On-box behavior-based traffic classification
In one embodiment, a networking device in a network detects a traffic flow conveyed in the network via the networking device. The networking device generates flow data for the traffic flow. The networking device performs a classification of the traffic flow using the flow data as input to a machine learning-based classifier. The networking device performs a mediation action based on the classification of the traffic flow.
1. A method, comprising:
detecting, at a networking device in a network, a traffic flow conveyed in the network via the networking device, wherein the traffic flow is encrypted;
generating, by the networking device, behavioral flow data for the traffic flow;
selecting, by the networking device, a machine learning-based classifier among a plurality of machine learning-based classifiers hosted by the networking device, wherein the selecting of the machine learning-based classifier is based on one or more characteristics of the behavioral flow data and one or more respective parameters of the plurality of machine learning-based classifiers;
after selecting the machine learning-based classifier among the plurality of machine learning-based classifiers;
performing, by the networking device, a classification of the traffic flow using the behavioral flow data as input to the selected machine learning-based classifier;
performing, by the networking device, a first mediation action when the classification of the traffic flow satisfies a threshold; and
performing, by the networking device, a second mediation action different from the first mediation action when the classification of the traffic flow satisfies the threshold and at least one other feature of the traffic flow satisfies a predefined condition.
2. The method as in claim 1 , wherein the behavioral flow data comprises one or more of: Transport Layer Security (TLS) metadata or Secure Socket Layer (SSL) metadata.
3. The method as in claim 1 , wherein the behavioral flow data comprises one or more of: sequence of packet lengths and time (SPLT) data regarding the traffic flow, sequence of application lengths and time (SALT) data regarding the traffic flow, or byte distribution (BD) data regarding the traffic flow.
4. The method as in claim 1 , wherein the selected machine learning-based classifier is selected based in part on a number of bytes or packets of the traffic flow observed by the networking device.
5. The method as in claim 1 , further comprising: receiving, at the networking device, an adjustment to the selected machine learning-based classifier from a supervisory device.
6. The method as in claim 1 , wherein the first or second mediation action comprises blocking the traffic flow or raising an alert.
7. The method as in claim 1 , wherein the networking device is internal to the network and in communication with one or more network edge devices located on an edge of the network.
8. An apparatus, comprising:
one or more network interfaces to communicate with a network;
a processor coupled to the 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 operable to:
detect a traffic flow conveyed in the network via the apparatus, wherein the traffic flow is encrypted;
generate behavioral flow data for the traffic flow; select a machine learning-based classifier among a plurality of machine learning-based classifiers hosted by the networking device, wherein the selecting of the machine learning-based classifier is based on one or more characteristics of the behavioral flow data and one or more respective parameters of the plurality of machine learning-based classifiers;
after selecting the machine learning-based classifier among the plurality of machine learning-based classifiers;
perform a classification of the traffic flow using the behavioral flow data as input to a machine the selected machine learning-based classifier;
perform a first mediation action when the classification of the traffic flow satisfies a threshold; and
perform a second mediation action different from the first mediation action when the classification of the traffic flow satisfies the threshold and at least one other feature of the traffic flow satisfies a predefined condition.
9. The apparatus as in claim 8 , wherein the apparatus is a network router or a network switch.
10. The apparatus as in claim 8 , wherein the behavioral flow data comprises one or more of: sequence of packet lengths and time (SPLT) data regarding the traffic flow, sequence of application lengths and time (SALT) data regarding the traffic flow, or byte distribution (BD) data regarding the traffic flow.
11. The apparatus as in claim 8 , wherein the selected machine learning-based classifier is selected based in part on a number of bytes or packets of the traffic flow observed by the apparatus.
12. The apparatus as in claim 8 , wherein the process when executed is further operable to: receive an adjustment to the selected machine learning-based classifier from a supervisory device.
13. The apparatus as in claim 8 , wherein the first or second mediation action comprises blocking the traffic flow or raising an alert.
14. A tangible, non-transitory, computer-readable medium storing program instructions that cause a networking device in a network to execute a process comprising:
detecting, at the networking device, a traffic flow conveyed in the network via the networking device, wherein the traffic flow is encrypted;
generating, by the networking device, behavioral flow data for the traffic flow; selecting, by the networking device, a machine learning-based classifier among a plurality of machine learning-based classifiers hosted by the networking device, wherein the selecting of the machine learning-based classifier is based on one or more characteristics of the behavioral flow data and one or more respective parameters of the plurality of machine learning-based classifiers;
after selecting the machine learning-based classifier among the plurality of machine learning-based classifiers;
performing, by the networking device, a classification of the traffic flow using the behavioral flow data as input to the selected machine learning-based classifier;
performing, by the networking device, a first mediation action when the classification of the traffic flow satisfies a threshold; and
performing, by the networking device, a second mediation action different from the first mediation action when the classification of the traffic flow satisfies the threshold and at least one other feature of the traffic flow satisfies a predefined condition.
15. The computer-readable medium as in claim 14 , wherein the networking device is a network router or a network switch.
16. The computer-readable medium as in claim 14 , wherein the behavioral flow data comprises one or more of: sequence of packet lengths and time (SPLT) data regarding the traffic flow or byte distribution (BD) data regarding the traffic flow.
17. The computer-readable medium as in claim 14 , wherein the selected machine learning-based classifier is selected based in part on a number of bytes or packets of the traffic flow observed by the networking device.