Hi,
In the last few years, I have been interested in cryptography. I took courses at the university on the subject, as well as MOOCs (Coursera Cyprtography I) or reading books. This allowed me to learn about cryptographic primitives and building blocks but not so much about protocols such as TLS or Signal protocol. There is sometimes a brief description with a general overview of how the protocol works and its security properties but nothing on how we can reason and proves these properties. After some research, I found this article https://galois.com/blog/2021/05/who-is-verifying-their-cryptographic-protocols/ or this one https://bblanche.gitlabpages.inria.fr/publications/BlanchetETAPS12.pdf and it seems that proving properties of protocols uses techniques similar to verification of software with software prover. I would like to know if any of you have good resources (courses, books,…) to recommend for learning about these topics. Thank you.
So?
I'm happy to share with you my greatest work yet "The best cryptographic protocol ever!". It contains all valuable hard-earned secret crypto knowledge that I've accumulated over the years!
https://cryptosubtlety.medium.com/the-best-cryptographic-protocol-ever-31ee5108aaa2
During the pandemic I come up with ideas for new software projects every few days, even if I don't really even start any of them and lately I was thinking about gaming and which problems could arise, if you program a open source game (besides the financial problems).
So I came up with a though experiment and want to ask you, what you think about it, and if someone of you maybe even knows about a program that realizes what I'm talking about.
Lets assume a game developed by a few voluntary devs, who maintain a small round based table top game, but the devs want the game to stay functional without themselves, or the devs of a fork being in need to maintain a server for the online multiplayer mode. So the multiplayer mode is planed to work with peer2peer connections without a neutral server controlling the game and maybe a discord or a subreddit where people can find their opponents. Most multiplayer games sooner or later get problems with hackers when they become popular and writing hacks gets much easier if the program is open source, because anyone could just create a fork of the game which just includes hacks. So I was thinking about some ways to define a protocol, which can be used to make at least those kinds of cheating impossible, I was able to come up with, like
- looking at the cards in your opponents hand or other informations you're not supposed to know
- manipulate the outcome of the rng of dice rolls or drawing cards
and I actually had some ideas to solve those problems. For the dice manipulation cheat maybe choosing a seed for the rng at the beginning of the game, sending the encrypted seed to the other player, making a protocol of everything that happens during the game and sending the key for decrypting the seed after the end of the game, so that the client of the other player can check, if everything what my rng did during the game is consistent with the seed chosen at the beginning.
Now I was interested, if games which work like this already exist and googling for a while didn't give me any results. So I'm asking you, if you know any games with such some kind of cryptographic anti-cheat system, or if the idea is to ridiculous to assume that anyone would ever implement it.
More specifically, Assume v1 has a closest vector in set V = {V1, V2, V3, V4} which is V2. I encrypt v1 with some small random noise, r1, and get enc(v1+r1) I encrypt V with some small random noise, and get new set {enc(V1+R1), enc(V2+R2), enc(V3+R3), enc(V4+R4)}. The encrypted vector should have the same closest vector in the encrypted set, in other words, enc(v1+r1) should have closest vector enc(V2+R2) if noise is small enough. Are there any schemes which have this homomorphism and are secure?
Did anyone else spot the update to this security advisory? https://fortiguard.com/psirt/FG-IR-18-100
Last week it was saying upgrading to 6.0.7 was suggested. Now it has changed to say we should be going to 6.2 branch with no mention of the other branches at all.
The news from last week https://www.zdnet.com/article/some-fortinet-products-shipped-with-hardcoded-encryption-keys/
