Quantum resistant algo for TLS


AWS already adopted quantum resistant encryption for TLS. What plans does Storj has?

This would be effective only for sites which uses TLS. The Storj protocol uses a different technology to encrypt data.
However, I passed your question to the team.

I was thinking to the connections to sats and clients. I saw in logs “TLS connection established…” or something like that.

It’s a little bit more complicated. The node’s identity is a certificate, so it uses a TLS some kind to establish a connection, but the data itself is encrypted in a different way, client-side. See How Encryption is Implemented - Storj DCS Docs
In other words, to break the TLS and intercept the encrypted data is not enough to get the data. You also need to collect at least 29 pieces (which are not exposed, where they are), and then - assemble them to the one segment and only then try to break an encryption. But it’s also not the end of the story. You need to find all segments of the file and combine them in a right sequence. For this you need also break the metadata information from the satellite, which is also encrypted. And also gather the API key, or guess it before our civilization would disappear…
So,… good luck to break all of this even with a quantum computing.


We may begin using something like Kyber in TLS key exchange at some point when it has seen a bit wider usage. But, as @Alexey says, cracking our TLS layer would only tell you things like “node N was asked to store piece P of segment S”, “node N asked for payment of X amount of bandwidth”, and so on. User objects and paths would still be encrypted with AES or SecretBox (XSalsa20-Poly1305). There are no known ways to attack those specifically with a quantum computer, except for the general application of Grover’s algorithm, which would cut down attack time to its square root (i.e., AES256 would be about as secure as AES128 is in our pre-quantum world).