Agenda IETF124: pearg
agenda-124-pearg-00

PEARG Agenda
11:30 Note Well and Status - Chairs, 5 mins

11:35 Chat Control - Current Status and Discussion - Mallory Knodel, 30 mins

12:05 Measuring and Understanding ECH Deployment, global trend and challenges - Arturo Filastò - Founder, ED & CTO @ OONI, 40 mins

12:45 Buffer and Wrap-Up Discussion of RG Horizons, 10 mins

Arturo Filastò Abstract
Encrypted Client Hello (ECH) is a IETF draft that adds support for encrypting the ClientHello message of the TLS handshake, improving privacy and incidentally, censorship resistance [1]. Despite browser support since 2023, server-side adoption remains limited, reducing its potential for privacy and censorship resistance. Additionally, some countries [3] have implemented blanket blocks affecting its availability and users have turned to disabling it server-side to avoid disrupting their services [4].

This talk presents OONI’s global measurements of ECH availability [5] and hope this leads to a discussion about the double-edged sword of how censorship resistance technologies may lead to unintended consequences.

[1] https://datatracker.ietf.org/doc/draft-ietf-tls-esni/
[2] https://support.mozilla.org/en-US/kb/understand-encrypted-client-hello
[3] https://github.com/net4people/bbs/issues/417
[4] https://community.cloudflare.com/t/access-issues-in-russia-unable-to-disable-ech-in-free-plan/733401
[5] https://github.com/ooni/spec/blob/master/nettests/ts-039-echcheck.md

I’d like to discuss how we can anticipate and address unintended consequences when designing censorship-resilient protocols, drawing lessons from ECH’s deployment trajectory.

It can be argued that what happened with ECH is analogous to what occurred with TLS, where network operators were put in a position to either block everything or nothing (i.e. everything with SNI == example.com, vs only http://example.com/content).