IoT application learning
A system and method for performing automated learning of an Internet-of-Things (IoT) application are disclosed. The automated learning is based on generation of application-agnostic events, allowing the automated learning to be performed without prior knowledge of the IoT application.
1. A method, comprising:
receiving a detected set of Internet of Things (IoT) application events, wherein the IoT application events are associated with activities of an IoT application executing on an IoT device;
identifying, from a predetermined set of different types of activities, one or more application-specific activities;
extracting one or more attributes from a plurality of payloads of IoT messages associated with the IoT application executing on the IoT device as a set of activity parameters and using extracted information to perform automated payload learning, wherein the extracting includes filtering out one or more confidential values;
predicting a set of activities of the IoT application in accordance with the set of activity parameters at least in part by using domain knowledge;
determining whether at least one of the IoT application events falls outside the predicted set of activities; and
generating an alert associated with the at least one of the IoT application events when it is determined the at least one of the IoT application events falls outside the predicted set of activities.
2. The method of claim 1 , wherein the IoT application events are detected via passive monitoring.
3. The method of claim 1 , wherein the IoT application events are detected using deep packet inspection (DPI).
4. The method of claim 1 , wherein the IoT application events are detected using subscription-based inspection.
5. The method of claim 1 , wherein using the domain knowledge to predict the set of activities includes identifying a repeating pattern.
6. The method of claim 1 , wherein using the domain knowledge to predict the set of activities includes utilizing tags or labels injected into activity fields of the IoT application events.
7. The method of claim 1 , wherein an IoT application event of the IoT application events comprises a raw event.
8. The method of claim 1 , wherein the IoT application events comprise one or more of network sessions, portions of network sessions, message transport events, and message log events.
9. A system, comprising:
a processor configured to:
receive a detected set of Internet of Things (IoT) application events, wherein the IoT application events are associated with activities of an IoT application executing on an IoT device;
identify, from a predetermined set of different types of activities, one or more application-specific activities;
extract one or more attributes from a plurality of payloads of IoT messages associated with the IoT application executing on the IoT device as a set of activity parameters and use extracted information to perform automated payload learning, wherein the extracting includes filtering out one or more confidential values;
predict a set of activities of the IoT application in accordance with the set of activity parameters at least in part by using domain knowledge;
determine whether at least one of the IoT application events falls outside the predicted set of activities; and
generate an alert associated with the at least one of the IoT application events when it is determined the at least one of the IoT application events falls outside the predicted set of activities; and
a memory coupled to the processor and configured to provide the processor with instructions.
10. The system of claim 9 , wherein the IoT application events are detected via passive monitoring.
11. The system of claim 9 , wherein the IoT application events are detected using deep packet inspection (DPI).
12. The system of claim 9 , wherein the IoT application events are detected using subscription-based inspection.
13. The system of claim 9 , wherein using the domain knowledge to predict the set of activities includes identifying a repeating pattern.
14. The system of claim 9 , wherein using the domain knowledge to predict the set of activities includes utilizing tags or labels injected into activity fields of the IoT application events.
15. The system of claim 9 , wherein an IoT application event of the IoT application events comprises a raw event.
16. The system of claim 9 , wherein the IoT application events comprise one or more of network sessions, portions of network sessions, message transport events, and message log events.
17. The method of claim 1 , wherein the detected set of IoT application events serve as a signature of the IoT application.
18. The system of claim 9 , wherein the detected set of IoT application events serve as a signature of the IoT application.