have some set of =
identifiers that it has bound to Alice, which can be associated with =
that policy.
>=20
> - How does the RS know about the AS? GNAP does offer some discovery =
for the technical connections =
(https://www.ietf.org/archive/id/draft-ietf-gnap-resource-servers-01.html#=
name-rs-facing-as-discovery =
<https://www.ietf.org/archive/id/draft-ietf-gnap-resource-servers-01.html#=
name-rs-facing-as-discovery>), but more important is the question of how =
the RS decides to trust the AS to protect any given resource. In your =
scenario this sounds like a configuration time decision, but it might =
not always be.
>=20
>=20
>> (If an AS is confused on this point, there's a risk that Bob can set =
policies for Alice's files.) It's possible that the Resource =
Registration API could be extended to model this kind of ownership, if =
both servers have some interoperable way to communicate user identities. =
I'm not sure if there are existing patterns to recommend here?
>=20
> - How does the AS know about what the RS is protecting? GNAP allows =
the RS to dynamically introduce a new resource set to the AS and get =
back a reference for it =
(https://www.ietf.org/archive/id/draft-ietf-gnap-resource-servers-01.html#=
name-registering-a-resource-set =
<https://www.ietf.org/archive/id/draft-ietf-gnap-resource-servers-01.html#=
name-registering-a-resource-set>), and like you, I could see this being =
expanded with an ability to say =E2=80=9Cthis RO / set of ROs controls =
this resource set=E2=80=9D as part of that registration step. As long as =
the identifiers aligned, it would let the AS do some enforcement at =
token issuance time. But that really isn=E2=80=99t meant to set =
policies, it=E2=80=99s meant to allow an RS to register a set of rights =
with a simple identifier (like a scope with a random value). Perhaps a =
different API is actually needed for policy management? And with that in =
mind, it would only be needed if we wanted the AS to make sure Bob is =
never prompted for one of Alice=E2=80=99s files at the AS, ever. If you =
don=E2=80=99t want that information to leak to the AS, you can allow Bob =
to get an ineffective token for Alice=E2=80=99s file resource, and count =
on the RS to say =E2=80=9Chey you aren=E2=80=99t Alice, go away=E2=80=9D. =
Again, this goes back to where you draw the enforcement lines, which =
often has more to do with the nature of the API you=E2=80=99re =
protecting than anything.
>=20
> - How does the RS know how to associate a specific resource with =
=E2=80=9CAlice=E2=80=9D, and not =E2=80=9CBob"? This requires the RS to =
know who =E2=80=9CAlice=E2=80=9D is and know that a given file is hers =
to control. Doing this dynamically feels really difficult
>=20
> - How do the RS and AS agree that =E2=80=9CAlice=E2=80=9D is the same =
person with the same identifiers? It=E2=80=99s obvious that simply =
trusting a global identifier here won=E2=80=99t solve it on its own =E2=80=
=94 confusion remains trivial, it would just take an RS to apply Bob=E2=80=
=99s global identifier to Alice=E2=80=99s data, or for Bob to get =
assigned Alice=E2=80=99s identifier by a bad AS. But GNAP does have a =
whole notion of =E2=80=9Csubject identifiers=E2=80=9D that are scoped to =
the AS=E2=80=99s trust domain that can be used here: if the RS refers to =
Alice by an identifier known to the AS, and it trusts the AS to enforce =
that binding for Alice, then it will work.
>=20
>=20
>> It's also very important for the RS to know, whenever it receives an =
access token, that the RS is an intended audience for that token. The =
RAR "locations" aren't enough to accomplish this, unless the AS has a =
reliable scheme in place to prevent one RS from crafting RARs that refer =
to resource locations within another RS.  And as the spec says, the AS =
introspection endpoint needs to ensure that access tokens are =
"appropriate for presentation at the identified RS". But that's =
especially challenging when a single access token can include multiple =
RARs, potentially intended for different RS's (which appears to be =
allowed by the base spec). And for structured token formats where =
introspection isn't required, the same questions hold; is there existing =
work on this?
>=20
> The =E2=80=98locations=E2=80=99 field on its own isn=E2=80=99t enough =
for this, but it can be a place to hold identifiers for each RS. These =
need not be URLs to the APIs in question, though in practice I=E2=80=99d =
expect them to be. It then also begs the question of how the RS =
identifiers came to be, and if they=E2=80=99re then required for this =
type of processing.=20
>=20
>>=20
>> Scenario 2. Now consider the case where 10 organizations each manage =
a File Sharing RS. Each RS hosts files for multiple ROs, and the ROs get =
to choose which AS should be trusted for their own files. In this =
context, I suppose the workflow that introduces the RS to the AS can be =
mediated by the RO, so that the RO is logged into both systems when =
making the introduction, and the RS can essentially register a new =
client instance with the AS for each RO who decides to establish a =
connection. This pattern has the advantage that there's no confusion =
about *who* controls a given resource -- but perhaps there are other, =
better ways to accomplish this.
>=20
> That was what UMA tried to do with resource set registration, in a =
way. It was awkward from a technology perspective, and it made RS owners =
nervous in a lot of ways to have a new AS introduced. I appreciate the =
ideology, but to this day I don=E2=80=99t believe it=E2=80=99s realistic =
to implement or deploy.
>=20
> I think a better pattern is to allow introduction of =E2=80=9CAlice=E2=80=
=9D at the AS level instead of at the RS level. In other words, Alice =
doesn=E2=80=99t bring her own =E2=80=9CAS=E2=80=9D, but she could bring =
her own identity and policy to the AS that protects the RS. That way the =
RS-AS relationship can be more tightly aligned, but the user =
capabilities are greatly expanded. It changes the fit of the components =
from what we might be used to, but I strongly believe there=E2=80=99s a =
lot of promise in this pattern. And in fact, we see a similar pattern in =
OAuth 2 today with API Gateways that can be configured to allow =
federated logins, for example.
>=20
> Finally, I think there=E2=80=99s an even more interesting scenario: =
when Alice says that Bob should be able to access her stuff. Who does =
she tell and who needs to care? There are a lot of ways to slice this, =
but this jumps to my mind immediately:
>=20
>  - Alice has a resource A at the RS, which is protected by the AS
>  - The AS has a policy that says Alice can authorize client instances =
to access resource A
>  - The AS also has an interface for Alice to set additional policies
>  - Alice says to the AS that a user identified as Bob can also access =
A
>  - When either Alice or Bob come to the AS, the AS can figure out from =
this policy to create an access token that can access A
>  - The RS doesn=E2=80=99t need to know about Alice or Bob at all, only =
whether the token is good for accessing A
>=20
>  =E2=80=94 Justin
>=20
>=20
>=20
>>=20
>>=20
>> --=20
>> TXAuth mailing list
>> TXAuth@ietf.org <mailto:TXAuth@ietf.org>
>> https://www.ietf.org/mailman/listinfo/txauth =
<https://www.ietf.org/mailman/listinfo/txauth>
>=20


--Apple-Mail=_E68346AB-D2A0-4279-9633-140650342DFB
Content-Transfer-Encoding: quoted-printable
Content-Type: text/html;
	charset=utf-8

<html><head><meta http-equiv=3D"Content-Type" content=3D"text/html; =
charset=3Dutf-8"></head><body style=3D"word-wrap: break-word; =
-webkit-nbsp-mode: space; line-break: after-white-space;" class=3D"">On =
Sep 28, 2021, at 3:20 PM, Josh Mandel &lt;<a =
href=3D"mailto:jmandel@gmail.com" class=3D"">jmandel@gmail.com</a>&gt; =
wrote:<br class=3D""><div><blockquote type=3D"cite" class=3D""><br =
class=3D"Apple-interchange-newline"><div class=3D""><meta =
http-equiv=3D"Content-Type" content=3D"text/html; charset=3Dutf-8" =
class=3D""><div dir=3D"ltr" class=3D"">Thanks Justin -- this whole =
response is extremely helpful. I particularly appreciate&nbsp;your very =
clear declaration of scope and intention:<div class=3D""><br =
class=3D""></div><div class=3D""><i class=3D"">&gt; That was what UMA =
tried to do with resource set registration, in a way. It was awkward =
from a technology perspective, and it made RS owners nervous in a lot of =
ways to have a new AS introduced. I appreciate the ideology, but to this =
day I don=E2=80=99t believe it=E2=80=99s realistic to implement or =
deploy.</i></div><div class=3D""><br class=3D""></div><div class=3D"">I =
agree entirely with this assessment, and I agree that some tighter =
degree of AS::RS coupling is needed to support real-world system design. =
Understanding that this UMA-like level of RS::AS flexibility is *not* =
the design intent for GNAP alleviates several of my concerns in one fell =
swoop. </div></div></div></blockquote><div><br =
class=3D""></div><div><div>To be clear, this is my own take on the goals =
of GNAP vs. experience with UMA, and the stance I=E2=80=99ve been =
advocating for here for some time. I think GNAP could allow that, in =
that GNAP doesn=E2=80=99t :require: a tight coupling per se, but it also =
doesn=E2=80=99t assume a completely loose coupling or force the protocol =
through those hoops for all other use cases. UMA tried to do the latter, =
but just about every UMA deployment I saw didn=E2=80=99t make use of =
that loose coupling, but you still had to go through the bits of the =
protocol that were built just in case that happened.&nbsp;</div><div><br =
class=3D""></div><div>It=E2=80=99s a goal of GNAP to learn from the good =
and bad of previous protocols, including UMA, and this is one area that =
I think we can improve.</div></div><br class=3D""><blockquote =
type=3D"cite" class=3D""><div class=3D""><div dir=3D"ltr" class=3D""><div =
class=3D"">I should mention that for readers familiar with UMA (or at =
least... for me) this was not clear from reading GNAP; I came in =
assuming GNAP intended to cover this space.</div></div><br =
class=3D""></div></blockquote><div><br class=3D""></div><div>I think =
what might be confusing here is that GNAP does intend to cover the use =
case space of having a user delegate to another user, and externalizing =
a lot of the user rights aspects of the ore delegation process too. But =
instead of doing it through =E2=80=9Ceveryone brings their own AS=E2=80=9D=
 type of things, GNAP is trying to do it through =E2=80=9Cthe AS can now =
be flexible enough to get a lot of different information=E2=80=9D. There =
were some slides on this a while back:</div><div><br =
class=3D""></div><div><a =
href=3D"https://datatracker.ietf.org/meeting/110/materials/slides-110-gnap=
-core-protocol-slides-00" =
class=3D"">https://datatracker.ietf.org/meeting/110/materials/slides-110-g=
nap-core-protocol-slides-00</a></div><div><br =
class=3D""></div><div>Starting at slide 35 you=E2=80=99ll see diagrams =
and points of some of the alternatives. I know the meeting is recorded =
but I=E2=80=9Dm not sure where that lives if you want to review =
it.</div><div><br class=3D""></div><div>&nbsp;- Justin</div><div><br =
class=3D""></div><br class=3D""><blockquote type=3D"cite" class=3D""><div =
class=3D""><div class=3D"gmail_quote"><div dir=3D"ltr" =
class=3D"gmail_attr">On Tue, Sep 28, 2021 at 1:57 PM Justin Richer =
&lt;<a href=3D"mailto:jricher@mit.edu" class=3D"">jricher@mit.edu</a>&gt; =
wrote:<br class=3D""></div><blockquote class=3D"gmail_quote" =
style=3D"margin:0px 0px 0px 0.8ex;border-left:1px solid =
rgb(204,204,204);padding-left:1ex"><div style=3D"overflow-wrap: =
break-word;" class=3D"">Hi Josh, thanks for bringing up this use case. =
An AS protecting several different unrelated RS=E2=80=99s is absolutely =
in scope, as this is a common deployment for OAuth 2 today. If I =
understand everything you=E2=80=99re asking, currently GNAP would allow =
everything here, but it doesn=E2=80=99t specify the details of every =
part of it. I=E2=80=99m going to reply to your scenarios inline below. =
Apologies in advance for the long email, there=E2=80=99s a lot to unpack =
here!<div class=3D""><div class=3D""><br class=3D""><blockquote =
type=3D"cite" class=3D""><div class=3D"">On Sep 27, 2021, at 1:56 PM, =
Josh Mandel &lt;<a href=3D"mailto:jmandel@gmail.com" target=3D"_blank" =
class=3D"">jmandel@gmail.com</a>&gt; wrote:</div><br class=3D""><div =
class=3D""><div dir=3D"ltr" class=3D""><div dir=3D"auto" class=3D"">Hi =
All,</div><div dir=3D"auto" class=3D""><br class=3D""></div><div =
dir=3D"auto" class=3D"">I'm interested in exploring how GNAP =
can/does/might account for the Resource Owner associated with specific =
protected resources, in scenarios where a single AS is trusted by =
multiple independent RS's (specifically in scenarios where these RS's =
don't know or trust one another). Justin suggestd I might reach out =
here.<br class=3D""></div><div dir=3D"auto" class=3D""><br =
class=3D""></div><div dir=3D"auto" class=3D""><div class=3D"">My sense =
is that there are security considerations related to these questions, =
and perhaps some best practices or even areas where the core spec's data =
models could be enhanced to support these scenarios. I'm not proposing =
anything in particular at this stage; just trying to get the lay of the =
land.</div><div class=3D""><br class=3D""></div><div class=3D"">See =
assumptions and scenarios below.</div><div class=3D""><br =
class=3D""></div><div class=3D"">-Josh</div><div class=3D""><br =
class=3D""></div><div class=3D"">--</div></div><div dir=3D"auto" =
class=3D""><div dir=3D"auto" class=3D""><br class=3D""></div><div =
dir=3D"auto" class=3D""><b class=3D"">Assumptions.&nbsp;</b>I'll focus =
on a very simple API from the RS perspective: a standardized file access =
API where resource servers would like to make access control decisions =
based on access tokens whose RARs look something like:</div><div =
dir=3D"auto" class=3D""><br class=3D""></div><div class=3D""><font =
size=3D"1" face=3D"monospace" class=3D"">{</font></div><div =
class=3D""><font size=3D"1" face=3D"monospace" class=3D"">&nbsp; "type": =
"file-access",</font></div><div class=3D""><font size=3D"1" =
face=3D"monospace" class=3D"">&nbsp; "actions": =
["read"],</font></div><div class=3D""><font size=3D"1" face=3D"monospace" =
class=3D"">&nbsp; "location": ["{{URL of one protected =
file}}"]</font></div><div class=3D""><font size=3D"1" face=3D"monospace" =
class=3D"">}</font></div><div dir=3D"auto" class=3D""><div dir=3D"auto" =
class=3D""><br class=3D""></div><div dir=3D"auto" class=3D"">---</div><div=
 dir=3D"auto" class=3D""><br class=3D""></div><div dir=3D"auto" =
class=3D""><b class=3D"">Scenario 1. </b>Consider the case where =
a&nbsp;single RS manages files owned by multiple end users. The RS uses =
a single AS for all files (but the AS also serves other untrusted RS's). =
Each RO is allowed to set authorization policies for their own files. In =
this scenario, how does the AS learn which RO is responsible for setting =
policies for each file,&nbsp;so it can present Alice with a list of =
current policies,&nbsp;allow her to review/tweak access, =
etc?</div></div></div></div></div></blockquote><div class=3D""><br =
class=3D""></div><div class=3D"">Setting and updating policies are =
outside the scope of GNAP core, including any interfaces for doing =
so.</div><div class=3D""><br class=3D""></div><div class=3D"">The =
policies are controlled at the AS and enforced by the RS. It=E2=80=99s =
really up to the individual deployments how and where to make that =
split: an AS could issue tokens that are simply tied to the specific =
data, the result of processing the security context of the delegation. =
Or the AS could send over all the context information to the RS to let =
it make its own decisions entirely, while the AS only has an idea of =
something existing and not who owns or controls it. Or something in =
between. And all of these are functionally the same whether you=E2=80=99re=
 introspecting or using structured tokens, so I=E2=80=99m not going to =
get into the differences here.</div><div class=3D""><br =
class=3D""></div><div class=3D"">Doing this statically would be pretty =
easy, since the decisions can just be set in code and executed. Doing it =
dynamically feels a lot harder and more interesting. There are a lot of =
pieces that need to be fulfilled to even get to this point, and I=E2=80=99=
ll try to add some thoughts on each:&nbsp;</div><div class=3D""><br =
class=3D""></div><div class=3D"">- How does the AS know who Alice is? In =
OAuth, it=E2=80=99s assumed that Alice can =E2=80=9Clog in=E2=80=9D to =
the AS using her account. In GNAP we also allow that but further let =
Alice have other ways of interacting with the AS, and doesn=E2=80=99t =
assume that Alice has an =E2=80=9Caccount=E2=80=9D at the AS in a =
traditional sense. Even so, it=E2=80=99s likely in this scenario that =
the AS will have some set of identifiers that it has bound to Alice, =
which can be associated with that policy.</div><div class=3D""><br =
class=3D""></div><div class=3D"">- How does the RS know about the AS? =
GNAP does offer some discovery for the technical connections (<a =
href=3D"https://www.ietf.org/archive/id/draft-ietf-gnap-resource-servers-0=
1.html#name-rs-facing-as-discovery" target=3D"_blank" =
class=3D"">https://www.ietf.org/archive/id/draft-ietf-gnap-resource-server=
s-01.html#name-rs-facing-as-discovery</a>), but more important is the =
question of how the RS decides to trust the AS to protect any given =
resource. In your scenario this sounds like a configuration time =
decision, but it might not always be.</div><div class=3D""><br =
class=3D""></div><div class=3D""><br class=3D""></div><blockquote =
type=3D"cite" class=3D""><div class=3D""><div dir=3D"ltr" class=3D""><div =
dir=3D"auto" class=3D""><div dir=3D"auto" class=3D""><div dir=3D"auto" =
class=3D""> (If an AS is confused on this point, there's a risk that Bob =
can set policies for Alice's files.) It's possible that the Resource =
Registration API could be extended to model this kind of ownership, if =
both servers have some interoperable way to communicate user identities. =
I'm not sure if there are existing patterns to recommend here? =
</div></div></div></div></div></blockquote><div class=3D""><br =
class=3D""></div><div class=3D""><div class=3D"">- How does the AS know =
about what the RS is protecting? GNAP allows the RS to dynamically =
introduce a new resource set to the AS and get back a reference for it =
(<a =
href=3D"https://www.ietf.org/archive/id/draft-ietf-gnap-resource-servers-0=
1.html#name-registering-a-resource-set" target=3D"_blank" =
class=3D"">https://www.ietf.org/archive/id/draft-ietf-gnap-resource-server=
s-01.html#name-registering-a-resource-set</a>), and like you, I could =
see this being expanded with an ability to say =E2=80=9Cthis RO / set of =
ROs controls this resource set=E2=80=9D as part of that registration =
step. As long as the identifiers aligned, it would let the AS do some =
enforcement at token issuance time. But that really isn=E2=80=99t meant =
to set policies, it=E2=80=99s meant to allow an RS to register a set of =
rights with a simple identifier (like a scope with a random value). =
Perhaps a different API is actually needed for policy management? And =
with that in mind, it would only be needed if we wanted the AS to make =
sure Bob is never prompted for one of Alice=E2=80=99s files at the AS, =
ever. If you don=E2=80=99t want that information to leak to the AS, you =
can allow Bob to get an ineffective token for Alice=E2=80=99s file =
resource, and count on the RS to say =E2=80=9Chey you aren=E2=80=99t =
Alice, go away=E2=80=9D. Again, this goes back to where you draw the =
enforcement lines, which often has more to do with the nature of the API =
you=E2=80=99re protecting than anything.</div><div class=3D""><br =
class=3D""></div></div><div class=3D""><div class=3D"">- How does the RS =
know how to associate a specific resource with =E2=80=9CAlice=E2=80=9D, =
and not =E2=80=9CBob"? This requires the RS to know who =E2=80=9CAlice=E2=80=
=9D is and know that a given file is hers to control. Doing this =
dynamically feels really difficult</div><div class=3D""><br =
class=3D""></div><div class=3D"">- How do the RS and AS agree that =
=E2=80=9CAlice=E2=80=9D is the same person with the same identifiers? =
It=E2=80=99s obvious that simply trusting a global identifier here =
won=E2=80=99t solve it on its own =E2=80=94 confusion remains trivial, =
it would just take an RS to apply Bob=E2=80=99s global identifier to =
Alice=E2=80=99s data, or for Bob to get assigned Alice=E2=80=99s =
identifier by a bad AS. But GNAP does have a whole notion of =E2=80=9Csubj=
ect identifiers=E2=80=9D that are scoped to the AS=E2=80=99s trust =
domain that can be used here: if the RS refers to Alice by an identifier =
known to the AS, and it trusts the AS to enforce that binding for Alice, =
then it will work.</div><div class=3D""><br class=3D""></div></div><br =
class=3D""><blockquote type=3D"cite" class=3D""><div class=3D""><div =
dir=3D"ltr" class=3D""><div dir=3D"auto" class=3D""><div dir=3D"auto" =
class=3D""><div dir=3D"auto" class=3D"">It's also very important for the =
RS to know, whenever it receives an access token, that the RS is an =
intended audience for that token. The RAR "locations" aren't enough to =
accomplish this, unless the AS has a reliable scheme in place to prevent =
one RS from crafting RARs that refer to resource locations within =
another RS.&nbsp; And as the spec says, the AS introspection endpoint =
needs to&nbsp;ensure that&nbsp;access tokens are "appropriate for =
presentation at the identified RS". But&nbsp;that's especially =
challenging when a single access token can include multiple RARs, =
potentially intended for different RS's (which appears to be allowed by =
the base spec). And for structured token formats where introspection =
isn't required, the same questions hold; is there existing work on =
this?<br class=3D""></div></div></div></div></div></blockquote><div =
class=3D""><br class=3D""></div><div class=3D"">The =E2=80=98locations=E2=80=
=99 field on its own isn=E2=80=99t enough for this, but it can be a =
place to hold identifiers for each RS. These need not be URLs to the =
APIs in question, though in practice I=E2=80=99d expect them to be. It =
then also begs the question of how the RS identifiers came to be, and if =
they=E2=80=99re then required for this type of =
processing.&nbsp;</div><br class=3D""><blockquote type=3D"cite" =
class=3D""><div class=3D""><div dir=3D"ltr" class=3D""><div dir=3D"auto" =
class=3D""><div dir=3D"auto" class=3D""><div dir=3D"auto" class=3D""><br =
class=3D""></div><div class=3D""><b class=3D"">Scenario 2.&nbsp;</b>Now =
consider the case where 10 organizations each manage a File Sharing RS. =
Each RS hosts files for multiple ROs, and the ROs get to choose which AS =
should be trusted for their own files. In this context, I suppose the =
workflow that introduces the RS to the AS can be mediated by the RO, so =
that the RO is logged into both systems when making the introduction, =
and the RS can essentially register a new client instance with the AS =
for each RO who decides to establish a connection. This pattern has the =
advantage that there's no confusion about *who* controls a given =
resource -- but perhaps there are other, better ways to accomplish =
this.</div></div></div></div></div></blockquote><div class=3D""><br =
class=3D""></div><div class=3D"">That was what UMA tried to do with =
resource set registration, in a way. It was awkward from a technology =
perspective, and it made RS owners nervous in a lot of ways to have a =
new AS introduced. I appreciate the ideology, but to this day I don=E2=80=99=
t believe it=E2=80=99s realistic to implement or deploy.</div><div =
class=3D""><br class=3D""></div><div class=3D"">I think a better pattern =
is to allow introduction of =E2=80=9CAlice=E2=80=9D at the AS level =
instead of at the RS level. In other words, Alice doesn=E2=80=99t bring =
her own =E2=80=9CAS=E2=80=9D, but she could bring her own identity and =
policy to the AS that protects the RS. That way the RS-AS relationship =
can be more tightly aligned, but the user capabilities are greatly =
expanded. It changes the fit of the components from what we might be =
used to, but I strongly believe there=E2=80=99s a lot of promise in this =
pattern. And in fact, we see a similar pattern in OAuth 2 today with API =
Gateways that can be configured to allow federated logins, for =
example.</div><div class=3D""><br class=3D""></div><div =
class=3D"">Finally, I think there=E2=80=99s an even more interesting =
scenario: when Alice says that Bob should be able to access her stuff. =
Who does she tell and who needs to care? There are a lot of ways to =
slice this, but this jumps to my mind immediately:</div><div =
class=3D""><br class=3D""></div><div class=3D"">&nbsp;- Alice has a =
resource A at the RS, which is protected by the AS</div><div =
class=3D"">&nbsp;- The AS has a policy that says Alice can authorize =
client instances to access resource A</div><div class=3D"">&nbsp;- The =
AS also has an interface for Alice to set additional policies</div><div =
class=3D"">&nbsp;- Alice says to the AS that a user identified as Bob =
can also access A</div><div class=3D"">&nbsp;- When either Alice or Bob =
come to the AS, the AS can figure out from this policy to create an =
access token that can access A</div><div class=3D"">&nbsp;- The RS =
doesn=E2=80=99t need to know about Alice or Bob at all, only whether the =
token is good for accessing A</div><div class=3D""><br =
class=3D""></div><div class=3D"">&nbsp;=E2=80=94 Justin</div><div =
class=3D""><br class=3D""></div><div class=3D""><br class=3D""></div><br =
class=3D""><blockquote type=3D"cite" class=3D""><div class=3D""><div =
dir=3D"ltr" class=3D""><div dir=3D"auto" class=3D""><div dir=3D"auto" =
class=3D""><div dir=3D"auto" class=3D""><div dir=3D"auto" class=3D""><br =
class=3D""></div><div dir=3D"auto" class=3D""><br =
class=3D""></div></div></div></div>
</div>
-- <br class=3D"">TXAuth mailing list<br class=3D""><a =
href=3D"mailto:TXAuth@ietf.org" target=3D"_blank" =
class=3D"">TXAuth@ietf.org</a><br class=3D""><a =
href=3D"https://www.ietf.org/mailman/listinfo/txauth" target=3D"_blank" =
class=3D"">https://www.ietf.org/mailman/listinfo/txauth</a><br =
class=3D""></div></blockquote></div><br =
class=3D""></div></div></blockquote></div>
</div></blockquote></div><br class=3D""></body></html>=

--Apple-Mail=_E68346AB-D2A0-4279-9633-140650342DFB--


From nobody Tue Sep 28 21:27:51 2021
Return-Path: <agropper@gmail.com>
X-Original-To: txauth@ietfa.amsl.com
Delivered-To: txauth@ietfa.amsl.com
Received: from localhost (localhost [127.0.0.1]) by ietfa.amsl.com (Postfix) with ESMTP id 539473A0CD4 for <txauth@ietfa.amsl.com>; Tue, 28 Sep 2021 21:27:48 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -1.648
X-Spam-Level: 
X-Spam-Status: No, score=-1.648 tagged_above=-999 required=5 tests=[BAYES_00=-1.9, FREEMAIL_FORGED_FROMDOMAIN=0.249, FREEMAIL_FROM=0.001, HEADER_FROM_DIFFERENT_DOMAINS=0.001, HTML_MESSAGE=0.001, RCVD_IN_MSPIKE_H2=-0.001, SPF_HELO_NONE=0.001, SPF_PASS=-0.001, URIBL_BLOCKED=0.001] autolearn=no autolearn_force=no
Received: from mail.ietf.org ([4.31.198.44]) by localhost (ietfa.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id woYjaRnY-gRj for <txauth@ietfa.amsl.com>; Tue, 28 Sep 2021 21:27:44 -0700 (PDT)
Received: from mail-qv1-f47.google.com (mail-qv1-f47.google.com [209.85.219.47]) (using TLSv1.2 with cipher ECDHE-RSA-AES128-GCM-SHA256 (128/128 bits)) (No client certificate requested) by ietfa.amsl.com (Postfix) with ESMTPS id D977A3A11BF for <txauth@ietf.org>; Tue, 28 Sep 2021 21:27:43 -0700 (PDT)
Received: by mail-qv1-f47.google.com with SMTP id n6so742432qvp.7 for <txauth@ietf.org>; Tue, 28 Sep 2021 21:27:43 -0700 (PDT)
X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20210112; h=x-gm-message-state:mime-version:references:in-reply-to:from:date :message-id:subject:to:cc; bh=XIAOub7Kb10P8ufRugWlW1FXKelLRA2KJEi7oSLFkcM=; b=ZkBJSEJLQkbrS2IpdXMdduJjLVeHrFvJdnV4XshDFp72ornl2OJ8xE0lAORKmC9kZE 5pg3ymOfbNXZwfsRXtAFefGk05t2UYxxqMuSxNKxkEIbJ9WbRjryC3HEDA5QxUQ89cr1 4TJmVQsJr2UbRwxFxl7bAmY435kf2O5xUgDULuft7icGxMvgnv6enJ5XBFUJ8SyP1Ipe 6eoR/KwaJlJ1HlvJiCXsNNkQ7Vszj38OpHT+w7jduzXUpiP0qXKR8uQtw79OTwoI8Xe0 GvEQgjZgxl95Y+0GHrRp0mSH8bitzD3CKVo2rM68EFPOUShAQF0ek6AL1o5UnXsWqXso u/uQ==
X-Gm-Message-State: AOAM531QNaTGZnads/6TlcjrVM3JEzNoAXKUDDMsiEB82oHrs2r07fOp mhubLjC3s4A3k0UjP3Ez4JMl9MS2qMo7yPhHfUc=
X-Google-Smtp-Source: ABdhPJzqnomRfuP0A8zNAeDB9avj3bMyLAhkjhmB/wPv1k8R3re6llE6Hg0NOOXUbQR1f1GGelaDQaBS5JefrgmeeKw=
X-Received: by 2002:a0c:d989:: with SMTP id y9mr1149532qvj.67.1632889662664; Tue, 28 Sep 2021 21:27:42 -0700 (PDT)
MIME-Version: 1.0
References: <CANSMLKHx4xKrR+c-KzdpxF0==c_=wqPexQJTVNTEXVmU5Usm9g@mail.gmail.com> <5AE1FBFF-C475-4B05-BB94-5C97B3B5CA22@mit.edu> <CANSMLKFR=9QH+ZCCVj3ePDwA9TnVrMALan3W=icg+pz0+fV4rA@mail.gmail.com> <7D238C14-3FEB-4166-91CD-D54086CFEFA1@mit.edu>
In-Reply-To: <7D238C14-3FEB-4166-91CD-D54086CFEFA1@mit.edu>
From: Adrian Gropper <agropper@healthurl.com>
Date: Wed, 29 Sep 2021 00:27:28 -0400
Message-ID: <CANYRo8iHLBX2Fxq7rZa=N8OeWGqYptUuzZu656B4VgWEtAy9Bw@mail.gmail.com>
To: Justin Richer <jricher@mit.edu>
Cc: Josh Mandel <jmandel@gmail.com>, GNAP Mailing List <txauth@ietf.org>
Content-Type: multipart/alternative; boundary="0000000000004cf4ab05cd1ac076"
Archived-At: <https://mailarchive.ietf.org/arch/msg/txauth/SI4l5TC94eEQ00-lOeZdECmP9rc>
Subject: Re: [GNAP] Accounting for Resource Owner in RS::AS interactions
X-BeenThere: txauth@ietf.org
X-Mailman-Version: 2.1.29
Precedence: list
List-Id: GNAP <txauth.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/options/txauth>, <mailto:txauth-request@ietf.org?subject=unsubscribe>
List-Archive: <https://mailarchive.ietf.org/arch/browse/txauth/>
List-Post: <mailto:txauth@ietf.org>
List-Help: <mailto:txauth-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/txauth>, <mailto:txauth-request@ietf.org?subject=subscribe>
X-List-Received-Date: Wed, 29 Sep 2021 04:27:49 -0000

--0000000000004cf4ab05cd1ac076
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable

Josh and Justin,

>From a privacy perspective, your glee over not having the RO specify the AS
to the RS is troubling. Why would the RS care about privacy unless mandated
in some way? (I'm assuming mandates are out of scope for GNAP :-).

The privacy problem is simply data minimization. As an RO, I should not
have to share my policies with the RS because all the RS needs is the
access token. If the RS is unwilling (or whatever you mean by "nervous",
above) to trust a random AS simply because the RO trusts it, then we're
left with capabilities. Imagine a typical RS serving millions of ROs. The
policy database would be a gold mine in so many ways. Traffic analysis by
the RO might still be a problem and will also need to be considered.

If specifying the AS is not an option, the RO associated with a resource
MUST be given a full set of capabilities on demand. The RO can then pass
some or all of those to whatever AS they trust with their policies and then
that AS can issue attenuated access tokens with or without further
delegation rights to any client they choose.

GNAP core must choose either or both arbitrary AS or capabilities and maybe
explain "nervous" in a way that people can understand.

- Adrian


On Tue, Sep 28, 2021 at 3:35 PM Justin Richer <jricher@mit.edu> wrote:

> On Sep 28, 2021, at 3:20 PM, Josh Mandel <jmandel@gmail.com> wrote:
>
>
> Thanks Justin -- this whole response is extremely helpful. I particularly
> appreciate your very clear declaration of scope and intention:
>
> *> That was what UMA tried to do with resource set registration, in a way=
.
> It was awkward from a technology perspective, and it made RS owners nervo=
us
> in a lot of ways to have a new AS introduced. I appreciate the ideology,
> but to this day I don=E2=80=99t believe it=E2=80=99s realistic to impleme=
nt or deploy.*
>
> I agree entirely with this assessment, and I agree that some tighter
> degree of AS::RS coupling is needed to support real-world system design.
> Understanding that this UMA-like level of RS::AS flexibility is *not* the
> design intent for GNAP alleviates several of my concerns in one fell swoo=
p.
>
>
> To be clear, this is my own take on the goals of GNAP vs. experience with
> UMA, and the stance I=E2=80=99ve been advocating for here for some time. =
I think
> GNAP could allow that, in that GNAP doesn=E2=80=99t :require: a tight cou=
pling per
> se, but it also doesn=E2=80=99t assume a completely loose coupling or for=
ce the
> protocol through those hoops for all other use cases. UMA tried to do the
> latter, but just about every UMA deployment I saw didn=E2=80=99t make use=
 of that
> loose coupling, but you still had to go through the bits of the protocol
> that were built just in case that happened.
>
> It=E2=80=99s a goal of GNAP to learn from the good and bad of previous pr=
otocols,
> including UMA, and this is one area that I think we can improve.
>
> I should mention that for readers familiar with UMA (or at least... for
> me) this was not clear from reading GNAP; I came in assuming GNAP intende=
d
> to cover this space.
>
>
> I think what might be confusing here is that GNAP does intend to cover th=
e
> use case space of having a user delegate to another user, and externalizi=
ng
> a lot of the user rights aspects of the ore delegation process too. But
> instead of doing it through =E2=80=9Ceveryone brings their own AS=E2=80=
=9D type of things,
> GNAP is trying to do it through =E2=80=9Cthe AS can now be flexible enoug=
h to get a
> lot of different information=E2=80=9D. There were some slides on this a w=
hile back:
>
>
> https://datatracker.ietf.org/meeting/110/materials/slides-110-gnap-core-p=
rotocol-slides-00
>
> Starting at slide 35 you=E2=80=99ll see diagrams and points of some of th=
e
> alternatives. I know the meeting is recorded but I=E2=80=9Dm not sure whe=
re that
> lives if you want to review it.
>
>  - Justin
>
>
> On Tue, Sep 28, 2021 at 1:57 PM Justin Richer <jricher@mit.edu> wrote:
>
>> Hi Josh, thanks for bringing up this use case. An AS protecting several
>> different unrelated RS=E2=80=99s is absolutely in scope, as this is a co=
mmon
>> deployment for OAuth 2 today. If I understand everything you=E2=80=99re =
asking,
>> currently GNAP would allow everything here, but it doesn=E2=80=99t speci=
fy the
>> details of every part of it. I=E2=80=99m going to reply to your scenario=
s inline
>> below. Apologies in advance for the long email, there=E2=80=99s a lot to=
 unpack
>> here!
>>
>> On Sep 27, 2021, at 1:56 PM, Josh Mandel <jmandel@gmail.com> wrote:
>>
>> Hi All,
>>
>> I'm interested in exploring how GNAP can/does/might account for the
>> Resource Owner associated with specific protected resources, in scenario=
s
>> where a single AS is trusted by multiple independent RS's (specifically =
in
>> scenarios where these RS's don't know or trust one another). Justin
>> suggestd I might reach out here.
>>
>> My sense is that there are security considerations related to these
>> questions, and perhaps some best practices or even areas where the core
>> spec's data models could be enhanced to support these scenarios. I'm not
>> proposing anything in particular at this stage; just trying to get the l=
ay
>> of the land.
>>
>> See assumptions and scenarios below.
>>
>> -Josh
>>
>> --
>>
>> *Assumptions. *I'll focus on a very simple API from the RS perspective:
>> a standardized file access API where resource servers would like to make
>> access control decisions based on access tokens whose RARs look somethin=
g
>> like:
>>
>> {
>>   "type": "file-access",
>>   "actions": ["read"],
>>   "location": ["{{URL of one protected file}}"]
>> }
>>
>> ---
>>
>> *Scenario 1. *Consider the case where a single RS manages files owned by
>> multiple end users. The RS uses a single AS for all files (but the AS al=
so
>> serves other untrusted RS's). Each RO is allowed to set authorization
>> policies for their own files. In this scenario, how does the AS learn wh=
ich
>> RO is responsible for setting policies for each file, so it can present
>> Alice with a list of current policies, allow her to review/tweak access,
>> etc?
>>
>>
>> Setting and updating policies are outside the scope of GNAP core,
>> including any interfaces for doing so.
>>
>> The policies are controlled at the AS and enforced by the RS. It=E2=80=
=99s really
>> up to the individual deployments how and where to make that split: an AS
>> could issue tokens that are simply tied to the specific data, the result=
 of
>> processing the security context of the delegation. Or the AS could send
>> over all the context information to the RS to let it make its own decisi=
ons
>> entirely, while the AS only has an idea of something existing and not wh=
o
>> owns or controls it. Or something in between. And all of these are
>> functionally the same whether you=E2=80=99re introspecting or using stru=
ctured
>> tokens, so I=E2=80=99m not going to get into the differences here.
>>
>> Doing this statically would be pretty easy, since the decisions can just
>> be set in code and executed. Doing it dynamically feels a lot harder and
>> more interesting. There are a lot of pieces that need to be fulfilled to
>> even get to this point, and I=E2=80=99ll try to add some thoughts on eac=
h:
>>
>> - How does the AS know who Alice is? In OAuth, it=E2=80=99s assumed that=
 Alice
>> can =E2=80=9Clog in=E2=80=9D to the AS using her account. In GNAP we als=
o allow that but
>> further let Alice have other ways of interacting with the AS, and doesn=
=E2=80=99t
>> assume that Alice has an =E2=80=9Caccount=E2=80=9D at the AS in a tradit=
ional sense. Even
>> so, it=E2=80=99s likely in this scenario that the AS will have some set =
of
>> identifiers that it has bound to Alice, which can be associated with tha=
t
>> policy.
>>
>> - How does the RS know about the AS? GNAP does offer some discovery for
>> the technical connections (
>> https://www.ietf.org/archive/id/draft-ietf-gnap-resource-servers-01.html=
#name-rs-facing-as-discovery),
>> but more important is the question of how the RS decides to trust the AS=
 to
>> protect any given resource. In your scenario this sounds like a
>> configuration time decision, but it might not always be.
>>
>>
>> (If an AS is confused on this point, there's a risk that Bob can set
>> policies for Alice's files.) It's possible that the Resource Registratio=
n
>> API could be extended to model this kind of ownership, if both servers h=
ave
>> some interoperable way to communicate user identities. I'm not sure if
>> there are existing patterns to recommend here?
>>
>>
>> - How does the AS know about what the RS is protecting? GNAP allows the
>> RS to dynamically introduce a new resource set to the AS and get back a
>> reference for it (
>> https://www.ietf.org/archive/id/draft-ietf-gnap-resource-servers-01.html=
#name-registering-a-resource-set),
>> and like you, I could see this being expanded with an ability to say =E2=
=80=9Cthis
>> RO / set of ROs controls this resource set=E2=80=9D as part of that regi=
stration
>> step. As long as the identifiers aligned, it would let the AS do some
>> enforcement at token issuance time. But that really isn=E2=80=99t meant =
to set
>> policies, it=E2=80=99s meant to allow an RS to register a set of rights =
with a
>> simple identifier (like a scope with a random value). Perhaps a differen=
t
>> API is actually needed for policy management? And with that in mind, it
>> would only be needed if we wanted the AS to make sure Bob is never promp=
ted
>> for one of Alice=E2=80=99s files at the AS, ever. If you don=E2=80=99t w=
ant that
>> information to leak to the AS, you can allow Bob to get an ineffective
>> token for Alice=E2=80=99s file resource, and count on the RS to say =E2=
=80=9Chey you aren=E2=80=99t
>> Alice, go away=E2=80=9D. Again, this goes back to where you draw the enf=
orcement
>> lines, which often has more to do with the nature of the API you=E2=80=
=99re
>> protecting than anything.
>>
>> - How does the RS know how to associate a specific resource with =E2=80=
=9CAlice=E2=80=9D,
>> and not =E2=80=9CBob"? This requires the RS to know who =E2=80=9CAlice=
=E2=80=9D is and know that a
>> given file is hers to control. Doing this dynamically feels really diffi=
cult
>>
>> - How do the RS and AS agree that =E2=80=9CAlice=E2=80=9D is the same pe=
rson with the
>> same identifiers? It=E2=80=99s obvious that simply trusting a global ide=
ntifier
>> here won=E2=80=99t solve it on its own =E2=80=94 confusion remains trivi=
al, it would just
>> take an RS to apply Bob=E2=80=99s global identifier to Alice=E2=80=99s d=
ata, or for Bob to
>> get assigned Alice=E2=80=99s identifier by a bad AS. But GNAP does have =
a whole
>> notion of =E2=80=9Csubject identifiers=E2=80=9D that are scoped to the A=
S=E2=80=99s trust domain
>> that can be used here: if the RS refers to Alice by an identifier known =
to
>> the AS, and it trusts the AS to enforce that binding for Alice, then it
>> will work.
>>
>>
>> It's also very important for the RS to know, whenever it receives an
>> access token, that the RS is an intended audience for that token. The RA=
R
>> "locations" aren't enough to accomplish this, unless the AS has a reliab=
le
>> scheme in place to prevent one RS from crafting RARs that refer to resou=
rce
>> locations within another RS.  And as the spec says, the AS introspection
>> endpoint needs to ensure that access tokens are "appropriate for
>> presentation at the identified RS". But that's especially challenging wh=
en
>> a single access token can include multiple RARs, potentially intended fo=
r
>> different RS's (which appears to be allowed by the base spec). And for
>> structured token formats where introspection isn't required, the same
>> questions hold; is there existing work on this?
>>
>>
>> The =E2=80=98locations=E2=80=99 field on its own isn=E2=80=99t enough fo=
r this, but it can be a
>> place to hold identifiers for each RS. These need not be URLs to the API=
s
>> in question, though in practice I=E2=80=99d expect them to be. It then a=
lso begs
>> the question of how the RS identifiers came to be, and if they=E2=80=99r=
e then
>> required for this type of processing.
>>
>>
>> *Scenario 2. *Now consider the case where 10 organizations each manage a
>> File Sharing RS. Each RS hosts files for multiple ROs, and the ROs get t=
o
>> choose which AS should be trusted for their own files. In this context, =
I
>> suppose the workflow that introduces the RS to the AS can be mediated by
>> the RO, so that the RO is logged into both systems when making the
>> introduction, and the RS can essentially register a new client instance
>> with the AS for each RO who decides to establish a connection. This patt=
ern
>> has the advantage that there's no confusion about *who* controls a given
>> resource -- but perhaps there are other, better ways to accomplish this.
>>
>>
>> That was what UMA tried to do with resource set registration, in a way.
>> It was awkward from a technology perspective, and it made RS owners nerv=
ous
>> in a lot of ways to have a new AS introduced. I appreciate the ideology,
>> but to this day I don=E2=80=99t believe it=E2=80=99s realistic to implem=
ent or deploy.
>>
>> I think a better pattern is to allow introduction of =E2=80=9CAlice=E2=
=80=9D at the AS
>> level instead of at the RS level. In other words, Alice doesn=E2=80=99t =
bring her
>> own =E2=80=9CAS=E2=80=9D, but she could bring her own identity and polic=
y to the AS that
>> protects the RS. That way the RS-AS relationship can be more tightly
>> aligned, but the user capabilities are greatly expanded. It changes the =
fit
>> of the components from what we might be used to, but I strongly believe
>> there=E2=80=99s a lot of promise in this pattern. And in fact, we see a =
similar
>> pattern in OAuth 2 today with API Gateways that can be configured to all=
ow
>> federated logins, for example.
>>
>> Finally, I think there=E2=80=99s an even more interesting scenario: when=
 Alice
>> says that Bob should be able to access her stuff. Who does she tell and =
who
>> needs to care? There are a lot of ways to slice this, but this jumps to =
my
>> mind immediately:
>>
>>  - Alice has a resource A at the RS, which is protected by the AS
>>  - The AS has a policy that says Alice can authorize client instances to
>> access resource A
>>  - The AS also has an interface for Alice to set additional policies
>>  - Alice says to the AS that a user identified as Bob can also access A
>>  - When either Alice or Bob come to the AS, the AS can figure out from
>> this policy to create an access token that can access A
>>  - The RS doesn=E2=80=99t need to know about Alice or Bob at all, only w=
hether
>> the token is good for accessing A
>>
>>  =E2=80=94 Justin
>>
>>
>>
>>
>>
>> --
>> TXAuth mailing list
>> TXAuth@ietf.org
>> https://www.ietf.org/mailman/listinfo/txauth
>>
>>
>>
> --
> TXAuth mailing list
> TXAuth@ietf.org
> https://www.ietf.org/mailman/listinfo/txauth
>

--0000000000004cf4ab05cd1ac076
Content-Type: text/html; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable

<div dir=3D"ltr"><div>Josh and Justin, <br></div><div><br></div><div>From a=
 privacy perspective, your glee over not having the RO specify the AS to th=
e RS is troubling. Why would the RS care about privacy unless mandated in s=
ome way? (I&#39;m assuming mandates are out of scope for GNAP :-).<br></div=
><div><br></div><div>The privacy problem is simply data minimization. As an=
 RO, I should not have to share my policies with the RS because all the RS =
needs is the access token. If the RS is unwilling (or whatever you mean by =
&quot;nervous&quot;, above) to trust a random AS simply because the RO trus=
ts it, then we&#39;re left with capabilities. Imagine a typical RS serving =
millions of ROs. The policy database would be a gold mine in so many ways. =
Traffic analysis by the RO might still be a problem and will also need to b=
e considered.<br></div><div><br></div><div>If specifying the AS is not an o=
ption, the RO associated with a resource MUST be given a full set of capabi=
lities on demand. The RO can then pass some or all of those to whatever AS =
they trust with their policies and then that AS can issue attenuated access=
 tokens with or without further delegation rights to any client they choose=
. <br></div><div><br></div><div>GNAP core must choose either or both arbitr=
ary AS or capabilities and maybe explain &quot;nervous&quot; in a way that =
people can understand.</div><div><br></div><div>- Adrian<br></div><div><br>=
</div></div><br><div class=3D"gmail_quote"><div dir=3D"ltr" class=3D"gmail_=
attr">On Tue, Sep 28, 2021 at 3:35 PM Justin Richer &lt;<a href=3D"mailto:j=
richer@mit.edu">jricher@mit.edu</a>&gt; wrote:<br></div><blockquote class=
=3D"gmail_quote" style=3D"margin:0px 0px 0px 0.8ex;border-left:1px solid rg=
b(204,204,204);padding-left:1ex"><div style=3D"overflow-wrap: break-word;">=
On Sep 28, 2021, at 3:20 PM, Josh Mandel &lt;<a href=3D"mailto:jmandel@gmai=
l.com" target=3D"_blank">jmandel@gmail.com</a>&gt; wrote:<br><div><blockquo=
te type=3D"cite"><br><div><div dir=3D"ltr">Thanks Justin -- this whole resp=
onse is extremely helpful. I particularly appreciate=C2=A0your very clear d=
eclaration of scope and intention:<div><br></div><div><i>&gt; That was what=
 UMA tried to do with resource set registration, in a way. It was awkward f=
rom a technology perspective, and it made RS owners nervous in a lot of way=
s to have a new AS introduced. I appreciate the ideology, but to this day I=
 don=E2=80=99t believe it=E2=80=99s realistic to implement or deploy.</i></=
div><div><br></div><div>I agree entirely with this assessment, and I agree =
that some tighter degree of AS::RS coupling is needed to support real-world=
 system design. Understanding that this UMA-like level of RS::AS flexibilit=
y is *not* the design intent for GNAP alleviates several of my concerns in =
one fell swoop. </div></div></div></blockquote><div><br></div><div><div>To =
be clear, this is my own take on the goals of GNAP vs. experience with UMA,=
 and the stance I=E2=80=99ve been advocating for here for some time. I thin=
k GNAP could allow that, in that GNAP doesn=E2=80=99t :require: a tight cou=
pling per se, but it also doesn=E2=80=99t assume a completely loose couplin=
g or force the protocol through those hoops for all other use cases. UMA tr=
ied to do the latter, but just about every UMA deployment I saw didn=E2=80=
=99t make use of that loose coupling, but you still had to go through the b=
its of the protocol that were built just in case that happened.=C2=A0</div>=
<div><br></div><div>It=E2=80=99s a goal of GNAP to learn from the good and =
bad of previous protocols, including UMA, and this is one area that I think=
 we can improve.</div></div><br><blockquote type=3D"cite"><div><div dir=3D"=
ltr"><div>I should mention that for readers familiar with UMA (or at least.=
.. for me) this was not clear from reading GNAP; I came in assuming GNAP in=
tended to cover this space.</div></div><br></div></blockquote><div><br></di=
v><div>I think what might be confusing here is that GNAP does intend to cov=
er the use case space of having a user delegate to another user, and extern=
alizing a lot of the user rights aspects of the ore delegation process too.=
 But instead of doing it through =E2=80=9Ceveryone brings their own AS=E2=
=80=9D type of things, GNAP is trying to do it through =E2=80=9Cthe AS can =
now be flexible enough to get a lot of different information=E2=80=9D. Ther=
e were some slides on this a while back:</div><div><br></div><div><a href=
=3D"https://datatracker.ietf.org/meeting/110/materials/slides-110-gnap-core=
-protocol-slides-00" target=3D"_blank">https://datatracker.ietf.org/meeting=
/110/materials/slides-110-gnap-core-protocol-slides-00</a></div><div><br></=
div><div>Starting at slide 35 you=E2=80=99ll see diagrams and points of som=
e of the alternatives. I know the meeting is recorded but I=E2=80=9Dm not s=
ure where that lives if you want to review it.</div><div><br></div><div>=C2=
=A0- Justin</div><div><br></div><br><blockquote type=3D"cite"><div><div cla=
ss=3D"gmail_quote"><div dir=3D"ltr" class=3D"gmail_attr">On Tue, Sep 28, 20=
21 at 1:57 PM Justin Richer &lt;<a href=3D"mailto:jricher@mit.edu" target=
=3D"_blank">jricher@mit.edu</a>&gt; wrote:<br></div><blockquote class=3D"gm=
ail_quote" style=3D"margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,=
204,204);padding-left:1ex"><div>Hi Josh, thanks for bringing up this use ca=
se. An AS protecting several different unrelated RS=E2=80=99s is absolutely=
 in scope, as this is a common deployment for OAuth 2 today. If I understan=
d everything you=E2=80=99re asking, currently GNAP would allow everything h=
ere, but it doesn=E2=80=99t specify the details of every part of it. I=E2=
=80=99m going to reply to your scenarios inline below. Apologies in advance=
 for the long email, there=E2=80=99s a lot to unpack here!<div><div><br><bl=
ockquote type=3D"cite"><div>On Sep 27, 2021, at 1:56 PM, Josh Mandel &lt;<a=
 href=3D"mailto:jmandel@gmail.com" target=3D"_blank">jmandel@gmail.com</a>&=
gt; wrote:</div><br><div><div dir=3D"ltr"><div dir=3D"auto">Hi All,</div><d=
iv dir=3D"auto"><br></div><div dir=3D"auto">I&#39;m interested in exploring=
 how GNAP can/does/might account for the Resource Owner associated with spe=
cific protected resources, in scenarios where a single AS is trusted by mul=
tiple independent RS&#39;s (specifically in scenarios where these RS&#39;s =
don&#39;t know or trust one another). Justin suggestd I might reach out her=
e.<br></div><div dir=3D"auto"><br></div><div dir=3D"auto"><div>My sense is =
that there are security considerations related to these questions, and perh=
aps some best practices or even areas where the core spec&#39;s data models=
 could be enhanced to support these scenarios. I&#39;m not proposing anythi=
ng in particular at this stage; just trying to get the lay of the land.</di=
v><div><br></div><div>See assumptions and scenarios below.</div><div><br></=
div><div>-Josh</div><div><br></div><div>--</div></div><div dir=3D"auto"><di=
v dir=3D"auto"><br></div><div dir=3D"auto"><b>Assumptions.=C2=A0</b>I&#39;l=
l focus on a very simple API from the RS perspective: a standardized file a=
ccess API where resource servers would like to make access control decision=
s based on access tokens whose RARs look something like:</div><div dir=3D"a=
uto"><br></div><div><font size=3D"1" face=3D"monospace">{</font></div><div>=
<font size=3D"1" face=3D"monospace">=C2=A0 &quot;type&quot;: &quot;file-acc=
ess&quot;,</font></div><div><font size=3D"1" face=3D"monospace">=C2=A0 &quo=
t;actions&quot;: [&quot;read&quot;],</font></div><div><font size=3D"1" face=
=3D"monospace">=C2=A0 &quot;location&quot;: [&quot;{{URL of one protected f=
ile}}&quot;]</font></div><div><font size=3D"1" face=3D"monospace">}</font><=
/div><div dir=3D"auto"><div dir=3D"auto"><br></div><div dir=3D"auto">---</d=
iv><div dir=3D"auto"><br></div><div dir=3D"auto"><b>Scenario 1. </b>Conside=
r the case where a=C2=A0single RS manages files owned by multiple end users=
. The RS uses a single AS for all files (but the AS also serves other untru=
sted RS&#39;s). Each RO is allowed to set authorization policies for their =
own files. In this scenario, how does the AS learn which RO is responsible =
for setting policies for each file,=C2=A0so it can present Alice with a lis=
t of current policies,=C2=A0allow her to review/tweak access, etc?</div></d=
iv></div></div></div></blockquote><div><br></div><div>Setting and updating =
policies are outside the scope of GNAP core, including any interfaces for d=
oing so.</div><div><br></div><div>The policies are controlled at the AS and=
 enforced by the RS. It=E2=80=99s really up to the individual deployments h=
ow and where to make that split: an AS could issue tokens that are simply t=
ied to the specific data, the result of processing the security context of =
the delegation. Or the AS could send over all the context information to th=
e RS to let it make its own decisions entirely, while the AS only has an id=
ea of something existing and not who owns or controls it. Or something in b=
etween. And all of these are functionally the same whether you=E2=80=99re i=
ntrospecting or using structured tokens, so I=E2=80=99m not going to get in=
to the differences here.</div><div><br></div><div>Doing this statically wou=
ld be pretty easy, since the decisions can just be set in code and executed=
. Doing it dynamically feels a lot harder and more interesting. There are a=
 lot of pieces that need to be fulfilled to even get to this point, and I=
=E2=80=99ll try to add some thoughts on each:=C2=A0</div><div><br></div><di=
v>- How does the AS know who Alice is? In OAuth, it=E2=80=99s assumed that =
Alice can =E2=80=9Clog in=E2=80=9D to the AS using her account. In GNAP we =
also allow that but further let Alice have other ways of interacting with t=
he AS, and doesn=E2=80=99t assume that Alice has an =E2=80=9Caccount=E2=80=
=9D at the AS in a traditional sense. Even so, it=E2=80=99s likely in this =
scenario that the AS will have some set of identifiers that it has bound to=
 Alice, which can be associated with that policy.</div><div><br></div><div>=
- How does the RS know about the AS? GNAP does offer some discovery for the=
 technical connections (<a href=3D"https://www.ietf.org/archive/id/draft-ie=
tf-gnap-resource-servers-01.html#name-rs-facing-as-discovery" target=3D"_bl=
ank">https://www.ietf.org/archive/id/draft-ietf-gnap-resource-servers-01.ht=
ml#name-rs-facing-as-discovery</a>), but more important is the question of =
how the RS decides to trust the AS to protect any given resource. In your s=
cenario this sounds like a configuration time decision, but it might not al=
ways be.</div><div><br></div><div><br></div><blockquote type=3D"cite"><div>=
<div dir=3D"ltr"><div dir=3D"auto"><div dir=3D"auto"><div dir=3D"auto"> (If=
 an AS is confused on this point, there&#39;s a risk that Bob can set polic=
ies for Alice&#39;s files.) It&#39;s possible that the Resource Registratio=
n API could be extended to model this kind of ownership, if both servers ha=
ve some interoperable way to communicate user identities. I&#39;m not sure =
if there are existing patterns to recommend here? </div></div></div></div><=
/div></blockquote><div><br></div><div><div>- How does the AS know about wha=
t the RS is protecting? GNAP allows the RS to dynamically introduce a new r=
esource set to the AS and get back a reference for it (<a href=3D"https://w=
ww.ietf.org/archive/id/draft-ietf-gnap-resource-servers-01.html#name-regist=
ering-a-resource-set" target=3D"_blank">https://www.ietf.org/archive/id/dra=
ft-ietf-gnap-resource-servers-01.html#name-registering-a-resource-set</a>),=
 and like you, I could see this being expanded with an ability to say =E2=
=80=9Cthis RO / set of ROs controls this resource set=E2=80=9D as part of t=
hat registration step. As long as the identifiers aligned, it would let the=
 AS do some enforcement at token issuance time. But that really isn=E2=80=
=99t meant to set policies, it=E2=80=99s meant to allow an RS to register a=
 set of rights with a simple identifier (like a scope with a random value).=
 Perhaps a different API is actually needed for policy management? And with=
 that in mind, it would only be needed if we wanted the AS to make sure Bob=
 is never prompted for one of Alice=E2=80=99s files at the AS, ever. If you=
 don=E2=80=99t want that information to leak to the AS, you can allow Bob t=
o get an ineffective token for Alice=E2=80=99s file resource, and count on =
the RS to say =E2=80=9Chey you aren=E2=80=99t Alice, go away=E2=80=9D. Agai=
n, this goes back to where you draw the enforcement lines, which often has =
more to do with the nature of the API you=E2=80=99re protecting than anythi=
ng.</div><div><br></div></div><div><div>- How does the RS know how to assoc=
iate a specific resource with =E2=80=9CAlice=E2=80=9D, and not =E2=80=9CBob=
&quot;? This requires the RS to know who =E2=80=9CAlice=E2=80=9D is and kno=
w that a given file is hers to control. Doing this dynamically feels really=
 difficult</div><div><br></div><div>- How do the RS and AS agree that =E2=
=80=9CAlice=E2=80=9D is the same person with the same identifiers? It=E2=80=
=99s obvious that simply trusting a global identifier here won=E2=80=99t so=
lve it on its own =E2=80=94 confusion remains trivial, it would just take a=
n RS to apply Bob=E2=80=99s global identifier to Alice=E2=80=99s data, or f=
or Bob to get assigned Alice=E2=80=99s identifier by a bad AS. But GNAP doe=
s have a whole notion of =E2=80=9Csubject identifiers=E2=80=9D that are sco=
ped to the AS=E2=80=99s trust domain that can be used here: if the RS refer=
s to Alice by an identifier known to the AS, and it trusts the AS to enforc=
e that binding for Alice, then it will work.</div><div><br></div></div><br>=
<blockquote type=3D"cite"><div><div dir=3D"ltr"><div dir=3D"auto"><div dir=
=3D"auto"><div dir=3D"auto">It&#39;s also very important for the RS to know=
, whenever it receives an access token, that the RS is an intended audience=
 for that token. The RAR &quot;locations&quot; aren&#39;t enough to accompl=
ish this, unless the AS has a reliable scheme in place to prevent one RS fr=
om crafting RARs that refer to resource locations within another RS.=C2=A0 =
And as the spec says, the AS introspection endpoint needs to=C2=A0ensure th=
at=C2=A0access tokens are &quot;appropriate for presentation at the identif=
ied RS&quot;. But=C2=A0that&#39;s especially challenging when a single acce=
ss token can include multiple RARs, potentially intended for different RS&#=
39;s (which appears to be allowed by the base spec). And for structured tok=
en formats where introspection isn&#39;t required, the same questions hold;=
 is there existing work on this?<br></div></div></div></div></div></blockqu=
ote><div><br></div><div>The =E2=80=98locations=E2=80=99 field on its own is=
n=E2=80=99t enough for this, but it can be a place to hold identifiers for =
each RS. These need not be URLs to the APIs in question, though in practice=
 I=E2=80=99d expect them to be. It then also begs the question of how the R=
S identifiers came to be, and if they=E2=80=99re then required for this typ=
e of processing.=C2=A0</div><br><blockquote type=3D"cite"><div><div dir=3D"=
ltr"><div dir=3D"auto"><div dir=3D"auto"><div dir=3D"auto"><br></div><div><=
b>Scenario 2.=C2=A0</b>Now consider the case where 10 organizations each ma=
nage a File Sharing RS. Each RS hosts files for multiple ROs, and the ROs g=
et to choose which AS should be trusted for their own files. In this contex=
t, I suppose the workflow that introduces the RS to the AS can be mediated =
by the RO, so that the RO is logged into both systems when making the intro=
duction, and the RS can essentially register a new client instance with the=
 AS for each RO who decides to establish a connection. This pattern has the=
 advantage that there&#39;s no confusion about *who* controls a given resou=
rce -- but perhaps there are other, better ways to accomplish this.</div></=
div></div></div></div></blockquote><div><br></div><div>That was what UMA tr=
ied to do with resource set registration, in a way. It was awkward from a t=
echnology perspective, and it made RS owners nervous in a lot of ways to ha=
ve a new AS introduced. I appreciate the ideology, but to this day I don=E2=
=80=99t believe it=E2=80=99s realistic to implement or deploy.</div><div><b=
r></div><div>I think a better pattern is to allow introduction of =E2=80=9C=
Alice=E2=80=9D at the AS level instead of at the RS level. In other words, =
Alice doesn=E2=80=99t bring her own =E2=80=9CAS=E2=80=9D, but she could bri=
ng her own identity and policy to the AS that protects the RS. That way the=
 RS-AS relationship can be more tightly aligned, but the user capabilities =
are greatly expanded. It changes the fit of the components from what we mig=
ht be used to, but I strongly believe there=E2=80=99s a lot of promise in t=
his pattern. And in fact, we see a similar pattern in OAuth 2 today with AP=
I Gateways that can be configured to allow federated logins, for example.</=
div><div><br></div><div>Finally, I think there=E2=80=99s an even more inter=
esting scenario: when Alice says that Bob should be able to access her stuf=
f. Who does she tell and who needs to care? There are a lot of ways to slic=
e this, but this jumps to my mind immediately:</div><div><br></div><div>=C2=
=A0- Alice has a resource A at the RS, which is protected by the AS</div><d=
iv>=C2=A0- The AS has a policy that says Alice can authorize client instanc=
es to access resource A</div><div>=C2=A0- The AS also has an interface for =
Alice to set additional policies</div><div>=C2=A0- Alice says to the AS tha=
t a user identified as Bob can also access A</div><div>=C2=A0- When either =
Alice or Bob come to the AS, the AS can figure out from this policy to crea=
te an access token that can access A</div><div>=C2=A0- The RS doesn=E2=80=
=99t need to know about Alice or Bob at all, only whether the token is good=
 for accessing A</div><div><br></div><div>=C2=A0=E2=80=94 Justin</div><div>=
<br></div><div><br></div><br><blockquote type=3D"cite"><div><div dir=3D"ltr=
"><div dir=3D"auto"><div dir=3D"auto"><div dir=3D"auto"><div dir=3D"auto"><=
br></div><div dir=3D"auto"><br></div></div></div></div>
</div>
-- <br>TXAuth mailing list<br><a href=3D"mailto:TXAuth@ietf.org" target=3D"=
_blank">TXAuth@ietf.org</a><br><a href=3D"https://www.ietf.org/mailman/list=
info/txauth" target=3D"_blank">https://www.ietf.org/mailman/listinfo/txauth=
</a><br></div></blockquote></div><br></div></div></blockquote></div>
</div></blockquote></div><br></div>-- <br>
TXAuth mailing list<br>
<a href=3D"mailto:TXAuth@ietf.org" target=3D"_blank">TXAuth@ietf.org</a><br=
>
<a href=3D"https://www.ietf.org/mailman/listinfo/txauth" rel=3D"noreferrer"=
 target=3D"_blank">https://www.ietf.org/mailman/listinfo/txauth</a><br>
</blockquote></div>

--0000000000004cf4ab05cd1ac076--


From nobody Wed Sep 29 05:17:00 2021
Return-Path: <jricher@mit.edu>
X-Original-To: txauth@ietfa.amsl.com
Delivered-To: txauth@ietfa.amsl.com
Received: from localhost (localhost [127.0.0.1]) by ietfa.amsl.com (Postfix) with ESMTP id 89B053A14CC for <txauth@ietfa.amsl.com>; Wed, 29 Sep 2021 05:16:58 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -1.898
X-Spam-Level: 
X-Spam-Status: No, score=-1.898 tagged_above=-999 required=5 tests=[BAYES_00=-1.9, RCVD_IN_MSPIKE_H2=-0.001, SPF_HELO_NONE=0.001, SPF_NONE=0.001, URIBL_BLOCKED=0.001] autolearn=ham autolearn_force=no
Received: from mail.ietf.org ([4.31.198.44]) by localhost (ietfa.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id o-10htaiDw9t for <txauth@ietfa.amsl.com>; Wed, 29 Sep 2021 05:16:53 -0700 (PDT)
Received: from outgoing-exchange-7.mit.edu (outgoing-exchange-7.mit.edu [18.9.28.58]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by ietfa.amsl.com (Postfix) with ESMTPS id 9E4B43A14F1 for <txauth@ietf.org>; Wed, 29 Sep 2021 05:16:52 -0700 (PDT)
Received: from w92exedge3.exchange.mit.edu (W92EXEDGE3.EXCHANGE.MIT.EDU [18.7.73.15]) by outgoing-exchange-7.mit.edu (8.14.7/8.12.4) with ESMTP id 18TCGQqc013956; Wed, 29 Sep 2021 08:16:46 -0400
Received: from oc11expo18.exchange.mit.edu (18.9.4.49) by w92exedge3.exchange.mit.edu (18.7.73.15) with Microsoft SMTP Server (TLS) id 15.0.1497.23; Wed, 29 Sep 2021 08:16:26 -0400
Received: from oc11expo18.exchange.mit.edu (18.9.4.49) by oc11expo18.exchange.mit.edu (18.9.4.49) with Microsoft SMTP Server (TLS) id 15.0.1497.23; Wed, 29 Sep 2021 08:16:29 -0400
Received: from oc11expo18.exchange.mit.edu ([18.9.4.49]) by oc11expo18.exchange.mit.edu ([18.9.4.49]) with mapi id 15.00.1497.023; Wed, 29 Sep 2021 08:16:29 -0400
From: Justin Richer <jricher@mit.edu>
To: Adrian Gropper <agropper@healthurl.com>
CC: "jmandel@gmail.com" <jmandel@gmail.com>, GNAP Mailing List <txauth@ietf.org>
Thread-Topic: [GNAP] Accounting for Resource Owner in RS::AS interactions
Thread-Index: AQHXs8kKpD6zmNxgu0KI6JYI9heXrau5zcsAgABJcID//8EiAIAA174AgAA8uyE=
Date: Wed, 29 Sep 2021 12:16:29 +0000
Message-ID: <7e4836211bae4dea8fec42a0236e6ba6@oc11expo18.exchange.mit.edu>
References: <CANSMLKHx4xKrR+c-KzdpxF0==c_=wqPexQJTVNTEXVmU5Usm9g@mail.gmail.com> <5AE1FBFF-C475-4B05-BB94-5C97B3B5CA22@mit.edu> <CANSMLKFR=9QH+ZCCVj3ePDwA9TnVrMALan3W=icg+pz0+fV4rA@mail.gmail.com> <7D238C14-3FEB-4166-91CD-D54086CFEFA1@mit.edu>, <CANYRo8iHLBX2Fxq7rZa=N8OeWGqYptUuzZu656B4VgWEtAy9Bw@mail.gmail.com>
In-Reply-To: <CANYRo8iHLBX2Fxq7rZa=N8OeWGqYptUuzZu656B4VgWEtAy9Bw@mail.gmail.com>
Accept-Language: en-US
Content-Language: en-US
X-MS-Has-Attach: 
X-MS-TNEF-Correlator: 
x-ms-exchange-transport-fromentityheader: Hosted
x-originating-ip: [71.174.62.56]
Content-Type: text/plain; charset="Windows-1252"
Content-Transfer-Encoding: quoted-printable
MIME-Version: 1.0
Archived-At: <https://mailarchive.ietf.org/arch/msg/txauth/eV8O7KjGnXE6tQn5DlSvCA7EmQw>
Subject: Re: [GNAP] Accounting for Resource Owner in RS::AS interactions
X-BeenThere: txauth@ietf.org
X-Mailman-Version: 2.1.29
Precedence: list
List-Id: GNAP <txauth.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/options/txauth>, <mailto:txauth-request@ietf.org?subject=unsubscribe>
List-Archive: <https://mailarchive.ietf.org/arch/browse/txauth/>
List-Post: <mailto:txauth@ietf.org>
List-Help: <mailto:txauth-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/txauth>, <mailto:txauth-request@ietf.org?subject=subscribe>
X-List-Received-Date: Wed, 29 Sep 2021 12:16:59 -0000

Adrian,=20

Once again, I believe you are conflating the technical roles of the protoco=
l (AS, RS) with the parties that would fulfill those roles. What you descri=
be as the RS entity in your solution is, in reality, an AS/RS pair. What yo=
u describe as the RO's chosen AS is what GNAP calls out as an input into th=
e AS. This is why I keep saying that it's the interaction and consent phase=
 of GNAP where things like SSI and policy descriptions and user-to-user del=
egation come into play. It's a different model from UMA, and I believe it's=
 a model that works much better and fits reality.=20

The AS issues tokens that are trusted, the RS trusts and executes those tok=
ens. Giving an unknown party complete security control over your API is ter=
rifying, and we've already seen that what actually happens is that people b=
uild in a internal AS to handle that case anyway. It's why we have things l=
ike security assertions and token exchange and application gateways.=20

Software delegation protocols, like OAuth and GNAP, aren't particularly wel=
l suited to the user delegation problem on their own, because you can't hav=
e a user present a technical input. You can however make explicit places wi=
thin the protocol for those inputs to be expressed, and that's what we are =
trying to do with GNAP.=20

You can absolutely build out the model I'm describing in privacy preserving=
 ways, and I will argue that it will have significantly better security pro=
perties when compared to letting a user pick their AS.=20

So, let the RS dictate who it will listen to tokens from, and allow the AS =
to get its information about what those tokens are for, and when to issue t=
hem, from a variety of places.=20

- Justin
________________________________________
From: Adrian Gropper [agropper@healthurl.com]
Sent: Wednesday, September 29, 2021 12:27 AM
To: Justin Richer
Cc: jmandel@gmail.com; GNAP Mailing List
Subject: Re: [GNAP] Accounting for Resource Owner in RS::AS interactions

Josh and Justin,

>From a privacy perspective, your glee over not having the RO specify the AS=
 to the RS is troubling. Why would the RS care about privacy unless mandate=
d in some way? (I'm assuming mandates are out of scope for GNAP :-).

The privacy problem is simply data minimization. As an RO, I should not hav=
e to share my policies with the RS because all the RS needs is the access t=
oken. If the RS is unwilling (or whatever you mean by "nervous", above) to =
trust a random AS simply because the RO trusts it, then we're left with cap=
abilities. Imagine a typical RS serving millions of ROs. The policy databas=
e would be a gold mine in so many ways. Traffic analysis by the RO might st=
ill be a problem and will also need to be considered.

If specifying the AS is not an option, the RO associated with a resource MU=
ST be given a full set of capabilities on demand. The RO can then pass some=
 or all of those to whatever AS they trust with their policies and then tha=
t AS can issue attenuated access tokens with or without further delegation =
rights to any client they choose.

GNAP core must choose either or both arbitrary AS or capabilities and maybe=
 explain "nervous" in a way that people can understand.

- Adrian


On Tue, Sep 28, 2021 at 3:35 PM Justin Richer <jricher@mit.edu<mailto:jrich=
er@mit.edu>> wrote:
On Sep 28, 2021, at 3:20 PM, Josh Mandel <jmandel@gmail.com<mailto:jmandel@=
gmail.com>> wrote:

Thanks Justin -- this whole response is extremely helpful. I particularly a=
ppreciate your very clear declaration of scope and intention:

> That was what UMA tried to do with resource set registration, in a way. I=
t was awkward from a technology perspective, and it made RS owners nervous =
in a lot of ways to have a new AS introduced. I appreciate the ideology, bu=
t to this day I don=92t believe it=92s realistic to implement or deploy.

I agree entirely with this assessment, and I agree that some tighter degree=
 of AS::RS coupling is needed to support real-world system design. Understa=
nding that this UMA-like level of RS::AS flexibility is *not* the design in=
tent for GNAP alleviates several of my concerns in one fell swoop.

To be clear, this is my own take on the goals of GNAP vs. experience with U=
MA, and the stance I=92ve been advocating for here for some time. I think G=
NAP could allow that, in that GNAP doesn=92t :require: a tight coupling per=
 se, but it also doesn=92t assume a completely loose coupling or force the =
protocol through those hoops for all other use cases. UMA tried to do the l=
atter, but just about every UMA deployment I saw didn=92t make use of that =
loose coupling, but you still had to go through the bits of the protocol th=
at were built just in case that happened.

It=92s a goal of GNAP to learn from the good and bad of previous protocols,=
 including UMA, and this is one area that I think we can improve.

I should mention that for readers familiar with UMA (or at least... for me)=
 this was not clear from reading GNAP; I came in assuming GNAP intended to =
cover this space.


I think what might be confusing here is that GNAP does intend to cover the =
use case space of having a user delegate to another user, and externalizing=
 a lot of the user rights aspects of the ore delegation process too. But in=
stead of doing it through =93everyone brings their own AS=94 type of things=
, GNAP is trying to do it through =93the AS can now be flexible enough to g=
et a lot of different information=94. There were some slides on this a whil=
e back:

https://datatracker.ietf.org/meeting/110/materials/slides-110-gnap-core-pro=
tocol-slides-00

Starting at slide 35 you=92ll see diagrams and points of some of the altern=
atives. I know the meeting is recorded but I=94m not sure where that lives =
if you want to review it.

 - Justin


On Tue, Sep 28, 2021 at 1:57 PM Justin Richer <jricher@mit.edu<mailto:jrich=
er@mit.edu>> wrote:
Hi Josh, thanks for bringing up this use case. An AS protecting several dif=
ferent unrelated RS=92s is absolutely in scope, as this is a common deploym=
ent for OAuth 2 today. If I understand everything you=92re asking, currentl=
y GNAP would allow everything here, but it doesn=92t specify the details of=
 every part of it. I=92m going to reply to your scenarios inline below. Apo=
logies in advance for the long email, there=92s a lot to unpack here!

On Sep 27, 2021, at 1:56 PM, Josh Mandel <jmandel@gmail.com<mailto:jmandel@=
gmail.com>> wrote:

Hi All,

I'm interested in exploring how GNAP can/does/might account for the Resourc=
e Owner associated with specific protected resources, in scenarios where a =
single AS is trusted by multiple independent RS's (specifically in scenario=
s where these RS's don't know or trust one another). Justin suggestd I migh=
t reach out here.

My sense is that there are security considerations related to these questio=
ns, and perhaps some best practices or even areas where the core spec's dat=
a models could be enhanced to support these scenarios. I'm not proposing an=
ything in particular at this stage; just trying to get the lay of the land.

See assumptions and scenarios below.

-Josh

--

Assumptions. I'll focus on a very simple API from the RS perspective: a sta=
ndardized file access API where resource servers would like to make access =
control decisions based on access tokens whose RARs look something like:

{
  "type": "file-access",
  "actions": ["read"],
  "location": ["{{URL of one protected file}}"]
}

---

Scenario 1. Consider the case where a single RS manages files owned by mult=
iple end users. The RS uses a single AS for all files (but the AS also serv=
es other untrusted RS's). Each RO is allowed to set authorization policies =
for their own files. In this scenario, how does the AS learn which RO is re=
sponsible for setting policies for each file, so it can present Alice with =
a list of current policies, allow her to review/tweak access, etc?

Setting and updating policies are outside the scope of GNAP core, including=
 any interfaces for doing so.

The policies are controlled at the AS and enforced by the RS. It=92s really=
 up to the individual deployments how and where to make that split: an AS c=
ould issue tokens that are simply tied to the specific data, the result of =
processing the security context of the delegation. Or the AS could send ove=
r all the context information to the RS to let it make its own decisions en=
tirely, while the AS only has an idea of something existing and not who own=
s or controls it. Or something in between. And all of these are functionall=
y the same whether you=92re introspecting or using structured tokens, so I=
=92m not going to get into the differences here.

Doing this statically would be pretty easy, since the decisions can just be=
 set in code and executed. Doing it dynamically feels a lot harder and more=
 interesting. There are a lot of pieces that need to be fulfilled to even g=
et to this point, and I=92ll try to add some thoughts on each:

- How does the AS know who Alice is? In OAuth, it=92s assumed that Alice ca=
n =93log in=94 to the AS using her account. In GNAP we also allow that but =
further let Alice have other ways of interacting with the AS, and doesn=92t=
 assume that Alice has an =93account=94 at the AS in a traditional sense. E=
ven so, it=92s likely in this scenario that the AS will have some set of id=
entifiers that it has bound to Alice, which can be associated with that pol=
icy.

- How does the RS know about the AS? GNAP does offer some discovery for the=
 technical connections (https://www.ietf.org/archive/id/draft-ietf-gnap-res=
ource-servers-01.html#name-rs-facing-as-discovery), but more important is t=
he question of how the RS decides to trust the AS to protect any given reso=
urce. In your scenario this sounds like a configuration time decision, but =
it might not always be.


(If an AS is confused on this point, there's a risk that Bob can set polici=
es for Alice's files.) It's possible that the Resource Registration API cou=
ld be extended to model this kind of ownership, if both servers have some i=
nteroperable way to communicate user identities. I'm not sure if there are =
existing patterns to recommend here?

- How does the AS know about what the RS is protecting? GNAP allows the RS =
to dynamically introduce a new resource set to the AS and get back a refere=
nce for it (https://www.ietf.org/archive/id/draft-ietf-gnap-resource-server=
s-01.html#name-registering-a-resource-set), and like you, I could see this =
being expanded with an ability to say =93this RO / set of ROs controls this=
 resource set=94 as part of that registration step. As long as the identifi=
ers aligned, it would let the AS do some enforcement at token issuance time=
. But that really isn=92t meant to set policies, it=92s meant to allow an R=
S to register a set of rights with a simple identifier (like a scope with a=
 random value). Perhaps a different API is actually needed for policy manag=
ement? And with that in mind, it would only be needed if we wanted the AS t=
o make sure Bob is never prompted for one of Alice=92s files at the AS, eve=
r. If you don=92t want that information to leak to the AS, you can allow Bo=
b to get an ineffective token for Alice=92s file resource, and count on the=
 RS to say =93hey you aren=92t Alice, go away=94. Again, this goes back to =
where you draw the enforcement lines, which often has more to do with the n=
ature of the API you=92re protecting than anything.

- How does the RS know how to associate a specific resource with =93Alice=
=94, and not =93Bob"? This requires the RS to know who =93Alice=94 is and k=
now that a given file is hers to control. Doing this dynamically feels real=
ly difficult

- How do the RS and AS agree that =93Alice=94 is the same person with the s=
ame identifiers? It=92s obvious that simply trusting a global identifier he=
re won=92t solve it on its own =97 confusion remains trivial, it would just=
 take an RS to apply Bob=92s global identifier to Alice=92s data, or for Bo=
b to get assigned Alice=92s identifier by a bad AS. But GNAP does have a wh=
ole notion of =93subject identifiers=94 that are scoped to the AS=92s trust=
 domain that can be used here: if the RS refers to Alice by an identifier k=
nown to the AS, and it trusts the AS to enforce that binding for Alice, the=
n it will work.


It's also very important for the RS to know, whenever it receives an access=
 token, that the RS is an intended audience for that token. The RAR "locati=
ons" aren't enough to accomplish this, unless the AS has a reliable scheme =
in place to prevent one RS from crafting RARs that refer to resource locati=
ons within another RS.  And as the spec says, the AS introspection endpoint=
 needs to ensure that access tokens are "appropriate for presentation at th=
e identified RS". But that's especially challenging when a single access to=
ken can include multiple RARs, potentially intended for different RS's (whi=
ch appears to be allowed by the base spec). And for structured token format=
s where introspection isn't required, the same questions hold; is there exi=
sting work on this?

The =91locations=92 field on its own isn=92t enough for this, but it can be=
 a place to hold identifiers for each RS. These need not be URLs to the API=
s in question, though in practice I=92d expect them to be. It then also beg=
s the question of how the RS identifiers came to be, and if they=92re then =
required for this type of processing.


Scenario 2. Now consider the case where 10 organizations each manage a File=
 Sharing RS. Each RS hosts files for multiple ROs, and the ROs get to choos=
e which AS should be trusted for their own files. In this context, I suppos=
e the workflow that introduces the RS to the AS can be mediated by the RO, =
so that the RO is logged into both systems when making the introduction, an=
d the RS can essentially register a new client instance with the AS for eac=
h RO who decides to establish a connection. This pattern has the advantage =
that there's no confusion about *who* controls a given resource -- but perh=
aps there are other, better ways to accomplish this.

That was what UMA tried to do with resource set registration, in a way. It =
was awkward from a technology perspective, and it made RS owners nervous in=
 a lot of ways to have a new AS introduced. I appreciate the ideology, but =
to this day I don=92t believe it=92s realistic to implement or deploy.

I think a better pattern is to allow introduction of =93Alice=94 at the AS =
level instead of at the RS level. In other words, Alice doesn=92t bring her=
 own =93AS=94, but she could bring her own identity and policy to the AS th=
at protects the RS. That way the RS-AS relationship can be more tightly ali=
gned, but the user capabilities are greatly expanded. It changes the fit of=
 the components from what we might be used to, but I strongly believe there=
=92s a lot of promise in this pattern. And in fact, we see a similar patter=
n in OAuth 2 today with API Gateways that can be configured to allow federa=
ted logins, for example.

Finally, I think there=92s an even more interesting scenario: when Alice sa=
ys that Bob should be able to access her stuff. Who does she tell and who n=
eeds to care? There are a lot of ways to slice this, but this jumps to my m=
ind immediately:

 - Alice has a resource A at the RS, which is protected by the AS
 - The AS has a policy that says Alice can authorize client instances to ac=
cess resource A
 - The AS also has an interface for Alice to set additional policies
 - Alice says to the AS that a user identified as Bob can also access A
 - When either Alice or Bob come to the AS, the AS can figure out from this=
 policy to create an access token that can access A
 - The RS doesn=92t need to know about Alice or Bob at all, only whether th=
e token is good for accessing A

 =97 Justin





--
TXAuth mailing list
TXAuth@ietf.org<mailto:TXAuth@ietf.org>
https://www.ietf.org/mailman/listinfo/txauth


--
TXAuth mailing list
TXAuth@ietf.org<mailto:TXAuth@ietf.org>
https://www.ietf.org/mailman/listinfo/txauth


From nobody Wed Sep 29 06:26:31 2021
Return-Path: <agropper@gmail.com>
X-Original-To: txauth@ietfa.amsl.com
Delivered-To: txauth@ietfa.amsl.com
Received: from localhost (localhost [127.0.0.1]) by ietfa.amsl.com (Postfix) with ESMTP id 749513A17B9 for <txauth@ietfa.amsl.com>; Wed, 29 Sep 2021 06:26:28 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -1.648
X-Spam-Level: 
X-Spam-Status: No, score=-1.648 tagged_above=-999 required=5 tests=[BAYES_00=-1.9, FREEMAIL_FORGED_FROMDOMAIN=0.249, FREEMAIL_FROM=0.001, HEADER_FROM_DIFFERENT_DOMAINS=0.001, HTML_MESSAGE=0.001, RCVD_IN_MSPIKE_H2=-0.001, SPF_HELO_NONE=0.001, SPF_PASS=-0.001, URIBL_BLOCKED=0.001] autolearn=no autolearn_force=no
Received: from mail.ietf.org ([4.31.198.44]) by localhost (ietfa.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id 7fwRGStMRSOX for <txauth@ietfa.amsl.com>; Wed, 29 Sep 2021 06:26:23 -0700 (PDT)
Received: from mail-qk1-f177.google.com (mail-qk1-f177.google.com [209.85.222.177]) (using TLSv1.2 with cipher ECDHE-RSA-AES128-GCM-SHA256 (128/128 bits)) (No client certificate requested) by ietfa.amsl.com (Postfix) with ESMTPS id 9F6BE3A17B7 for <txauth@ietf.org>; Wed, 29 Sep 2021 06:26:23 -0700 (PDT)
Received: by mail-qk1-f177.google.com with SMTP id 138so2221647qko.10 for <txauth@ietf.org>; Wed, 29 Sep 2021 06:26:23 -0700 (PDT)
X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20210112; h=x-gm-message-state:mime-version:references:in-reply-to:from:date :message-id:subject:to:cc; bh=BXRm8OkpqQMaiPZu26U90Lx5xJwbqnOOGRKGZkAlTW8=; b=V11P74VRvgfc6PGLPl3l6Ga8fEUx4vVYJ1l2qXctbkPeoYl/rTqUK8GHtKjwEXDnME j+Cuo9dbp2G77ySh6gChaJ8uJDe3rV1WnrgbzG3peCJd3+GRjLJapQSsF/OvKXrTMD5B aMlMGKc+Ot3+k3ysG5G2yxXwLz5S/Gch29rUZ2JRPsKnJkZ/c/w8jCR5Nx1iY7Mar9uv pxeLGVb3Z+tWajwq8GnrixYybuh+q/sM45tgyikLFrAWuPyA9b/G/RN8RrUNlANet+4N omOxStOTs5j3t02iX0pvmv6kA5ymx8L8qiN/qmt0Rsn+BwZuNeDiaRSk1h+OvO2rEewH nrDA==
X-Gm-Message-State: AOAM533oxFmCHo3tFP3ZrcMezUI7Oc1bylhSI9zSTN31LZwX6mBixYPo 4neB/PE1pqqpU0XO6ii+yBpnKnkMu/Jv3P4Mu5E=
X-Google-Smtp-Source: ABdhPJzmnEJhioT/cSsqndTxODSLcEaD+17OFH4Q/RSQqSGH+UrwCeMF2XTPaY/ITHT/rdqrFMkgfHcVWzcs5nTG5Qw=
X-Received: by 2002:a05:620a:62d:: with SMTP id 13mr5367442qkv.216.1632921982463;  Wed, 29 Sep 2021 06:26:22 -0700 (PDT)
MIME-Version: 1.0
References: <CANSMLKHx4xKrR+c-KzdpxF0==c_=wqPexQJTVNTEXVmU5Usm9g@mail.gmail.com> <5AE1FBFF-C475-4B05-BB94-5C97B3B5CA22@mit.edu> <CANSMLKFR=9QH+ZCCVj3ePDwA9TnVrMALan3W=icg+pz0+fV4rA@mail.gmail.com> <7D238C14-3FEB-4166-91CD-D54086CFEFA1@mit.edu> <CANYRo8iHLBX2Fxq7rZa=N8OeWGqYptUuzZu656B4VgWEtAy9Bw@mail.gmail.com> <7e4836211bae4dea8fec42a0236e6ba6@oc11expo18.exchange.mit.edu>
In-Reply-To: <7e4836211bae4dea8fec42a0236e6ba6@oc11expo18.exchange.mit.edu>
From: Adrian Gropper <agropper@healthurl.com>
Date: Wed, 29 Sep 2021 09:26:08 -0400
Message-ID: <CANYRo8h=+weoZx_NnV2eNwvVkYsCWqZ=9ZC7z_Pnij9c5z+XkQ@mail.gmail.com>
To: Justin Richer <jricher@mit.edu>
Cc: "jmandel@gmail.com" <jmandel@gmail.com>, GNAP Mailing List <txauth@ietf.org>, Alan Karp <alanhkarp@gmail.com>
Content-Type: multipart/alternative; boundary="000000000000b5f35a05cd22465f"
Archived-At: <https://mailarchive.ietf.org/arch/msg/txauth/CmdSiduOx5Gs3hDV-0IeNvI9UuI>
Subject: Re: [GNAP] Accounting for Resource Owner in RS::AS interactions
X-BeenThere: txauth@ietf.org
X-Mailman-Version: 2.1.29
Precedence: list
List-Id: GNAP <txauth.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/options/txauth>, <mailto:txauth-request@ietf.org?subject=unsubscribe>
List-Archive: <https://mailarchive.ietf.org/arch/browse/txauth/>
List-Post: <mailto:txauth@ietf.org>
List-Help: <mailto:txauth-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/txauth>, <mailto:txauth-request@ietf.org?subject=subscribe>
X-List-Received-Date: Wed, 29 Sep 2021 13:26:29 -0000

--000000000000b5f35a05cd22465f
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable

Please help me understand what's happening in privacy terms so that I might
be able to explain it to others.

inline... and apologies for the long email

On Wed, Sep 29, 2021 at 8:16 AM Justin Richer <jricher@mit.edu> wrote:

> Adrian,
>
> Once again, I believe you are conflating the technical roles of the
> protocol (AS, RS) with the parties that would fulfill those roles. What y=
ou
> describe as the RS entity in your solution is, in reality, an AS/RS pair.


Could be, but from a privacy perspective what matters is trust domains. If
the AS/RS pair is operated by the same entity or federation, then they are
just an RS from a privacy perspective. If that trust domain can aggregate
and make inferences from millions of policies belonging to ROs then we have
a privacy problem and are introducing a compromise in privacy in the name
of security. Let's try to avoid that.

What you describe as the RO's chosen AS is what GNAP calls out as an input
> into the AS.


It sounds like this "input" above is a policy management input rather than
a request input. My point just above is that the RS has no valid reason to
aggregate policies across ROs. It's not just an information minimization
issue but also separation of concerns that helps audits and zero-trust
architectures.


> This is why I keep saying that it's the interaction and consent phase of
> GNAP where things like SSI and policy descriptions and user-to-user
> delegation come into play. It's a different model from UMA, and I believe
> it's a model that works much better and fits reality.
>

I don't understand the basis of this statement. You're now introducing
user-to-user delegation. If User 1 is making requests to the RS (as RO)
then is User 2 making requests to what you're calling the AS/RS or to an AS
chosen by User 1 (RO)?

If I understand delegation via capabilities, User 3 would be making a
request to an AS chosen by User 2.

>
> The AS issues tokens that are trusted, the RS trusts and executes those
> tokens.


That's one way to do it, by defining the RO as the party that chooses and
specifies which AS is trusted. The other way is capabilities where the
token structure allows attenuation as determined by AS policy.

Giving an unknown party complete security control over your API is
> terrifying,


SSI is terrifying but the choice of the RO to delegate to an agent they
choose is what makes it self-sovereign and less terrifying. Agency can
always be pro se or fiduciary (to use the terms that apply to legal
representation). The difference between self-sovereign and fiduciary is
regulation and GNAP must not presume regulation is available or effective.


> and we've already seen that what actually happens is that people build in
> a internal AS to handle that case anyway. It's why we have things like
> security assertions and token exchange and application gateways.
>

I probably agree with you here but, as just above, this seems out-of-scope
for GNAP.

>
> Software delegation protocols, like OAuth and GNAP, aren't particularly
> well suited to the user delegation problem on their own, because you can'=
t
> have a user present a technical input.


If the "technical input" is what I'm calling the policy management input to
the AS, I have no problem with GNAP keeping it out of scope. That technical
input could be a web form or an API or, heaven forbid code or XACML. It's a
directive, not a request. What mitigates the privacy issue is that the RO
can choose the AS which tends to keep both the RS and the AS honest.
Obviously, there is a cost to the RO if switching ASs involves digesting
some other policy framework for the same RS and delegation goals but
standardizing the policy management data model is not a task we should take
up in GNAP core.

You can however make explicit places within the protocol for those inputs
> to be expressed, and that's what we are trying to do with GNAP.
>

Sure. Policy management inputs by the RO is what defines the AS from my
privacy perspective. How else could it be?

>
> You can absolutely build out the model I'm describing in privacy
> preserving ways, and I will argue that it will have significantly better
> security properties when compared to letting a user pick their AS.
>

I'm trying to understand how.

>
> So, let the RS dictate who it will listen to tokens from, and allow the A=
S
> to get its information about what those tokens are for, and when to issue
> them, from a variety of places.
>

Sorry. It's this "variety of places" that needs to be addressed as a
separation of concerns. Otherwise, we have a model that looks like the No
Fly List or a Credit Bureau where the policies and access rights of the RO
must be heavily regulated and security is a matter of "trust me" because
I've been audited by someone.

Security through zero trust architecture, as I understand it, requires
separation of concerns in the design phase, privilege minimization at the
transaction phase, as well as audit capability in the post-transaction
phase for every RS access by any end-user. I hope GNAP can describe its
role in ZTA as well as it's role in protecting privacy and human rights.

- Adrian



> - Justin
> ________________________________________
> From: Adrian Gropper [agropper@healthurl.com]
> Sent: Wednesday, September 29, 2021 12:27 AM
> To: Justin Richer
> Cc: jmandel@gmail.com; GNAP Mailing List
> Subject: Re: [GNAP] Accounting for Resource Owner in RS::AS interactions
>
> Josh and Justin,
>
> From a privacy perspective, your glee over not having the RO specify the
> AS to the RS is troubling. Why would the RS care about privacy unless
> mandated in some way? (I'm assuming mandates are out of scope for GNAP :-=
).
>
> The privacy problem is simply data minimization. As an RO, I should not
> have to share my policies with the RS because all the RS needs is the
> access token. If the RS is unwilling (or whatever you mean by "nervous",
> above) to trust a random AS simply because the RO trusts it, then we're
> left with capabilities. Imagine a typical RS serving millions of ROs. The
> policy database would be a gold mine in so many ways. Traffic analysis by
> the RO might still be a problem and will also need to be considered.
>
> If specifying the AS is not an option, the RO associated with a resource
> MUST be given a full set of capabilities on demand. The RO can then pass
> some or all of those to whatever AS they trust with their policies and th=
en
> that AS can issue attenuated access tokens with or without further
> delegation rights to any client they choose.
>
> GNAP core must choose either or both arbitrary AS or capabilities and
> maybe explain "nervous" in a way that people can understand.
>
> - Adrian
>
>
> On Tue, Sep 28, 2021 at 3:35 PM Justin Richer <jricher@mit.edu<mailto:
> jricher@mit.edu>> wrote:
> On Sep 28, 2021, at 3:20 PM, Josh Mandel <jmandel@gmail.com<mailto:
> jmandel@gmail.com>> wrote:
>
> Thanks Justin -- this whole response is extremely helpful. I particularly
> appreciate your very clear declaration of scope and intention:
>
> > That was what UMA tried to do with resource set registration, in a way.
> It was awkward from a technology perspective, and it made RS owners nervo=
us
> in a lot of ways to have a new AS introduced. I appreciate the ideology,
> but to this day I don=E2=80=99t believe it=E2=80=99s realistic to impleme=
nt or deploy.
>
> I agree entirely with this assessment, and I agree that some tighter
> degree of AS::RS coupling is needed to support real-world system design.
> Understanding that this UMA-like level of RS::AS flexibility is *not* the
> design intent for GNAP alleviates several of my concerns in one fell swoo=
p.
>
> To be clear, this is my own take on the goals of GNAP vs. experience with
> UMA, and the stance I=E2=80=99ve been advocating for here for some time. =
I think
> GNAP could allow that, in that GNAP doesn=E2=80=99t :require: a tight cou=
pling per
> se, but it also doesn=E2=80=99t assume a completely loose coupling or for=
ce the
> protocol through those hoops for all other use cases. UMA tried to do the
> latter, but just about every UMA deployment I saw didn=E2=80=99t make use=
 of that
> loose coupling, but you still had to go through the bits of the protocol
> that were built just in case that happened.
>
> It=E2=80=99s a goal of GNAP to learn from the good and bad of previous pr=
otocols,
> including UMA, and this is one area that I think we can improve.
>
> I should mention that for readers familiar with UMA (or at least... for
> me) this was not clear from reading GNAP; I came in assuming GNAP intende=
d
> to cover this space.
>
>
> I think what might be confusing here is that GNAP does intend to cover th=
e
> use case space of having a user delegate to another user, and externalizi=
ng
> a lot of the user rights aspects of the ore delegation process too. But
> instead of doing it through =E2=80=9Ceveryone brings their own AS=E2=80=
=9D type of things,
> GNAP is trying to do it through =E2=80=9Cthe AS can now be flexible enoug=
h to get a
> lot of different information=E2=80=9D. There were some slides on this a w=
hile back:
>
>
> https://datatracker.ietf.org/meeting/110/materials/slides-110-gnap-core-p=
rotocol-slides-00
>
> Starting at slide 35 you=E2=80=99ll see diagrams and points of some of th=
e
> alternatives. I know the meeting is recorded but I=E2=80=9Dm not sure whe=
re that
> lives if you want to review it.
>
>  - Justin
>
>
> On Tue, Sep 28, 2021 at 1:57 PM Justin Richer <jricher@mit.edu<mailto:
> jricher@mit.edu>> wrote:
> Hi Josh, thanks for bringing up this use case. An AS protecting several
> different unrelated RS=E2=80=99s is absolutely in scope, as this is a com=
mon
> deployment for OAuth 2 today. If I understand everything you=E2=80=99re a=
sking,
> currently GNAP would allow everything here, but it doesn=E2=80=99t specif=
y the
> details of every part of it. I=E2=80=99m going to reply to your scenarios=
 inline
> below. Apologies in advance for the long email, there=E2=80=99s a lot to =
unpack
> here!
>
> On Sep 27, 2021, at 1:56 PM, Josh Mandel <jmandel@gmail.com<mailto:
> jmandel@gmail.com>> wrote:
>
> Hi All,
>
> I'm interested in exploring how GNAP can/does/might account for the
> Resource Owner associated with specific protected resources, in scenarios
> where a single AS is trusted by multiple independent RS's (specifically i=
n
> scenarios where these RS's don't know or trust one another). Justin
> suggestd I might reach out here.
>
> My sense is that there are security considerations related to these
> questions, and perhaps some best practices or even areas where the core
> spec's data models could be enhanced to support these scenarios. I'm not
> proposing anything in particular at this stage; just trying to get the la=
y
> of the land.
>
> See assumptions and scenarios below.
>
> -Josh
>
> --
>
> Assumptions. I'll focus on a very simple API from the RS perspective: a
> standardized file access API where resource servers would like to make
> access control decisions based on access tokens whose RARs look something
> like:
>
> {
>   "type": "file-access",
>   "actions": ["read"],
>   "location": ["{{URL of one protected file}}"]
> }
>
> ---
>
> Scenario 1. Consider the case where a single RS manages files owned by
> multiple end users. The RS uses a single AS for all files (but the AS als=
o
> serves other untrusted RS's). Each RO is allowed to set authorization
> policies for their own files. In this scenario, how does the AS learn whi=
ch
> RO is responsible for setting policies for each file, so it can present
> Alice with a list of current policies, allow her to review/tweak access,
> etc?
>
> Setting and updating policies are outside the scope of GNAP core,
> including any interfaces for doing so.
>
> The policies are controlled at the AS and enforced by the RS. It=E2=80=99=
s really
> up to the individual deployments how and where to make that split: an AS
> could issue tokens that are simply tied to the specific data, the result =
of
> processing the security context of the delegation. Or the AS could send
> over all the context information to the RS to let it make its own decisio=
ns
> entirely, while the AS only has an idea of something existing and not who
> owns or controls it. Or something in between. And all of these are
> functionally the same whether you=E2=80=99re introspecting or using struc=
tured
> tokens, so I=E2=80=99m not going to get into the differences here.
>
> Doing this statically would be pretty easy, since the decisions can just
> be set in code and executed. Doing it dynamically feels a lot harder and
> more interesting. There are a lot of pieces that need to be fulfilled to
> even get to this point, and I=E2=80=99ll try to add some thoughts on each=
:
>
> - How does the AS know who Alice is? In OAuth, it=E2=80=99s assumed that =
Alice can
> =E2=80=9Clog in=E2=80=9D to the AS using her account. In GNAP we also all=
ow that but
> further let Alice have other ways of interacting with the AS, and doesn=
=E2=80=99t
> assume that Alice has an =E2=80=9Caccount=E2=80=9D at the AS in a traditi=
onal sense. Even
> so, it=E2=80=99s likely in this scenario that the AS will have some set o=
f
> identifiers that it has bound to Alice, which can be associated with that
> policy.
>
> - How does the RS know about the AS? GNAP does offer some discovery for
> the technical connections (
> https://www.ietf.org/archive/id/draft-ietf-gnap-resource-servers-01.html#=
name-rs-facing-as-discovery),
> but more important is the question of how the RS decides to trust the AS =
to
> protect any given resource. In your scenario this sounds like a
> configuration time decision, but it might not always be.
>
>
> (If an AS is confused on this point, there's a risk that Bob can set
> policies for Alice's files.) It's possible that the Resource Registration
> API could be extended to model this kind of ownership, if both servers ha=
ve
> some interoperable way to communicate user identities. I'm not sure if
> there are existing patterns to recommend here?
>
> - How does the AS know about what the RS is protecting? GNAP allows the R=
S
> to dynamically introduce a new resource set to the AS and get back a
> reference for it (
> https://www.ietf.org/archive/id/draft-ietf-gnap-resource-servers-01.html#=
name-registering-a-resource-set),
> and like you, I could see this being expanded with an ability to say =E2=
=80=9Cthis
> RO / set of ROs controls this resource set=E2=80=9D as part of that regis=
tration
> step. As long as the identifiers aligned, it would let the AS do some
> enforcement at token issuance time. But that really isn=E2=80=99t meant t=
o set
> policies, it=E2=80=99s meant to allow an RS to register a set of rights w=
ith a
> simple identifier (like a scope with a random value). Perhaps a different
> API is actually needed for policy management? And with that in mind, it
> would only be needed if we wanted the AS to make sure Bob is never prompt=
ed
> for one of Alice=E2=80=99s files at the AS, ever. If you don=E2=80=99t wa=
nt that
> information to leak to the AS, you can allow Bob to get an ineffective
> token for Alice=E2=80=99s file resource, and count on the RS to say =E2=
=80=9Chey you aren=E2=80=99t
> Alice, go away=E2=80=9D. Again, this goes back to where you draw the enfo=
rcement
> lines, which often has more to do with the nature of the API you=E2=80=99=
re
> protecting than anything.
>
> - How does the RS know how to associate a specific resource with =E2=80=
=9CAlice=E2=80=9D,
> and not =E2=80=9CBob"? This requires the RS to know who =E2=80=9CAlice=E2=
=80=9D is and know that a
> given file is hers to control. Doing this dynamically feels really diffic=
ult
>
> - How do the RS and AS agree that =E2=80=9CAlice=E2=80=9D is the same per=
son with the same
> identifiers? It=E2=80=99s obvious that simply trusting a global identifie=
r here
> won=E2=80=99t solve it on its own =E2=80=94 confusion remains trivial, it=
 would just take
> an RS to apply Bob=E2=80=99s global identifier to Alice=E2=80=99s data, o=
r for Bob to get
> assigned Alice=E2=80=99s identifier by a bad AS. But GNAP does have a who=
le notion
> of =E2=80=9Csubject identifiers=E2=80=9D that are scoped to the AS=E2=80=
=99s trust domain that can
> be used here: if the RS refers to Alice by an identifier known to the AS,
> and it trusts the AS to enforce that binding for Alice, then it will work=
.
>
>
> It's also very important for the RS to know, whenever it receives an
> access token, that the RS is an intended audience for that token. The RAR
> "locations" aren't enough to accomplish this, unless the AS has a reliabl=
e
> scheme in place to prevent one RS from crafting RARs that refer to resour=
ce
> locations within another RS.  And as the spec says, the AS introspection
> endpoint needs to ensure that access tokens are "appropriate for
> presentation at the identified RS". But that's especially challenging whe=
n
> a single access token can include multiple RARs, potentially intended for
> different RS's (which appears to be allowed by the base spec). And for
> structured token formats where introspection isn't required, the same
> questions hold; is there existing work on this?
>
> The =E2=80=98locations=E2=80=99 field on its own isn=E2=80=99t enough for=
 this, but it can be a
> place to hold identifiers for each RS. These need not be URLs to the APIs
> in question, though in practice I=E2=80=99d expect them to be. It then al=
so begs
> the question of how the RS identifiers came to be, and if they=E2=80=99re=
 then
> required for this type of processing.
>
>
> Scenario 2. Now consider the case where 10 organizations each manage a
> File Sharing RS. Each RS hosts files for multiple ROs, and the ROs get to
> choose which AS should be trusted for their own files. In this context, I
> suppose the workflow that introduces the RS to the AS can be mediated by
> the RO, so that the RO is logged into both systems when making the
> introduction, and the RS can essentially register a new client instance
> with the AS for each RO who decides to establish a connection. This patte=
rn
> has the advantage that there's no confusion about *who* controls a given
> resource -- but perhaps there are other, better ways to accomplish this.
>
> That was what UMA tried to do with resource set registration, in a way. I=
t
> was awkward from a technology perspective, and it made RS owners nervous =
in
> a lot of ways to have a new AS introduced. I appreciate the ideology, but
> to this day I don=E2=80=99t believe it=E2=80=99s realistic to implement o=
r deploy.
>
> I think a better pattern is to allow introduction of =E2=80=9CAlice=E2=80=
=9D at the AS
> level instead of at the RS level. In other words, Alice doesn=E2=80=99t b=
ring her
> own =E2=80=9CAS=E2=80=9D, but she could bring her own identity and policy=
 to the AS that
> protects the RS. That way the RS-AS relationship can be more tightly
> aligned, but the user capabilities are greatly expanded. It changes the f=
it
> of the components from what we might be used to, but I strongly believe
> there=E2=80=99s a lot of promise in this pattern. And in fact, we see a s=
imilar
> pattern in OAuth 2 today with API Gateways that can be configured to allo=
w
> federated logins, for example.
>
> Finally, I think there=E2=80=99s an even more interesting scenario: when =
Alice
> says that Bob should be able to access her stuff. Who does she tell and w=
ho
> needs to care? There are a lot of ways to slice this, but this jumps to m=
y
> mind immediately:
>
>  - Alice has a resource A at the RS, which is protected by the AS
>  - The AS has a policy that says Alice can authorize client instances to
> access resource A
>  - The AS also has an interface for Alice to set additional policies
>  - Alice says to the AS that a user identified as Bob can also access A
>  - When either Alice or Bob come to the AS, the AS can figure out from
> this policy to create an access token that can access A
>  - The RS doesn=E2=80=99t need to know about Alice or Bob at all, only wh=
ether the
> token is good for accessing A
>
>  =E2=80=94 Justin
>
>
>
>
>
> --
> TXAuth mailing list
> TXAuth@ietf.org<mailto:TXAuth@ietf.org>
> https://www.ietf.org/mailman/listinfo/txauth
>
>
> --
> TXAuth mailing list
> TXAuth@ietf.org<mailto:TXAuth@ietf.org>
> https://www.ietf.org/mailman/listinfo/txauth
>

--000000000000b5f35a05cd22465f
Content-Type: text/html; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable

<div dir=3D"ltr"><div>Please help me understand what&#39;s happening=C2=A0i=
n privacy terms so that I might be able to explain it to others.</div><div>=
<br></div><div>inline... and apologies=C2=A0for the long email</div><br><di=
v class=3D"gmail_quote"><div dir=3D"ltr" class=3D"gmail_attr">On Wed, Sep 2=
9, 2021 at 8:16 AM Justin Richer &lt;<a href=3D"mailto:jricher@mit.edu">jri=
cher@mit.edu</a>&gt; wrote:<br></div><blockquote class=3D"gmail_quote" styl=
e=3D"margin:0px 0px 0px 0.8ex;border-left-width:1px;border-left-style:solid=
;border-left-color:rgb(204,204,204);padding-left:1ex">Adrian, <br>
<br>
Once again, I believe you are conflating the technical roles of the protoco=
l (AS, RS) with the parties that would fulfill those roles. What you descri=
be as the RS entity in your solution is, in reality, an AS/RS pair. </block=
quote><div><br></div><div>Could be, but from a privacy perspective what mat=
ters is trust domains. If the AS/RS pair is operated by the same entity or =
federation, then they are just an RS from a privacy perspective. If that tr=
ust domain can aggregate and make inferences from millions of policies belo=
nging to ROs then we have a privacy problem and are introducing=C2=A0a comp=
romise in privacy in the name of security. Let&#39;s try to avoid that.</di=
v><div><br></div><blockquote class=3D"gmail_quote" style=3D"margin:0px 0px =
0px 0.8ex;border-left-width:1px;border-left-style:solid;border-left-color:r=
gb(204,204,204);padding-left:1ex">What you describe as the RO&#39;s chosen =
AS is what GNAP calls out as an input into the AS. </blockquote><div><br></=
div><div>It sounds like this &quot;input&quot; above is a policy management=
 input rather than a request input. My point just above is that the RS has =
no valid reason to aggregate policies across ROs. It&#39;s not just an info=
rmation minimization issue but also separation of concerns that helps audit=
s and zero-trust architectures.</div><div>=C2=A0</div><blockquote class=3D"=
gmail_quote" style=3D"margin:0px 0px 0px 0.8ex;border-left-width:1px;border=
-left-style:solid;border-left-color:rgb(204,204,204);padding-left:1ex">This=
 is why I keep saying that it&#39;s the interaction and consent phase of GN=
AP where things like SSI and policy descriptions and user-to-user delegatio=
n come into play. It&#39;s a different model from UMA, and I believe it&#39=
;s a model that works much better and fits reality. <br></blockquote><div><=
br></div><div>I don&#39;t understand the basis of this statement. You&#39;r=
e now introducing user-to-user delegation. If User 1 is making requests to =
the RS (as RO) then is User 2 making requests to what you&#39;re calling th=
e AS/RS or to an AS chosen by User 1 (RO)?<br></div><div><br></div><div>If =
I understand delegation via capabilities, User 3 would be making a request =
to an AS chosen by User 2.</div><blockquote class=3D"gmail_quote" style=3D"=
margin:0px 0px 0px 0.8ex;border-left-width:1px;border-left-style:solid;bord=
er-left-color:rgb(204,204,204);padding-left:1ex">
<br>
The AS issues tokens that are trusted, the RS trusts and executes those tok=
ens. </blockquote><div><br></div><div>That&#39;s one way to do it, by defin=
ing the RO as the party that chooses and specifies which AS is trusted. The=
 other way is capabilities where the token structure allows attenuation as =
determined by AS policy.=C2=A0</div><div><br></div><blockquote class=3D"gma=
il_quote" style=3D"margin:0px 0px 0px 0.8ex;border-left-width:1px;border-le=
ft-style:solid;border-left-color:rgb(204,204,204);padding-left:1ex">Giving =
an unknown party complete security control over your API is terrifying, </b=
lockquote><div><br></div><div>SSI is terrifying but the choice of the RO to=
 delegate to an agent they choose is what makes it self-sovereign and less =
terrifying. Agency can always be pro se or fiduciary (to use the terms that=
 apply=C2=A0to legal representation). The difference between self-sovereign=
 and fiduciary is regulation and GNAP must not presume regulation is availa=
ble or effective.</div><div>=C2=A0</div><blockquote class=3D"gmail_quote" s=
tyle=3D"margin:0px 0px 0px 0.8ex;border-left-width:1px;border-left-style:so=
lid;border-left-color:rgb(204,204,204);padding-left:1ex">and we&#39;ve alre=
ady seen that what actually happens is that people build in a internal AS t=
o handle that case anyway. It&#39;s why we have things like security assert=
ions and token exchange and application gateways. <br></blockquote><div><br=
></div><div>I probably agree with you here but, as just above, this seems o=
ut-of-scope for GNAP.=C2=A0</div><blockquote class=3D"gmail_quote" style=3D=
"margin:0px 0px 0px 0.8ex;border-left-width:1px;border-left-style:solid;bor=
der-left-color:rgb(204,204,204);padding-left:1ex">
<br>
Software delegation protocols, like OAuth and GNAP, aren&#39;t particularly=
 well suited to the user delegation problem on their own, because you can&#=
39;t have a user present a technical input.</blockquote><div><br></div><div=
>If the &quot;technical input&quot; is what I&#39;m calling the policy mana=
gement input to the AS, I have no problem with GNAP keeping it out of scope=
. That technical input could be a web form or an API or, heaven forbid code=
=C2=A0or XACML. It&#39;s a directive, not a request. What mitigates the pri=
vacy issue is that the RO can choose the AS which tends to keep both the RS=
 and the AS honest. Obviously, there is a cost to the RO if=C2=A0switching =
ASs involves digesting some other policy framework for the same RS and dele=
gation goals but standardizing the policy management data model is not a ta=
sk we should take up in GNAP core.</div><div><br></div><blockquote class=3D=
"gmail_quote" style=3D"margin:0px 0px 0px 0.8ex;border-left-width:1px;borde=
r-left-style:solid;border-left-color:rgb(204,204,204);padding-left:1ex"> Yo=
u can however make explicit places within the protocol for those inputs to =
be expressed, and that&#39;s what we are trying to do with GNAP. <br></bloc=
kquote><div><br></div><div>Sure. Policy management inputs by the RO is what=
 defines the AS from my privacy perspective. How else could it be?</div><bl=
ockquote class=3D"gmail_quote" style=3D"margin:0px 0px 0px 0.8ex;border-lef=
t-width:1px;border-left-style:solid;border-left-color:rgb(204,204,204);padd=
ing-left:1ex">
<br>
You can absolutely build out the model I&#39;m describing in privacy preser=
ving ways, and I will argue that it will have significantly better security=
 properties when compared to letting a user pick their AS. <br></blockquote=
><div><br></div><div>I&#39;m trying to understand how. =C2=A0</div><blockqu=
ote class=3D"gmail_quote" style=3D"margin:0px 0px 0px 0.8ex;border-left-wid=
th:1px;border-left-style:solid;border-left-color:rgb(204,204,204);padding-l=
eft:1ex">
<br>
So, let the RS dictate who it will listen to tokens from, and allow the AS =
to get its information about what those tokens are for, and when to issue t=
hem, from a variety of places. <br></blockquote><div><br></div><div>Sorry. =
It&#39;s this &quot;variety of places&quot; that needs to be addressed as a=
 separation of concerns. Otherwise, we have a model that looks like the No =
Fly List or a Credit Bureau where the policies and access rights of the RO =
must be heavily regulated and security is a matter of &quot;trust me&quot; =
because I&#39;ve been audited by someone.=C2=A0</div><div><br></div><div>Se=
curity through zero trust architecture, as I understand it, requires separa=
tion of concerns in the design phase, privilege minimization at the transac=
tion phase, as well as audit capability in the post-transaction phase for e=
very RS access by any end-user. I hope GNAP can describe its role in ZTA as=
 well as it&#39;s role in protecting privacy and human rights.</div><div><b=
r></div><div>- Adrian</div><div><br></div><div><br></div><blockquote class=
=3D"gmail_quote" style=3D"margin:0px 0px 0px 0.8ex;border-left-width:1px;bo=
rder-left-style:solid;border-left-color:rgb(204,204,204);padding-left:1ex">
<br>
- Justin<br>
________________________________________<br>
From: Adrian Gropper [<a href=3D"mailto:agropper@healthurl.com" target=3D"_=
blank">agropper@healthurl.com</a>]<br>
Sent: Wednesday, September 29, 2021 12:27 AM<br>
To: Justin Richer<br>
Cc: <a href=3D"mailto:jmandel@gmail.com" target=3D"_blank">jmandel@gmail.co=
m</a>; GNAP Mailing List<br>
Subject: Re: [GNAP] Accounting for Resource Owner in RS::AS interactions<br=
>
<br>
Josh and Justin,<br>
<br>
>From a privacy perspective, your glee over not having the RO specify the AS=
 to the RS is troubling. Why would the RS care about privacy unless mandate=
d in some way? (I&#39;m assuming mandates are out of scope for GNAP :-).<br=
>
<br>
The privacy problem is simply data minimization. As an RO, I should not hav=
e to share my policies with the RS because all the RS needs is the access t=
oken. If the RS is unwilling (or whatever you mean by &quot;nervous&quot;, =
above) to trust a random AS simply because the RO trusts it, then we&#39;re=
 left with capabilities. Imagine a typical RS serving millions of ROs. The =
policy database would be a gold mine in so many ways. Traffic analysis by t=
he RO might still be a problem and will also need to be considered.<br>
<br>
If specifying the AS is not an option, the RO associated with a resource MU=
ST be given a full set of capabilities on demand. The RO can then pass some=
 or all of those to whatever AS they trust with their policies and then tha=
t AS can issue attenuated access tokens with or without further delegation =
rights to any client they choose.<br>
<br>
GNAP core must choose either or both arbitrary AS or capabilities and maybe=
 explain &quot;nervous&quot; in a way that people can understand.<br>
<br>
- Adrian<br>
<br>
<br>
On Tue, Sep 28, 2021 at 3:35 PM Justin Richer &lt;<a href=3D"mailto:jricher=
@mit.edu" target=3D"_blank">jricher@mit.edu</a>&lt;mailto:<a href=3D"mailto=
:jricher@mit.edu" target=3D"_blank">jricher@mit.edu</a>&gt;&gt; wrote:<br>
On Sep 28, 2021, at 3:20 PM, Josh Mandel &lt;<a href=3D"mailto:jmandel@gmai=
l.com" target=3D"_blank">jmandel@gmail.com</a>&lt;mailto:<a href=3D"mailto:=
jmandel@gmail.com" target=3D"_blank">jmandel@gmail.com</a>&gt;&gt; wrote:<b=
r>
<br>
Thanks Justin -- this whole response is extremely helpful. I particularly a=
ppreciate your very clear declaration of scope and intention:<br>
<br>
&gt; That was what UMA tried to do with resource set registration, in a way=
. It was awkward from a technology perspective, and it made RS owners nervo=
us in a lot of ways to have a new AS introduced. I appreciate the ideology,=
 but to this day I don=E2=80=99t believe it=E2=80=99s realistic to implemen=
t or deploy.<br>
<br>
I agree entirely with this assessment, and I agree that some tighter degree=
 of AS::RS coupling is needed to support real-world system design. Understa=
nding that this UMA-like level of RS::AS flexibility is *not* the design in=
tent for GNAP alleviates several of my concerns in one fell swoop.<br>
<br>
To be clear, this is my own take on the goals of GNAP vs. experience with U=
MA, and the stance I=E2=80=99ve been advocating for here for some time. I t=
hink GNAP could allow that, in that GNAP doesn=E2=80=99t :require: a tight =
coupling per se, but it also doesn=E2=80=99t assume a completely loose coup=
ling or force the protocol through those hoops for all other use cases. UMA=
 tried to do the latter, but just about every UMA deployment I saw didn=E2=
=80=99t make use of that loose coupling, but you still had to go through th=
e bits of the protocol that were built just in case that happened.<br>
<br>
It=E2=80=99s a goal of GNAP to learn from the good and bad of previous prot=
ocols, including UMA, and this is one area that I think we can improve.<br>
<br>
I should mention that for readers familiar with UMA (or at least... for me)=
 this was not clear from reading GNAP; I came in assuming GNAP intended to =
cover this space.<br>
<br>
<br>
I think what might be confusing here is that GNAP does intend to cover the =
use case space of having a user delegate to another user, and externalizing=
 a lot of the user rights aspects of the ore delegation process too. But in=
stead of doing it through =E2=80=9Ceveryone brings their own AS=E2=80=9D ty=
pe of things, GNAP is trying to do it through =E2=80=9Cthe AS can now be fl=
exible enough to get a lot of different information=E2=80=9D. There were so=
me slides on this a while back:<br>
<br>
<a href=3D"https://datatracker.ietf.org/meeting/110/materials/slides-110-gn=
ap-core-protocol-slides-00" rel=3D"noreferrer" target=3D"_blank">https://da=
tatracker.ietf.org/meeting/110/materials/slides-110-gnap-core-protocol-slid=
es-00</a><br>
<br>
Starting at slide 35 you=E2=80=99ll see diagrams and points of some of the =
alternatives. I know the meeting is recorded but I=E2=80=9Dm not sure where=
 that lives if you want to review it.<br>
<br>
=C2=A0- Justin<br>
<br>
<br>
On Tue, Sep 28, 2021 at 1:57 PM Justin Richer &lt;<a href=3D"mailto:jricher=
@mit.edu" target=3D"_blank">jricher@mit.edu</a>&lt;mailto:<a href=3D"mailto=
:jricher@mit.edu" target=3D"_blank">jricher@mit.edu</a>&gt;&gt; wrote:<br>
Hi Josh, thanks for bringing up this use case. An AS protecting several dif=
ferent unrelated RS=E2=80=99s is absolutely in scope, as this is a common d=
eployment for OAuth 2 today. If I understand everything you=E2=80=99re aski=
ng, currently GNAP would allow everything here, but it doesn=E2=80=99t spec=
ify the details of every part of it. I=E2=80=99m going to reply to your sce=
narios inline below. Apologies in advance for the long email, there=E2=80=
=99s a lot to unpack here!<br>
<br>
On Sep 27, 2021, at 1:56 PM, Josh Mandel &lt;<a href=3D"mailto:jmandel@gmai=
l.com" target=3D"_blank">jmandel@gmail.com</a>&lt;mailto:<a href=3D"mailto:=
jmandel@gmail.com" target=3D"_blank">jmandel@gmail.com</a>&gt;&gt; wrote:<b=
r>
<br>
Hi All,<br>
<br>
I&#39;m interested in exploring how GNAP can/does/might account for the Res=
ource Owner associated with specific protected resources, in scenarios wher=
e a single AS is trusted by multiple independent RS&#39;s (specifically in =
scenarios where these RS&#39;s don&#39;t know or trust one another). Justin=
 suggestd I might reach out here.<br>
<br>
My sense is that there are security considerations related to these questio=
ns, and perhaps some best practices or even areas where the core spec&#39;s=
 data models could be enhanced to support these scenarios. I&#39;m not prop=
osing anything in particular at this stage; just trying to get the lay of t=
he land.<br>
<br>
See assumptions and scenarios below.<br>
<br>
-Josh<br>
<br>
--<br>
<br>
Assumptions. I&#39;ll focus on a very simple API from the RS perspective: a=
 standardized file access API where resource servers would like to make acc=
ess control decisions based on access tokens whose RARs look something like=
:<br>
<br>
{<br>
=C2=A0 &quot;type&quot;: &quot;file-access&quot;,<br>
=C2=A0 &quot;actions&quot;: [&quot;read&quot;],<br>
=C2=A0 &quot;location&quot;: [&quot;{{URL of one protected file}}&quot;]<br=
>
}<br>
<br>
---<br>
<br>
Scenario 1. Consider the case where a single RS manages files owned by mult=
iple end users. The RS uses a single AS for all files (but the AS also serv=
es other untrusted RS&#39;s). Each RO is allowed to set authorization polic=
ies for their own files. In this scenario, how does the AS learn which RO i=
s responsible for setting policies for each file, so it can present Alice w=
ith a list of current policies, allow her to review/tweak access, etc?<br>
<br>
Setting and updating policies are outside the scope of GNAP core, including=
 any interfaces for doing so.<br>
<br>
The policies are controlled at the AS and enforced by the RS. It=E2=80=99s =
really up to the individual deployments how and where to make that split: a=
n AS could issue tokens that are simply tied to the specific data, the resu=
lt of processing the security context of the delegation. Or the AS could se=
nd over all the context information to the RS to let it make its own decisi=
ons entirely, while the AS only has an idea of something existing and not w=
ho owns or controls it. Or something in between. And all of these are funct=
ionally the same whether you=E2=80=99re introspecting or using structured t=
okens, so I=E2=80=99m not going to get into the differences here.<br>
<br>
Doing this statically would be pretty easy, since the decisions can just be=
 set in code and executed. Doing it dynamically feels a lot harder and more=
 interesting. There are a lot of pieces that need to be fulfilled to even g=
et to this point, and I=E2=80=99ll try to add some thoughts on each:<br>
<br>
- How does the AS know who Alice is? In OAuth, it=E2=80=99s assumed that Al=
ice can =E2=80=9Clog in=E2=80=9D to the AS using her account. In GNAP we al=
so allow that but further let Alice have other ways of interacting with the=
 AS, and doesn=E2=80=99t assume that Alice has an =E2=80=9Caccount=E2=80=9D=
 at the AS in a traditional sense. Even so, it=E2=80=99s likely in this sce=
nario that the AS will have some set of identifiers that it has bound to Al=
ice, which can be associated with that policy.<br>
<br>
- How does the RS know about the AS? GNAP does offer some discovery for the=
 technical connections (<a href=3D"https://www.ietf.org/archive/id/draft-ie=
tf-gnap-resource-servers-01.html#name-rs-facing-as-discovery" rel=3D"norefe=
rrer" target=3D"_blank">https://www.ietf.org/archive/id/draft-ietf-gnap-res=
ource-servers-01.html#name-rs-facing-as-discovery</a>), but more important =
is the question of how the RS decides to trust the AS to protect any given =
resource. In your scenario this sounds like a configuration time decision, =
but it might not always be.<br>
<br>
<br>
(If an AS is confused on this point, there&#39;s a risk that Bob can set po=
licies for Alice&#39;s files.) It&#39;s possible that the Resource Registra=
tion API could be extended to model this kind of ownership, if both servers=
 have some interoperable way to communicate user identities. I&#39;m not su=
re if there are existing patterns to recommend here?<br>
<br>
- How does the AS know about what the RS is protecting? GNAP allows the RS =
to dynamically introduce a new resource set to the AS and get back a refere=
nce for it (<a href=3D"https://www.ietf.org/archive/id/draft-ietf-gnap-reso=
urce-servers-01.html#name-registering-a-resource-set" rel=3D"noreferrer" ta=
rget=3D"_blank">https://www.ietf.org/archive/id/draft-ietf-gnap-resource-se=
rvers-01.html#name-registering-a-resource-set</a>), and like you, I could s=
ee this being expanded with an ability to say =E2=80=9Cthis RO / set of ROs=
 controls this resource set=E2=80=9D as part of that registration step. As =
long as the identifiers aligned, it would let the AS do some enforcement at=
 token issuance time. But that really isn=E2=80=99t meant to set policies, =
it=E2=80=99s meant to allow an RS to register a set of rights with a simple=
 identifier (like a scope with a random value). Perhaps a different API is =
actually needed for policy management? And with that in mind, it would only=
 be needed if we wanted the AS to make sure Bob is never prompted for one o=
f Alice=E2=80=99s files at the AS, ever. If you don=E2=80=99t want that inf=
ormation to leak to the AS, you can allow Bob to get an ineffective token f=
or Alice=E2=80=99s file resource, and count on the RS to say =E2=80=9Chey y=
ou aren=E2=80=99t Alice, go away=E2=80=9D. Again, this goes back to where y=
ou draw the enforcement lines, which often has more to do with the nature o=
f the API you=E2=80=99re protecting than anything.<br>
<br>
- How does the RS know how to associate a specific resource with =E2=80=9CA=
lice=E2=80=9D, and not =E2=80=9CBob&quot;? This requires the RS to know who=
 =E2=80=9CAlice=E2=80=9D is and know that a given file is hers to control. =
Doing this dynamically feels really difficult<br>
<br>
- How do the RS and AS agree that =E2=80=9CAlice=E2=80=9D is the same perso=
n with the same identifiers? It=E2=80=99s obvious that simply trusting a gl=
obal identifier here won=E2=80=99t solve it on its own =E2=80=94 confusion =
remains trivial, it would just take an RS to apply Bob=E2=80=99s global ide=
ntifier to Alice=E2=80=99s data, or for Bob to get assigned Alice=E2=80=99s=
 identifier by a bad AS. But GNAP does have a whole notion of =E2=80=9Csubj=
ect identifiers=E2=80=9D that are scoped to the AS=E2=80=99s trust domain t=
hat can be used here: if the RS refers to Alice by an identifier known to t=
he AS, and it trusts the AS to enforce that binding for Alice, then it will=
 work.<br>
<br>
<br>
It&#39;s also very important for the RS to know, whenever it receives an ac=
cess token, that the RS is an intended audience for that token. The RAR &qu=
ot;locations&quot; aren&#39;t enough to accomplish this, unless the AS has =
a reliable scheme in place to prevent one RS from crafting RARs that refer =
to resource locations within another RS.=C2=A0 And as the spec says, the AS=
 introspection endpoint needs to ensure that access tokens are &quot;approp=
riate for presentation at the identified RS&quot;. But that&#39;s especiall=
y challenging when a single access token can include multiple RARs, potenti=
ally intended for different RS&#39;s (which appears to be allowed by the ba=
se spec). And for structured token formats where introspection isn&#39;t re=
quired, the same questions hold; is there existing work on this?<br>
<br>
The =E2=80=98locations=E2=80=99 field on its own isn=E2=80=99t enough for t=
his, but it can be a place to hold identifiers for each RS. These need not =
be URLs to the APIs in question, though in practice I=E2=80=99d expect them=
 to be. It then also begs the question of how the RS identifiers came to be=
, and if they=E2=80=99re then required for this type of processing.<br>
<br>
<br>
Scenario 2. Now consider the case where 10 organizations each manage a File=
 Sharing RS. Each RS hosts files for multiple ROs, and the ROs get to choos=
e which AS should be trusted for their own files. In this context, I suppos=
e the workflow that introduces the RS to the AS can be mediated by the RO, =
so that the RO is logged into both systems when making the introduction, an=
d the RS can essentially register a new client instance with the AS for eac=
h RO who decides to establish a connection. This pattern has the advantage =
that there&#39;s no confusion about *who* controls a given resource -- but =
perhaps there are other, better ways to accomplish this.<br>
<br>
That was what UMA tried to do with resource set registration, in a way. It =
was awkward from a technology perspective, and it made RS owners nervous in=
 a lot of ways to have a new AS introduced. I appreciate the ideology, but =
to this day I don=E2=80=99t believe it=E2=80=99s realistic to implement or =
deploy.<br>
<br>
I think a better pattern is to allow introduction of =E2=80=9CAlice=E2=80=
=9D at the AS level instead of at the RS level. In other words, Alice doesn=
=E2=80=99t bring her own =E2=80=9CAS=E2=80=9D, but she could bring her own =
identity and policy to the AS that protects the RS. That way the RS-AS rela=
tionship can be more tightly aligned, but the user capabilities are greatly=
 expanded. It changes the fit of the components from what we might be used =
to, but I strongly believe there=E2=80=99s a lot of promise in this pattern=
. And in fact, we see a similar pattern in OAuth 2 today with API Gateways =
that can be configured to allow federated logins, for example.<br>
<br>
Finally, I think there=E2=80=99s an even more interesting scenario: when Al=
ice says that Bob should be able to access her stuff. Who does she tell and=
 who needs to care? There are a lot of ways to slice this, but this jumps t=
o my mind immediately:<br>
<br>
=C2=A0- Alice has a resource A at the RS, which is protected by the AS<br>
=C2=A0- The AS has a policy that says Alice can authorize client instances =
to access resource A<br>
=C2=A0- The AS also has an interface for Alice to set additional policies<b=
r>
=C2=A0- Alice says to the AS that a user identified as Bob can also access =
A<br>
=C2=A0- When either Alice or Bob come to the AS, the AS can figure out from=
 this policy to create an access token that can access A<br>
=C2=A0- The RS doesn=E2=80=99t need to know about Alice or Bob at all, only=
 whether the token is good for accessing A<br>
<br>
=C2=A0=E2=80=94 Justin<br>
<br>
<br>
<br>
<br>
<br>
--<br>
TXAuth mailing list<br>
<a href=3D"mailto:TXAuth@ietf.org" target=3D"_blank">TXAuth@ietf.org</a>&lt=
;mailto:<a href=3D"mailto:TXAuth@ietf.org" target=3D"_blank">TXAuth@ietf.or=
g</a>&gt;<br>
<a href=3D"https://www.ietf.org/mailman/listinfo/txauth" rel=3D"noreferrer"=
 target=3D"_blank">https://www.ietf.org/mailman/listinfo/txauth</a><br>
<br>
<br>
--<br>
TXAuth mailing list<br>
<a href=3D"mailto:TXAuth@ietf.org" target=3D"_blank">TXAuth@ietf.org</a>&lt=
;mailto:<a href=3D"mailto:TXAuth@ietf.org" target=3D"_blank">TXAuth@ietf.or=
g</a>&gt;<br>
<a href=3D"https://www.ietf.org/mailman/listinfo/txauth" rel=3D"noreferrer"=
 target=3D"_blank">https://www.ietf.org/mailman/listinfo/txauth</a><br>
</blockquote></div></div>

--000000000000b5f35a05cd22465f--

