On Mon, 20 May 2002 13:35:31 -0700 (PDT)
Jeffrey Goldberg <jeffrey@goldmark.org> wrote:
> On Sun, 19 May 2002, J C Lawrence wrote:
>> Start out by extending TLS such that message contents are encluded in
>> the negotiation and the resultant signature is embedded in a
>> header. In this manner extend TLS chain-of-transcription to both
>> message bodies and the Received: path. Then, just to wrap, start
>> rejecting all mail which doesn't have end-to-end TLS containment.
> That is overkill for the effect that you get. All you are saying is
> that each SMTP server should only talk to clients that authenticate
> well and that the authentication information should be passed on.
> Furthermore, each subsequence server should only accept mail from
> servers that (recursively) follow the same requirement that their
> clients provide authentication information.
No. That is specifically not what I'm saying. I'm saying that MTAs
should by default negotiate TLS connections as above, and that those
sessions should be recorded in usefully auditable fashions as above.
The filtering based on that data occurs at the edge, in your mail
filters for your account, should you so wish.
> This is just a generalization (stronger restriction) of the sort that
> says that we shouldn't have open relays on the net nor relays that
> accept mail from dynamic IP addresses without authentication.
Almost. Its a generalisation of the mail filter I run which checks all
the Received: headers for IP addresses on known open relays.
> Since people don't even agree on blocking mail in the RSS and DUL
> lists, I hardly see how your stronger proposal could ever come to
> pass.
Frankly, I have little hope, but I do have a little hope.
--
J C Lawrence
---------(*) Satan, oscillate my metallic sonatas.
claw@kanga.nu He lived as a devil, eh?
http://www.kanga.nu/~claw/ Evil is a name of a foeman, as I live.
References:
|
|
