Authors: Carlos R Aguayo González Jeffrey H Reed
Publish Date: 2011/09/29
Volume: 69, Issue: 2-3, Pages: 307-
Abstract
SoftwareDefined Radio SDR provides a flexible platform that facilitates radio resource management and enables new technologies and applications Unfortunately their reliance on software implementations makes them vulnerable to malicious software attacks that could impact their spectral emissions and disclose sensitive information It is of critical importance for the widespread deployment of SDR to develop technologies that enable effective integrity assessment of communications platforms and timely detection of malicious intrusions We provide further evidence of the feasibility of a novel approach called Power Fingerprinting PFP that enables an effective mechanism to perform integrity assessment of SDR PFP relies on an external monitor that captures finegrained measurements of the processor’s power consumption and compares them against stored signatures from trusted software by applying pattern recognition and signal detection techniques Because it is implemented by an external monitor PFP causes minimal disruption on the target system and also provides the necessary isolation to protect against malicious attacks to the monitor itself Finegranularity measurements deliver improved visibility into the execution status and make the PFP monitor difficult to evade while the reliance on anomaly detection from trusted references makes it effective against zeroday attacks We present the results of different feasibility experiments that support the applicability of PFP to SDR integrity assessment In the first experiment a PFP monitor is able to effectively detect the execution of a tampered routine that misconfigures the operational mode of a PICDEM Z radio platform affecting the resulting spectral emission In a second experiment our monitor effectively identifies when a transmission routine is modified affecting encryption settings We also present an approach to improve the performance of PFP by characterizing the way a specific platform consumes power This platform characterization which can be done using principal component analysis or linear discriminant analysis allows a PFP monitor to work only on the features that carry the most information As a result the PFP monitor is able to detect execution deviations resulting from a difference of a single bit transition the smallest possible disruptionThe authors would like to thank Wireless Virginia Tech Affiliates for their support This work was supported in part by the National Science Foundation under Grant CNS0910531 Any opinions findings and conclusions or recommendations expressed in this material are those of the authors and do not necessarily reflect the views of the National Science Foundation or Wireless Virginia Tech Affiliates
Keywords: