An Effective Simulation Analysis of Transient Electromagnetic Multiple Faults.

Area of Science:

• Cryptography and Information Security
• Hardware Security and Embedded Systems

Background:

• Embedded encryption devices and smart sensors are susceptible to physical attacks, including laser, particle radiation, and electromagnetic transient injection.
• Adversaries often inject multiple transient faults, making traditional single-nibble fault analysis insufficient.
• The exact number of injected faults is typically unknown to the attacker, complicating cryptanalysis.

Purpose of the Study:

• To propose a novel ciphertext-only impossible differential analysis method.
• To enable the analysis of random multiple faults affecting block ciphers.
• To enhance the security of embedded encryption devices against physical fault injection attacks.

Main Methods:

• Development of ciphertext-only impossible differentials capable of analyzing up to six random faults.
• Utilizing impossible differentials to eliminate non-existent secret keys.
• Employing inverse difference equations to progressively derive the unique secret key.

Main Results:

• Successful software simulation of 32,000 random multiple fault attacks on the Midori block cipher.
• Verification of the theoretical model for multiple fault attacks and establishment of fault injection-to-information content relationships.
• Demonstrated that the optimized fault attack method can recover the secret key at least 11 times, reducing the number of required attacks.

Conclusions:

• The proposed ciphertext-only impossible differential analysis is an effective strategy for random multiple fault analysis.
• This method significantly improves the security of block ciphers against sophisticated physical attacks.
• The findings contribute to developing more robust embedded security solutions.