I'm trying to practice and to understand possible scenarios and the basics of symmetric encryption schemes. For the scenarios I listed below, I am not 100% sure of scenarios 1 and 3.

    1. A symmetric encryption algorithm using non-random/predictable IV that is CPA secure. I do not think this scheme is possible. For example, there is the BEAST attack for the TLS 1.0 scheme on AESCBC encryption schemes. Random IVs are probably needed to make them CPA-secure.
    1. A symmetric encryption algorithm with equal constant size key, message, and ciphertext space that is semantically secure. It is possible because of OTP One Time Pad, isn't it?
    1. A symmetric encryption algorithm with equal constant size key and message space of magnitude X, with cipher text space of 2^X that is CPA-secure. I think this is possible as the AES-CBC encryption scheme follows these requirements.

I will be thankful for clarification and help with these scenarios.

  • $\begingroup$ 1) It is possible, see the counter mode. 2) What do you mean by constant size? the key message and ciphertext have the same length? 3) Does magnitude mean cardinality? $\endgroup$ Dec 6, 2022 at 8:59
  • $\begingroup$ Please clarify your specific problem or provide additional details to highlight exactly what you need. As it's currently written, it's hard to tell exactly what you're asking. $\endgroup$
    – Community Bot
    Dec 6, 2022 at 12:03

1 Answer 1

  1. Counter-example: counter mode. (Pun intended in an afterthought). Most of the attacks like BEAST can be deterred by using "authenticated" encryption.

  2. I suppose yes? I don't fully get your thinking about this part.

  3. True. Almost all modern ciphers are capable of encrypting data that's orders of magnitude big. (e.g. various mode of operations of blockciphers intended for guaranteeing confidentiality, special designs such as Salsa20/ChaCha20, etc.)

  • $\begingroup$ 2. is about semantic security so, I am not sure bitfliping attacks are appropriate or relevant for the question. $\endgroup$ Dec 6, 2022 at 12:33
  • $\begingroup$ @MarcIlunga I'm weak on crypto maths, but I think there are different types (e.g. KMA, CPA, CCA) of adversaries against which semantic security should be achieved. $\endgroup$
    – DannyNiu
    Dec 6, 2022 at 12:53
  • $\begingroup$ Semantic security has a precise definition outlining the capabilities and restrictions on the adversary (so we can't necessarily apply a CCA attack to a scheme that we only expected to have CPA or semantic security). $\endgroup$ Dec 6, 2022 at 13:25
  • $\begingroup$ In fact the OTP can be shown to be semantically secure: see cs.princeton.edu/~mzhandry/courses/2018-Spring-COS433/LN/… or crypto.stanford.edu/~dabo/courses/OnlineCrypto/slides/… $\endgroup$ Dec 6, 2022 at 13:26
  • $\begingroup$ @MarcIlunga I rolled back my answer. $\endgroup$
    – DannyNiu
    Dec 7, 2022 at 0:45

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.