Looking at the spec, the problem is you can still only authorize one api at a 
time.  You couldn't specify multiple audience apis and match them up with 
scopes.

A while ago I did start to write some stuff up about a structured scope 
specification where scope becomes a JSON multi-value structure so that multiple 
scopes and end-points could be defined.

I abandoned it after seeing that the same thing could be accomplished (as 
Google did) by defining the relationship during registration. I agree, it 
precludes being able to change on the fly, but came to the conclusion that is 
pretty rare (we're not talking about browsers for the most part).

Aside…this is why I include "target" in the SCIM Registration draft.

Phil

@independentid
www.independentid.com
[email protected]







On 2013-08-21, at 9:49 AM, Justin Richer <[email protected]> wrote:

> I think it makes sense to have both, and this would parallel what we've got 
> with the "scope" parameter today. At registration, the client is saying "this 
> is what I want to be able to use" and the server is saying "this is what 
> you're allowed to use". At auth time, the client is saying "this is what I'm 
> using now" and the user is saying "this is what you're authorized to use now".
> 
> If there were a standardized "audience" parameter at the auth endpoint, it 
> could easily be added to the registration's client model in parallel.
> 
> -- Justin
> 
> On 08/21/2013 12:46 PM, Phil Hunt wrote:
>> Yes.  The trade off is that each client_id becomes associated with a target.
>> 
>> Phil
>> 
>> @independentid
>> www.independentid.com
>> [email protected]
>> 
>> 
>> 
>> 
>> 
>> 
>> 
>> On 2013-08-21, at 9:45 AM, Anthony Nadalin <[email protected]> wrote:
>> 
>>> I think binding audience at registration time is to limiting as we see 
>>> audience being on a per token request level and also see the audience being 
>>> part of the restrictions for "act as" or "on behalf of" support
>>> 
>>> -----Original Message-----
>>> From: [email protected] [mailto:[email protected]] On Behalf Of 
>>> Hannes Tschofenig
>>> Sent: Wednesday, August 21, 2013 9:41 AM
>>> To: Phil Hunt
>>> Cc: <[email protected]>
>>> Subject: Re: [OAUTH-WG] Audience parameter in authorization flow
>>> 
>>> That's certainly true although the referenced document did not talk about 
>>> the registration phase but rather about the time when the client talks to 
>>> the authorization server to obtain an access token.
>>> 
>>> Maybe UMA has provided a story for this already...
>>> 
>>> On 08/21/2013 06:35 PM, Phil Hunt wrote:
>>>> This could be bound up in the client registration process since oauth 
>>>> clients don't authorize for random "targets".
>>>> 
>>>> Phil
>>>> 
>>>> @independentid
>>>> www.independentid.com
>>>> [email protected]
>>>> 
>>>> 
>>>> 
>>>> 
>>>> 
>>>> 
>>>> 
>>>> On 2013-08-21, at 9:30 AM, "Tschofenig, Hannes (NSN - FI/Espoo)" 
>>>> <[email protected]> wrote:
>>>> 
>>>>> Hi Sergey,
>>>>> 
>>>>> The idea of the audience was to provide a way for the client to indicate 
>>>>> the resource server it wants to talk to explicitly rather than 
>>>>> overloading the scope field. We certainly need that capability for the 
>>>>> MAC token work.
>>>>> 
>>>>> The audience information is provided when the client interacts with the 
>>>>> AS.
>>>>> 
>>>>> Ciao
>>>>> Hannes
>>>>> 
>>>>> 
>>>>>> -----Original Message-----
>>>>>> From: [email protected] [mailto:[email protected]] On
>>>>>> Behalf Of ext Sergey Beryozkin
>>>>>> Sent: Sunday, August 18, 2013 6:32 PM
>>>>>> To: <[email protected]>
>>>>>> Subject: [OAUTH-WG] Audience parameter in authorization flow
>>>>>> 
>>>>>> Hi Hannes, All,
>>>>>> 
>>>>>> Regarding [1], where would you expect an audience parameter be
>>>>>> provided during the authorization flow ?
>>>>>> 
>>>>>> It appears to me it should be provided during the initial redirect
>>>>>> (similarly to a parameter like redirect_uri).
>>>>>> 
>>>>>> Also, would it make sense to support pre-registered audience values,
>>>>>> example, a client registers and specifies an audience during the
>>>>>> registration ?
>>>>>> 
>>>>>> Thanks, Sergey
>>>>>> 
>>>>>> [1] http://tools.ietf.org/html/draft-tschofenig-oauth-audience-00
>>>>>> _______________________________________________
>>>>>> OAuth mailing list
>>>>>> [email protected]
>>>>>> https://www.ietf.org/mailman/listinfo/oauth
>>>>> _______________________________________________
>>>>> OAuth mailing list
>>>>> [email protected]
>>>>> https://www.ietf.org/mailman/listinfo/oauth
>>>> _______________________________________________
>>>> OAuth mailing list
>>>> [email protected]
>>>> https://www.ietf.org/mailman/listinfo/oauth
>>>> 
>>> _______________________________________________
>>> OAuth mailing list
>>> [email protected]
>>> https://www.ietf.org/mailman/listinfo/oauth
>> _______________________________________________
>> OAuth mailing list
>> [email protected]
>> https://www.ietf.org/mailman/listinfo/oauth
> 

_______________________________________________
OAuth mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/oauth

Reply via email to