Cyber-Physical System Attack Detection via Segmented Sensor Networks
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for detecting attacks on cyber-physical systems are ineffective as sophisticated attackers can spoof telemetry from sensors, even when redundant sensors are used, as these sensors often communicate over the same network, allowing hackers to deceive the control station and pose safety hazards.
Innovation Solution
A system utilizing redundant sensors operating on separate networks with security measures, where smart contracts are executed by distributed entities on a blockchain to compare readings from primary and redundant sensors, allowing for immediate remedial actions if discrepancies are detected, ensuring that any single entity can identify and address potential attacks.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If redundant sensors are used to detect attacks, then detection reliability is improved, but the system becomes vulnerable to network-based spoofing attacks
Solution Approach 1:
The system divides the sensor network into multiple independent networks, where each network operates separately with its own sensors. This segmentation prevents a single network compromise from affecting all sensors, as attackers would need to compromise multiple independent networks simultaneously to spoof all sensor readings.
Solution Approach 2:
The patent introduces blockchain technology as an intermediary layer between sensors and the control system. Smart contracts on the blockchain verify sensor readings from multiple networks before accepting them, creating a trustless verification mechanism that prevents spoofed readings from being processed even if some sensors are compromised.
2Device complexity
If sensors communicate over the same network, then system complexity is reduced, but security against spoofing attacks deteriorates
Solution Approach 1:
The system implements multiple separate communication networks for sensor data transmission. Each network is independent and can be operated by different entities, reducing the attack surface for any single network while maintaining manageable complexity through modular architecture.
Solution Approach 2:
Different networks have different security characteristics and trust levels. The system allows each network to be configured with appropriate security measures for its specific requirements, rather than applying a uniform security model across all communications.
3Ease of operation
If a single control station processes all sensor readings, then ease of operation is improved, but vulnerability to centralized attacks increases
Solution Approach 1:
The control function is divided into multiple independent control stations, each capable of processing sensor readings and executing control decisions. This distribution eliminates the single point of failure in centralized systems, as attackers would need to compromise multiple geographically dispersed control stations simultaneously.
Solution Approach 2:
Each control station is designed to be universally capable of processing readings from multiple sensor networks and making independent control decisions. This multi-functionality allows any control station to take over if others are compromised, maintaining system operation under attack conditions.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
One embodiment facilitates detection of attacks in a cyber-physical system of interacting elements with physical inputs and outputs. During operation, the system receives, by a first entity of a plurality of entities, a first reading from a first set of sensors of the cyber-physical system via a first network. The system receives, by the first entity, a second reading from a second set of sensors of the cyber-physical system via a second network, wherein the second network includes security measures which prevent access by any external entity or any of the plurality of entities. The system executes a set of instructions based on the first reading and the second reading. The system determines that a result of the executed instructions does not match an expected condition. The system performs a remedial action based on the result.