Predicting the intent of a network operator for making config changes
Systems and methods described herein are adapted to predict the intent of a network operator when it is determined that configuration changes (or config changes) are made to a network device. A method, according to one implementation, includes the step of collecting raw data related to a Network Element (NE). The method also includes the step of pre-processing the raw data to obtain a config change associated with the NE. Also, the method includes the step of applying the config change as input to a ML model to predict a user intent representing a desired network outcome.
1 . A non-transitory computer-readable medium configured to store computer logic having instructions that, when executed, enable a processing device to:
collect raw data related to a Network Element (NE);
pre-process the raw data by detecting a configuration difference between successive configuration files of the NE to obtain a config change associated with the NE, wherein the configuration difference is stored as a unified difference and converted into feature vectors using a text vectorizer process; and
apply the config change as input to a Machine Learning (ML) model to predict a user intent representing a desired network outcome, wherein the ML model is adapted to process the configuration difference in real-time across a multi-domain, multi-vendor enterprise network using raw configuration information as an input data source to collect the raw data in lieu of requiring Resource Adapters wherein applying the config change includes providing the feature vectors to a supervised ML classifier, and wherein, in response to predicting the user intent, the processing device automatically triggers a workflow change in the enterprise network according to a closed loop automation process,
wherein the ML model is trained using supervised training procedures based on historical data and labels, where historical config change information is used as input data to train the ML model, wherein the labels include one or more of 1) device logs and/or records identifying execution of a specific service template on a given network element at a given time, and 2) information from device logs and/or from ticketing systems or Information Technology Service Management (ITSM) systems correlated with any config change at a given time.
2 . The non-transitory computer-readable medium of claim 1 , wherein the user intent is related to one or more of a provisioning of a new network service, an addition or removal of a network function, a rebooting of the NE, and an upgrade to software and/or firmware on the NE.
3 . The non-transitory computer-readable medium of claim 1 , wherein the instructions further enable the processing device to execute one or more functions selected from the group consisting of a) providing options to a network operator, b) providing recommendations to a network operator, c) providing notifications to a network operator, and d) automatically triggering workflow changes in a network according to a closed-loop automation process.
4 . The non-transitory computer-readable medium of claim 1 , wherein the NE is adapted to operate on a multi-domain, multi-vendor enterprise network and the non-transitory computer-readable medium is implemented on an orchestrator of a Network Operations Center (NOC) in communication with the enterprise network.
5 . The non-transitory computer-readable medium of claim 1 wherein the instructions further enable the processing device to re-train the ML model in a self-evolving manner based on the historical data and new data.
6 . The non-transitory computer-readable medium of claim 1 , wherein collecting the raw data includes collecting data from device logs, configuration management systems, configuration back-up servers, ticketing systems, third party configuration management tools, and Information Technology Service Management (ITSM) systems.
7 . The non-transitory computer-readable medium of claim 1 , wherein collecting the raw data includes obtaining commands entered by a network operator using a Command Line Interface (CLI).
8 . The non-transitory computer-readable medium of claim 7 , wherein the config change is applied to a config file using one or more config templates on the CLI, each of the one or more config templates including one or more of a list of config commands, a template description, a name of the config template, a specifically identified tool, and a difference between config versions corresponding to incremental change resulting from execution of commands in the config template.
9 . The non-transitory computer-readable medium of claim 1 , wherein the ML model is adapted to predict user intent in real-time.
10 . A system comprising:
a processing device, and
a memory device configured to store a computer program having instructions that, when executed, enable the processing device to
collect raw data related to a Network Element (NE);
pre-process the raw data by detecting a configuration difference between successive configuration files of the NE to obtain a config change associated with the NE, wherein the configuration difference is stored as a unified difference and converted into feature vectors using a text vectorizer process; and
apply the config change as input to a Machine Learning (ML) model to predict a user intent representing a desired network outcome, wherein the ML model is adapted to process the configuration difference in real-time across a multi-domain, multi-vendor enterprise network using raw configuration information as an input data source to collect the raw data in lieu of requiring Resource Adapters, wherein applying the config change includes providing the feature vectors to a supervised ML classifier, and wherein, in response to predicting the user intent, the processing device automatically triggers a workflow change in the enterprise network according to a closed-loop automation process,
wherein the ML model is trained using supervised training procedures based on historical data and labels, where historical config change information is used as input data to train the ML model, wherein the labels include one or more of 1) device logs and/or records identifying execution of a specific service template on a given network element at a given time, and 2) information from device logs and/or from ticketing systems or Information Technology Service Management (ITSM) systems correlated with any config change at a given time.
11 . The system of claim 10 , wherein the user intent is related to one or more of a provisioning of a new network service, an addition or removal of a network function, a rebooting of the NE, and an upgrade to software and/or firmware on the NE.
12 . The system of claim 10 , wherein the instructions further enable the processing device to execute one or more functions selected from the group consisting of a) providing options to a network operator, b) providing recommendations to a network operator, c) providing notifications to a network operator, and d) automatically triggering workflow changes in a network according to a closed-loop automation process.
13 . The system of claim 10 , wherein the NE is adapted to operate on a multi-domain, multi-vendor enterprise network and the non-transitory computer-readable medium is implemented on an orchestrator of a Network Operations Center (NOC) in communication with the enterprise network.
14 . The system of claim 10 , wherein the instructions further enable the processing device to re-train the ML model in a self-evolving manner based on the historical data and new data.
15 . A method comprising the steps of:
collecting raw data related to a Network Element (NE);
pre-processing the raw data by detecting a configuration difference between successive configuration files of the NE to obtain a config change associated with the NE, wherein the configuration difference is stored as a unified difference and converted into feature vectors using a text vectorizer process; and
applying the config change as input to a Machine Learning (ML) model to predict a user intent representing a desired network outcome, wherein the ML model is adapted to process the configuration difference in real-time across a multi-domain, multi-vendor enterprise network using raw configuration information as an input data source to collect the raw data i in lieu of requiring Resource Adapters, wherein applying the config change includes providing the feature vectors to a supervised ML classifier, and wherein, in response to predicting the user intent, the processing device automatically triggers a workflow change in the enterprise network according to a closed-loop automation process,
wherein the ML model is trained using supervised training procedures based on historical data and labels, where historical config change information is used as input data to train the ML model, wherein the labels include one or more of 1) device logs and/or records identifying execution of a specific service template on a given network element at a given time, and 2) information from device logs and/or from ticketing systems or Information Technology Service Management (ITSM) systems correlated with any config change at a given time.
16 . The method of claim 15 , wherein collecting the raw data includes collecting data from device logs, configuration management systems, configuration back-up servers, ticketing systems, third party configuration management tools, and Information Technology Service Management (ITSM) systems.
17 . The method of claim 15 , wherein collecting the raw data includes obtaining commands entered by a network operator using a Command Line Interface (CLI).
18 . The method of claim 17 , wherein the config change is applied to a config file using one or more config templates on the CLI, each of the one or more config templates including one or more of a list of config commands, a template description, a name of the config template, a specifically identified tool, and a difference between config versions corresponding to incremental change resulting from execution of commands in the config template.