RE: DISCUSS: MUST reject in OCSPv1

["Deacon, Alex" <alex@xxxxxxxxxxxx>]

Mike,

> -----Original Message-----
> From: Michael Myers [mailto:mmyers@xxxxxxxxx] 
[snip]
> 
> Tom,
> 
> We may be saying the same thing:
> 
> if( req->sent_nonce != rsp->rcvd_nonce )
>    fail( BADNONCE )
> else
>    proceed( rsp );
> 
> The more interesting question along these lines is the one 
> regarding Florian's server-unilateral nonces.  I.e. a client 
> receives a nonce extension in the response even though the 
> client didn't provide a nonce extension in its request.  
> While I'm not yet convinced of this practice's security 
> value, I see no reason to preclude it in our forthcoming 
> clarification regarding nonces.

As I mentioned earlier, it will be important to clarify this case as clients
using piggybacked OCSPResponses (such as those implementing the TLS
extension) may receive a response that contains a nonce (the one the server
generated) eventhough they did not send one.

Alex