Alan DeKok writes:
>
> "Vic Abell" <[EMAIL PROTECTED]> wrote:
> > I don't think someone should be authorized before the claimed identity
> > has been authenticated.  Otherwise authorization might be given to
> > someone falsely claiming an identity.
> 
>   Nonsense.  The authorization isn't returned to the caller until
> after they've been authenticated.

No, it's not nonsense.  The secretary's telling me that Vic
Abell has an appointment gives away potentially useful
information.

It's like a login program telling an intruder that the login
exists, but the intruder hasn't yet specified the correct
authentication.  That's an invitation to try again.

However, because the RADIUS protocol doesn't seem to have
separate returns for authorization and authentication
results, it's not giving the client a hint that the attempt
is worth retrying with a different password.

You gave an explanation much more acceptable to me in separate
e-mail when you said:

>   The issue is that one of the authorization parameters is which
> authentication methods you are allowed to use.  So the control flow
> should really go like:
>
>   authorize which authentication type
>   do authentication
>   if authenticated, give additional authorization.

I completely overlooked the feature of the RADIUS protocol and
server that allows multiple authentication methods to be applied.
In that case it is clearly useful to authorize the authentication
type first.

Thanks for your useful contributions to my understanding of the
RADIUS protocol.

Vic

- 
List info/subscribe/unsubscribe? See http://www.freeradius.org/list/users.html

Reply via email to