Machine learned network vulnerability assessment
Methods and systems for identifying a network vulnerability. The system may gather data regarding a new or previously unknown network device, and compare the gathered data to one or more known devices that are scanned by a vulnerability assessment device. The vulnerability assessment device may then scan the previously unknown device upon a processor determining the previously unknown device shares at least one feature with a known device that is scanned.
1. A method comprising:
performing, by a vulnerability assessment system implemented by one or more computing devices:
receiving scan configuration data that indicates:
a plurality of known devices in a computer network known to the vulnerability assessment system, wherein each known device has respective features, and
whether individual ones of the known devices are to be scanned for vulnerabilities;
receiving network activity data from the computer network, wherein the network activity data indicates an unknown device in the network that is unknown to the vulnerability assessment system;
determining, using a machine learned classifier, to scan the unknown device for vulnerabilities, wherein the machine learned classified is trained based on the respective features of the known devices, and the determination is made based on features of the unknown device;
in response to the determination, generating a message to a user interface indicating the determination of the machine learned classifier to scan the unknown device; and
in response to user input, updating the scan configuration data to enable scanning of the unknown device.
2. The method of claim 1 , the determining to scan the unknown device includes determining, by the machine learned classifier, that the unknown device is sufficiently similar to one or more known devices that are scanned.
3. The method of claim 1 , wherein the message indicates a confidence value of the machine learned classifier for the determination to scan the unknown device.
4. The method of claim 1 , wherein the machine learned classifier comprises a machine learning model that implements one or more of a random forest, a support vector machine, a decision trees, a Bayesian network, a neural network, and a logistic regression model.
5. The method of claim 1 , further comprising the vulnerability assessment system training the machine learned classifier.
6. The method of claim 5 , further comprising the vulnerability assessment system randomly creating training and test datasets to train the machine learned classifier.
7. The method of claim 1 , wherein the determination to scan the unknown devices is made based at least in part on one or more rules that exclude one or more types of devices from scanning.
8. The method of claim 7 , wherein the one or more rules exclude a type of printer or a type of healthcare device.
9. The method of claim 1 , wherein the respective features of the known devices include one or more of a set of ports receiving data on the known devices, a number of peers connected to the known devices, a volume of data sent from the known devices, a volume of data received by the known devices, a busiest time period of day of the known devices, and a least busy time period of day of the known devices.
10. The method of claim 1 , wherein the network activity data is received from at least one of a Netflow record, an address resolution protocol announcement, a web proxy, a domain name server, and a firewall.
11. A system comprising:
one or more computing devices that a vulnerability assessment system, configured to:
receive scan configuration data that indicates:
a plurality of known devices in a computer network known to the vulnerability assessment system, wherein each known device has respective features, and
whether individual ones of the known devices are to be scanned for vulnerabilities;
receive network activity data from the computer network, wherein the network activity data indicates an unknown device in the network that is unknown to the vulnerability assessment system;
determine, using a machine learned classifier, to scan the unknown device for vulnerabilities, wherein the machine learned classified is trained based on the respective features of the known devices, and the determination is made based on features of the unknown device;
in response to the determination, generate a message to a user interface indicating the determination of the machine learned classifier to scan the unknown device; and
in response to user input, update the scan configuration data to enable scanning of the unknown device.
12. The system of claim 11 , to determine to scan the unknown device, the machine learned classifier is configured to determine that the unknown device is sufficiently similar to one or more known devices that are scanned.
13. The system of claim 11 , wherein the message indicates a confidence value of the machine learned classifier for the determination to scan the unknown device.
14. The system of claim 11 , wherein the message indicates a reason for the determination to scan the unknown device.
15. The system of claim 11 , wherein the machine learned classifier comprises a machine learning model that implements one or more of a random forest, a support vector machine, a decision trees, a Bayesian network, a neural network, and a logistic regression model.
16. The system of claim 11 , wherein the vulnerability assessment system is configured to:
randomly create training and test datasets; and
train the machine learned classifier using the training and test datasets.
17. The system of claim 11 , wherein the machine learned classifier is configured to make the determination to scan the unknown device based at least in part on one or more rules.
18. The system of claim 11 , wherein the vulnerability assessment system is configured to store the scan configuration data and the respective features of the known devices in a database.
19. One or more non-transitory computer-readable media storing program instructions that when executed on one or more processors implement a vulnerability assessment system and cause the vulnerability assessment system to:
receive scan configuration data that indicates:
a plurality of known devices in a computer network known to the vulnerability assessment system, wherein each known device has respective features, and
whether individual ones of the known devices are to be scanned for vulnerabilities;
receive network activity data from the computer network, wherein the network activity data indicates an unknown device in the network that is unknown to the vulnerability assessment system;
determine, using a machine learned classifier, to scan the unknown device for vulnerabilities, wherein the machine learned classified is trained based on the respective features of the known devices, and the determination is made based on features of the unknown device;
in response to the determination, generate a message to a user interface indicating the determination of the machine learned classifier to scan the unknown device; and
in response to user input, update the scan configuration data to enable scanning of the unknown device.
20. The one or more non-transitory computer-readable media of claim 19 , wherein the program instructions when executed on the one or more processors cause the vulnerability assessment system to scan the known devices according to the scan configuration data.