# US 12,015,533 B2 · System for machine learning (ML) based network resilience and steering > Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt - Number: US 12,015,533 B2 - Document ID: US12015533B2 - Status: Granted - Kind: B2 - Application number: 17/956,035 - Filed: 2022-09-29 - Published: 2024-03-21 - Granted: 2024-06-18 - Earliest claimed: 2022-09-19 - Assignees: Mellanox Technologies, Ltd. - Applicants: Mellanox Technologies, Ltd. - Inventors: Ioannis (Giannis) Patronas; Tamar Viclizki Cohen; Vadim Gechman; Dimitrios Syrivelis; Paraskevas Bakopoulos; Nikolaos Argyris; Elad Mentovich - Web page: https://patentgrep.com/patent/US12015533B2 ## CPC - H04L 43/065: related to network devices - H04L 41/0668: by dynamic selection of recovery network elements, e.g. replacement by the most appropriate element after failure - H04L 41/0677: Localisation of faults - H04L 41/16: using machine learning or artificial intelligence - H04L 43/0817: by checking functioning - H04L 43/16: Threshold monitoring - H04L 45/28: using route fault recovery - H04L: TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION (arrangements common to telegraphic and telephonic communication H04M) ## Abstract Systems, computer program products, and methods are described herein for machine learning (ML) based system for network resilience and steering. An example system monitors data movement across one or more network ports; extracts network performance indicators associated with the data movement; determines, via a machine learning (ML) subsystem, that a status of a first network port is indicative of operational failure based on at least the network performance indicators; determines that the first network port is associated with a first network port cluster; determines a redundant network port and an intermediate network switch associated with the first network port cluster; and triggers the intermediate network switch to reroute a portion of network traffic from the first network port to the redundant network port in response to the status of the first network port. ## Claims ### Claim 1 A machine learning (ML) based system for network resilience and steering comprising: a non-transitory storage device; and a processor coupled to the non-transitory storage device, wherein the processor is to: monitor data movement across one or more network ports; extract network performance indicators associated with the data movement; determine, via a machine learning (ML) subsystem, a likelihood of failure for an operational first network port based on at least the network performance indicators; determine that a status of the first network port is indicative of operational failure in an instance in which the likelihood of failure for the first network port satisfies a failure threshold; determine that the first network port is associated with a first network port cluster; determine a redundant network port and an intermediate network switch associated with the first network port cluster; trigger the intermediate network switch to reroute a portion of network traffic from the first network port to the redundant network port in response to the status of the first network port, by terminating a communication link to the first network port; and re-trigger the intermediate switch to reroute the portion of network traffic back to the first network port upon detecting recovery of the first network port, by re-establishing the communication link. ### Claim 2 (depends on 1) The system of claim 1, wherein, in determining that the status of the first network port is indicative of operational failure, the processor is to: deploy, via the ML subsystem, a trained ML model on the network performance indicators extracted from the data movement; and determine, using the trained ML model, the likelihood of failure for the first network port. ### Claim 3 (depends on 2) The system of claim 2, wherein the processor is to: monitor data movement across the one or more network ports for a first time period after triggering the rerouting of the portion of network traffic from the first network port to the redundant network port; determine, using the trained ML model, the likelihood of failure for the first network port; and trigger the intermediate network switch to reroute the portion of network traffic from the redundant network port to the first network port in an instance in which the likelihood of failure of the first network port fails to satisfy the failure threshold. ### Claim 4 (depends on 3) The system of claim 3, wherein, in triggering the intermediate network switch to reroute the portion of network traffic from the redundant network port to the first network port, the processor is to trigger the intermediate network switch to terminate a communication link to the redundant network port. ### Claim 5 (depends on 1) The system of claim 1, wherein the processor is to: monitor data movement across the one or more network ports for a second time period; extract the network performance indicators associated with the data movement for the second time period; determine the status of the one or more network ports during the second time period; generate a feature set using the network performance indicators and the status of the one or more network ports; and train, using the ML subsystem, an ML model using the feature set to generate the trained ML model. ### Claim 6 (depends on 1) The system of claim 1, wherein, prior to monitoring the data movement, the processor is to supply a stimulus to the first network port cluster. ### Claim 7 A computer program product for machine learning (ML) based network resilience and steering, the computer program product comprising a non-transitory computer-readable medium comprising code causing an apparatus to: monitor data movement across one or more network ports; extract network performance indicators associated with the data movement; determine, via a machine learning (ML) subsystem, a likelihood of failure for an operational first network port based on at least the network performance indicators; determine that a status of the first network port is indicative of operational failure in an instance in which the likelihood of failure for the first network port satisfies a failure threshold; determine that the first network port is associated with a first network port cluster; determine a redundant network port and an intermediate network switch associated with the first network port cluster; trigger the intermediate network switch to reroute a portion of network traffic from the first network port to the redundant network port in response to the status of the first network port, by terminating a communication link to the first network port; and re-trigger the intermediate switch to reroute the portion of network traffic back to the first network port upon detecting recovery of the first network port, by re-establishing the communication link. ### Claim 8 (depends on 7) The computer program product of claim 7, wherein, in determining that the status of the first network port is indicative of operational failure, the apparatus is to: deploy, via the ML subsystem, a trained ML model on the network performance indicators extracted from the data movement; and determine, using the trained ML model, the likelihood of failure for the first network port. ### Claim 9 (depends on 8) The computer program product of claim 8, wherein the apparatus is to: monitor data movement across the one or more network ports for a first time period after triggering the rerouting of the portion of network traffic from the first network port to the redundant network port; determine, using the trained ML model, the likelihood of failure for the first network port; and trigger the intermediate network switch to reroute the portion of network traffic from the redundant network port to the first network port in an instance in which the likelihood of failure of the first network port fails to satisfy the failure threshold. ### Claim 10 (depends on 9) The computer program product of claim 9, wherein, in triggering the intermediate network switch to reroute the portion of network traffic from the redundant network port to the first network port, the apparatus is to trigger the intermediate network switch to terminate a communication link to the redundant network port. ### Claim 11 (depends on 7) The computer program product of claim 7, wherein the apparatus is to: monitor data movement across the one or more network ports for a second time period; extract the network performance indicators associated with the data movement for the second time period; determine the status of the one or more network ports during the second time period; generate a feature set using the network performance indicators and the status of the one or more network ports; and train, using the ML subsystem, an ML model using the feature set to generate the trained ML model. ### Claim 12 (depends on 7) The computer program product of claim 7, wherein, prior to monitoring the data movement, the apparatus is to supply a stimulus to the first network port cluster. ### Claim 13 A method for machine learning (ML) based network resilience and steering, the method comprising: monitoring data movement across one or more network ports; extracting network performance indicators associated with the data movement; determining, via a machine learning (ML) subsystem, a likelihood of failure for an operational first network port based on at least the network performance indicators; determining that a status of the first network port is indicative of operational failure in an instance in which the likelihood of failure for the first network port satisfies a failure threshold; determining that the first network port is associated with a first network port cluster; determining a redundant network port and an intermediate network switch associated with the first network port cluster; and triggering the intermediate network switch to reroute a portion of network traffic from the first network port to the redundant network port in response to the status of the first network port, by terminating a communication link to the first network port; and re-triggering the intermediate switch to reroute the portion of network traffic back to the first network port upon detecting recovery of the first network port, by re-establishing the communication link. ### Claim 14 (depends on 13) The method of claim 13, wherein determining that the status of the first network port is indicative of operational failure further comprises: deploying, via the ML subsystem, a trained ML model on the network performance indicators extracted from the data movement; and determining, using the trained ML model, the likelihood of failure for the first network port. ### Claim 15 (depends on 14) The method of claim 14, further comprising: monitoring data movement across the one or more network ports for a first time period after triggering the rerouting of the portion of network traffic from the first network port to the redundant network port; determining, using the trained ML model, the likelihood of failure for the first network port; and triggering the intermediate network switch to reroute the portion of network traffic from the redundant network port to the first network port in an instance in which the likelihood of failure of the first network port fails to satisfy the failure threshold. ### Claim 16 (depends on 13) The method of claim 13, wherein triggering the intermediate network switch to reroute the portion of network traffic from the redundant network port to the first network port further comprises triggering the intermediate network switch to terminate a communication link to the redundant network port. ### Claim 17 (depends on 13) The method of claim 13, further comprising: monitoring data movement across the one or more network ports for a second time period; extracting the network performance indicators associated with the data movement for the second time period; determining the status of the one or more network ports during the second time period; generating a feature set using the network performance indicators and the status of the one or more network ports; and training, using the ML subsystem, an ML model using the feature set to generate the trained ML model. ### Claim 18 (depends on 1) The system of claim 1, wherein the first network port is operational during the determination of the likelihood of failure for the first network port based on at least the network performance indicators. ### Claim 19 (depends on 7) The computer program product of claim 7, wherein the first network port is operational during the determination of the likelihood of failure for the first network port based on at least the network performance indicators. ### Claim 20 (depends on 13) The method of claim 13, wherein the first network port is operational during the determination of the likelihood of failure for the first network port based on at least the network performance indicators.