Home // SECURWARE 2025, The Nineteenth International Conference on Emerging Security Information, Systems and Technologies // View article

Identification of Dual Processes Using Power Side-channels

Authors:
Jakob Sternby
Niklas Lindskog
Håkan Englund

Keywords: Security; Side-channel Monitoring; Dual-core.

Abstract:
Malware is one of the main threats against electronic devices, as malicious software can damage the device, disrupt network communications and provide an entry point for additional attacks. While software-based countermeasures such as antivirus can be effective, they require presence on the device and can furthermore be disabled or fooled by advanced malware. Monitoring of physical side-channels, on the other hand, provides a non-invasive and hard-to-spoof method to detect unauthorized software being executed on a device. However, in a modern device, several processes may execute at once, making detection of alterations difficult, especially in the case where more than one process is security sensitive and should be monitored. In this paper, we present a solution for enabling granular side-channel monitoring for complex, multi-core devices. We apply new machine-learning enhanced methodology, focused on efficiently representing the measurements in latent space, to enable classification of two simultaneously executing processes. The classification training is based on labeled power side-channel traces of dual-core measurements. Our results show that it is feasible to classify two processes on separate cores having observed a single power trace obtained from a single probe.

Pages: 27 to 32

Copyright: Copyright (c) IARIA, 2025

Publication date: October 26, 2025

Published in: conference

ISSN: 2162-2116

ISBN: 978-1-68558-306-4

Location: Barcelona, Spain

Dates: from October 26, 2025 to October 30, 2025