The message flows within OAuth 2.0 are virtually identical to the Kerberos 
protocol (RFC4120 and RFC1510). There is also significant overlap in 
functionality (e.g. OAuth token matches the Kerberos ticket, the Refresh Token 
matches the TGT, etc).

Since there is a huge install-base of Kerberos out there, we want to figure out 
how to get the best interoperability between OAuth 2.0 and Kerberos. The 
scenario we think most probable is for organizations to stand-up an OAuth 
front-end, while retaining their Kerberos infra in the back-end.

What is the process in this Working Group to suggest ideas?  Should I just send 
to this list or write-up a draft?

Thanks.

/thomas/


__________________________________________

From: Dick Hardt [mailto:dick.ha...@gmail.com]
Sent: Tuesday, May 25, 2010 1:53 AM
To: Thomas Hardjono
Cc: OAuth WG
Subject: Re: [OAUTH-WG] Access token opaqueness question

Not sure why you want to pull the OAuth token parameters from the Kerberos 
token.

Are you envisioning the Protected Resource is a Kerberos Client?


On Mon, May 24, 2010 at 9:31 AM, Thomas Hardjono 
<hardj...@mit.edu<mailto:hardj...@mit.edu>> wrote:

I'm still a newbie to the OAuth and WRAP discussions, so please bear with me.

In Section 2.2, draft-ietf-oauth-v2-05 states that the access token is basicaly 
an opaque structure to the client, but which can have some "internal structure" 
(having meaning to the authz server and resource server):

] Access token strings can use any internal structure agreed upon
] between the authorization server and the resource server, but their
] structure is opaque to the client.  Since the access token provides
] the client access to the protected resource for the life of the
] access token (or until revoked), the authorization server should
] issue access tokens which expire within an appropriate time, usually
] much shorter than the duration of the access grant.

I was wondering how true this is?

I want to use a Kerberos v5 ticket as the token (base encoded). The OAuth 
Client need not understand this structure, but could simply pass it down to the 
Kerberos-client (in the OS or in the browser).

Then within this token (now a Kerberos ticket), the ticket fields would contain 
matching entries (as far as possible) to the OAuth token parameters (eg. 
expires_in, token_secret, etc).

For the cryptographic tokens (Section 5.3), I would then use the session-key in 
the ticket to compute he HMAC.

Would this work?  Any suggestions?

/thomas/


_______________________________________________
OAuth mailing list
OAuth@ietf.org<mailto:OAuth@ietf.org>
https://www.ietf.org/mailman/listinfo/oauth

_______________________________________________
OAuth mailing list
OAuth@ietf.org
https://www.ietf.org/mailman/listinfo/oauth

Reply via email to