On 10/22/2018 8:24 PM, Salz, Rich wrote:
* My suggestion with something similar to DANE and DKIM (in
utilizing DNS and DNSSEC), DNS TXT record is already been used by
acme protocol to pass a challenge, so why not use similar
implementation to authenticate the server itself for the client,
so the client can verify the certificate and the chain, without a
third-party.
No third-party is needed. The client has to trust **something** out
of band. In a browser, this is typically the root store, and any CA
trust chain should end up being signed by something in that store.
Other clients have different methods, but they are ultimately a set of
trusted “root anchors.” If you put data in DNS, then the client has
to trust DNS and the chain that signed that data.
I think we’re struggling to understand the issue you are trying to raise.
I am not trying to raise an issue, client can only trust acme server
simply by providing a key (public key) along with the ACME URL, you
trust example.com then put in your library this url and this key
supplied by example.com service, thus the client doesn't need to trust
anything else, by such key any MitM can be detected by both client and
server or just one, that depends on how you tweak this protocol, that
key can be supplied to the library by implementer, user or DNSSEC (in
this case you need a key) .
Acme server is CA server and shouldn't need a root store to be validated
or trusted, that root store can be easily manipulated even by a
software, even without locally manipulation the MitM can issue a
certificate to the client by simply hijacking the connection and having
certificate issued by trusted CA, and the client will validate and trust
that certificate.
Again i am sorry that you feel i am raising an issue, i am not, only
suggesting a concerning matter to discuss.
Best regards and sorry for wasting your time,
K. Obaideen
_______________________________________________
Acme mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/acme
