(Un)breakable Curses – Re-Encryption in the Fujisaki-Okamoto Transform

The Fujisaki-Okamoto transform (FO) is the go-to method for achieving chosen-ciphertext (CCA) security for post-quantum key encapsulation mechanisms (KEMs). An important step in FO is augmenting the decryption/ decapsulation algorithm with a re-encryption step -- the decrypted message is re-encrypted to check whether the correct encryption randomness was used. While solving a security problem (ciphertext-malleability), re-encryption has turned out to introduce side-channel vulnerabilities and is computationally expensive, which has lead designers to searching for alternatives. In this work, we perform a comprehensive study of such alternatives. We formalize a central security property, computational rigidity, and show that it is sufficient  for obtaining CCA security. We present a framework for analyzing algorithms that can replace re-encryption and still achieve rigidity, and analyze existing proposals in this framework.

Along the way, we pick up a novel QROM security statement for explicitly rejecting KEMs based on  deterministic PKE schemes, something that so far only was possible when requiring a hard-to-ensure quantum property for the base PKE scheme.