Email security with diagram and protocols

The e-mail security method discussed here is PGP / S/MIME-style end-to-end protection, backed by transport security (SMTP over TLS, STARTTLS). Plain e-mail (SMTP, POP3) is sent as readable text, so it can be read, altered or forged on the way.

Security services required

• Confidentiality: only the receiver can read the message.
• Authentication: the sender is who he claims to be.
• Integrity: the message was not changed.
• Non-repudiation: the sender cannot deny sending it.

Pretty Good Privacy (PGP)

PGP uses a hash, a public/private key pair and a one-time symmetric session key.

Sender

1. Hash the message and encrypt the hash with the sender's private key (digital signature).
2. Compress the message + signature.
3. Encrypt it with a random session key (e.g. AES).
4. Encrypt the session key with the receiver's public key.
5. Convert to ASCII (Radix-64) and send via SMTP.

Receiver: reverses the steps. He decrypts the session key with his private key, decrypts the message, decompresses it, and verifies the signature using the sender's public key.

 SENDER                                   RECEIVER
 Msg -> Hash -> Sign(KR_A) -+
                            v
 Msg || Sig -> ZIP -> Enc(KS) --> [SMTP/Internet]
 KS -> Enc(KU_B) ----------------^        |
                                          v
        Dec(KR_B)->KS -> Dec(KS) -> UNZIP -> Msg || Sig
        Verify Sig with KU_A -> Authentic + Intact

Protocols involved

Key management

• PGP uses a web of trust: users sign each other's public keys.
• S/MIME uses a certificate authority (CA) hierarchy.

Limitations

TLS protects only individual hops, while PGP/S/MIME protect the message end to end (but headers such as sender and subject stay visible).