Re: [Gen-art] Gen-ART Last Call review of draft-ietf-ace-dtls-authorize-12

Jim Schaad <> Sun, 19 July 2020 22:30 UTC

Return-Path: <>
Received: from localhost (localhost []) by (Postfix) with ESMTP id 453383A08FD; Sun, 19 Jul 2020 15:30:00 -0700 (PDT)
X-Virus-Scanned: amavisd-new at
X-Spam-Flag: NO
X-Spam-Score: -1.899
X-Spam-Status: No, score=-1.899 tagged_above=-999 required=5 tests=[BAYES_00=-1.9, SPF_HELO_NONE=0.001, SPF_PASS=-0.001, URIBL_BLOCKED=0.001] autolearn=ham autolearn_force=no
Received: from ([]) by localhost ( []) (amavisd-new, port 10024) with ESMTP id wBgYR-9EaAxV; Sun, 19 Jul 2020 15:29:58 -0700 (PDT)
Received: from ( []) (using TLSv1.2 with cipher ECDHE-RSA-AES256-SHA384 (256/256 bits)) (No client certificate requested) by (Postfix) with ESMTPS id F1CA33A08FA; Sun, 19 Jul 2020 15:29:57 -0700 (PDT)
Received: from Jude ( by ( with Microsoft SMTP Server (TLS) id 15.0.1395.4; Sun, 19 Jul 2020 15:29:51 -0700
From: Jim Schaad <>
To: 'Paul Kyzivat' <>,
CC: 'General Area Review Team' <>,
References: <>
In-Reply-To: <>
Date: Sun, 19 Jul 2020 15:29:48 -0700
Message-ID: <000001d65e1c$1eed9430$5cc8bc90$>
MIME-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Content-Transfer-Encoding: 7bit
X-Mailer: Microsoft Outlook 16.0
Thread-Index: AQHl0TW2xBQNibu5hPYR+rhuY3INLqjwer3g
Content-Language: en-us
X-Originating-IP: []
Archived-At: <>
Subject: Re: [Gen-art] Gen-ART Last Call review of draft-ietf-ace-dtls-authorize-12
X-Mailman-Version: 2.1.29
Precedence: list
List-Id: "GEN-ART: General Area Review Team" <>
List-Unsubscribe: <>, <>
List-Archive: <>
List-Post: <>
List-Help: <>
List-Subscribe: <>, <>
X-List-Received-Date: Sun, 19 Jul 2020 22:30:00 -0000

> -----Original Message-----
> From: Paul Kyzivat <>
> Sent: Sunday, July 19, 2020 1:24 PM
> To:
> Cc: General Area Review Team <>
> Subject: Gen-ART Last Call review of draft-ietf-ace-dtls-authorize-12
> I am the assigned Gen-ART reviewer for this draft. The General Area Review
> Team (Gen-ART) reviews all IETF documents being processed by the IESG for
> the IETF Chair.  Please treat these comments just like any other last call
> comments.
> For more information, please see the FAQ at
> <>.
> Document: draft-ietf-ace-dtls-authorize-12
> Reviewer: Paul Kyzivat
> Review Date: 2020-07-19
> IETF LC End Date: 2020-07-20
> IESG Telechat date: ?
> Summary:
> This draft is on the right track but has open issues, described in the
> General:
> Issues:
> Major: 2
> Minor: 1
> Nits:  1
> 1) MAJOR: Management of token storage in RS
> There seems to be an expectation that when the client uploads an access
> that the RS will retain it until the client attempts to establish a DTLS
> association. This seems to require some sort of management of token
> in the RS. The only discussion I can find of this issue is the following
in section 7:
>     ... A similar issue exists with the
>     unprotected authorization information endpoint where the resource
>     server needs to keep valid access tokens until their expiry.
>     Adversaries can fill up the constrained resource server's internal
>     storage for a very long time with interjected or otherwise retrieved
>     valid access tokens.
> This seems to imply a normative requirement to keep tokens until their
> But I find no supporting normative requirements about this. And, this
> only presents it as a DoS attack, rather than a potential problem with
> usage.
> ISTM that there is an implied requirement that the RS have the capacity to
> store one access token for every PoP key of every authorized client.
> If so, that ought to be stated. If not, then some other way of bounding
> ought to be discussed.

In section 5.8.1 of draft-ietf-ace-oauth-authz is the sentence "The RS MUST
be prepared to store at least one access token for future use."   When this
was put in, this is exactly what we were discussing.  There is no
requirement that an RS needs to store two access tokens for future use.  I
think this means that there is a strongly bounded requirement on storage.

Authors - It might be worthwhile to re-iterate this requirement in both of
the profile documents.

> 2) MAJOR: Missing normative language
> I found several places where the text seems to suggest required behavior
> fails to do so using normative language:
> * In section 3.3:
>     ... Instead of
>     providing the keying material in the access token, the authorization
>     server includes the key identifier in the "kid" parameter, see
>     Figure 7.  This key identifier enables the resource server to
>     calculate the symmetric key used for the communication with the
>     client using the key derivation key and a KDF to be defined by the
>     application, for example HKDF-SHA-256.  The key identifier picked by
>     the authorization server needs to be unique for each access token
>     where a unique symmetric key is required.
>     ...
>     Use of a unique (per resource server) "kid" and the use of a key
>     derivation IKM that is unique per authorization server/resource
>     server pair as specified above will ensure that the derived key is
>     not shared across multiple clients.
> The uniqueness seems to be a requirement. Perhaps "needs to be unique"
> should be "MUST be unique". (And something similar for the IKM.)
> * Also in section 3.3:
>     All CBOR data types are encoded in CBOR using preferred serialization
>     and deterministic encoding as specified in Section 4 of
>     [I-D.ietf-cbor-7049bis].  This implies in particular that the "type"
>     and "L" components use the minimum length encoding.  The content of
>     the "access_token" field is treated as opaque data for the purpose of
>     key derivation.
> IIUC the type of serialization and encoding is a requirement. Will need
> rewording to make it so.
> * In section 3.3.1:
>     ... To
>     be consistent with the recommendations in [RFC7252] a client is
>     expected to offer at least the ciphersuite TLS_PSK_WITH_AES_128_CCM_8
>     [RFC6655] to the resource server.
> I think "is expected" should be "MUST".

The rule would be MUST implement not MUST offer.  A client could offer a
completely different ciphersuite that is consistently used in the group and
that would be fine.  

> * Also in section 3.3.1:
>     ... This
>     specification assumes that the access token is a PoP token as
>     described in [I-D.ietf-ace-oauth-authz] unless specifically stated
>     otherwise.
> I think "assumes ... unless" should be "MUST ... unless".
> * Also in section 3.3.1:
>     ... New access tokens issued by the
>     authorization server are supposed to replace previously issued access
>     tokens for the respective client.
> Is this normative? Should "are supposed to" be "MUST"?
> 3) MINOR: Insufficient specification
> (I'm waffling whether this is minor or major.)
> There are a couple of places where what seem to be requirements are stated
> too vaguely to be implemented consistently:
> * In the previously mentioned paragraph in 3.3.1:
>     ... This
>     specification assumes that the access token is a PoP token as
>     described in [I-D.ietf-ace-oauth-authz] unless specifically stated
>     otherwise.
> The "unless specifically stated otherwise" is too vague to be normative.
> How would the alternative be indicated? Is this an escape hatch for future
> extensions? If so, it needs more work to make that clear and to open a
path for
> that future work.
> * Also in section 3.3.1:
>     ... The resource server therefore must
>     have a common understanding with the authorization server how access
>     tokens are ordered.
> The last statement ("must have a common understanding") is mysterious.
> IIUC section 4 is covering the same topic in a less mysterious way.

I am seeing this in section 3.4 - did I miss something?

> 4) NIT: Subsection organization
> Both 3.2 and 3.3 share a common structure:
> * The section begins with discussion of the interaction between the client
> the AS.
> * it is followed by a subsection discussing the interaction between the
> and the RS.
> It is odd to have a section with a single subsection. And the structure
isn't easily
> discerned from the TOC.
> I suggest it would be clearer if each of these sections had *two*
> one covering the AS interactions and the other the RS interactions. IOW,
put the
> material covering the AS interactions into a new subsection.