Improved Timing Attacks against the Secret Permutation in the McEliece PKC

Abstract

In this paper, we detail two side-channel attacks against the McEliece public-key cryptosystem. They are exploiting timing differences on the Patterson decoding algorithm in order to reveal one part of the secret key: the support permutation. The first one is improving two existing timing attacks and uses the correlation between two different steps of the decoding algorithm. This improvement can be deployed on all error-vectors with Hamming weight smaller than a quarter of the minimum distance of the code. The second attack targets the evaluation of the error locator polynomial and succeeds on several different decoding algorithms. We also give an appropriate countermeasure.

Key-Recovery by Side-Channel Information on the Matrix-Vector Product in Code-Based Cryptosystems

“…As in the first attack, this approach only works for Hamming weights of 2, 4, or 6 of the error vector e to recover the secret permutation matrix P . Despite the improvements in [6], the main problem with these two previous attacks is the number of cases (depending on the Hamming weight of e) that can be exploited to find the secret. In our attack, we have no constraints on the Hamming weights of the ciphertext (or error vector e) to find the secret matrix Q in the case that we correctly construct the matrix Z * with a random forest.…”

Defense Scheme to Protect IoT from Cyber Attacks using AI Principles

“…Nevertheless, as in any type of security system, total prevention of intrusion is impossible. The node compromise and intrusion heads to secret information like security keys being disclosed to the intruders in the system, which results collapse of the security mechanism [10]. Consequently, IDSs are developed and designed to make intrusions public, before disclosing the system resources which are secured.…”

“…• Rule Based: This technique compares data against signatures with state transition analysis. Every data packet is practiced to FSM (Finite State Machine) and follows transitions till the ultimate state reaches, resulted in the detection an attack [8,10].…”

“…The data which violates the defined standards is considered to be malicious. This approach has good success rate in lucrative systems but the flaws are producing erroneous positives for undefined protocols [10].…”