Re: Re: Comments on draft-morris-geopriv-core-01

Hi Henning,

>I have to admit that "Element K" (express permission) worries me, as I
>have a hard time picturing a generic mechanism. In the presence case, this
>is easy - SUBSCRIBE usually requires an explicit permission. Outside that
>context, this seems to span an almost infinite variety of mechanisms,
>including issues such as what counts as permission. Does 'clicking' on a
>button count? How much information needs to be presented? Etc.

One mechanism for implementing this is the following:
think of this as a challenge-response: the LS ask the RM (usually = Target)
to send a confirmation that LR may receive the LO, as indicated by the police,
but requiring explicit consent. This request from the LS to the RM also has
something like a time-stamp and /or a counter and/or another nonce. This is the
"challenge", let us call it Nonce. Then the RM sends as response a message signing
or otherwise authenticating "1;Nonce" (say, using a keyed hash).

How this is implemented at the terminal of the RM is irrelevant, of
course it may be just clicking on a button.

This is important is for situations like the following:

I want to let my boss to know my location except on some
situations (times, places, etc) that are impossible to express in
simple privacy rules. If the user is asked if his position may be
sent to the boss in a certain situation, the user is in full control
of the situation.

Or suppose you want the AAA automobile club to obtain your location
if you are having trouble with your car. As such, this
is impossible to write down on policy rules, but you may want to
contact the AAA club, and ask them to ask the LS for your location.
This may be a plain telefone call. Now the RM looks at the rules and
finds that the AAA has authorization to get the LO in full resolution,
but only if there is explicit consent. The LS asks the RM via some
messaging mechanism and the RM answers using the same protocol.

There are other ways of dealing with the situation (send a token to the AAA
club, for instance) but they are not equivalent.

--Jorge

------------------------------------
Dr Jorge R Cuellar   Office: Tel +49 89 636 47585
Siemens CT IC 3              Fax +49 89 636 48000
Otto-Hahn Ring 6    jorge.cuellar@mchp.siemens.de
81730 München
---------  Out of town: -------------
    Fax & Voicebox: +49 (0)  1212-5-130-30-549  
    JorgeRicardo@web.de

______________________________________________________________________________
Viren? Wir wissen nicht was Ihr Arzt empfiehlt. Wir empfehlen den
Virencheck für Ihre E-Mail-Anhänge! http://freemail.web.de/features/?mc=021159
Received on Mon Mar 31 04:07:11 2003