Spencer Dawkins wrote:
our inability to come up with a certificate for
an arbitrary person somewhere in the Internet,
so we can verify signatures?
I said nothing about certificates for "arbitrary persons". The
subject discussed was how do you prove that a message was sent by a
specific *server*. The key to simplification here is that the server
will have a DNS record and that DNS record can have SVR records
associated with it. What you do is set up an SVR record that points to
the certificate server for the mail server. Then, you pick any of the
many available certificate formats and say: "This certificate format is
the one to use for mail server identification."
Locating certificates for arbitrary people is a different
problem since we don't have anything like a DNS service for them. i.e.
there is no way to make a connection to an "arbitrary person" and ask it
questions. In any case, certificates for "arbitrary persons" are only
needed if you are trying to ensure that mail came from "arbitrary
persons." In this case, we're trying to determine if mail came from a
specific server -- not a person.
The kind of certificate needed in a system like I'm suggesting
here can be very minimal and even self-generated. You don't, for
instance, have to worry about having any CA that issues the things. You
can just create an arbitrary public key and publish it via the
certificate server of your choice. The reason for this is that you are
asking the server's DNS server itself to tell you where to find the
server's key. As long as you trust that you are speaking to the server's
DNS, you can invest reasonable trust in any statements it makes.
(Yes, I realize that one can spoof DNS data, etc. however, I
believe that will not be a significant problem in *this* specific
application. The worst case that results from any spoofed data is the
present case: i.e. faked addresses. However, I believe that the vast
majority of servers will have management that ensures that their data is
properly maintained and monitored. The fake source data will diminish as
more and more server managers adopt the method even though some small,
and insignificant portion of them will end up with hacked data.)
bob wyman
_______________________________________________
Asrg mailing list
Asrg(_at_)ietf(_dot_)org
https://www1.ietf.org/mailman/listinfo/asrg