If you look at my manuscript of the Artifact Binding (
http://www.sakimura.org/specs/ab/1.0 ) , it is clear that the Artifact is
created and consumed by OP/AuthzServer and is opaque to RP/WebAppClient. It
has no restriction other than that it should be below 400 bytes. (NB: It is
much bigger than SAML's 30bytes limit).

Or, if you are talking about the Access Token in fig. 1 of
http://www.sakimura.org/en/modules/wordpress/oauth-wrap-mobile-web-app-profile/,
then, it is also completely opaque. It does not even have the size
limit.
Only the condition is that AuthzServer and Resource has the common
understanding of what it is.

=nat


On Thu, Feb 11, 2010 at 7:56 AM, John Bradley <[email protected]>wrote:

> In principal it would work.  The only downside is that the artifact/token
> might be smaller if it were a simple SHA256 XORd with the association secret
> or something like that.
>
> I like the concept in principal if it doesn't compromise the ability to
> have a small response via GET.
>
> I had questions around the token format returned by the protected
> resource.(artifact resolution)
>
> John B.
>
> On 2010-02-10, at 7:27 PM, Allen Tom wrote:
>
> + [sp...@openid]
>
> Nat – this is exactly what I had in mind. In many ways Oauth and Oauth-WRAP
> are  similar to artifact binding – the user approves a token, which is then
> passed back to the RP via a browser redirect. The token is then used by the
> RP to make web service calls on the OP to access a Protected Resource.
>
> The token is kind of like an artifact, and the Protected Resource can be an
> OpenID assertion.
>
> Would we be able to combine the OpenID Artifact Binding Extension with
> OAuth WRAP? If so, that would be great.
>
> Allen
>
>
> On 2/8/10 7:29 PM, "Nat Sakimura" <[email protected]> wrote:
>
> Hi
>
> I was wondering if we could define an Artifact Binding/Mobile Profile for
> Wrap.
>
> The way I would do is pretty simple because Wrap Web App Profile is an
> Artifact Binding to some extent.
> Just send Verification Code Request directly from WebAppClient to
> AuthzServer
> and get an Artifact back and bring that to AuthzServer through UA.
> After PoP, another artifact is created at AuthzServer and
> it is brough back to the WebAppClient through UA redirect.
> Then, the verification Code Response can be obtained from AuthzServer
>  directly using the artifact.
> The rest is the same.
>
> I created an blog entry with pretty diagram at
>
> http://www.sakimura.org/en/modules/wordpress/oauth-wrap-mobile-web-app-profile/
>
> It may be easier to see the page instead of the above description.
>
> (Instead of using response artifact, Verification Code Response can be sent
> directly,
>  but then we would be introducing AuthzServer -> WebAppClient
> communication, which would have
>  some implication on firewall configuration.)
>
> For those of you who say that "Artifact is Complex", see the original Web
> App Profile here:
>
>
> http://www.sakimura.org/en/modules/wordpress/oauth-wrap-web-app-profile-summary/
>
> It is almost identical.
>
> Added value is that is is more "mobile" friendly, and is actually more
> secure if the
> Request Artifact and Response Artifact (wrap_verification_code) is
> generated cryptographically
> strongly.
>
> What would you think?
>
>  _______________________________________________
> specs mailing list
> [email protected]
> http://lists.openid.net/mailman/listinfo/openid-specs
>
>
>
> _______________________________________________
> specs mailing list
> [email protected]
> http://lists.openid.net/mailman/listinfo/openid-specs
>
>


-- 
Nat Sakimura (=nat)
http://www.sakimura.org/en/
http://twitter.com/_nat_en
_______________________________________________
specs mailing list
[email protected]
http://lists.openid.net/mailman/listinfo/openid-specs

Reply via email to