Hi Viktor,

Viktor Dukhovni wrote:
> Yes, that's the document.  It still needs some work, but it can be
> a stop-gap for the larger providers while they gear up to implement
> DNSSEC (a few years work).

For large hosting providers this proposal makes sense, as I've noted.
For the run-of-the-mill mail-server it won't really help that much in my
opinion. See below.

That said I've also noted my issues with DNSSEC in my GitHub issue.

> 
>> And I also think I've broken your proposal in this GitHub issue:
>> https://github.com/mrisher/smtp-sts/issues/1
> 
> No.  Neither DEEP nor TACK can protect MTA-to-MTA SMTP.  The reason
> is MX indirection.  DEEP and TACK pin server properties, not domain
> properties.  The MITM will just forge the MX RRset and bypass DEEP
> and TACK.  In any case, there are domains to which I send email
> very infrequently, but still want the transport to be secure, none
> of DEEP, TACK or STS address that.

I'm not saying that TACK is the only and ultimate way to go. It served
as an example in this case. The major problem I see with DNS is that a
lot of small mail-servers do not get a lot of attention, some of their
operators might not even have direct DNS access (think semi-hosted setup
where you have to either go through a web interface of your provider or
actually write an e-mail. Another case is where the Postmaster has no
real influence on DNS record because of company/university/whatever
bureaucracy). I know you heavily support DANE and I do like DANE in
general but see major problems in DNSSEC still (not only deployment).

Aaron

Attachment: signature.asc
Description: OpenPGP digital signature

_______________________________________________
Uta mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/uta

Reply via email to