On Sun, Oct 5, 2014 at 12:26 PM, Yaron Sheffer <[email protected]> wrote:
> 1. Say explicitly that opportunistic TLS is out of scope.
> 2. Or say explicitly that it is in scope, and with the same requirements as
> "regular" TLS.
> 3. Or come up with a different set of requirements for opportunistic TLS.

I think you can just avoid mentioning opportunistic TLS in the BCP at all.

How could unauthenticated TLS be in scope for a TLS BCP when it has
not even been deployed yet? It isn't a current practice.

> - With channel bindings, you can convert an unauthenticated TLS channel into
> an authenticated one, after the fact.

There's no commonly-accepted way of doing that. Regardless, it isn't a
common current practice.

> - Also, because we do not want to fragment the TLS ecosystem.

That's exactly what unauthenticated/opportunistic TLS is doing, and
that is in fact one of the biggest arguments against unauthenticated
TLS.

> - Lastly, an opportunistic deployment can eventually become authenticated
> TLS, when DANE is introduced.

DANE-authenticated TLS would be authenticated TLS, not
unauthenticated/opportunistic TLS. Regardless, it is also out of scope
for this BCP because it isn't a common current practice.

Cheers,
Brian

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

Reply via email to