2
$\begingroup$

Does it work if we try to construct a device (e.g., for counting purpose in several remote branches of a company) using a self-containing digital signature key enclosed in a tamper-proof hardware to make sure that the operations transcript of that function could be verified? (the transmitted report/transcript of the counting will be signed by the internal tamper-proof hardware)

Is there any concern by giving the adversary (possibly, if s/he owns the device environment) the chance of using that 'engine' to sign, even if we consider that the inviolability of the tamper proof hardware is absolute? (the whole device is not inviolable)

$\endgroup$
2
  • $\begingroup$ Obvious question: how would the tamper-proof hardware validate that the transcript report it has been given is actually correct? $\endgroup$
    – poncho
    Dec 6, 2022 at 22:55
  • $\begingroup$ @poncho, we know that the usual place of a tamper-proof hardware is in the 'server side'/'administrators'/'owner' side (am I correct?). So, sure my application example is obvious, but my point is: we can't expect no security when we place digital certificate in the "client", even in a tamper-proof hw; right? $\endgroup$ Dec 9, 2022 at 20:40

1 Answer 1

1
$\begingroup$

I believe that when you can hypothetically give the adversary a signature engine, even under the assumption that the engine is physically protected and the adversary could not access the internal keys (remember: side-channel attacks), you are giving him much power: the reference is the way we can prove the security of signature schemes - The Exact Security of Digital Signatures How to Sign with RSA and Rabin.

I would say that the general idea of that kind of proof is based on the message transcript view the adversary has. S/he tries to challenge the key owner by sending and receiving messages to be signed. If you place the adversary in a position where s/he can control the signature process, that doesn't make sense: s/he always can win the challenge; or, you're giving much more than a simple transcript view. Which contradicts the proof model described in the literature.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.