Constructing a CPA-Secure: Pseudo-Random Permutations and Block Ciphers, Modes of Operation, Security Against Chosen-Ciphertext Attacks (CCA)

Abstract

This paper examines symmetric-key encryption as a layered security construction, tracing how cryptographic guarantees propagate from primitive to protocol. The study is conducted as a structured literature review of foundational and recent (2021–2025) cryptographic research on block cipher modes of operation and their resistance to chosen-plaintext and chosen-ciphertext attacks. At the foundation lies the block cipher, modeled as a pseudorandom permutation (PRP) whose security rests on cryptanalytic conjecture rather than provable hardness. Building on this, modes of operation including ECB, CBC, CFB, OFB, and CTR combine block-cipher calls to encrypt arbitrary-length messages, with security formally reduced to the underlying PRP assumption under indistinguishability against chosen-plaintext attack (IND-CPA). The review finds that while CPA security is necessary, it is insufficient for real-world deployment, since adversaries in network settings routinely gain oracle-like access to decryption; this is evidenced by recurring vulnerabilities such as padding-oracle attacks and related exploits against CBC-mode TLS. It further finds that the stronger requirement of indistinguishability under chosen-ciphertext attack (IND-CCA) is achieved through authenticated constructions such as Encrypt-then-MAC and, increasingly in current practice, integrated Authenticated Encryption with Associated Data (AEAD) schemes such as AES-GCM. The paper concludes that authenticated encryption should be the default standard in protocol design, closing the theoretical–practical gap that has historically enabled real-world cryptographic exploits.