Re: [OAUTH-WG] Access Token Response without expires_in

["Richer, Justin P." <[e-mail]>]

Information inside the token is outside of OAuth completely, so I like Eran's suggestion that doesn't mention how this information would be communicated. However, I would like to leave it open for different expiration semantics beyond expiration time. I suggest the following text (which could probably use some wordsmithing):

expires_in
         OPTIONAL.  The lifetime in seconds of the access token.  For
         example, the value "3600" denotes that the access token will
         expire in one hour from the time the response was generated. If
                omitted, the authorization server SHOULD provide the token expiration behavior
                via other means or documentation.

________________________________
From: [e-mail] [[e-mail]] on behalf of John Bradley [[e-mail]]
Sent: Tuesday, January 17, 2012 7:47 AM
To: Mike Jones
Cc: wolter.eldering; OAuth WG
Subject: Re: [OAUTH-WG] Access Token Response without expires_in

I am OK with that.

The expiration time in the token is intended for the protected resource.
The client inspecting the token is a potential optimization in cases where the JWT is not encrypted to the
protected resource.

I think leaving it open to inspect the token or otherwise provide it in configuration information is flexible enough.

John B.


On 2012-01-17, at 5:54 AM, Mike Jones wrote:

Your new wording is better, as it doesn’t conflict with the possibility of the expiration time being in the token.

                                                            -- Mike

From: Eran Hammer [mailto:[e-mail]]
Sent: Tuesday, January 17, 2012 12:30 AM
To: Mike Jones; Aaron Parecki; OAuth WG
Cc: wolter.eldering
Subject: RE: [OAUTH-WG] Access Token Response without expires_in

This is clearly not a solution as actual implementation feedback raised this issue. We have to document the meaning of this parameter missing. Also, the example of a self-contained token does not conflict with also providing this information via the parameter whenever possible to improve interop.

I’m going to go with adding: If omitted, the authorization server SHOULD provide the expiration time via other means or document the default value.

EHL

From: Mike Jones [mailto:[e-mail]]<mailto:[mailto:[e-mail]]>
Sent: Tuesday, January 17, 2012 12:02 AM
To: Eran Hammer; Aaron Parecki; OAuth WG
Cc: wolter.eldering
Subject: RE: [OAUTH-WG] Access Token Response without expires_in

This doesn’t work for me, as it doesn’t mesh well with the case of the token containing the expiration time.  For instance, both SAML and JWT tokens can contain expiration times.  In this case, the expires_in time is unnecessary and the token may have no default expiration time and will expire even though not explicitly invoked.

I would recommend no change to the current text, which is:
   expires_in
         OPTIONAL.  The lifetime in seconds of the access token.  For
         example, the value "3600" denotes that the access token will
         expire in one hour from the time the response was generated.

                                                            -- Mike

From: [e-mail]<mailto:[e-mail]> [mailto:[e-mail]]<mailto:[mailto:[e-mail]]> On Behalf Of Eran Hammer
Sent: Monday, January 16, 2012 11:20 PM
To: Aaron Parecki; OAuth WG
Cc: wolter.eldering
Subject: Re: [OAUTH-WG] Access Token Response without expires_in

WFM.

From: Aaron Parecki [mailto:[e-mail]]<mailto:[mailto:[e-mail]]>
Sent: Monday, January 16, 2012 11:08 PM
To: OAuth WG
Cc: Eran Hammer; Richer, Justin P.; wolter.eldering
Subject: Re: [OAUTH-WG] Access Token Response without expires_in

Actually now I'm having second thoughts about making expires_in RECOMMENDED. Here's another attempt at a clarification:

expires_in
         OPTIONAL.  The lifetime in seconds of the access token.  For
         example, the value "3600" denotes that the access token will
         expire in one hour from the time the response was generated.
         If omitted, the authorization server SHOULD document the
         default expiration time or indicate that the token will not
         expire until explicitly revoked.

-aaronpk

On Mon, Jan 16, 2012 at 10:37 PM, Eran Hammer <[e-mail]<mailto:[e-mail]>> wrote:
Hmm. This might become too much work at this stage…

Happy for suggestions but I won’t pursue it on my own for now.

EHL

From: Aaron Parecki [mailto:[e-mail]<mailto:[e-mail]>]
Sent: Monday, January 16, 2012 10:36 PM
To: OAuth WG
Cc: Richer, Justin P.; wolter.eldering; Eran Hammer

Subject: Re: [OAUTH-WG] Access Token Response without expires_in

That seems like a good idea, but then it should also be explicitly stated what to do if the server issues non-expiring tokens.

aaronpk

On Mon, Jan 16, 2012 at 10:29 PM, Eran Hammer <[e-mail]<mailto:[e-mail]>> wrote:
How do you feel about changing expires_in from OPTIONAL to RECOMMENDED?

EHL

> -----Original Message-----
> From: Richer, Justin P. [mailto:[e-mail]<mailto:[e-mail]>]
> Sent: Monday, January 16, 2012 7:29 PM
> To: Eran Hammer
> Cc: OAuth WG; wolter.eldering
> Subject: Re: [OAUTH-WG] Access Token Response without expires_in
>
> I think #3.
>
> #1 will be a common instance, and #2 (or its variant, a limited number of
> uses) is a different expiration pattern than time that would want to have its
> own expiration parameter name. I haven't seen enough concrete use of this
> pattern to warrant its own extension though.
>
> Which is why I vote #3 - it's a configuration issue. Perhaps we should rather
> say that the AS "SHOULD document the token behavior in the absence of this
> parameter, which may include the token not expiring until explicitly revoked,
> expiring after a set number of uses, or other expiration behavior." That's a lot
> of words here though.
>
>  -- Justin
>
> On Jan 16, 2012, at 1:53 PM, Eran Hammer wrote:
>
> > A question came up about the access token expiration when expires_in is
> not included in the response. This should probably be made clearer in the
> spec. The three options are:
> >
> > 1. Does not expire (but can be revoked) 2. Single use token 3.
> > Defaults to whatever the authorization server decides and until
> > revoked
> >
> > #3 is the assumed answer given the WG history. I'll note that in the spec,
> but wanted to make sure this is the explicit WG consensus.
> >
> > EHL
> >
> >
> > _______________________________________________
> > OAuth mailing list
> > [e-mail]<mailto:[e-mail]>
> > https://www.ietf.org/mailman/listinfo/oauth

_______________________________________________
OAuth mailing list
[e-mail]<mailto:[e-mail]>
https://www.ietf.org/mailman/listinfo/oauth


_______________________________________________
OAuth mailing list
[e-mail]<mailto:[e-mail]>
https://www.ietf.org/mailman/listinfo/oauth