Analysis of an Optimal Fault Attack on the LED-64 Lightweight Cryptosystem

Abstract

This paper presents an optimal method for recovering the secret keys of the light encryption device (LED) by combining the impossible differential fault attack with the algebraic differential fault attack. The proposed optimal method effectively improves the performance of fault attacks. A fault attack model, named the redundant random nibble fault model (RRNFM), is proposed to simulate a multiple fault injection in a real-world environment. Using the optimal method and the RRNFM, the 64-bit secret key of LED can be recovered by injecting no more than six faults in 1300 experiments. We establish an equation of inverse proportionality between the number of faults injected and the remaining amount of the secret key. Using the equation, the number of fault injection can be accurately predicted, providing an effective way of evaluating fault attacks. INDEX TERMS Impossible differential fault attack, LED, algebraic differential attack, redundant faults.

Low area field‐programmable gate array implementation of PRESENT image encryption with key rotation and substitution

“…Dong et al [18] proposed a light encryption device for the cryptosystem with the optimal fault attack. This research combined two fault attacks, namely, various impossible attacks and algebraic fault attacks.…”

General Differential Fault Attack on PRESENT and GIFT Cipher With Nibble

“…DFA [1] is a new cryptanalysis method proposed by E. Beniham and A. Hamir based on a combination of mathematical and physical methods in 1997. This method has been applied to many block ciphers, like FOX [2], SMS4 [3], AES [4], LED [5], SIMON [6] etc. Meanwhile, with the development of the Internet of things(IoT), a large number of various lightweight block ciphers have emerged, and have proved its efficiency in resource-constrained environments such as RFID tags and wireless sensor networks(WSN).…”

Breaking Trivium Stream Cipher Implemented in ASIC Using Experimental Attacks and DFA

“…In [26][27][28][29][30] theoretical analyses of fault injections in different versions of AES are presented. This type of analysis has also been theoretically applied to a multitude of both block and stream ciphers, such as Midori, PRESENT, SIMON, SPECK, LED, MICKEY, GRAIN and PLANTLET [3,4,[31][32][33][34][35][36][37]. All these works use the DFA method to theoretically recover the ciphers secret key, using transient fault injections and exploiting the obtained information.…”

“…In the case of the stream ciphers the number of works is even lower, being all of them presented using FPGAs implementations as [44] and as far as we know none of them applied to ASIC implementations. [3,4], [26][27][28], [30][31][32][33][34][35][36][37][38][39][40][41] [25], [29], [44][45][46] [ [42][43][44] In the case of the Trivium stream cipher the most important DFA analyses are the mentioned above [38][39][40][41]. These works are in Group A, and they base their analyses on ideal fault models or scenarios.…”