Zero-trust architecture for industrial automation
According to one or more embodiments of the disclosure, a device in a network obtains parameters for entropy testing of industrial equipment that controls a physical process. Entropy is added to commands sent to the industrial equipment during the entropy testing. The device receives packets that were generated during the entropy testing of the industrial equipment and include sensor data regarding the physical process. The device determines whether the sensor data is inconsistent by analyzing the sensor data using a machine learning model that models the physical process. The device initiates a corrective measure, when the sensor data is determined to be inconsistent.
1. A method comprising:
obtaining, by a device in a network, parameters for entropy testing of industrial equipment that controls a physical process, wherein entropy is added to commands sent to the industrial equipment during the entropy testing;
receiving, at the device, packets that were generated during the entropy testing of the industrial equipment and include sensor data regarding the physical process;
determining, by the device, whether the sensor data is inconsistent by analyzing the sensor data using a machine learning model that models the physical process; and
initiating, by the device, a corrective measure, when the sensor data is determined to be inconsistent.
2. The method as in claim 1 , wherein the corrective measure comprises shutting down the physical process or sending an alert to a user interface.
3. The method as in claim 1 , further comprising:
extracting, by the device, the sensor data from the packets by performing deep packet inspection of the packets.
4. The method as in claim 1 , wherein the device comprises a router or network switch.
5. The method as in claim 1 , wherein the parameters for the entropy testing are indicative of a start time for the entropy testing and the commands sent to the industrial equipment during the entropy testing.
6. The method as in claim 1 , wherein the machine learning model comprises a cluster-based classifier.
7. The method as in claim 1 , wherein the machine learning model is trained in part by using principal component analysis to identify correlated variables associated with the physical process.
8. The method as in claim 1 , wherein the sensor data regarding the physical process is indicative of at least one of: a temperature measurement, a pressure measurement, or a vibration measurement.
9. The method as in claim 1 , wherein a start time for the entropy testing is selected randomly by a security service for the network.
10. The method as in claim 1 , wherein the entropy added to the commands sent to the industrial equipment during the entropy testing is selected based in part on a safe operating range for the industrial equipment.
11. An apparatus, comprising:
one or more network interfaces to communicate with a network;
a processor coupled to the one or more network interfaces and configured to execute one or more processes; and
a memory configured to store a process that is executable by the processor, the process when executed configured to:
obtain parameters for entropy testing of industrial equipment that controls a physical process, wherein entropy is added to commands sent to the industrial equipment during the entropy testing;
receive packets that were generated during the entropy testing of the industrial equipment and include sensor data regarding the physical process;
determine whether the sensor data is inconsistent by analyzing the sensor data using a machine learning model that models the physical process; and
initiate a corrective measure, when the sensor data is determined to be inconsistent.
12. The apparatus as in claim 11 , wherein the corrective measure comprises shutting down the physical process or sending an alert to a user interface.
13. The apparatus as in claim 11 , wherein the process when executed is further configured to:
extract the sensor data from the packets by performing deep packet inspection of the packets.
14. The apparatus as in claim 11 , wherein the apparatus comprises a router or network switch.
15. The apparatus as in claim 11 , wherein the parameters for the entropy testing are indicative of a start time for the entropy testing and the commands sent to the industrial equipment during the entropy testing.
16. The apparatus as in claim 11 , wherein the machine learning model comprises a cluster-based classifier.
17. The apparatus as in claim 11 , wherein the sensor data regarding the physical process is indicative of at least one of: a temperature measurement, a pressure measurement, or a vibration measurement.
18. The apparatus as in claim 11 , wherein a start time for the entropy testing is selected randomly by a security service for the network.
19. The apparatus as in claim 11 , wherein the entropy added to the commands sent to the industrial equipment during the entropy testing is selected based in part on a safe operating range for the industrial equipment.
20. A tangible, non-transitory, computer-readable medium storing program instructions that cause a device in a network to execute a process comprising:
obtaining, by the device in the network, parameters for entropy testing of industrial equipment that controls a physical process, wherein entropy is added to commands sent to the industrial equipment during the entropy testing;
receiving, at the device, packets that were generated during the entropy testing of the industrial equipment and include sensor data regarding the physical process;
determining, by the device, whether the sensor data is inconsistent by analyzing the sensor data using a machine learning model that models the physical process; and
initiating, by the device, a corrective measure, when the sensor data is determined to be inconsistent.