Signal puts a lot of effort into their threat model that assumes a hostile host (i.e. AWS). That's the whole point of end to end encryption, even if the host is compromised the attackers do not get any information. They even go as far as padding out the lengths of encrypted messages so everyone looks like they are sending identical blocks of data

or federated server

I guess the research doesn't have to be limited to signal. If other apps can benefit from it the more resilient "private communications over the internet" get.

So that's why Signal didn't send my messages very quickly today then, maybe.

I think the biggest thing here is that beyond just a session key (to make sessions secure from each other), this approach uses a rotating session key. That means each transaction in a sesssion is unique ensuring forward and backward secrecy.

I may have read it wrong plus cybersecurity is not my forte.

TLS can do key rotation too, it's just not needed in practice. The QUIC RFCs talk about it a bit, e.g. when using ChaCha20 the key usage limit is actually larger than the defined limit for transmitted data, making it only potentially relevant for AES.

Isn't asymmetric used for the handshake only? And then like AES or something which have evolving keys (and are quite quantum resistant).

I believe it's the same thing I noted in another comment - https://lemmy.world/comment/19924290

Basically session key vs rotating session key, but to be honest I only skimmed your link as it got dense fast.

Only for account creation, afterwards you can hide it and use username.

I understand what you're saying here, and I agree, BUT...

I think it's also important to understand that signal is also a company that, at some ooint, needs to make money from somewhere to do this awesome thing and they won't get that limiting themselves to an obscure app store that maybe 1 in 100 users even know about.

I also highly doubt that Google would modify the signal binary with a backdoor. When that happens, truth will come out within days and it would severely damage Google's reputation, and at least the EU would be screaming bloody murder to push new stores from independent providers immediately. It would end bmvery badly for Google, so I think there is honest doubt about that argument.

So yeah, they should at least ALSO host a version in fdroid, give users the option where to install it from

The entire thing of limited app stores is just absolutely removed in the first place, it's nice and easy (taken the idea from open source there) but implemented in a way that only AND ONLY benefits the vendor Google. Or apple, same shit, different name.

Phones should have an operating system where you can add app stores to. Linux phone, here I come!

Fortunately, mathematical formulas tend to be surprisingly rugged if implemented well.

The underlying problem (Shortest Vector Problem) is known to be NP-Hard.

We know, via mathematical proof, that if there is a solution to SVP that's solvable in polynomial time then it would necessarily mean that it is possible to solve any NP problem in polynomial time.

This would be astonishing, to put it mildly. It would be as if physics suddenly discovered that things could move faster than light and have negative mass. Physics would get wormholes and computer science would get "Arthur C. Clark magic"* programs.

(*“Any sufficiently advanced technology is indistinguishable from magic.” )

Yeah wireguard has something also, as a post quantum security aid.

You have the private and public keypair, but then an addition preshared key.

I was mainly sharing it for those that are interested in the details of the changes. I agree they're not the first, and groups have been making these changes for a few years now. Signal did another change a few years ago:

https://arstechnica.com/security/2023/09/signal-preps-its-encryption-engine-for-the-quantum-doomsday-inevitability/