Predicting wireless access point radio failures using machine learning
AssigneeCisco Technology, Inc.
InventorsJean-Philippe Vasseur, Vinay Kumar Kolar
Life of this patent
Estimating the term…
Abstract
In one embodiment, a network assurance system that monitors a network forms a cluster of similarly behaving wireless access points (APs). The cluster includes APs associated with different software versions. The network assurance system trains a machine learning-based failure prediction model for the cluster based on a set of features of the APs in the cluster. The network assurance system proactively triggers a client in the network to roam from a first AP to a second AP, based on the failure prediction model predicting a failure of the first AP. The network assurance system quarantines the failure prediction model when a new software version is associated with one or more of the APs.
Claims20
A method comprising:
- forming, by a device of network assurance system that monitors a network, a cluster of similarly behaving wireless access points (APs), wherein the cluster comprises APs associated with different software versions;
- training, by the device, a machine learning-based failure prediction model for the cluster based on a set of features of the APs in the cluster;
- proactively triggering, by the device, a client in the network to roam from a first AP to a second AP, based on the machine learning-based failure prediction model predicting a failure of the first AP;
- and
- quarantining, by the device, the machine learning-based failure prediction model when a new software version is associated with one or more of the APs,
- wherein quarantining the machine learning-based failure prediction model comprises:
- assessing, by the device, whether to initiate retraining of the machine learning-based failure prediction model based on the new software version.
An apparatus, comprising:
- one or more network interfaces to communicate with a network that comprises a plurality of sensors and actuators;
- a processor coupled to the one or more network interfaces, the processor to execute one or more processes;
- and
- a memory to store a process executable by the processor, the process when executed comprising:
- forming a cluster of similarly behaving wireless access points (APs), wherein the cluster comprises APs associated with different software versions;
- training a machine learning-based failure prediction model for the cluster based on a set of features of the APs in the cluster;
- proactively triggering a client in the network to roam from a first AP to a second AP, based on the machine learning-based failure prediction model predicting a failure of the first AP;
- and
- quarantining the machine learning-based failure prediction model when a new software version is associated with one or more of the APs,
- wherein the apparatus quarantines the machine learning-based failure prediction model by:
- assessing whether to initiate retraining of the machine learning-based failure prediction model based on the new software version.
A tangible, non-transitory, computer-readable medium storing program instructions that cause a device of a network assurance system that monitors a network to execute a process comprising:
- forming, by the device, a cluster of similarly behaving wireless access points (APs), wherein the cluster comprises APs associated with different software versions;
- training, by the device, a machine learning-based failure prediction model for the cluster based on a set of features of the APs in the cluster;
- proactively triggering, by the device, a client in the network to roam from a first AP to a second AP, based on the machine learning-based failure prediction model predicting a failure of the first AP;
- and
- quarantining, by the device, the machine learning-based failure prediction model when a new software version is associated with one or more of the APs,
- wherein quarantining the machine learning-based failure prediction model comprises:
- assessing, by the device, whether to initiate retraining of the machine learning-based failure prediction model based on the new software version.
Dependencies are inferred from the claim text.