# US 11,121,952 B2 · Device health assessment data summarization using machine learning > Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt - Number: US 11,121,952 B2 - Document ID: US11121952B2 - Status: Granted - Kind: B2 - Application number: 16/162,549 - Filed: 2018-10-17 - Published: 2020-04-23 - Granted: 2021-09-14 - Earliest claimed: 2018-10-17 - Assignees: Cisco Technology, Inc. - Applicants: Cisco Technology, Inc. - Inventors: Mohammed Hamzeh; Sam Grimee; Cedric Van Labeke; Peter Van Eynde; Raphaël Wouters - Web page: https://patentgrep.com/patent/US11121952B2 ## CPC - H04L 43/0817: by checking functioning - H04L 63/1425: {Traffic logging, e.g. anomaly detection} - G06N 3/09: Supervised learning - G06N 5/01: Dynamic search techniques; Heuristics; Dynamic trees; Branch-and-bound - G06N 20/00: Machine learning - G06N 20/20: Ensemble learning - H04L 41/22: comprising specially adapted graphical user interfaces [GUI] - H04L 43/14: {using software, i.e. software packages (network security related monitoring H04L63/1408)} - G06N 3/044: Recurrent networks, e.g. Hopfield networks - G06N 3/049: Temporal neural networks, e.g. delay elements, oscillating neurons or pulsed inputs - G06N 3/08: Learning methods - G06N 7/01: Probabilistic graphical models, e.g. probabilistic networks - G06N 20/10: using kernel methods, e.g. support vector machines [SVM] - H04L: TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION (arrangements common to telegraphic and telephonic communication H04M) - G06N: COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS ## Abstract In one embodiment, a device health assessment service extracts device health status indicators from health assessment data that the service uses to determine a device health status of a networking device. The service forms, using the extracted set of device health status indicators, a health status signature for a particular device health status. The service trains a machine learning-based model to classify whether a networking device has the particular device health status, based in part on the health status signature. The service deploys the machine learning-based model to a target network for local device health assessment of one or more networking devices in the target network. ## Claims ### Claim 1 A method comprising: extracting, by a device health assessment service, device health status indicators from health assessment data that the service uses to determine a device health status of a networking device; forming, by the device health assessment service and using the extracted device health status indicators, a health status signature for a particular device health status by labeling each of the extracted device health status indicators with a corresponding health status; training, by the device health assessment service and based in part on the health status signature, a machine learning-based model to classify whether the networking device has the particular device health status, wherein training data for the machine learning-based model is generated synthetically by converting logic inside the health status signature into a truth table representation of the logic; deploying, by the device health assessment service, the machine learning-based model to a target network for local device health assessment of one or more networking devices in the target network; generating, by the device health assessment service, a command line output parser configured to form a device signature from one or more command line outputs of the networking device; and deploying, by the device health assessment service, the command line output parser to the target network, wherein the machine learning-based model takes the device signature from the command line output parser as input to classify whether a networking device with the device signature has the particular device health status. ### Claim 2 (depends on 1) The method as in claim 1, wherein the one or more networking devices in the target network comprise at least one of: a network router or a network switch. ### Claim 3 (depends on 1) The method as in claim 1, wherein the health assessment data comprises a set of one or more health assessment scripts that assess the one or more command line outputs of the networking device. ### Claim 4 (depends on 1) The method as in claim 1, wherein the machine learning-based model comprises a random forest classifier. ### Claim 5 (depends on 1) The method as in claim 1, wherein the machine learning-based model is executed in the target network by the networking device whose device health status the machine learning-based model is to assess. ### Claim 6 (depends on 1) The method as in claim 1, wherein the machine learning-based model is configured to provide an indication as to whether the networking device has the particular device health status to a user interface. ### Claim 7 (depends on 1) The method as in claim 1, further comprising: using, by the device health assessment service, the health assessment data to determine a device health status of the networking device, based on the one or more command line outputs of the networking device. ### Claim 8 An apparatus, comprising: one or more network interfaces to communicate with a network; a processor coupled to the network interfaces and configured to execute one or more processes; and a memory configured to store a process executable by the processor, the process when executed configured to: extract device health status indicators from health assessment data that the apparatus uses to determine a device health status of a networking device; form, using the extracted device health status indicators, a health status signature for a particular device health status by labeling each of the extracted device health status indicators with a corresponding health status; train, based in part on the health status signature, a machine learning-based model to classify whether the networking device has the particular device health status, wherein training data for the machine learning-based model is generated synthetically by converting logic inside the health status signature into a truth table representation of the logic; deploy the machine learning-based model to a target network for local device health assessment of one or more networking devices in the target network; generating, by the device health assessment service, a command line output parser configured to form a device signature from one or more command line outputs of the networking device; and deploying, by the device health assessment service, the command line output parser to the target network, wherein the machine learning-based model takes the device signature from the command line output parser as input to classify whether a networking device with the device signature has the particular device health status. ### Claim 9 (depends on 8) The apparatus as in claim 8, wherein the one or more networking devices in the target network comprise at least one of: a network router or a network switch. ### Claim 10 (depends on 8) The apparatus as in claim 8, wherein the health assessment data comprises a set of one or more health assessment scripts that assess the one or more command line outputs of the networking device. ### Claim 11 (depends on 8) The apparatus as in claim 8, wherein the machine learning-based model comprises a random forest classifier. ### Claim 12 (depends on 8) The apparatus as in claim 8, wherein the machine learning-based model is executed in the target network by the networking device whose device health status the machine learning-based model is to assess. ### Claim 13 (depends on 8) The apparatus as in claim 8, wherein the machine learning-based model is configured to provide an indication as to whether the networking device has the particular device health status to a user interface. ### Claim 14 (depends on 8) The apparatus as in claim 8, wherein the process when executed is further configured to: use the health assessment data to determine a device health status of the networking device, based on the one or more command line outputs of the networking device. ### Claim 15 A tangible, non-transitory, computer-readable medium storing program instructions that cause a device health assessment service to execute a process comprising: extracting, by the device health assessment service, device health status indicators from health assessment data that the service uses to determine a device health status of a networking device; forming, by the device health assessment service and using the extracted device health status indicators, a health status signature for a particular device health status by labeling each of the extracted device health status indicators with a corresponding health status; training, by the device health assessment service and based in part on the health status signature, a machine learning-based model to classify whether the networking device has the particular device health status, wherein training data for the machine learning-based model is generated synthetically by converting logic inside the health status signature into a truth table representation of the logic; deploying, by the device health assessment service, the machine learning-based model to a target network for local device health assessment of one or more networking devices in the target network; generating, by the device health assessment service, a command line output parser configured to form a device signature from one or more command line outputs of the networking device; and deploying, by the device health assessment service, the command line output parser to the target network, wherein the machine learning-based model takes the device signature from the command line output parser as input to classify whether a networking device with the device signature has the particular device health status.