Seeking guidance on the implementation of native/rich client flow
Todd W Lainhart
lainhart at us.ibm.com
Mon Oct 28 14:29:06 UTC 2013
> I would need to know more about the other credentials. I am assuming a
desktop app if it is Eclipse based.
That's right - an Eclipse-based desktop app. The login dialog raised by
the app ultimately uses Apache httpclient to negotiate auth with the AS.
That dialog allows for the entry of username/password, a client-based
certificate, or smart-card. On the feature list is SPENGO (Kerberos).
Also a CLI, similarly using httpclient.
> If it is going to authenticate based on some other credential already
in the browser then use the code flow.
This hadn't occurred to me. Basically the user needs to know how the AS
is configured for auth. If cert-based or Kerberos, they enter their
certificate and the native client (with auto-redirection turned off)
initiates the code flow, assuming no challenges, until its redirect_uri
returns in the Location header. If the AS is configured for
username/password, use the password flow.
Ideally I was thinking that the client wouldn't have to switch flows based
on how the AS is configured, but perhaps that's unrealistic. (In our
OAuth extension, we implemented a proprietary sign-in endpoint that
optionally took username/password, and returned a token).
Todd Lainhart
Rational software
IBM Corporation
550 King Street, Littleton, MA 01460-1250
1-978-899-4705
2-276-4705 (T/L)
lainhart at us.ibm.com
From: John Bradley <john.bradley at wingaa.com>
To: Torsten Lodderstedt <torsten at lodderstedt.net>,
Cc: Todd W Lainhart/Lexington/IBM at IBMUS,
"openid-specs at lists.openid.net" <openid-specs at lists.openid.net>
Date: 10/26/2013 08:10 AM
Subject: Re: Seeking guidance on the implementation of native/rich
client flow
I would have the client use the resource owner credentials flow if it has
the password. If it is going to authenticate based on some other
credential already in the browser then use the code flow.
I would need to know more about the other credentials. I am assuming a
desktop app if it is Eclipse based.
John B.
Sent from my iPhone
On Oct 26, 2013, at 8:52 AM, Torsten Lodderstedt <torsten at lodderstedt.net>
wrote:
We use OIDC in conjunction with resource owner password credential grant
for native apps (no 3rd party apps, just our own apps)
Todd W Lainhart <lainhart at us.ibm.com> schrieb:
I'm referencing http://openid.net/specs/openid-connect-core-1_0.html
We have an Authorization Server that supports SSO via session extensions
to OAuth 2.0. We're looking to replace that protocol w/ OIDC. There's a
couple of sticky points that I'm not sure how to translate.
1) Rich/Native Client login
Imagine an Eclipse-based rich client accepts user credentials and receives
a bearer token in return. The negotiation may be basic,
credentials-based, SPENGO. The client is anonymous. Rather than using
the Resource Owner Password Credentials Grant (where username/password are
REQUIRED parameters), we opted for a custom endpoint so that the AS could
determine if the request was authenticated in the absence of
username/password. Similar to Resource Owner Password Credentials Grant.
I'm wondering what the guidance is for such a setup in OIDC. Implicit
requires the native client to follow (presumably) 302s with the AS until
it gets the final 302 to the callback location. Seems messy for this
setup.
In the absence of guidance/precedent, I'm inclined to think that a
Resource Owner Password Credentials Grant style extension is the way to go
for this scenario.
Todd Lainhart
Rational software
IBM Corporation
550 King Street, Littleton, MA 01460-1250
1-978-899-4705
2-276-4705 (T/L)
lainhart at us.ibm.com
specs mailing list
specs at lists.openid.net
http://lists.openid.net/mailman/listinfo/openid-specs
_______________________________________________
specs mailing list
specs at lists.openid.net
http://lists.openid.net/mailman/listinfo/openid-specs
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openid.net/pipermail/openid-specs/attachments/20131028/a24c7a84/attachment.html>
More information about the specs
mailing list