A Critical Analysis on the Security Attacks and Relevant Countermeasures using ML

In Side Channel Analysis (SCA), machine learning algorithms have shown their efficacy, offering several upgrades over the well-established profiling procedure of Template Attacks. A design concept and method that may be successfully utilised as a foundation for adding SCA safeguards on the scalar Elliptic Cryptography (ECC) integrators is provided, with a focus on the necessity to limit their impact on embedded systems. The deconstruction of the round computations of the Scalar Multiplication (SM) and Lafayette Power Ladder (MPL) algorithms into the highlighted finite site activities and their reorganisation into single phase operation sets form the foundation of the suggested design approach. We demonstrate how sophisticated SCA measures may be quickly implemented, focused on randomising the projective values of the ECC point outputs from the MPL round, utilising the suggested design technique as a foundation. A roadmap for the assault is presented, and numerous simple ML-based SCAs are conducted in order to assess the approach and associated SCA responses. The suggested roadmap makes the assumption that attacker would have limited funds and won't gain access to a large amount of leaks trails in order to launch Deep Learning assaults. When SCA remedies are included into the suggested design method, The performance of the trained models shows a strong resilience to ML-based SCAs.