Back to results
US 2020/0259701 A1Published application
System and method for root cause analysis of call failures in a communication network
AssigneeT-Mobile USA, Inc.
InventorMarcelo Povoa
Life of this patent
Estimating the term…
Abstract
The claimed system and method describes a root cause analysis system for a radio access network. Some aspects include automatic identification of possible causes for network issues, their ranking, determination of the root (main) cause and execution of related best actions, alerts and reporting in order to automatically identify, mitigate or eliminate the problem.
Claims20
A system for root cause analysis of wireless network failures comprising:
- an aggregator which receives data on network anomalies from at least one network node;
- a correlator which receives data on network issues from a group comprising
- an aggregator;
- an anomaly detector;
- a machine learning processor which receives network status data from a group comprising:
- the correlators;
- a configuration management watchdog;
- and
- a network watchdog;
- and
- wherein the machine learning processor is physically configured to analyze the network status data to determine and select a most likely cause or solution from a plurality of possible solutions based on past problems and previous solutions;
- a root cause analysis operator in communication with the machine learning processor comprising a root cause processor physically configured to analyze the possible root cause and create a response selected from at least one of the following:
- report;
- actions;
- alerts;
- and
- optimization.
14 more elements
A method for root cause analysis of wireless network failures comprising:
- collecting network data at network nodes;
- communicating the network data to an aggregator where the network data is aggregated into aggregate data;
- communicating the aggregate data to a correlator;
- communicating anomaly data to the correlator;
- analyzing the aggregate data and anomaly data to determine root cause data;
- receiving at a machine learning processor network data wherein network status data comprises,
- the root cause data;
- configuration data from a configuration management watchdog;
- and
- network watchdog data;
- and
- analyzing the network data to determine and select a most likely solution from a plurality of solutions based on past problems and previous solutions;
- analyzing in a root cause analysis operator possible solutions;
- determining a ranking of possible solutions;
- create a response to the root cause by following the ranking of possible solutions;
- communicating the response to the network;
- monitoring if the response has a desired result;
- if the response does not have the desired result, selecting a subsequent response;
- and
- if the subsequent response does not have the desired result, communicating an alert to a supervisor.
17 more elements
A tangible computer readable medium comprising computer executable instructions for executing a root cause analysis of wireless network failures comprising instructions for:
- collecting network data at network nodes;
- communicating the network data to an aggregator where the network data is aggregated into aggregate data;
- communicating the aggregate data to a correlator;
- communicating anomaly data to the correlator;
- analyzing the aggregate data and anomaly data to determine root cause data;
- receiving at a machine learning processor network status data wherein network status data comprises,
- the root cause data;
- configuration data from a configuration management watchdog;
- and
- a network watchdog;
- and
- analyzing the network status data to determine and select a most likely solution from a plurality of solutions based on past problems and previous solutions;
- analyzing in a root cause analysis operator possible solutions;
- determining a ranking of possible solutions;
- create a response to the root cause by following the ranking of possible solutions;
- communicating the response to the network;
- monitoring if the response has a desired result;
- if the response does not have the desired result, selecting a subsequent response;
- and
- if the subsequent response does not have the desired result, communicating an alert to a supervisor.
17 more elements
Dependencies are inferred from the claim text.