Back to results
US 11,570,038 B2Granted
Network system fault resolution via a machine learning model
AssigneeJuniper Networks, Inc.
InventorsJisheng Wang, Xiaoying Wu, Shmuel Shaffer, David Jea
Life of this patent
Estimating the term…
Abstract
Disclosed are embodiments for automatically resolving faults in a complex network system. Some embodiments monitor one or more of system operational parameter values and message exchanges between network components. A machine learning model detects a fault in the complex network system, and an action is selected based on a cause of the fault. After the action is applied to the complex network system, additional monitoring is performed to either determine the fault has been resolved or additional actions are to be applied to further resolve the fault.
Claims20
A method, comprising:
- receiving, from one or more devices of a network system, a time series of operational parameter values;
- providing the time series of operational parameter values to a machine learning model;
- receiving, from the machine learning model, an indication of a cause of a fault in operation of the network system;
- selecting a first rectifying action to perform on the network system based on the cause;
- evaluating a confidence that the selected first rectifying action will resolve the fault;
- in response to determining that the confidence does not satisfy a first threshold, selecting a diagnostic action;
- in response to determining a cost of performing the diagnostic action does not satisfy a diagnostic action cost threshold, invoking the first rectifying action;
- in response to determining a first cost of performing a first diagnostic action satisfies the diagnostic action cost threshold, first injecting the first diagnostic action;
- monitoring operational parameter values after the first injecting of the first diagnostic action;
- determining a first root cause and associated first probability that the first root cause is the cause of the fault based on the monitored operational parameters;
- second injecting the first diagnostic action based on a determination that the first probability is below a predetermined threshold;
- monitoring operational parameter values after the second injecting of the first diagnostic action;
- determining a second probability that the first root cause is the cause of the fault associated with the first root cause;
- adjusting the diagnostic cost threshold based on the first and second probabilities;
- and
- determining whether to inject an additional diagnostic action based on the adjusted diagnostic cost threshold.
13 more elements
A non-transitory computer readable storage medium comprising instructions that when executed configure hardware processing circuitry to perform operations comprising:
- receiving, from one or more devices of a network system, a time series of operational parameter values;
- providing the time series of operational parameter values to a machine learning model;
- receiving, from the machine learning model, an indication of a cause of a fault in operation of the network system;
- selecting a first rectifying action to perform on the network system based on the cause;
- evaluating a confidence that the selected first rectifying action will resolve the fault;
- in response to determining that the confidence does not satisfy a first threshold, selecting a diagnostic action;
- in response to determining a cost of performing the diagnostic action does not satisfy a diagnostic action cost threshold, invoking the first rectifying action;
- in response to determining a first cost of performing a first diagnostic action satisfies the diagnostic action cost threshold, first injecting the first diagnostic action;
- monitoring operational parameter values after the first injecting of the first diagnostic action;
- determining a first root cause and associated first probability that the first root cause is the cause of the fault based on the monitored operational parameters;
- second injecting the first diagnostic action based on a determination that the first probability is below a predetermined threshold;
- monitoring operational parameter values after the second injecting of the first diagnostic action;
- determining a second probability that the first root cause is the cause of the fault associated with the first root cause;
- adjusting the diagnostic cost threshold based on the first and second probabilities;
- and
- determining whether to inject an additional diagnostic action based on the adjusted diagnostic cost threshold.
13 more elements
A system, comprising:
- hardware processing circuitry;
- one or more hardware memories storing instructions that when executed configure the hardware processing circuitry to perform operations comprising:
- receiving, from one or more devices of a network system, a time series of operational parameter values;
- providing the time series of operational parameter values to a machine learning model;
- receiving, from the machine learning model, an indication of a cause of a fault in operation of the network system;
- selecting a first rectifying action to perform on the network system based on the cause;
- evaluating a confidence that the selected first rectifying action will resolve the fault;
- in response to determining that the confidence does not satisfy a first threshold, selecting a diagnostic action;
- in response to determining a cost of performing the diagnostic action does not satisfy a diagnostic action cost threshold, invoking the first rectifying action;
- in response to determining a first cost of performing a first diagnostic action satisfies the diagnostic action cost threshold, first injecting the first diagnostic action;
- monitoring operational parameter values after the first injecting of the first diagnostic action;
- determining a first root cause and associated first probability that the first root cause is the cause of the fault based on the monitored operational parameters;
- second injecting the first diagnostic action based on a determination that the first probability is below a predetermined threshold;
- monitoring operational parameter values after the second injecting of the first diagnostic action;
- determining a second probability that the first root cause is the cause of the fault associated with the first root cause;
- adjusting the diagnostic cost threshold based on the first and second probabilities;
- and
- determining whether to inject an additional diagnostic action based on the adjusted diagnostic cost threshold.
15 more elements
Dependencies are inferred from the claim text.