FYI, I have been writing HoK for JWT/JWS Token by introducing a new claim
'cid'.
=nat via iPhone
Dec 14, 2012 11:56、"zhou.suj...@zte.com.cn" のメッセ�`ジ:
Yep, could do it soon later.
Currently, I suggest a modification for
"The token service is the assertion issuer; its role is to fulfill
requests from clients, which present various
credentials, and mint assertions as requested, fill them with appropriate
information, and sign them." (in section 3 )
into
"The token service is the assertion issuer, *it could be implemented in
any entity besides client, e.g., Resource Owner, Authorization Server;* its
role is to fulfill requests from clients, which present various
credentials, and mint assertions as requested, fill them with appropriate
information, and sign them."
Chuck Mortimore 写于 2012-12-14 10:44:05:
> Correct. That said no one has yet profiled it for holder of key
>
> - cmort
>
> On Dec 13, 2012, at 6:39 PM, "zhou.suj...@zte.com.cn" <
zhou.suj...@zte.com.cn
> > wrote:
>
> Oh, But the description of assertion generation in the document
> should not be limited by bear assertion, right?
>
>
> Chuck Mortimore 写于 2012-12-14 10:34:13:
>
> > You want a holder of key pattern. The draft touches on it
> >
> >
> >The protocol parameters and processing rules defined in this document
> >are intended to support a client presenting a bearer assertion to an
> >authorization server. The use of holder-of-key assertions are not
> >precluded by this document, but additional protocol details would
> >need to be specified.
>
> >
> > So - if you want this, you should put forth a holder of key
> > profiling of this draft and see if there are any issues. The only
> > profiles we have thus far are saml and jwt bearer assertions.
> >
> >
> > - cmort
> >
> > On Dec 13, 2012, at 6:27 PM, "zhou.suj...@zte.com.cn" suj...@zte.com.cn
> > > wrote:
>
> >
> > I am not sure if the following usecase http://www.ietf.org/mail-
> > archive/web/oauth/current/msg10233.html
> > could be supported by assertion framework,
> > We have some discussion in
> > http://www.ietf.org/mail-archive/web/oauth/current/msg10203.html
> > http://www.ietf.org/mail-archive/web/oauth/current/msg10198.html
> >
> > In my use case or in some other cases, assertions don't need
> > confidential protection,
> > basically STS don't have to authenticate a client before issueing
> > "assertion", if it could be called assertion here.
> >
> > Example,I trust my laywer, I may issue an "assertion" stating
> > delegation in advance, and send to the lawyer when it is needed,
> > it could be I give the assertion to a false lawyer, but it does not
> > matter, because the lawyer has to prove he knows some credential
> > corresponding to his ID,
> > who is delegated some rights.
> >
> > If assertion framework want to support this use case, then
> > generation of assertion should be relaxed,
> > otherwise new work is required to support the use case.
> >
> >
> >
> > Chuck Mortimore 写于 2012-12-14 10:08:34:
> >
> > > On Dec 13, 2012, at 4:30 PM, "zhou.suj...@zte.com.cn" > suj...@zte.com.cn
> > > > wrote:
> >
> > >
> > > From the language, I got an impression that assertion is only
> > > generated by token service after clients presenting some credentials,
> > > there are may be some cases that "assertion" don't need client's
> > credential.
> > > e.g., Resource owner as a token service could generate "assertion"
> > > to a client he trusts, bu signing a statement that "This delegation
> > > is given to a client called clinet-id
> > > for doing something for me".
> > >
> > > So how does the STS trust the client? Presumably if it is trusted
> > > it has some level of authentication, yes?
> > >
> > > -cmort
> > >
> > >
> > >
> > >
> > >
> > > Chuck Mortimore 写于 2012-12-14 00:39:03:
> > >
> > > > The language is simply meant to help illustrate how the framework
> > > > might be used. How do you think it will restrict usage? How
> > > > could it be improved?
> > > >
> > > > -cmort
> > > >
> > > > On Dec 12, 2012, at 11:04 PM, wrote:
> > > >
> > > >
> > > > In "section 3
> > > > The token service is the assertion issuer; its
> > > > role is to fulfill requests from clients, which present various
> > > > credentials, and mint assertions as requested, fill them with
> > > > appropriate information, and sign them."
> > > >
> > > >
> > > > As I understand, an assertion generated by a STS, is done flollowing
> > > > thess steps:
> > > > 1. Client presents credential and requests an assertion
> > > > 2. STS generates assertion and sends to Client
> > > > Correct?
> > > >
> > > > That may restrict the use cases that this assertion framework
> > > could support,
> > > > is it a must?
> > > >
> > > >
> > > >
> > > >
> > > > oauth-boun...@ietf.org 写于 2012-12-11 02:33:57:
> > > >
> > > > >
> > > > > The IESG has received a request from the Web Authorization
Protocol WG
> > > > > (oauth) to consider the following document:
> > > > > - 'Asser