zinstack625 / gossip Goto Github PK

This opens the possibility of an attack, resulting in discovery of encryption data by an unwanted figures. Solution is to renew iv with every new message.
IV could be randomly generated and sent with the whole message, but that poses a problem in desynchronization in case of network fragmentation. It could also be pseudo-randomly generated with initial IV as seed, in which case fragmentation could probably be easier to recover from, but still remains an issue. In any case, message transport function will suffer severe reworking and receiving messages will be a nightmare.
One other solution I can think of is using multiple nonces and choosing between them randomly before the transmission, then brute-forcing through those at decryption. This shouldn't be too much of an issue, considering most CPU's have hardware-accelerated encryption and the worst that could happen with incorrect nonce is decryptor throwing an error, but still seems like a hack to me