Published April 12, 2023 | Version v1

Metior: A Comprehensive Model to Evaluate Obfuscating Side-Channel Defense Schemes

Description

Our artifact is comprised of a cache model, a cache-occupancy attack implementation, and a Maximal Leakage calculator. The cache model can be configured to simulate various different secure cache architectures (including fully-associative random replacement caches and MIRAGE) under varying sizing configurations. The cache occupancy attack implementation leverages this cache model to simulate the attacker’s (probabilistic) observations for a given attacker/victim scenario. The top-level script (runSweep.py) samples a large number of observations from this attack implementation, sweeping across varying attacker strategies for a given victim space. Using these samples, our artifact provides an estimate of an attack’s Maximal Leakage.

We have provided an end-to-end flow to replicate the results of Case Study I/Figure 7, demonstrating the leakage of cache occupancy attacks waged against an AES implementation for a globally random replacement cache architecture. The provided scripts sweep varying cache-occupancy attacker strategies (including the number of lines primed, and the number of priming iterations). This flow can be extended by researchers to easily derive the Maximal Leakage values for their own simulations.

Files

metior_artifact.zip

Files (19.8 kB)

Name Size
md5:6a0c6391512a276763fe3fe5934a28fa
19.8 kB Preview Download