公开(公告)号：US20200099709A1

公开(公告)日：2020-03-26

申请号：US16141007

申请日：2018-09-25

Applicant: Cisco Technology, Inc.

Inventor： Jean-Philippe Vasseur ,  Grégory Mermoud ,  Santosh Ghanshyam Pandey

IPC: H04L29/06 ,  H04L12/24 ,  H04L12/707 ,  H04L12/747 ,  H04W88/12 ,  G06N99/00

Abstract: In one embodiment, a network assurance service that monitors a network detects, using a machine learning-based anomaly detector, network anomalies associated with source nodes in the monitored network. The network assurance service identifies, for each of the detected anomalies, a set of network paths between the source nodes associated with the anomaly and one or more potential destinations of traffic for that source node. The network assurance service correlates networking devices along the network paths in the identified sets of network paths with the detected network anomalies. The network assurance service adjusts the machine learning-based anomaly detector to use a performance measurement for a particular one of the networking devices as an input feature, based on the correlation between the particular networking device and the detected network anomalies.

公开(公告)号：US20190342195A1

公开(公告)日：2019-11-07

申请号：US15972306

申请日：2018-05-07

Applicant: Cisco Technology, Inc.

Inventor： Grégory Mermoud ,  Jean-Philippe Vasseur ,  Erwan Zerhouni

IPC: H04L12/26 ,  H04L29/06 ,  G06F15/18 ,  G06F17/27

Abstract: In one embodiment, a network assurance service that monitors a network detects anomalies in the network by applying one or more machine learning-based anomaly detectors to telemetry data from the network. The network assurance service receives ranking feedback from a plurality of anomaly rankers regarding relevancy of the detected anomalies. The network assurance service calculates a rescaling factor and quantile parameter by applying an objective function to the ranking feedback, in order to optimize the rescaling factor and quantile parameter of the one or more anomaly detectors. The network assurance service adjusts the rescaling factor and quantile parameter of the one or more anomaly detectors using the calculated rescaling factor and quantile parameter.

公开(公告)号：US10440577B1

公开(公告)日：2019-10-08

申请号：US16183940

申请日：2018-11-08

Applicant: Cisco Technology, Inc.

Inventor： Jean-Philippe Vasseur ,  Pierre-André Savalle ,  Grégory Mermoud

IPC: H04W12/08 ,  H04W4/70 ,  H04W8/00 ,  H04W8/24 ,  H04W24/08 ,  H04W48/02 ,  G05B19/045 ,  G06F9/448 ,  H04L29/06 ,  H04W64/00 ,  G06N20/00

Abstract: In one embodiment, a device classification service receives a first set of telemetry data captured by one or more networking devices in a network regarding traffic associated with an endpoint device in the network. The service classifies the endpoint device as being of an unknown device type, by applying a machine learning-based classifier to the first set of telemetry data. The service instructs the one or more networking devices in the network to reset a finite state machine (FSM) of the traffic associated with the endpoint device. The device classification service receives a second set of telemetry data regarding traffic associated with the endpoint device and captured after reset of the FSM. The service reclassifies the endpoint device as being of a particular device type, by applying the machine learning-based classifier to the second set of telemetry data.

公开(公告)号：US10389613B2

公开(公告)日：2019-08-20

申请号：US15872359

申请日：2018-01-16

Applicant: Cisco Technology, Inc.

Inventor： Sukrit Dasgupta ,  Jean-Philippe Vasseur ,  Grégory Mermoud

IPC: H04L12/26

Abstract: In one embodiment, a device in a network receives data indicative of traffic characteristics of traffic associated with a particular application. The device identifies one or more paths in the network via which the traffic associated with the particular application was sent, based on the traffic characteristics. The device determines a probing schedule based on the traffic characteristics. The probing schedule simulates the traffic associated with the particular application. The device sends probes along the one or more identified paths according to the determined probing schedule.

公开(公告)号：US20190238396A1

公开(公告)日：2019-08-01

申请号：US15881909

申请日：2018-01-29

Applicant: Cisco Technology, Inc.

Inventor： David Tedaldi ,  Grégory Mermoud ,  Jean-Philippe Vasseur

IPC: H04L12/24

Abstract: In one embodiment, a network assurance service receives one or more sets of network characteristics of a network, each network characteristic forming a different feature dimension in a multi-dimensional feature space. The network assurance service applies machine learning-based anomaly detection to the one or more sets of network characteristics, to label each set of network characteristics as anomalous or non-anomalous. The network assurance service identifies, based on the labeled one or more sets of network characteristics, an anomaly pattern as a collection of unidimensional cutoffs in the feature space. The network assurance service initiates a change to the network based on the identified anomaly pattern.

公开(公告)号：US20190138938A1

公开(公告)日：2019-05-09

申请号：US15803968

申请日：2017-11-06

Applicant: Cisco Technology, Inc.

Inventor： Jean-Philippe Vasseur ,  Grégory Mermoud ,  Vinay Kumar Kolar

IPC: G06N99/00 ,  H04L12/26

Abstract: In one embodiment, a service receives relevancy feedback regarding anomalies detected in a network by one or more unsupervised learning-based anomaly detectors. The service generates a set of rules based on those of the anomalies deemed relevant by the relevancy feedback. The service uses the set of rules to trigger collection of data features from the network. The service trains a supervised learning-based classifier using the data features collected from the network.

公开(公告)号：US10277476B2

公开(公告)日：2019-04-30

申请号：US14164876

申请日：2014-01-27

Applicant: Cisco Technology, Inc.

Inventor： Grégory Mermoud ,  Jean-Philippe Vasseur ,  Sukrit Dasgupta

IPC: G06N99/00 ,  H04L12/24 ,  H04L12/805 ,  H04L12/751 ,  H04L12/803 ,  H04W24/04 ,  H04L12/721

Abstract: In one embodiment, a predictive model is constructed by mapping multiple network characteristics to multiple network performance metrics. Then, a network performance metric pertaining to a node in a network is predicted based on the constructed predictive model and one or more network characteristics relevant to the node. Also, a local parameter of the node is optimized based on the predicted network performance metric.

公开(公告)号：US20180365581A1

公开(公告)日：2018-12-20

申请号：US15704595

申请日：2017-09-14

Applicant: Cisco Technology, Inc.

Inventor： Jean-Philippe Vasseur ,  Grégory Mermoud ,  Pierre-André Savalle ,  Javier Cruz Mota

IPC: G06N7/00 ,  H04L12/24 ,  G06N99/00

Abstract: In one embodiment, a service uses a set of collected characteristics of a client device in a network as input to a machine learning-based model that predicts a quality score for an online conference in which the client device is a participant. The service determines a resource consumption by the client device or the network that is associated with collecting the characteristics of the client device. The service determines an efficacy of the machine learning-based model as a function of the set of collected characteristics of the client device. The service adjusts the set of collected characteristics of the client device to optimize the efficacy of the model and the resource consumption associated with collecting the characteristics of the client device.

公开(公告)号：US20180357560A1

公开(公告)日：2018-12-13

申请号：US15620116

申请日：2017-06-12

Applicant: Cisco Technology, Inc.

Inventor： Andrea Di Pietro ,  Grégory Mermoud ,  Sukrit Dasgupta ,  Jean-Philippe Vasseur

IPC: G06N99/00 ,  H04L12/26 ,  G06N5/04

CPC classification number: G06N99/005 ,  G06N5/04 ,  H04L43/06

Abstract: In one embodiment, a device identifies a new data source of characteristics data for a monitored network. The device initiates a quarantine period for the characteristic data from the new data source. The characteristic data from the new data source is quarantined from input to a machine learning-based analyzer during the quarantine period. The device models the characteristic data from the new data source during the quarantine period, to determine whether the characteristic data from the new data source is reliable for input to the machine learning-based analyzer. After the quarantine period, the device provides the characteristic data from the new data source to the machine learning-based analyzer based on a determination that the characteristic data from the new data source is reliable.

公开(公告)号：US10103970B2

公开(公告)日：2018-10-16

申请号：US15146397

申请日：2016-05-04

Applicant: Cisco Technology, Inc.

Inventor： Sukrit Dasgupta ,  Jean-Philippe Vasseur ,  Grégory Mermoud

IPC: H04L12/721 ,  H04L12/851 ,  H04W40/14 ,  H04W40/24 ,  H04L12/725 ,  H04L12/753 ,  H04L12/24 ,  H04L12/26

Abstract: Statistical and historical values of performance metrics are actively used to influence routing decisions for optimum topologies in a constrained network. Traffic service level is constantly monitored and compared with a service level agreement. If deviation exists between the monitored traffic service level and the terms of the service level agreement, stability metrics are used to maintain paths through the network that meet the terms of the traffic service level agreement or that improve the traffic flow through the network. Backup parent selection for a node in the network is performed based on previous performance of backup parents for the node.