Re: [TLS] the idea of using multiple keys with multiple certificate authorities in a TLS session.

[Hanno Böck <[e-mail]>]

On Sat, 7 Feb 2015 11:59:20 +0330
Shahin Noursalehi <[e-mail]> wrote:

> Anyway, this would be an optional process - so abides by the policies
> of QoS too. After 2014, enterprises may look for a method that reduces
> the risk of being victim of "fake SSL certificate issuers" or those
> who "put a backdoor in RSA key-pairs". When we have sensitive data,
> the cost among benefits is nothing. Sometimes a simple email in
> internet among a patient and a physician that contains some critical
> health information may jeopardize his/her life style and can not
> roll-back.

I don't understand how your proposal makes "fake ssl certificate
issuers" less likely.

You propose to use two keys signed from different CAs. However, in what
way does this make it less likely that a completely different CA may
issue a rogue cert for the same domain? The attacked user will never
see your two keys.

If you want to make fake ssl certificates less likely you should use
key pinning (right now only available for https, unfortunately not for
other protocols) and certificate transparency. I don't see what your
proposal adds to that.

Also I don't understand your threat scenario "put a backdoor in RSA
key-pairs". What do you mean by that? I know that there are some CAs
that provide to create the key for you. If you are worried about that
just don't do that. Use CSRs and create the key yourself.

-- 
Hanno Böck
http://hboeck.de/

mail/jabber: [e-mail]
GPG: BBB51E42