Home // ICIW 2025, The Twentieth International Conference on Internet and Web Applications and Services // View article

A Real-Time Cache Side Channel Attack Detection and Mitigation Framework Based on Machine Learning

Authors:
Youssra Ghabbara
Zied Trifa

Keywords: Cache side-channel attack; Hardware performance counters; Machine learning; Cloud Computing.

Abstract:
Cloud computing has transformed data management by offering scalable and flexible services to organizations. However, its multi-tenant architecture, where users share physical hardware, introduces critical security risks related to data isolation and resource sharing. Among these, cache side-channel attacks (CSCAs) represent a particularly dangerous threat. These attacks exploit shared cache memory, using variations in cache access times to extract sensitive information from co-located virtual machines. The shared nature of cloud hardware and the difficulty of detecting such attacks in real-time make cache side-channel attacks especially challenging to address. Traditional security measures, such as encryption and access control, fall short against these threats, as they do not target vulnerabilities in the cloud underlying hardware architecture. This research proposes a real- time detection and mitigation framework leveraging machine learning to address the pressing issue of cache side-channel attacks. Key focuses of this study include designing an efficient feature extraction mechanism to identify malicious cache behaviors and selecting machine learning algorithms that provide high detection accuracy with minimal latency. Additionally, the framework incorporates real-time mitigation strategies designed to minimize performance degradation in cloud environments.

Pages: 1 to 6

Copyright: Copyright (c) IARIA, 2025

Publication date: April 6, 2025

Published in: conference

ISSN: 2308-3972

ISBN: 978-1-68558-249-4

Location: Valencia, Spain

Dates: from April 6, 2025 to April 10, 2025