Methods and apparatus to analyze telemetry data of a network device for malicious activity

1. An apparatus comprising:

interface circuitry to obtain first telemetry data including a first transport layer security (TLS) parameter;

a gateway;

at least one memory;

machine readable instructions; and

programmable circuitry to at least one of instantiate or execute the machine readable instructions to:

generate a global block list using a first machine learning model, the first machine learning model generated based on a device specific block list, a device specific allow list, and the first TLS parameter;

transmit the global block list to the gateway, the gateway to perform on-path classification of a second TLS parameter of second telemetry data based on a score indicative of a likelihood that the second telemetry data is malicious, the on-path classification based on a comparison of the second TLS parameter to the device specific block list, a comparison of the second TLS parameter to the device specific allow list, and the score, the score output by a second machine learning model; and

when the score fails to satisfy a threshold, transmit the second telemetry data for off-path analysis using the first machine learning model.

2. The apparatus of claim 1 , wherein the programmable circuitry is to at least one of instantiate or execute the machine readable instructions to:

generate the second machine learning model using the first machine learning model, the device specific allow list, and the device specific block list.

3. The apparatus of claim 2 , wherein the second machine learning model is a light-weight machine learning model, and the first machine learning model is a full neural network machine learning model.

4. The apparatus of claim 2 , wherein the interface circuitry is to obtain third telemetry data from the gateway, the third telemetry data obtained when the gateway is unsuccessful in performing on-path classification.

5. The apparatus of claim 4 , wherein the programmable circuitry is to at least one of instantiate or execute the machine readable instructions to use the first machine learning model to facilitate off-path classification of the third telemetry data.

6. The apparatus of claim 1 , wherein the programmable circuitry is to at least one of instantiate or execute the machine readable instructions to:

generate the device specific block list, the device specific block list including first parameters of the first telemetry data known to be malicious; and

generate the device specific allow list, the device specific allow list including second parameters of the first telemetry data known to be benign.

7. The apparatus of claim 1 , wherein the programmable circuitry is to at least one of instantiate or execute the machine readable instructions to store the first machine learning model, the global block list, the device specific allow list, and the device specific block list for use in performing off-path classification.

8. A non-transitory computer readable storage medium comprising instructions to cause programmable circuitry to at least:

obtain first telemetry data including a first transport layer security (TLS) parameter;

generate a global block list using a first machine learning model, the first machine learning model based on a device specific block list, a device specific allow list, and the first TLS parameter;

transmit the global block list to a gateway, wherein the instructions cause the gateway to perform on-path classification of a second TLS parameter of second telemetry data based on a score indicative of a likelihood that the second telemetry data is malicious, the on-path classification based on a comparison of the second TLS parameter to the device specific block list, the device specific allow list, and an output of a second machine learning model; and

when the score fails to satisfy a threshold, transmit the second telemetry data for off-path analysis using the first machine learning model.

9. The non-transitory computer readable storage medium of claim 8 , wherein the instructions are to cause the programmable circuitry to:

generate the device specific block list, the device specific block list including firsts parameters of the first telemetry data known to be malicious;

generate the device specific allow list, the device specific allow list including second parameters of the first telemetry data known to be benign;

generate the first machine learning model based on the device specific block list and the device specific allow list; and

generate the second machine learning model using the first machine learning model, the device specific allow list, and the device specific block list.

10. The non-transitory computer readable storage medium of claim 9 , wherein the second machine learning model is a light-weight machine learning model, and the first machine learning model is a full neural network machine learning model.

11. The non-transitory computer readable storage medium of claim 9 , wherein the instructions are to cause the programmable circuitry to obtain third telemetry data from the gateway, the third telemetry data obtained when the gateway is unsuccessful in performing on-path classification.

12. The non-transitory computer readable storage medium of claim 11 , wherein the instructions are to cause the programmable circuitry to use the first machine learning model to facilitate off-path classification of the third telemetry data.

13. The non-transitory computer readable storage medium of claim 12 , wherein the instructions are to cause the programmable circuitry to transmit a result of the off-path classification to the gateway.

14. The non-transitory computer readable storage medium of claim 8 , wherein the instructions are to cause the programmable circuitry to store the first machine learning model, the global block list, the device specific allow list, and the device specific block list for use in performing off-path classification.

15. A method comprising:

obtaining first telemetry data including a first transport layer security (TLS) parameter;

generating a global block list using a first machine learning model, the first machine learning model based on a device specific block list, a device specific allow list and the first TLS parameter;

transmitting the global block list to a gateway,

causing the gateway to perform on-path classification of a second TLS parameter of second telemetry data based on a score indicative of a likelihood that the second telemetry data is malicious, the on-path classification based on a comparison of the second TLS parameter to the device specific block list, a comparison of the second TLS parameter to the device specific allow list, and the score output by a second machine learning model; and

when the score fails to satisfy a threshold, transmitting the second telemetry data for off-path analysis using the first machine learning model.

16. The method of claim 15 , further including: generating the second machine learning model using the first machine learning model, the device specific allow list, and the device specific block list.

17. The method of claim 16 , wherein the second machine learning model is a light-weight machine learning model, and the first machine learning model is a full neural network model.

18. The method of claim 16 , further including obtaining third telemetry data from the gateway, the third telemetry data obtained when the gateway is unsuccessful in performing on-path classification.

19. The method of claim 18 , further including using the first machine learning model to facilitate off-path classification of the third telemetry data.

20. The method of claim 19 , further including:

generating the device specific block list, the device specific block list including first parameters of the first telemetry data known to be malicious;

generating the device specific allow list, the device specific allow list including second parameters of the first telemetry data known to be benign; and

generating the second machine learning model based on the device specific block list and the device specific allow list.

21. The method of claim 15 , further including storing the first machine learning model, the global block list, the device specific allow list, and the device specific block list for use in performing off-path classification.