Real-Time Detection and Categorization of Cache Side-Channel Attacks Using Deep Learning and Morlet Wavelet Assistance
Lakshminatha Reddy C, Malathi Kanagasabai · International Journal of Electronics and Communication Engineering · 2024
Cache Side Channel (CSC) attacks track user’s cache activity to get private data. This attack often targets the L3 cache that the user and the intruder share. As a result, an intruder can gather private data without tipping off the target. The approach for real-time detection of CSC attacks is developed in this research. The proposed method quickly detects CSC attacks after detecting the change in the CPU numbers. To do this, the value of the CPU counters is measured using a Deep Learning (DL) network with Intel Morlet Wavelet assistance and Weighted Mean of Vectors (WMOV) optimization. Understanding the kind of data that is released is crucial for distinguishing and categorizing cache-based side-channel attacks. Timing attacks, commonly referred to as time-driven attacks, produce quantitative implementation data pertaining to timing. Multiple counters captured changes throughout the attack through the course of the research. The work also presents a categorization of these attacks according to the information-leaking source. After that, a qualitative examination of the effectiveness, complexity, and vulnerabilities of the target cryptosystems of these attacks is conducted. The results of the experiments demonstrate that the proposed detection method performs well for real-time detection in a variety of situations.