A Classical and Machine Learning-Based Reliability Analysis on Catalan Object Encryption Scheme

Designing
lightweight
secure cryptographic schemes for Internet of Things (IoT) and radio frequency identification (RFID)-based devices is challenging as a designer needs to address resource-constraints along with physical and classical security notions. Even though plethora of such lightweight encryption schemes have been proposed in the literature, computationally efficient attacks on some has also been published, which stresses the fact that robust and reliable design paradigms to consider during design process. In this article, using classical methods, we launched the full-key attack on a Catalan-object-based encryption scheme proposed by Saracevic
et al.
(2020) in

IEEE Transactions on Reliability

. Their proposed encryption scheme is efficient and lightweight specifically meant for IoT applications and is based upon combinatorial structure of Catalan key objects. We proved data privacy violation and recover the full encryption-key with computationally efficient algorithm. We also proposed the machine-learning-based regression model to efficiently predict all bits of ciphertext using just a single plaintext without any key. A high correlation between plaintext and ciphertext and full-key recovery poses the encryption scheme to be unreliable to be used for IoT-based applications. To the best of authors’ knowledge, this is the first work on reliability analysis on Catalan-based encryption schemes.