Cisco Technology, Inc.

Grégory Mermoud

Jean-Philippe Vasseur

Pierre-André Savalle

Vinay Kumar Kolar

Sukrit Dasgupta

David Tedaldi

Eduard Schornig

Laurent Sartran

Jonathan W Hui

Javier Cruz Mota

Andrea Di Pietro

Michal Wladyslaw Garcarz

Sébastien Gay

Fabien Flacher

Mukund Yelahanka Raghuprasad

Jürg Nicolaus Diemand

Diane Bouchacourt

Sharon Shoshana Wulff

Abhishek Kumar

Vikram Kumaran

Alexandre Honoré

Grégoire Magendie

Santosh Ghanshyam Pandey

Erwan Barry Tarik Zerhouni

Romain Kakko-Chiloff

Sambarta Dasgupta

Yannick Weibel

Stéphane Bernard Martin

Kevin Gagnon

Gad Miller

Xav Laumonier

Petar Stupar

Jean-Phillippe Vasseur

Jean-Philippe J Vasseur

Javier Mota

In one embodiment, a device obtains information regarding temporary routing patches applied to a network. Each temporary routing patch implements a routing change in the network for a specified amount of time to avoid or mitigate against a service level agreement violation. The device evaluates, using the information regarding the temporary routing patches applied to the network, a plurality of replay scenarios for the network. The device determines, based on the plurality of replay scenarios, a long-term configuration change for the network. The device provides an indication of the long-term configuration change for display.

Generating long-term network changes from SLA violations

In one implementation, a device performs, using a using a large language model-based troubleshooting agent, troubleshooting of a plurality of issues in a network indicated by a plurality of support tickets opened by users of the network. The device aggregates, based on results of the troubleshooting, the plurality of support tickets into a master support ticket. The device obtains a resolution to the master support ticket from a support engineer. The device uses the resolution to the master support ticket in conjunction with the large language model-based troubleshooting agent to troubleshoot a new support ticket.

LLM-based agent as a back-office virtual network troubleshooting assistant

In one embodiment, a device obtains information regarding a plurality of network nodes of a network via which an online application is accessed. The device uses a prediction model to predict a quality of experience metric for the online application. The device determines, a topology change for the network, based on the quality of experience metric predicted for the online application and the information regarding the plurality of network nodes. The device causes the topology change to be implemented in the network.

QoE optimized elastic cloud networks

In one implementation, a device receives an input request for a large language model-based network troubleshooting agent regarding an issue in a network. The large language model-based network troubleshooting agent performs a lookup of a recipe based on the input request, wherein the recipe comprises contextual information for the issue. The device generates, by the large language model-based network troubleshooting agent, a prompt for a large language model based on the input request and on the recipe. The device provides, by the large language model-based network troubleshooting agent, the prompt to the large language model to troubleshoot the issue in the network.

LLM-based network troubleshooting using expert-curated recipes

In one embodiment, a device discovers one or more network interfaces that an endpoint in a local network could use to access an online application. The device identifies a plurality of connectivity options available to the endpoint to access the online application via an external network. The device makes a prediction that a path that comprises a particular connectivity option from among the plurality of connectivity options and a particular network interface from among the one or more network interfaces will provide an optimal quality of experience metric associated with the online application. The device causes the endpoint to use the path to connect to the online application.

Predictive application-aware routing for remote work

In one embodiment, a device uses a multi-armed bandit model to select different network paths over time via which traffic associated with an online application is routed. The device obtains, from a provider of the online application, application experience metrics associated with the different network paths and indicative of user satisfaction with the online application. The device learns, by the multi-armed bandit model, which of the different network paths will provide satisfactory application experience metrics, based on the application experience metrics associated with the different network paths. The device causes the traffic associated with the online application to be routed via a set of one or more paths expected by the multi-armed bandit model to provide satisfactory application experience metrics for the online application.

Avoiding user experience disruptions using contextual multi-armed bandits

In one embodiment, a device obtains media data captured by an endpoint of an online application that results from conducting a test in a network that subjects traffic of the online application to a particular network path degradation pattern. The device receives user feedback regarding the media data from one or more user interfaces. The device computes, based on the user feedback, a quality of experience metric for the online application associated with the particular network path degradation pattern. The device causes a configuration change to be made with respect to the online application, based on the quality of experience metric.

Quantifying application quality of experience under different path performance motifs

In one embodiment, a recommendation service of a device provides a recommended action to a client of an online application predicted to improve a quality of experience metric for the online application. The device receives feedback from the client indicative of the recommended action not being implemented by a user of the client. The device determines, based on the feedback, a reason for the recommended action not being implemented. The device updates the recommendation service based on the reason for the recommended action not being implemented.

End user recommendation service to ensure satisfactory application QoE

In one embodiment, a device makes a prediction regarding service level agreement violations by a network transport available between a site and an online application. The device associates a risk metric with the prediction, based in part on a type of the network transport. The device computes a longevity metric for the prediction that indicates an expected validity period for the prediction. The device cause traffic to be routed between the site and the online application using the network transport, based on the prediction and its associated risk metric and its longevity metric.

Predictive routing using risk and longevity metrics

In one implementation, a troubleshooting agent executed by a device receives an indication of an anomaly detected in a network by an anomaly detection agent. The troubleshooting agent uses a language model to determine a root cause of the anomaly. The troubleshooting agent determines whether automated remediation for the root cause is allowed by a policy. The troubleshooting agent provides a remediation request to a remediation agent in the network to remediate the root cause of the anomaly, when the policy allows automated remediation for the root cause.

Multi-agent coordination for network anomaly detection, troubleshooting, and remediation using language models

In one embodiment, a device predicts, for each of a plurality of applications accessible via a network, quality metrics for different network paths where traffic for that application be routed via one or more paths among the different network paths. The device generates a congestion risk prediction model that predicts a risk of traffic congestion for a particular combination of: applications from among the plurality of applications, traffic flows associated with those applications, and paths among the network paths via which those traffic flows may be routed. The device performs a constrained optimization based on the predicted quality metrics and on the risk of traffic congestion predicted by the model, to assign traffic flows for the applications to a selected subset of the different paths. The device causes the traffic flows to be routed in the network via the selected subset of the different paths to which they are assigned.

Cross-application predictive routing

In one embodiment, a device uses a quality of experience model to make predictions comprising probabilities of a quality of experience metric for an online application accessed via a network being acceptable. The device determines, for each of a subset of the predictions having lowest probabilities of acceptable user experience from amongst the predictions, whether that prediction corresponds to a negative experience rating by a user of the online application. The device computes a negative retrieval accuracy for the quality of experience model based on a count of negative experience ratings associated with the subset and a size of the subset. The device uses the negative retrieval accuracy to detect an issue in the network.

Measuring performance in cognitive networks

In one embodiment, a device obtains data regarding routing decisions made by a machine learning-based predictive routing engine for a network. The device determines, based on the data regarding the routing decisions, a behavior of the machine learning-based predictive routing engine. The device compares the behavior of the machine learning-based predictive routing engine to a behavioral policy for the machine learning-based predictive routing engine. The device adjusts operation of the machine learning-based predictive routing engine, when the behavior of the machine learning-based predictive routing engine violates the behavioral policy.

Safety net engine for machine learning-based network automation

In one embodiment, a device deploys short timescale prediction model and a long timescale prediction model to one or more hosts in a network, whereby the short timescale prediction model predicts failure conditions for an online application that are attributable to the network on a timescale that is shorter than that of the long timescale prediction model. The device configures a trigger that causes the long timescale prediction model to activate predictions by the short timescale prediction model. The device evaluates performance of the short timescale prediction model. The device adjusts the trigger, when the performance of the short timescale prediction model is unacceptable.

Capturing network dynamics and forecasting across multiple timescales

In one embodiment, a device identifies a plurality of online applications whose traffic is conveyed via an interface of a networking entity of a network. The device computes a relationship between a quality of experience metric for a particular online application in the plurality of online applications and a traffic volume associated with the interface for the plurality of online applications. The device makes, based on the relationship, a determination that degradation of the quality of experience metric for the particular online application is due to the traffic volume associated with the interface for the plurality of online applications. The device reconfigures, based on the determination, the networking entity to prioritize traffic for the particular online application.

Growing community of inventors

Veyras, Switzerland

Grégory Mermoud