Native application SSO Working Group

Anthony Nadalin tonynad at microsoft.com
Tue Jul 2 17:03:38 UTC 2013


Since this is slated to be an OpenID WG, it’s what the WG wants to do.

From: openid-specs-bounces at lists.openid.net [mailto:openid-specs-bounces at lists.openid.net] On Behalf Of Torsten Lodderstedt
Sent: Tuesday, July 2, 2013 9:53 AM
To: Paul Madsen
Cc: John Bradley; ashishjain at vmware.com; openid-specs at lists.openid.net
Subject: Re: Native application SSO Working Group

Hi Paul,

got it :-) Would it make sense to add this assumption to the charter?

Does this mean:
- a single AZA manages access to multiple authz servers?
- an app needs to be able to register its authz server/idp at the AZA?

Thanks,
Torsten.


Paul Madsen <pmadsen at pingidentity.com<mailto:pmadsen at pingidentity.com>> schrieb:
Hi Torsten, wrt the possibility of an id_token being used against a 'home' IdP, the current model is that it would be the AZA that would perform this exchange, not the native app itself - this because the overarching assumption being that the AZA should do as much of the heavy lifting as possible - and thereby simplify life for the native apps.

But that is separate I think from the use case of an native app wanting to consume an id_token directly (for access control, customization etc) and so i will look at charter to make sure this scenario is supported.

paul

On 7/2/13 11:31 AM, Torsten Lodderstedt wrote:
Hi,

I agree with Nat on this use case. Another one is that the app wants to use the id_token as credential on its "home" IDP (probably via JWT bearer token profile). This is more or less 3rd party login for apps.

regards,
Torsten.


Nat Sakimura <sakimura at gmail.com><mailto:sakimura at gmail.com> schrieb:
Yes. If the app wants the identity information to evaluate its own access control, then it would probably want to know about the user identity (i.e., set of attributes related to the entity), and id_token is the right thing.

When I was talking to some law enforcement people in EU, they were talking similar things. Right now, we do not have any location data defined in the claims, but we may also want to do so in such cases.

Nat

2013/7/3 Paul Madsen <paulmadsen at rogers.com<mailto:paulmadsen at rogers.com>>
Hi Nat, the current AZA model does not preclude an access token being formatted as an id_token.

I believe Torsten was conjecturing that there was potential value in an id_token being delivered to a native app in addition to an access token (whether formatted as id_token or not)

Regards

paul


On 7/2/13 10:53 AM, Nat Sakimura wrote:
I actually do see some utility in the access token in the format of ID Token.
It can give appropriate audience restriction etc.

2013/7/2 Paul Madsen <paulmadsen at rogers.com<mailto:paulmadsen at rogers.com>>
Hi Torsten, the current model is that the Authorization Agent (AZA) may itself obtain an id_token and use it to obtain an access token, but that only access tokens would be 'handed over' by the AZA to its constituent native apps.

Are you proposing that there may be value in allowing the AZA to also hand over id_tokens (suitably targeted) as well?

paul
On 7/1/13 1:38 PM, Torsten Lodderstedt wrote:
Hi John,

I interpreted the text of the charter the other way around, so a client would be able to use an(y) id_token (as a credential) to obtain an access token. I'm fine if the mechanism is intended to support id_token issuance.

regards,
Torsten.

 Am 01.07.2013 15:06, schrieb John Bradley:

Hi Torsten,

In point 3 the charter talks about using id_tokens to get access tokens.

So it is imagined that the mechanism would issue id_tokens likely along the lines that Google is doing for the play store by having a 3rd party as an audience and using "azp" to indicate the client the token was issued to.   We don't want to be too specific on the solution in the charter.

If you think something needs to be added let me know.

John B.

On 2013-07-01, at 2:17 AM, Torsten Lodderstedt <torsten at lodderstedt.net<mailto:torsten at lodderstedt.net>> wrote:


Hi,

it would be great to have such a mechanism across platforms!

I'm wondering whether the mechanism should issue id tokens as well. Right now it seems to focus on access tokens.

Regards,
Torsten.


John Bradley <ve7jtb at ve7jtb.com<mailto:ve7jtb at ve7jtb.com>> schrieb:

The enclosed Work Group Charter is being sent to the Specs Council for review in anticipation of chartering the Group.



It is best have this activity under the foundation IPR as soon as possible.



Regards

John B.






________________________________



specs mailing list

specs at lists.openid.net<mailto:specs at lists.openid.net>

http://lists.openid.net/mailman/listinfo/openid-specs





_______________________________________________

specs mailing list

specs at lists.openid.net<mailto:specs at lists.openid.net>

http://lists.openid.net/mailman/listinfo/openid-specs


_______________________________________________
specs mailing list
specs at lists.openid.net<mailto:specs at lists.openid.net>
http://lists.openid.net/mailman/listinfo/openid-specs



--
Nat Sakimura (=nat)
Chairman, OpenID Foundation
http://nat.sakimura.org/
@_nat_en




--
Nat Sakimura (=nat)
Chairman, OpenID Foundation
http://nat.sakimura.org/
@_nat_en

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openid.net/pipermail/openid-specs/attachments/20130702/ec6e6b1f/attachment.html>


More information about the specs mailing list