Detecting attacks and quarantining malware infected devices
Attacks on a network device, e.g. an IoT device, are detected by analyzing network traffic and subsequently quarantining or blocking the network device on the network to prevent lateral movement of malware. The techniques described herein relate to developing a baseline of network device activity corresponding with a network device during a learning period and comparing the baseline of network device activity with new network activity by the network device in order to identify potentially unusual network device activity by the network device. If unusual network activity is found, remedial actions such as quarantining the network device or restricting some access to a network may be initiated.
1. A method comprising:
receiving, from a first device on a network, incoming network traffic;
comparing a network traffic baseline associated with the first device with the incoming network traffic;
determining, based on the comparison, an anomaly type associated with the incoming network traffic; and
changing access to the network by the first device according to an access restriction selected from a set of access restrictions, the access restriction corresponding to the anomaly type,
wherein:
the access restriction is associated with a role of the first device; and
changing access to the network by the first device according to the access restriction comprises changing the role of the first device.
2. The method of claim 1 , further comprising:
determining, based on analyzing the incoming network traffic, a device type of the first device,
wherein the network traffic baseline is associated with a plurality of devices having the device type.
3. The method of claim 1 , further comprising:
determining, based on analyzing the incoming network traffic, a device class of the first device,
wherein the network traffic baseline is associated with a plurality of devices having the device class.
4. The method of claim 1 , further comprising:
determining, by a machine learning engine and from previously received network traffic, a machine learning model,
wherein comparing the network traffic baseline to the incoming network traffic comprises applying the machine learning model to the incoming network traffic to determine an anomaly prediction, and wherein the anomaly type is determined according to the anomaly prediction.
5. The method of claim 1 , further comprising:
receiving previous network traffic from a second device having the same device type as the first device on an isolated network for a predetermined time period,
creating the network traffic baseline based on the previous network traffic.
6. The method of claim 1 , wherein changing access to the network by the first device according to the access restriction comprises further changing the access restriction corresponding to an anomaly severity.
7. The method of claim 1 , wherein changing access to the network by the first device further comprises one or more or a combination of:
quarantining the first device from the network;
removing, from the first device, access to a subset of devices on the network; and
logging network traffic of the first device within the network.
8. The method of claim 1 , wherein the network traffic baseline is associated with a device type of the first device, wherein the method further comprises:
determining the device type based on the incoming network traffic;
querying a data store for a network traffic baseline matching the device type; and
retrieving the network traffic baseline from the data store.
9. The method of claim 1 , wherein comparing the network traffic baseline with the incoming network traffic comprises comparing one or more or a combination of: a domain name, packet rate, or data rate of the network traffic baseline with a corresponding domain name, packet rate, or data rate of the incoming network traffic.
10. A system comprising:
one or more processors;
a non-transitory computer-readable medium including:
a data collector executable by the one or more processors to receive, from a first device on a network, incoming network traffic; and
a traffic analyzer executable by the one or more processors to:
receive, from the data collector, the incoming network traffic;
compare a network traffic baseline associated with the first device with the incoming network traffic;
determine, based on the comparison, an anomaly type associated with the incoming network traffic; and
change access to the network by the first device according to an access restriction selected from a set of access restrictions, the access restriction corresponding to the anomaly type,
wherein:
the access restriction is associated with a role of the first device; and
changing access to the network by the first device according to the access restriction comprises changing the role of the first device.
11. The system of claim 10 , further comprising:
a device profiler executable by the one or more processors to determine, based on analyzing the incoming network traffic, a device type of the first device,
wherein the network traffic baseline is associated with a plurality of devices having the device type.
12. The system of claim 11 , wherein the device profiler is further executable by the one or more processors to determine, based on analyzing the incoming network traffic, a device class of the first device,
wherein the network traffic baseline is associated with a plurality of devices having the device class.
13. The system of claim 10 , wherein:
the non-transitory computer-readable medium includes a machine learning engine;
the machine learning engine is executable by the one or more processors to determine a machine learning model based on previously received network traffic;
the traffic analyzer is further configured to apply the machine learning model to the incoming network traffic to determine an anomaly prediction; and
the anomaly type is determined according to the anomaly prediction.
14. The system of claim 10 , wherein the non-transitory computer-readable medium includes a data broker executable by the one or more processors to route the incoming network traffic from the data collector to the traffic analyzer and to store the incoming network traffic in a data store.
15. The system of claim 10 , wherein the non-transitory computer-readable medium includes a policy manager is executable by the one or more processors to enforce the access restriction according to the role assigned to the first device.
16. The system of claim 10 , wherein the set of access restrictions further comprises quarantining the first device from the network, removing, from the first device, access to a subset of devices on the network, and logging network traffic of the first device within the network.
17. The system of claim 10 , wherein the traffic analyzer is executable by the one or more processors to compare the network traffic baseline with the incoming network traffic received from the first device and comprises a traffic manager configured to compare one or more of a domain name, packet rate, or data rate of the network traffic baseline with a corresponding domain name, packet rate, or data rate of the incoming network traffic.
18. A non-transitory computer-readable medium comprising instructions for causing one or more processors to:
receive, from a first device on a network, incoming network traffic;
compare a network traffic baseline associated with the first device with the incoming network traffic;
determine, based on the comparison, an anomaly type associated with the incoming network traffic; and
change access to the network by the first device according to an access restriction selected from a set of access restrictions, the access restriction corresponding to the anomaly type
wherein:
the access restriction is associated with a role of the first device; and
changing access to the network by the first device according to the access restriction comprises changing the role of the first device.
19. The non-transitory computer-readable medium of claim 18 , wherein the instructions cause the one or more processors to:
determine, based on analyzing the incoming network traffic, a device type or a device class of the first device,
wherein the network traffic baseline is associated with a plurality of devices having the device type or the device class.
20. The non-transitory computer-readable medium of claim 18 , wherein the instructions cause the one or more processors to:
determine, by a machine learning engine and from previously received network traffic, a machine learning model,
wherein:
comparing the network traffic baseline to the incoming network traffic comprises applying the machine learning model to the incoming network traffic to determine an anomaly prediction, and
the anomaly type is determined according to the anomaly prediction.