Re: [TLS] Avoiding first use of RI in ClientHello

[Bodo Moeller <[e-mail]>]

On Nov 26, 2009, at 6:34 PM, Stefan Santesson wrote:

> Reading the draft I realize I misunderstood one aspect.
>
> I don't understand the benefit from having different signaling in the
> initial handshakes and renegotiation handshakes.
>
> Especially since I think it is better security design to not send  
> the verify
> data from previous handshakes.

You didn't misunderstand this here -- Eric focussed on the signaling  
issue here, saying the issue of whether the verify_data should be  
included in extensions is a separate one.  (Which, I guess, is also  
why you changed the "Subject" line for further discussion of that  
topic :-) )

Anyway, I'm really writing now to confirm that the signaling approach  
as used in draft-ietf-tls-renegotiation-01.txt looks fine to me.

Bodo