Re: [OAUTH-WG] oauth - Requested sessions have been scheduled for IETF 98

[Denis <[e-mail]>]

Hi Nat,


I have several comments on draft-sakimura-oauth-jpop-01 related to 
security or privacy.


1.The abstract states:

Only the party in possession of a corresponding cryptographic key with 
the Jpop token can use it to get access
to the associated resources unlike in the case of the bearer token 
described in [RFC6750] where any party
in possession of the access token can access the resource.

This is incorrect.

Replace with:

Any party able to use a corresponding private cryptographic key with the 
Jpop token can use it to get access
to the associated resources unlike in the case of the bearer token 
described in [RFC6750] where any party
in possession of the access token can access the resource.

2.In section 3, the text states:

audThe identifier of the resource server.

According to the content of RFC 7800:

The "aud" (audience) claim identifies the recipients that the JWT is 
intended for. The interpretation of audience values is application specific.

Replace with:

audThe recipients that the JWT is intended for (the interpretation of 
audience values is application specific).

3.In section 3, the text states:

cnfThe confirmation method.

Their semantics are defined in [RFC7519] and [RFC7800]


This is incorrect: cnf is neither defined in [RFC7519] nor in [RFC7800].

4.In section 6.2, the text states:

For this, the following steps are taken:

1.The client prepares a nonce.

2.The client creates JWS compact serialization over the nonce


    JSON Web Token Claims are listed at:
    https://www.iana.org/assignments/jwt/jwt.xhtml

"nonce" has not been defined by the IANA, but is mentioned in OpenID 
Connect Core 1.0 incorporating errata set 1. It is described as :

nonce

String value used to associate a Client session with an ID Token, and to 
mitigate replay attacks. The value is passed through
unmodified from the Authentication Request to the ID Token. If present 
in the ID Token, Clients MUST verify that the nonce
Claim Value is equal to the value of the nonceparameter sent in the 
Authentication Request. If present in the Authentication Request,
Authorization Servers MUST include a nonceClaim in the ID Token with the 
Claim Value being the nonce value sent in the Authentication Request.
Authorization Servers SHOULD perform no other processing on noncevalues 
used. The noncevalue is a case sensitive string.

I have several observations:

a)there is some difficulty to mandate the use of a parameter that is not 
registered by IANA.

b)the further processing of the nonce is not indicated in the text

c) The last sentence from the above description states: "Authorization 
Servers SHOULD perform no other processing on noncevalues used"
There is a practical problem with such a sentence since Authorization 
Servers would need to remember nonces for ever.
Either that sentence should be deleted or the nonce shall be only used 
with a UTC time parameter included in the Authentication Request.

In any case, the definition of a nonce as specified in OpenID Connect 
Core 1.0 incorporating errata set 1 should not be used and another 
parameter
(e.g. rdn for random) should be defined and registered by IANA and used 
in combination with a UTC time parameter included in the Authentication 
Request.
In this way, only the rdn received during the last X minutes will need 
to be remembered by the Authorization Servers.


5.The title of section 9.1 is: "Certificate validation"

Change the title of this section into :

"9.1. Common Name Constrained Token"

6.In section 9.1, the text states:

The "cn" JWT confirmation method relies its security property on the

X.509 client certificate authentication.

Replace with:

The "cn" JWT confirmation method relies its security property by the 
inclusion of the Common Name (CN)
that is part of the Distinguished Name (DN) of an X.509 certificate. The 
JWT is linked to the common name
included in the certificate. Such a method is not privacy friendly since 
it allows an easy linkage between
all the accounts of a given user on different resource servers.

7.Add a new section 9.2 to deal with the case of the cid.

Proposed text:

9.2. Client ID Constrained Token

The "cid" JWT confirmation method relies its security property on the 
assumption that the cid legitimately
used by one server cannot be used by another user. It also relies on the 
assumption that the authentication data
associated with "cid" combined with the "iss" will only be used by the 
legitimate user. This method is ineffective
in case of a collusion between two users, since one user can perform all 
the computations needed by the other user.

8.In section 9.2, the text states:

The client’s secret key must be kept securely. Otherwise, the notion of 
PoP breaks down.

The PKIX group from the IETF is using the vocabulary private key / 
public key when asymmetric cryptography is being used
and secret key when symmetric algorithms are being used (let us call a 
spade a spade).

However, keeping a client's private key securely is not the right 
wording either. If the key is kept securely in a secure element
(e.g. smart card), this is not enough, since the holder of the secure 
element may use this key for himself ... or worse for the benefit of 
someone else.

Proposed change :

9.3. Key Constrained Token

This method has four variants.

When the JWT contains a jwk, the JWT confirmation method relies its 
security property on the assumption that the private key
associated with the public key contained in the access token will only 
be used by the legitimate user. In order to avoid an easy linkage
between user's accounts, this method presents the advantage that the key 
pair can be changed for every JWT. However, this method
is ineffective in case of a collusion between two users, since one user 
can perform all the computations needed by the other user.

When the JWT contains a jwkt#s256, the server must have a prior 
knowledge of the public key and the method relies its security property
on the assumption that the private key associated with the public key 
contained in the access token will only be used by the legitimate user.
Hence, this method is ineffective in case of a collusion between two 
users, since one user can perform all the computations needed
by the other user.

When the JWT contains a x5t#s256, the server must have a prior knowledge 
of the public key certificate. The JWT is then linked to a hash value
of a certificate included in the JWT. The server knows a unique 
identifier of the user. Such a method is not privacy friendly since it 
allows
an easy linkage between all the accounts of a given user on different 
resource servers.

When the JWT contains a jwe, the JWT confirmation method relies its 
security property on the assumption that the secret key included
in the JWT will only be used by the legitimate user. In order to avoid 
an easy linkage between user's accounts, this method presents
the advantage that the secret key can be changed for every JWT. However, 
this method is ineffective in case of a collusion between two users,
since one user can perform all the computations needed by the other user.

9.The text states in section 9.3:

9.3.Audi_a_nce Restriction

When using the signature method the client must specify to the AS the 
aud it intends to send the token to, so that it can be included in the AT.

A malicious RS could receive a AT with no aud or a logical audience and 
then replay the AT and jws-on-nonce to the actual server.


Proposed change in order to address privacy concerns :

9.4.Audi_e_nce Restriction

When using the signature method, the client must specify to the AS the 
aud it intends to send the token to, so that it can be included in the AT.

RFC 7800 states that the interpretation of audience values is 
application specific. If a fixed value is being used, e.g. a URL of the 
server,
then the authorization server can easily know where the access tokens 
will be used and thus is in a position to act as Big Brother.
It is thus recommended to use a different value in the aud claims for 
each access token that contains no semantics in it but that the resource 
server
can easily recognize.//

If a malicious RS receives an AT with no aud or a logical audience in it 
then it can replay the AT and jws-on-nonce to another server.

Denis


> HI Chairs,
>
> I would also like to ask 5 min. on Monday (as I cannot be on Friday) for
> The OAuth 2.0 Authorization Framework: JWT Pop Token Usage [1].
>
> [1] https://tools.ietf.org/html/draft-sakimura-oauth-jpop-01
>
> It is capturing strong and rather urgent demands from the financial 
> sector and would be great if it can be considered in the WG.
>
> Best,
>
> Nat Sakimura
>
> On Tue, Mar 21, 2017 at 10:28 PM Antonio Sanso <[e-mail] 
> <mailto:[e-mail]>> wrote:
>
>     hi Torsten,
>
>     good one. I personally I am looking forward to see this particular
>     document find its way.
>
>     IMHO this is something much needed.
>
>     regards
>
>     antonio
>
>     On Mar 21, 2017, at 2:08 PM, Torsten Lodderstedt
>     <[e-mail] <mailto:[e-mail]>> wrote:
>
>>     Hi Chairs,
>>
>>     I would like to request 5 minutes on Monday to briefly present
>>     the status of the security document. This is mainly to raise
>>     awareness in the group since I didn’t get that much input on it
>>     since Seoul.
>>
>>     kind regards,
>>     Torsten.
>>
>>>     Am 18.03.2017 um 01:52 schrieb Mike Jones
>>>     <[e-mail] <mailto:[e-mail]>>:
>>>
>>>     Hi Chairs,
>>>
>>>     I'd like to request that the following presentations be added to
>>>     the agenda:
>>>
>>>     OAuth Token Exchange (draft-ietf-oauth-token-exchange) - Mike
>>>     Jones - 15 minutes
>>>     OAuth Authorization Server Metadata (draft-ietf-oauth-discovery)
>>>     - Mike Jones - 15 minutes
>>>
>>>     I'd also talked with Brian Campbell and I think he wants to lead
>>>     this discussion, in part based on his implementation experience:
>>>
>>>     OAuth Token Binding (draft-ietf-oauth-token-binding) - Brian
>>>     Campbell - 30 minutes
>>>
>>>     (Brian may suggest a different amount of time)
>>>
>>>     I agree that William Dennis should present about the OAuth
>>>     Device Flow (draft-ietf-oauth-device-flow).
>>>
>>>     For completeness, I don't think a presentation is needed about
>>>     OAuth AMR Values (draft-ietf-oauth-amr-values) because it's now
>>>     completed its IESG review.
>>>
>>>     I'll look forward to seeing many of you in just over a week!
>>>
>>>     -- Mike
>>>
>>>     -----Original Message-----
>>>     From: OAuth [mailto:[e-mail]] On Behalf Of "IETF
>>>     Secretariat"
>>>     Sent: Friday, March 3, 2017 3:55 PM
>>>     To: [e-mail] <mailto:[e-mail]>;
>>>     [e-mail] <mailto:[e-mail]>
>>>     Cc: [e-mail] <mailto:[e-mail]>
>>>     Subject: [OAUTH-WG] oauth - Requested sessions have been
>>>     scheduled for IETF 98
>>>
>>>     Dear Stephanie McCammon,
>>>
>>>     The session(s) that you have requested have been scheduled.
>>>     Below is the scheduled session information followed by the
>>>     original request.
>>>
>>>     oauth Session 1 (2:30:00)
>>>       Friday, Morning Session I 0900-1130
>>>       Room Name: Zurich C size: 100
>>>       ---------------------------------------------
>>>       oauth Session 2 (1:00:00)
>>>       Monday, Afternoon Session III 1710-1810
>>>       Room Name: Zurich C size: 100
>>>       ---------------------------------------------
>>>
>>>
>>>
>>>     Request Information:
>>>
>>>
>>>     ---------------------------------------------------------
>>>     Working Group Name: Web Authorization Protocol Area Name:
>>>     Security Area Session Requester: Stephanie McCammon
>>>
>>>     Number of Sessions: 2
>>>     Length of Session(s):  2.5 Hours, 1 Hour Number of Attendees: 50
>>>     Conflicts to Avoid:
>>>     First Priority: saag core tls tokbind
>>>
>>>
>>>
>>>
>>>     People who must be present:
>>>     Hannes Tschofenig
>>>     Kathleen Moriarty
>>>     Derek Atkins
>>>
>>>     Resources Requested:
>>>     Projector in room
>>>
>>>     Special Requests:
>>>     Please avoid conflict with sec area BoFs.
>>>     ---------------------------------------------------------
>>>
>>>     _______________________________________________
>>>     OAuth mailing list
>>>     [e-mail] <mailto:[e-mail]>
>>>     https://na01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fwww.ietf.org%2Fmailman%2Flistinfo%2Foauth&data=02%7C01%7C%7C254d07b9729a4cfc8dd408d4705b73a2%7Cfa7b1b5a7b34438794aed2c178decee1%7C0%7C0%7C636256985463058106&sdata=FYIqTvgn1%2Fpjyqw%2BtGhDWgiB0G0ATuL30ap%2B3bLX6aQ%3D&reserved=0
>>>
>>>
>>>     _______________________________________________
>>>     OAuth mailing list
>>>     [e-mail] <mailto:[e-mail]>
>>>     https://na01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fwww.ietf.org%2Fmailman%2Flistinfo%2Foauth&data=02%7C01%7C%7C254d07b9729a4cfc8dd408d4705b73a2%7Cfa7b1b5a7b34438794aed2c178decee1%7C0%7C0%7C636256985463058106&sdata=FYIqTvgn1%2Fpjyqw%2BtGhDWgiB0G0ATuL30ap%2B3bLX6aQ%3D&reserved=0
>>
>>     _______________________________________________
>>     OAuth mailing list
>>     [e-mail] <mailto:[e-mail]>
>>     https://na01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fwww.ietf.org%2Fmailman%2Flistinfo%2Foauth&data=02%7C01%7C%7C254d07b9729a4cfc8dd408d4705b73a2%7Cfa7b1b5a7b34438794aed2c178decee1%7C0%7C0%7C636256985463068122&sdata=5CIJnWs2VdLM9FUWt%2FWlOxIilp5N2vfr7b9elwhL%2BA4%3D&reserved=0
>
>     _______________________________________________
>     OAuth mailing list
>     [e-mail] <mailto:[e-mail]>
>     https://www.ietf.org/mailman/listinfo/oauth
>
> -- 
>
> Nat Sakimura
>
> Chairman of the Board, OpenID Foundation
>
>
>
> _______________________________________________
> OAuth mailing list
> [e-mail]
> https://www.ietf.org/mailman/listinfo/oauth