On Thu, Jul 8, 2010 at 11:53 AM, Kris Selden <[email protected]> wrote:
> The refresh token is key to separating concerns of the protected resource 
> from the auth server.
>
> I think this is succinctly explained the original rationale (refresh token = 
> refresh secret):
> http://wiki.oauth.net/ScalableOAuth#Appendix

That suggests security issues. If I use tokens that can be used to
access protected resources for extended periods of time, why can't I
just revoke the compromised tokens (possibly all of them) when a leak
is discovered? It seems somewhat convoluted to introduce a new system
and then just wait until a timeout happens (although that timeout may
be acceptably small)...

I do not understand how it separates concerns more than they were
already separated under OA1.0a. Under OA1.0a, if I understand
correctly, the auth server produces tokens, and the protected resource
consumes them. That feels like two separate concerns to me.

cheers
Laurens
_______________________________________________
OAuth mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/oauth

Reply via email to