Protecting device classification systems from adversarial endpoints
In various embodiments, a device classification service clusters devices in a network into a device type cluster based on attributes associated with the devices. The device classification service tracks changes to the device type cluster over time. The device classification service detects an attack on the device classification service by one or more of the devices based on the tracked changes to the device type cluster. The device classification service initiates a mitigation action for the detected attack on the device classification service.
1. A method comprising:
clustering, by a device classification service, devices in a network into a device type cluster based on attributes associated with the devices;
tracking, by the device classification service, changes to the device type cluster over time;
detecting, by the device classification service, an attack on the device classification service by one or more of the devices based on the tracked changes to the device type cluster; and
initiating, by the device classification service, a mitigation action for the detected attack on the device classification service.
2. The method as in claim 1 , wherein the mitigation action comprises:
flagging the one or more of the devices as untrusted; and
obtaining confirmation from a user interface that the one or more of the devices are untrusted.
3. The method as in claim 1 , wherein detecting the attack on the device classification service comprises:
comparing the attributes associated with the devices to expected attributes of the devices.
4. The method as in claim 1 , further comprising:
generating, by the device classification service, a device classification rule based on the device type cluster.
5. The method as in claim 1 , wherein the mitigation action comprises generating a firewall rule to block the one or more devices.
6. The method as in claim 5 , wherein firewall rule specifies the attributes associated with the one or more devices.
7. The method as in claim 1 , further comprising:
using an anomaly detection model to assign a trust index to each of the devices based on their associated attributes; and
determining that the one or more of the devices are associated with the attack based on their assigned trust indexes.
8. The method as in claim 7 , wherein the anomaly detection model assigns a low trust index to devices that are added to the cluster in a burst.
9. The method as in claim 7 , further comprising:
updating a threshold of the anomaly detection model based on feedback from a user interface.
10. An apparatus, comprising:
one or more network interfaces to communicate with one or more networks;
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 executable by the processor, the process when executed configured to:
cluster devices in a network into a device type cluster based on attributes associated with the devices, wherein the apparatus uses the device type cluster to provide a device classification service to the one or more networks;
track changes to the device type cluster over time;
detect an attack on the device classification service by one or more of the devices based on the tracked changes to the device type cluster; and
initiate a mitigation action for the detected attack on the device classification service.
11. The apparatus as in claim 10 , wherein the mitigation action comprises:
flagging the one or more of the devices as untrusted; and
obtaining confirmation from a user interface that the one or more of the devices are untrusted.
12. The apparatus as in claim 10 , wherein the apparatus detects the attack on the device classification service by:
comparing the attributes associated with the devices to expected attributes of the devices.
13. The apparatus as in claim 10 , wherein the process when executed is further configured to:
generate a device classification rule based on the device type cluster.
14. The apparatus as in claim 10 , wherein the mitigation action comprises generating a firewall rule to block the one or more devices.
15. The apparatus as in claim 14 , wherein firewall rule specifies the attributes associated with the one or more devices.
16. The apparatus as in claim 10 , wherein the process when executed is further configured to:
use an anomaly detection model to assign a trust index to each of the devices based on their associated attributes; and
determine that the one or more of the devices are associated with the attack based on their assigned trust indexes.
17. The apparatus as in claim 16 , wherein the anomaly detection model assigns a low trust index to devices that are added to the cluster in a burst.
18. The apparatus as in claim 16 , wherein the process when executed is further configured to:
update a threshold of the anomaly detection model based on feedback from a user interface.
19. A tangible, non-transitory, computer-readable medium storing program instructions that cause a device classification service to execute a process comprising:
clustering, by the device classification service, devices in a network into a device type cluster based on attributes associated with the devices;
tracking, by the device classification service, changes to the device type cluster over time;
detecting, by the device classification service, an attack on the device classification service by one or more of the devices based on the tracked changes to the device type cluster; and
initiating, by the device classification service, a mitigation action for the detected attack on the device classification service.
20. The computer-readable medium as in claim 19 , wherein the device type cluster further comprises devices located in one or more other networks, and wherein the mitigation action comprises generating a firewall rule to block the one or more devices and comprises propagating that firewall rule to the network and to the one or more other networks.