@TjerandSilde @StefanoMTessaro @guruvamsip @bwesterb @armfazh @RealWorldCrypto For what it’s worth, OPRFs are not the most widely used primitive for Privacy Pass today — it’s Blind RSA.
English
Christopher Wood
285 posts
@FiloSottile @kevinriggle @pwnallthethings @durumcrustulum @dadrian @cronokirby @hdevalence Do you mean HPKP instead of HPKE?
English
@kevinriggle @pwnallthethings @durumcrustulum @dadrian @cronokirby @hdevalence tlogged signed HTTP exchanges with an HPKE-like permanent opt-in, I guess. Considering how badly HPKE failed, I am not convinced that's deployable, as it sounds like it would be strictly harder.
English
Ah jesus christ why do we have to have the same conversation every other day on here
To E2E encrypt DMs, you will likely have to abandon support for DMs in the web clients.
English
Today...today was a great day! #trailrunning #mountainrunning #hokaoneone #anysurfaceanydistance hokaoneone @ Mount Antero instagram.com/p/CBUJHwLpw4M/…
Colorado, USA 🇺🇸 English
@cryptodavidw From the draft: “The name OPAQUE: A homonym of O-PAKE where O is for Oblivious (the name OPAKE was taken).”
English
what's more annoying on my twitter newsfeed: people baking or people playing animal crossing?
English
@grittygrease @PowerDNS_Bert @satefan @fosdem @bagder @bortzmeyer @jpmens @kazuho +1 — the PR is coming ASAP!
English
@PowerDNS_Bert @satefan @fosdem @bagder @bortzmeyer @jpmens @kazuho As currently written, it's dangerous to implement without a fallback (which hurts security). There are new proposed mechanisms to make it safer.
English
Christopher Wood retweetledi
Mathias Hall-Andersen, @alicharizard, @grittygrease and I just released our design mixing the QUIC protocol with the Noise protocol framework (instead of TLS 1.3) dl.acm.org/citation.cfm?i…
English
R.I.P. Mom.
On so many levels, I would be nothing were it not for you.
Love you.
English
@grittygrease @colmmacc We probably shouldn’t continue propping up the antiquated sockets API.
English
@colmmacc I'm talking about upstreaming the idea to IEEE for consideration in the POSIX standard.
Linux has already moved the record layer to the Kernel (lwn.net/Articles/66650…). Restricting this to TLS 1.3's simpler state machine might reduce the risk somewhat, but I agree it's iffy.
English
The Secure Socket API: TLS as an Operating System Service.
It took 317 lines of code to download something securely with OpenSSL APIs. The authors propose replacing this with a POSIX socket style API for TLS. #usesec18 usenix.org/conference/use…
English
@grittygrease For reference and comparison, take a look at the Network framework for simple-to-use TLS on Apple platforms (developer.apple.com/documentation/…).
English
@rootlabs @Kryptoblog @BerndPaysan @brouhaha Looking forward to them. Public comments on the list are also fine.
English
Haven’t thought this through. But using your private for anything but key operations (signing) makes my hair in the neck stand up. Why on earth is this a good strategy?
datatracker.ietf.org/doc/draft-crem…
English
@rootlabs @grittygrease @Kryptoblog @BerndPaysan @brouhaha That'd work, too. Though, since the signature must not be exposed and is ultimately fed through a PRF, risk here "seems small." I'm happy to chat about this more.
English
@grittygrease @Kryptoblog @BerndPaysan @brouhaha I don’t like that the tag is up to the implementation to generate. You end up signing arbitrary data, where the signature could be reused elsewhere. Better to derive a key via a KDF and a unique but exact tag specified in the standard.
English
Christopher Wood retweetledi
The PostSockets cabal, and friends, starting work on the combined IETF TAPS API - but mostly in need of more coffee to get started!
Milton, England 🇬🇧 English
English
Christopher Wood retweetledi
Prague, Czech Republic 🇨🇿 English
Christopher Wood retweetledi
Christopher Wood retweetledi
Nederlands