[TLS] Re: Implicit ECH Config for TLS 1.3 – addressing public_name fingerprinting

[Stephen Farrell <[e-mail]>]

Hiya,

On 06/03/2025 14:47, Martin Thomson wrote:
> On the broader topic, Marwan and I have a draft that looks at a
> different angle on this problem.  That has a bunch of complicated
> stuff in there, but those pieces aren't necessarily core to the
> idea.  I'm also aware of ongoing conversations about this that might
> lead to iterative improvements on these ideas.  It would be really
> nice if we could have some time to talk through some of the
> rationale behind these different ideas and see if we can tease out
> the real constraints.
> 
> Here's the draft that Marwan and I put together:https://
> datatracker.ietf.org/doc/draft-thomson-tls-ech-pnmasq/

I agree that that's complicated:-)

If we end up needing something that complex, (and I'm not
saying I'm convinced we do), then we should also consider
a new TLS extension code point and a new SvcParamKey, as
that might a) be simpler, and/or b) easier to deploy in
parallel with ECH as-is. And in that case, I suspect it'd
also make sense to see how things play out with ECH for a
year or so after we start to see significant numbers of
independent server deployments of ECH, before doing all
that.

Cheers,
S.