Paper on Microarchitectural Leakage Templates accepted at CCS 2022

Our paper Microarchitectural Leakage Templates and Their Application to Cache-Based Side Channels has been accepted at ACM CCS 2022. This is particularly exciting for me because it comes with a number of “firsts”: It is the first paper at a tier-1 conference that I contributed to, and it is also the first paper that I will present at an in-person conference.

The publication can be seen as a follow-up to the paper Validation of Abstract Side-Channel Models for Computer Architectures by Nemati et al.; in that work, the authors present an approach to find microarchitectural side channels using observational models. The idea is to generate programs and predict their behavior on a CPU by means of a model. This prediction is compared to the actual behavior, and any deviations indicate potential side channels. This approach produces concrete (potentially leaky) code samples, but it remained a hard problem to abstract from these samples and to describe the underlying side channel in detail.

In the new paper, we propose a solution to that problem. First, we show how a side channel can be described in a generic way, or, in our terminology, as a Leakage Template. Second, we describe how such a Leakage Template can be created for a specific side channel (such as a channel that was identified as a violation of an observational model). And lastly, we show how this abstract representation can be used to identify instances of the side channel in real-world binaries.

If you want to learn more, have a look at the paper or join my presentation at CCS 2022 (Session 7G on Thursday, November 10, 2pm). I will publish my slides on the publication page after the talk. We also plan to produce a video that I will link there.