In industrial and critical infrastructure, on-premise computational resources are the focal point of security efforts. Systems are hardened, third-party tools developed, perimeters secured, standards defined, tests conducted, and the public reassured.
Yet a frequently-overlooked and crucial element of decision-making is actually the data on which decisions are based. Today’s critical infrastructure and industrial systems rely on tens of thousands of sensors to control physical assets. Data from these sensors forms the basis for decision-making by humans and machines alike.
And despite their centrality to the operation of critical infrastructure and industrial facilities, sensors are notoriously under-protected. Subtle manipulation of data by attackers can leave operators effectively blind to the true state of physical systems.
Decision-making at large-scale industrial and infrastructure facilities is based nearly entirely on data fed to the ICS system from thousands of sensors. These sensors range from legacy devices to brand-new IoT monitors. Yet what happens when sensors malfunction, and the data received is inaccurate?
Existing safety and sensor fault detection mechanisms cannot detect subtly misleading sensor data owing to partial sensor malfunction. Such inaccurate data misleads control systems, masks the actual physical state of assets, reports false information, and leaves the operators in the control room blind to the true situation.
To avoid both damage and actual physical danger, potentially bogus data must be exposed and the true state of operations revealed.
Aperio Systems' Data Forgery Protection (DFP) solutions act as a data polygraph to protect against data inaccuracy owing to malfunction. Like the law enforcement version, DFP-based solutions enable the security teams to discover the truth – the integrity of the data received by the control room.
Using advanced algorithms, APERIO Systems’ DFP solutions identify and track the unique signal “fingerprints” of each sensor. These fingerprints are manifested within the exact fluctuations of reported signals, the physical micro-noise, and the specific system behavior within and between modes of operation. Once a fingerprint baseline exists for each sensor, deviations can be characterized and investigated, and the “truth” of a given data set can be determined.