One of the main features of SMT is that the message encryption scheme is “forward secure”. This means that if the loss or compromize of the recipient’s encryption private key does not allow the attacker to decrypt “past” messages the attacker is assumed to have. The attacker can only decrypt the messages going “forward” with the stolen key. The forward security is achieved generally by “changing the key” with time. For instance with TLS/SSL’s ECDHE provides forward security through the sender and receiver agreeing a new session key for each new TLS session. In SMT, I proposed that the receiver publishes new session keys over time ( as often as it wants ) and that these are back-propagated to the senders. I was worried if there was some “vulnerability” in re-using the ECDH public key of the recipient for an entire session ( in contrast to TLS which uses a new “ephemeral” key for each session ).
Thankfully i came across a question on Crypto Stackoverflow which pointed me to the information about Integrated Encryption Schemes, where the scheme i propose in SMT is basically ECIES which is standardised in IEEE P1363a, but where the static EC public key of the session is changing with time.
SMT could also support a variety of DiffieHellmann based schemes, like ECIES-KEM, PSEC-KEM, ACE-KEM like mentioned in A Proposal for an ISO Standard for Public Key Encryption.
Given that many very clever people have been thinking about this encryption scheme for a long time gives me more confidence in the security of SMT.