Re: [OAUTH-WG] draft-ietf-oauth-jwt-bearer != access tokens (was Re: draft-ietf-oauth-jwt-bearer Shepherd Write-up)

[Eve Maler <[e-mail]>]

I agree with Mike's point about flexibility. If OpenID Connect expects JWT and specific claim semantics, that makes sense in the context of its use case, but there are lots of other choices that can be made. For example, UMA's MTI token profile* defines a mechanism for conveying JSON-formatted permissions (à la Anil Saldhana's "entitlement" model** vs. "enforcement" for cloud authorization), and other token profiles could be created to convey JWT claims, the results of policy decisions, the applicable policies themselves, etc.

	Eve
* http://tools.ietf.org/html/draft-hardjono-oauth-umacore-09#section-3.3.2
** http://anil-identity.blogspot.com/2013/05/access-control-best-practices.html

On 25 Apr 2014, at 1:18 PM, Mike Jones <[e-mail]> wrote:

> To be clear, access tokens are opaque in OAuth and I don’t see any of us trying to change that in the general case.  Particular authorization servers may use JWTs as an access token format, but that’s their private choice.  I know of other authorization servers that have the access token value be an index into a local database table, which is just as legitimate a choice as using a structured access token.
>  
>                                                                 -- Mike
>  
> From: OAuth [mailto:[e-mail]] On Behalf Of Brian Campbell
> Sent: Friday, April 25, 2014 1:12 PM
> To: Bill Burke
> Cc: oauth
> Subject: Re: [OAUTH-WG] draft-ietf-oauth-jwt-bearer != access tokens (was Re: draft-ietf-oauth-jwt-bearer Shepherd Write-up)
>  
> I think it is kind of assumed, yeah. And JWT as it is gives you everything you need for that as long as the AS and RS can agree on keys, JWE and/or JWS, and how the claims will look. I suspect that's what most deployments are doing with JWT access tokens today. We are, or offer JWS + JWT access tokens as an option in product anyway, and I believe many others are doing the same.
> 
> IHMO getting everyone to agree on the specific claims etc. needed for a standardized JWT access token is a bit of a rat's nest, which is why there's not been much progress in that area.
> 
> 
> 
> 
>  
> 
> On Fri, Apr 25, 2014 at 1:51 PM, Bill Burke <[e-mail]> wrote:
> Thank you.  Thats what I thought.  Is it just assumed JWT would/might be used an access token format for Bearer token auth?  Or is there another draft somewhere for that?  Is anybody out there using JWS + JWT as a access token format?
> 
> 
> On 4/25/2014 2:59 PM, Brian Campbell wrote:
> draft-ietf-oauth-jwt-bearer is only about interactions (client
> authentication and JWT as an authorization grant) with the token
> endpoint and doesn't define JWT style access tokens.
> 
> 
> On Fri, Apr 25, 2014 at 12:51 PM, Bill Burke <[e-mail]
> <mailto:[e-mail]>> wrote:
> 
>     Red Hat Keycloak [1] only supports basic auth for client
>     authentication as suggested in the OAuth 2 spec.  But our access
>     tokens are JWS signed JWTs.
> 
>     Does draft-ietf-oauth-jwt-bearer relate to OAuth Bearer token auth
>     [2]?  Or is there another document I should be following?  I'd like
>     to see what other claims are being discussed related to JWT-based
>     access tokens and may have some additional access token claims we've
>     been experimenting with others might be interested in.
> 
>     Also, I'm not sure yet if we'll implement
>     draft-ietf-oauth-jwt-bearer to authenticate clients.  A lot of our
>     initial users are more interested in public clients and/or the
>     implicit flow as they are writing a lot of pure javascript apps
>     served up by simple static web servers.
> 
>     [1] http://keycloak.org
>     [2] http://tools.ietf.org/html/__rfc6750
>     <http://tools.ietf.org/html/rfc6750>
> 
> 
> -- 
> Bill Burke
> JBoss, a division of Red Hat
> http://bill.burkecentral.com
>  
> _______________________________________________
> OAuth mailing list
> [e-mail]
> https://www.ietf.org/mailman/listinfo/oauth


Eve Maler                                  http://www.xmlgrrl.com/blog
[phone]                         http://www.twitter.com/xmlgrrl