Re: [Uta] Éric Vyncke's No Objection on draft-ietf-uta-rfc7525bis-09: (with COMMENT)

[Peter Saint-Andre <[e-mail]>]

Hi Éric, see further thoughts on ECH below.

On 7/13/22 12:12 PM, Peter Saint-Andre wrote:
> On 7/12/22 1:50 AM, Éric Vyncke via Datatracker wrote:
>> Éric Vyncke has entered the following ballot position for
>> draft-ietf-uta-rfc7525bis-09: No Objection

<snip/>

>> ### Section 3.7 ESNI as a SHOULD ?
>>
>> Shouldn't ESNI be a normative "SHOULD" ? Or is the non-normative text 
>> "just" to
>> avoid forming a cluster with ESNI draft ? Which would be sad...
> 
> There is always debate about references to works in progress. It's not 
> clear that a protocol still being hashed out in a WG (even a protocol as 
> important as ESNI) can be reasonably considered a best current practice, 
> given that it's far from stable. Eventually, we assume that the IETF 
> will publish rfc7525ter to replace rfc7525bis, at which time that BCP 
> will include updated recommendations (including, I'd expect, ESNI and 
> various other things that are developed between now and then).

The implication of your original comment is that the authors are trying 
to avoid a cluster and push this document through without a normative 
reference to the ECH I-D being worked on in the TLS WG.

This is not what we're trying to do. Although personally I think that 
ECH is consistent with RFC 3365, there are questions and concerns about 
ECH that need to be worked out, the protocol needs to be finalized, 
implementations need to be written or finished, administrators of 
various TLS-based systems need figure out how to deploy it properly, 
etc. For these reasons, I think it is simply too early to recommend ECH 
as a best *current* practice. Hopefully it will be so when this BCP is 
updated again a few years from now.

Nevertheless and in light of the above considerations, I grant that the 
current text might be too strong:

    Implementers are strongly encouraged to support TLS
    Encrypted Client Hello once [I-D.ietf-tls-esni] has been
    standardized.

Something like this might be more accurate:

    At the time of writing, a technology for encrypting the SNI
    (called Encrypted Client Hello) is being worked on in the TLS
    Working Group [I-D.ietf-tls-esni].  Once that method has been
    standardized and widely implemented, it will likely be appropriate
    to recommend or mandate its usage in a future version of this
    BCP.

I'd also suggest changing "this information leak is closed by use of TLS 
Encrypted Client Hello" to "this information leak will be closed by use 
of TLS Encrypted Client Hello once that method has been standardized".

I've incorporated the foregoing suggestions into a PR:

https://github.com/yaronf/I-D/pull/463

Peter