Constrained Concealment Attacks against Reconstruction-based Anomaly Detectors in Industrial Control Systems

Constrained Concealment Attacks against Reconstruction-based Anomaly Detectors in Industrial Control Systems

Erba, Alessandro (CISPA Helmholtz Center (i.G.))
Taormina, R. (TU Delft Sanitary Engineering)
Galelli, Stefano (Singapore University of Technology and Design)
Pogliani, Marcello (Politecnico di Milano)
Carminati, Michele (Politecnico di Milano)
Zanero, Stefano (Politecnico di Milano)
Tippenhauer, Nils Ole (CISPA Helmholtz Center (i.G.))

2020

Recently, reconstruction-based anomaly detection was proposed as an effective technique to detect attacks in dynamic industrial control networks. Unlike classical network anomaly detectors that observe the network traffic, reconstruction-based detectors operate on the measured sensor data, leveraging physical process models learned a priori. In this work, we investigate different approaches to evade prior-work reconstruction-based anomaly detectors by manipulating sensor data so that the attack is concealed. We find that replay attacks (commonly assumed to be very strong) show bad performance (i.e., increasing the number of alarms) if the attacker is constrained to manipulate less than 95% of all features in the system, as hidden correlations between the features are not replicated well. To address this, we propose two novel attacks that manipulate a subset of the sensor readings, leveraging learned physical constraints of the system. Our attacks feature two different attacker models: A white box attacker, which uses an optimization approach with a detection oracle, and a black box attacker, which uses an autoencoder to translate anomalous data into normal data. We evaluate our implementation on two different datasets from the water distribution domain, showing that the detector's Recall drops from 0.68 to 0.12 by manipulating 4 sensors out of 82 in WADI dataset. In addition, we show that our black box attacks are transferable to different detectors: They work against autoencoder-, LSTM-, and CNN-based detectors. Finally, we implement and demonstrate our attacks on a real industrial testbed to demonstrate their feasibility in real-time.

Adversarial Machine Learning
Autoencoder
Classifier Evasion
Deep Learning
Evasion Attack
Industrial Control System
Intrusion Detection
Mean Squared Error
Multivariate Time Series

http://resolver.tudelft.nl/uuid:7e872bc3-1ee6-4330-821f-d22bfd2e9b6b

https://doi.org/10.1145/3427228.3427660

Association for Computing Machinery (ACM)

2021-07-01

9781450388580

Proceedings - 36th Annual Computer Security Applications Conference, ACSAC 2020

36th Annual Computer Security Applications Conference, ACSAC 2020, 2020-12-07 → 2020-12-11, Virtual, Online, United States

ACM International Conference Proceeding Series

Accepted Author Manuscript

Institutional Repository

conference paper

© 2020 Alessandro Erba, R. Taormina, Stefano Galelli, Marcello Pogliani, Michele Carminati, Stefano Zanero, Nils Ole Tippenhauer