[OAUTH-WG] RT treatment in Token Exchange

[Brian Campbell]

I gave a short presentation about OAuth 2.0 Token Exchange
<http://www.slideshare.net/briandavidcampbell/oauth-20-token-exchange-an-sts-for-the-rest-of-us>
at a recent identity conference, which seemed well received and a lot of
folks expressed their support for it and a desire to see support for it in
offerings out in the wild.

However, I did get some feedback from Vittorio Bertocci of Microsoft that
he felt that the language around issuing refresh tokens was overly
restrictive based on real world deployment experience they have with a
token exchange like interaction that they are currently offering. He
described it as follows:

"To summarize our main use case: we use [a token exchange like protocol]
for achieving user impersonation thru tiers, or delegation for confidential
clients which do not have any web UX.

Those cases do call for the ability of the middle tier to access the
resource also when the original credential is no longer valid (e.g. there
is no longer any user entertaining an active session with the middle tier).

Think of a hypothetical new feature of the Uber mobile app, which can
monitor your Exchange calendar and book a ride whenever a suitable slot
opens up within a certain time range. The Uber app would call an Uber-owned
middle tier, in form of Web API, and the Web API would itself be a client
of the Exchange API. Given that the Uber API doesn’t offer any browser
ready surface, as it is meant to be accessed only via oauth2 bearer token,
the [token exchange] is the only way through which it can obtain a token
for the Exchange API; and the scenario definitely call for offline access,
as the uber API should be able to monitor the Exchange calendar of the user
even if the user closed the app on his/her phone."

While the current text in the draft allows for this kind of thing, it
stigmatizes it somewhat with a "NOT RECOMMENDED" on sending an RT in the
response. So the proposal here is to make refresh tokens OPTIONAL and
change the accompanying text in sec 2.2.1 to the following:

refresh_token
OPTIONAL. A refresh token will typically not be issued when the
the exchange is of one temporary credential (the subject_token)
for a different temporary credential (the issued token) for use in
some other context. A refresh token can be issued in cases where
the client of the token exchange needs the ability to access a
resource even when the original credential is no longer valid
(e.g. user-not-present or offline scenarios where there is no
longer any user entertaining an active session with the client).
Profiles or deployments of this specification should clearly
document the conditions under which a client should expect a
refresh token in response to "urn:ietf:params:oauth:grant-
type:token-exchange" grant type requests.

I plan to make the aforementioned change and get a new draft published with
that and some other updates later this week before the Internet Draft
submission cut-off on Friday.

[Anthony Nadalin]

So I think the proposed wording is still too specific and limits the use case , I also don’t understand the usage of “credential” in your description as this does not have to be a credential. So suggest that this be simple and if you want you can explain in the security considerations section why one may not want to return a refresh_token

OPTIONAL. The refresh_token for the issued token in the token_exchange request

access_token (response), this does not have to be an access_token so this is a bad name


Also it’s unclear why the subject_token is required, is it being suggested that the subject_token is also being used as an access_token ? The actor_token if specified could also be the subject_token

The term “security_token” I assume is not limited to access_token since there is a subject_token_type.

want_composite is not really the effect we are looking for since it provides for a single token, the use case we have is where you want the ability to use the subject_token and the actor_token in combination and not as a composite of only the claims.

From: OAuth [mailto:oauth-***@ietf.org] On Behalf Of Brian Campbell
Sent: Tuesday, July 5, 2016 11:16 AM
To: oauth <***@ietf.org>
Subject: [OAUTH-WG] RT treatment in Token Exchange

I gave a short presentation about OAuth 2.0 Token Exchange<https://na01.safelinks.protection.outlook.com/?url=http%3a%2f%2fwww.slideshare.net%2fbriandavidcampbell%2foauth-20-token-exchange-an-sts-for-the-rest-of-us&data=01%7c01%7ctonynad%40microsoft.com%7c36da9d499be34762740a08d3a5007e31%7c72f988bf86f141af91ab2d7cd011db47%7c1&sdata=5a5Ca7pHESmgzq36BmJDtQH%2f564QTbk1JQzFNfasnJU%3d> at a recent identity conference, which seemed well received and a lot of folks expressed their support for it and a desire to see support for it in offerings out in the wild.
However, I did get some feedback from Vittorio Bertocci of Microsoft that he felt that the language around issuing refresh tokens was overly restrictive based on real world deployment experience they have with a token exchange like interaction that they are currently offering. He described it as follows:
"To summarize our main use case: we use [a token exchange like protocol] for achieving user impersonation thru tiers, or delegation for confidential clients which do not have any web UX.

Those cases do call for the ability of the middle tier to access the resource also when the original credential is no longer valid (e.g. there is no longer any user entertaining an active session with the middle tier).

Think of a hypothetical new feature of the Uber mobile app, which can monitor your Exchange calendar and book a ride whenever a suitable slot opens up within a certain time range. The Uber app would call an Uber-owned middle tier, in form of Web API, and the Web API would itself be a client of the Exchange API. Given that the Uber API doesn’t offer any browser ready surface, as it is meant to be accessed only via oauth2 bearer token, the [token exchange] is the only way through which it can obtain a token for the Exchange API; and the scenario definitely call for offline access, as the uber API should be able to monitor the Exchange calendar of the user even if the user closed the app on his/her phone."

While the current text in the draft allows for this kind of thing, it stigmatizes it somewhat with a "NOT RECOMMENDED" on sending an RT in the response. So the proposal here is to make refresh tokens OPTIONAL and change the accompanying text in sec 2.2.1 to the following:

refresh_token
OPTIONAL. A refresh token will typically not be issued when the
the exchange is of one temporary credential (the subject_token)
for a different temporary credential (the issued token) for use in
some other context. A refresh token can be issued in cases where
the client of the token exchange needs the ability to access a
resource even when the original credential is no longer valid
(e.g. user-not-present or offline scenarios where there is no
longer any user entertaining an active session with the client).
Profiles or deployments of this specification should clearly
document the conditions under which a client should expect a
refresh token in response to "urn:ietf:params:oauth:grant-
type:token-exchange" grant type requests.
I plan to make the aforementioned change and get a new draft published with that and some other updates later this week before the Internet Draft submission cut-off on Friday.

[Justin Richer]

+1 to the proposed wording change. It's clear and allows for use cases
where your inputs aren't as replayable as you might otherwise expect.

-- Justin

Post by Brian Campbell
I gave a short presentation about OAuth 2.0 Token Exchange
<http://www.slideshare.net/briandavidcampbell/oauth-20-token-exchange-an-sts-for-the-rest-of-us>
at a recent identity conference, which seemed well received and a lot
of folks expressed their support for it and a desire to see support
for it in offerings out in the wild.
However, I did get some feedback from Vittorio Bertocci of Microsoft
that he felt that the language around issuing refresh tokens was
overly restrictive based on real world deployment experience they have
with a token exchange like interaction that they are currently
"To summarize our main use case: we use [a token exchange like
protocol] for achieving user impersonation thru tiers, or delegation
for confidential clients which do not have any web UX.
Those cases do call for the ability of the middle tier to access the
resource also when the original credential is no longer valid (e.g.
there is no longer any user entertaining an active session with the
middle tier).
Think of a hypothetical new feature of the Uber mobile app, which can
monitor your Exchange calendar and book a ride whenever a suitable
slot opens up within a certain time range. The Uber app would call an
Uber-owned middle tier, in form of Web API, and the Web API would
itself be a client of the Exchange API. Given that the Uber API
doesn’t offer any browser ready surface, as it is meant to be accessed
only via oauth2 bearer token, the [token exchange] is the only way
through which it can obtain a token for the Exchange API; and the
scenario definitely call for offline access, as the uber API should be
able to monitor the Exchange calendar of the user even if the user
closed the app on his/her phone."
While the current text in the draft allows for this kind of thing, it
stigmatizes it somewhat with a "NOT RECOMMENDED" on sending an RT in
the response. So the proposal here is to make refresh tokens OPTIONAL
refresh_token
OPTIONAL. A refresh token will typically not be issued when the
the exchange is of one temporary credential (the subject_token)
for a different temporary credential (the issued token) for use in
some other context. A refresh token can be issued in cases where
the client of the token exchange needs the ability to access a
resource even when the original credential is no longer valid
(e.g. user-not-present or offline scenarios where there is no
longer any user entertaining an active session with the client).
Profiles or deployments of this specification should clearly
document the conditions under which a client should expect a
refresh token in response to "urn:ietf:params:oauth:grant-
type:token-exchange" grant type requests.
I plan to make the aforementioned change and get a new draft published
with that and some other updates later this week before the Internet
Draft submission cut-off on Friday.
_______________________________________________
OAuth mailing list
https://www.ietf.org/mailman/listinfo/oauth