Quote from: samuel999 on November 18, 2015, 01:31:42 AM

The only issue I can see with both of these schemes is that the recipient must check every single message he receives to see if any of the encrypted messages will be decrypted by his key.

My old core2duo cpu able to decrypt ECC + AES at 100messages/sec (first solution) and AES over 10000messages/sec (second solution), so I don't think its going to be a problem. Every site in zeronet is totally independent, so if the updating/verification become problematic anyone able to clone the site and creating a new network of messages.

Quote from: samuel999 on November 18, 2015, 01:31:42 AM

However, someone could be encrypting random data to send to the recipient, which when decrypted, will look just like any other ciphertext that is encrypted with another key since decrypting a message not meant for you will also result in random data.

To solve this we could add fixed prefix to the encrypted text, eg.: os.urandom(16)+"ZNE1"+message
The random needed because according to stackexchange it should be avoided to start the messages with same text. (we going to use AES CBC mode)