Abstract
Industrial Control Systems are under increased scrutiny. Their security is historically sub-par, and although measures are being taken by the manufacturers to remedy this, the large installed base of legacy systems cannot easily be updated with state-of-the-art security measures. In these publications we use a technique from cryptographic side-channel analysis, multivariate templating, to detect anomalous behaviour in Programmable Logic Controllers. Our solution uses side-channel measurements of the electromagnetic emissions of an industrial control system to detect behavioural changes of the software running on them. To demonstrate the feasibility of this method, we show it is possible to profile and distinguish between even small changes in programs on Siemens S7-317 PLCs, using methods from cryptographic side-channel analysis.
This dataset consists of raw electromagnetic trace files captured on 2017-05-27. It can be used to reproduce the results in the aforementioned publications. For three different programs, 16 different inputs were captured. Each input has 100.000 traces.
Analysis (& capturing) code is available as a separate dataset at doi:10.17026/dans-x7m-6222.
Please note that although this dataset of EM traces is published under CC0 -- effectively waiving all copyright and related or neighbouring rights -- the *code* dataset is licensed under GPLv3 and thus subject to the limitations of that license.
