A Survey on Harnessing High-Performance Computing and Machine Learning for Detecting Stealthy Cache-Based Side-Channel Attacks

This survey investigates different approaches to improve the ability of cloud systems to withstand cache-based side-channel attacks, with a primary emphasis on real-time detection capabilities and the ability to identify stealth attacks, incorporating machine learning and High-Performance Computing (HPC) techniques. The study categorises the security measures into hardware, hypervisor, and two-stage strategies. Additionally, the research assesses the effectiveness, accuracy, and impact of various defence mechanisms on cloud systems, specifically considering High-Performance Computing (HPC). The main objectives include analysing the occurrence of false-positive results, evaluating the impact of defences on cloud security and performance, and assessing the feasibility of employing machine learning and HPC for detecting stealth attacks in real-time.