Re: [MLS] Robert Wilton's Abstain on draft-ietf-mls-protocol-17: (with COMMENT)

[Daniel Kahn Gillmor <[e-mail]>]

On Thu 2023-02-02 06:18:57 -0800, Robert Wilton via Datatracker wrote:
> In the end I decided to ballot abstain because I'm unsure whether
> standardizing MLS in really the right thing to do for end users (which
> may just be my ignorance).  I'm strongly supportive of efforts to make
> messaging platforms interoperate cleanly (e.g., I'm supportive of the
> MIMI WG being chartered) and I appreciate that MLS is likely to
> underpin some of that work, but I also question whether the IETF
> standardizing this will ultimately be beneficial for societies and
> humanity (and note - I'm not advocating for pervasive monitoring, but
> preventing any ability for judicially supported interceptions for
> criminal investigations concerns me). In the limited cases where IETF
> standardization and technology choices are could directly impact the
> effectiveness of law enforcement

This is a pretty disappointing perspective to hear brought into the
IETF.  Does this mean that you're comparably dubious about the work done
in TLS, QUIC, OpenPGP, LAMPS, MASQUE, TEEP, OHAI, etc?  Those
cryptographic and operational frameworks could also "directly impact the
effectiveness of law enforcement."

> or conflict with democratically elected government policies then I
> think that it would be great if the IETF was able to receive and
> consider a wider range of views in the consensus process to ensure
> that we really are making the right choices.

Democratically elected governments (like all governments) can come up
with arbitrarily strange policies.  Indeed, two different democratically
elected governments could come up with two different and mutually
incompatible policies, such that it's not possible to avoid being in
conflict with at least one of them.  (i'll set aside the question of
which governments qualify as "democratically elected")

We are standardizing mechanisms with specific properties, that aim to
facilitate specific goals, however imperfectly.  It sounds to me like
you're raising concernas about the goals of the MLS WG entirely!  Even
if you believe that *some* actors should be able to have extraordinary
access, i think you'd be hard pressed to find (even within the still
relatively homogenous confines of IETF participation) a clear agreement
on *which* actors should be able to have extraordinary access and how we
can be confident that access is limited to them.

As such, we are better off standardizing mechanisms that effectively
lock out everyone but the intended parties to a communication.  If any
participating party wants to share their side of the conversation with
someone else, that's a question that the protocols can't directly
address.  But we should not be contemplating baked-in backdoors if we
want society in general to be able to rely on any of the mechanisms that
we specify.

I would have thought that none of this would need to be said on an IETF
list in 2023, and especially in response to a comment from an AD.  but i
guess the Crypto Wars never did actually stop, did they?

     --dkg