[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]


To: Edward Lewis <lewis@tislabs.com>
Cc: <dnssec@cafax.se>
From: Simon Josefsson <simon+dnssec@josefsson.org>
Date: Tue, 04 Sep 2001 22:22:57 +0200
In-Reply-To: <v03130301b7bae0979aeb@[199.171.39.4]> (Edward Lewis's messageof "Tue, 4 Sep 2001 15:54:21 -0400")
Sender: owner-dnssec@cafax.se
User-Agent: Gnus/5.090004 (Oort Gnus v0.04) Emacs/21.0.105
Subject: Re: CERTificates and public keys

Edward Lewis <lewis@tislabs.com> writes:

>>It is already possible to put a public key that is not signed by a CA
>>in a CERT record.  Let's use it.
>>
>>(Ok, if SSH wants a certificate type number in CERT of its own, they
>>need to register one.  But the same holds for KEY.)
>
> Why force SSH to use a certificate?
>
> There are two things a client needs to make a secure connection - the IP
> address and server host key.  The IP address is coming out of DNS, so
> implicitly there is trust in the DNS admin to enter the right data.  Why is
> the server host key any different?
>
> What are the "requirements" motivating the use of a certificate structure
> instead of a raw public key?  Why do the extra processing needed to get the
> server host key into a certificate, ship a certificate, and then parse and
> extract the key from the certificate in the client?

Oh.  I have been unclear, this is not what I want.  I am sorry, I will
try to explain what I'm after.

I'm suggesting to let SSH put a raw public key in a CERT record, using
a "SSH" Certificate type.

In my mind, this isn't even abusing [my understanding of] the
intention of the CERT record, a raw public key somehow signed by
someone else (be it DNSSEC, TSIG or whatever) _is_ a certificate and
thus OK to store in a CERT RR.

And even if this is abusing the intention of the CERT record, I don't
see any technical arguments against this.

Using APPKEY would work equally fine to me, but having both CERT and
APPKEY for similar purposes is not good.


Home | Date list | Subject list