On Feb 16, 2021, at 6:41 PM, Paul Wouters <[email protected]> wrote:
> 
> On Mon, 15 Feb 2021, Paul Wouters wrote:
> 
>> Here is a different sentinel:
>> 
>> _53._dns.ns0.example.com. IN TLSA x y z <base64ofCert>
>> 
>> Then do (D)TLS
>> 
>> Now you can choose:
>> 
>> 1) Use DNS(SEC) for validation
>> 2) Use WebPKI[*] for validation
>> 3) TOFU
>> 4) Take at face value
> 
> as PaulH pointed out, the TLSA RFC does not allow one to accept a TLSA
> RRset without DNSSEC signature protection. To allow for deployment
> without DNSSEC, you could instead use the CERT RRtype that has no such
> requirement.

Well, not exactly. (Yes, I said that to PaulW when we were talking, but I was 
not being precise.)

RFC 6698 both defines the TLSA RRtype and a protocol to use it. The protocol 
defined in RFC 6698 indeed requires the TLSA record to be validated with 
DNSSSEC. However, a new protocol, such as one here, could use the TLSA record 
type in a different protocol, such as the one PaulW has above. The document 
that defines that protocol would have to be completely clear about what it was 
doing and why.

--PaulH

Attachment: smime.p7s
Description: S/MIME cryptographic signature

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

Reply via email to