Phillip Hallam-Baker wrote:
> 
> In practice most email that is sent encrypted is encrypted using TLS.
> If we had an infrastructure that allowed mail servers to know that
> their corresponding servers required use of TLS, the man in the middle
> downgrade attack could be defeated.


I'm sorry Phillip, but MTA<->MTA delivery with STARTTLS is thoroughly
broken and effectively unfixable at the moment.

Not only is there no secure algorithm to determine which domains use
a TLS-enabled mail relay and which do not, but PKIX path validation
can not be done because plenty of mail relays are using certs that
do not validate under the (questionable) TLS X.509 PKI used by browsers,
and server endpoint validation can not be done because exactly noone
is carrying the Email domains in their SMTP Server certs for which
these servers are authorized to receive mail, and several SMTP fanciers
seem to be strongly attached to the idea that matching to the
*result* of an MX lookup rather than to the EMail target domain
would make sense security-wise (it doesn't).

And then there are SMTP servers out there (e.g. @gmail.com), that,
while being issued by a CA that is recognized under TLS X.509 PKI
of browsers, neither matches the EMail target domain, nor does
it match the insecure target of the MX record.


In theory, DNSSEC could be used to solve several problems (indicating
that a domain offers STARTTLS *plus* secure identification of acceptable
MTA servers.  But in the near term I expect a wide adoption of DNSSEC
not more likely or faster than the wide adoption of IPv6 to solve
the IPv4 address depletion...


-Martin
_______________________________________________
therightkey mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/therightkey

Reply via email to