Differential Fault Attack on ChaosForge

In this paper we propose a Differential Fault Attack (DFA) on ChaosForge. ChaosForge is a stream cipher based on three nonlinear feedback shift registers introduced for secure Internet of Things (IoT) communication. We analyze the security of this stream cipher against differential fault attack. To the best of our knowledge this is the first cryptanalysis on this cipher which recovers the 80 -bit secret key in practical time. We mount our differential fault attack on ChaosForge in 11094 seconds by injecting only 6 faults in random locations of the state of cipher during the keystream generation. We initially show how the location of the injected faults can be detected. We recover the initial state of the cipher by solving the system of equations generated using normal and faulty keystreams. The secret key is recovered further by inverting the key scheduling algorithm of the cipher.