 So I'm here to talk about a new RSA variant that I have been developing with Brian Lemakia So RSA we've all seen the shirt. We all know how it works You know the primes Modulus, etc. You got it But RSA is becoming obsolete. It's not post-quantum secure and a lot of side channel attacks fault attacks Keys are too big so You know to to make our variant of RSA first thing you need to do is generate keys, right? So this is the traditional RSA key generation algorithm. You start with your exponent You pick your primes and you just loop to make sure that The public exponent is relatively prime to E of n So with just a single character, we can change that to make sure that fee of n is in fact divisible by the public exponent and then we have a single letter change to the name and Indeed It doesn't make any sense to calculate the decrypt exponent any longer so we can just throw that out Shirt still works we just need to get rid of that one line so What's going on here with our new RSA and corruption algorithm is that if the public exponent divides fee of n Then you are going to have Elements mod n integers mod n that have order e and then in which case all the integers mod n can be written as a product of an order e element and an order you have n divided by e element so if you do public the public exponent Encryption what you end up with is you lose that factor of L And also there's no decrypt exponent, so it's not immediately clear how you would get anything back from this, but you know So alright, so the strength of RSA encryption is that is in fact a one character fix to the RSA algorithm Both in the code and name But we've got some problems. It's Still not post quantum secure the keys are still too big and you also can't decrypt But oh no, what if you're really excited by the beginning of this talk and you encrafted some of your really important stuff In which case we've got a decryption algorithm for you So assuming that the message has been padded with some known padding algorithm and that the public exponent divides fee of n But e squared doesn't divide fee of n and then if we're given the ciphertext and the prime factors p and q What we can do is then calculate d as e inverse mod fee of n divided by e and then we pick a g a Generator g of the e-order elements mod n initialize an empty set Set z equal to c to the d We just calculated mod n and then we'll just loop over all one up to the public exponent minus one and try each one of those times z and check the padding on that and If it does have valid padding then we add it to the set and when this is all set and done You should have either one or just a very small number of possible decryptions Decryptions that you can then check and see if that's the one that you're looking for So message here is don't crap yourself. All right But if you do it's okay We've got a tool that we can use to decrap things So please contact us if you're interested or if you've accidentally crapped yourself or your important stuff Thank you very much