At 8:56 AM 11/9/95, Ali Bahreman wrote:
Am I making sense saying that non-repudiation of "message receipt" is a
different service than non-repudiation of "receipt". The latter means that the
"receipt" is digitally signed, whereas the former implies that the recipient
cannot repudiate having received the message. These are two different services
and need two different mechanisms to support them.
Ali
Ali,
Based on first principles and the 7498-2 definition,
"non-repudiation with proof of message receipt" ought to imply that if the
recipient read the message, then the originator gets a receipt to that
effect. In contrast, "non-repudiation with proof of deliver" merely
provide a receipt to the originator stating that the message was delivered
to the recipient, irrespective of whether the recipient read the message or
not. These are legitimate, distinct security services. The latter is
provided by X.400, but the receipt is computed on the ciphertext (if the
message is encrypted), so the effect is a lot like certified mail with
return receipt, i.e., you get proof that something was delivered, but not
proof of what content was delivered. One can argue whether our existing
message transport system ought ot be trusted to provide this sort of proof,
but that is what the X.400 model envisions. ("Trust me, I from the PTT!")
Proof of receipt (vs. delivery) is hard to do with high assurance, without
resorting to the sort of multi-message exchange using a trusted third
party, as described by Tygar et al.
Steve