# US 10,375,565 B2 · Dynamic rerouting of wireless traffic based on input from machine learning-based mobility path analysis > Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt - Number: US 10,375,565 B2 - Document ID: US10375565B2 - Status: Granted - Kind: B2 - Application number: 15/783,342 - Filed: 2017-10-13 - Published: 2019-04-18 - Granted: 2019-08-06 - Earliest claimed: 2017-10-13 - Assignees: Cisco Technology, Inc. - Applicants: Cisco Technology, Inc. - Inventors: Jean-Philippe Vasseur; Vinay Kumar Kolar; Santosh Pandey - Web page: https://patentgrep.com/patent/US10375565B2 ## CPC - H04W 8/082: {for traffic bypassing of mobility servers, e.g. location registers, home PLMNs or home agents} - H04W 8/005: {Discovery of network devices, e.g. terminals} - H04W 24/04: Arrangements for maintaining operational condition - H04W 28/04: Error control - H04W 36/008375: {based on historical data} - H04W: WIRELESS COMMUNICATION NETWORKS (broadcast communication H04H; communication systems using wireless links for non-selective communication, e.g. wireless extensions H04M1/72) ## Abstract In one embodiment, a service receives data indicative of roaming failures along mobility paths in a network. The mobility paths represent ordered series of wireless access points via which wireless clients have accessed the network over time. The service uses, based on the data indicative of the roaming failures, a machine learning-based model to associate mobility path failure metrics with portions of the mobility paths. The service identifies, for a first mobility path, an alternate mobility path that has a lower mobility path failure metric than that of the first mobility path. The service triggers a mobility path reroute for a particular client device in the network on the first mobility path to the alternate mobility path. ## Claims ### Claim 1 A method comprising: receiving, at a service, data indicative of roaming failures along mobility paths in a network, wherein the mobility paths represent ordered series of wireless access points via which wireless clients have accessed the network over time; using, by the service and based on the data indicative of the roaming failures, a machine learning-based model to associate mobility path failure metrics with portions of the mobility paths; receiving, at the service, a threshold mobility path failure metric from the particular client device; identifying, by the service and for a first mobility path, an alternate mobility path that has a lower mobility path failure metric than that of the first mobility path based on an output from the machine learning model; and triggering, by the service, a mobility path reroute for a particular client device in the network on the first mobility path to the alternate mobility path based in part on a determination that the mobility path metric associated with the alternate mobility path is lower than the received threshold. ### Claim 2 (depends on 1) The method as in claim 1, wherein triggering the mobility path reroute comprises: sending, by the service, data indicative of the alternate mobility path to the particular client device via an access point in the network, wherein the data indicative of the alternate mobility path causes the particular client device to roam to an access point along the alternate mobility path. ### Claim 3 (depends on 2) The method as in claim 2, wherein the data indicative of the alternate mobility path comprises at least one mobility path metric associated with the alternate mobility path. ### Claim 4 (depends on 1) The method as in claim 1, further comprising: receiving, by the service and from the particular client device, an indication that the particular client device experienced a roaming failure along the alternate mobility path; and modifying, by the service, the machine learning-based model based on the received indication that the particular client device experienced a roaming failure along the alternate mobility path. ### Claim 5 (depends on 1) The method as in claim 1, wherein a mobility failure corresponds to a traffic disruption experienced by one of the client devices caused by a failure to roam between wireless access points. ### Claim 6 (depends on 1) The method as in claim 1, wherein the mobility path reroute is triggered in part based on one or more service level agreements (SLAs) associated with one or more active applications on the particular client device. ### Claim 7 (depends on 1) The method as in claim 1, wherein the wireless access points are Wi-Fi access points. ### Claim 8 (depends on 1) The method as in claim 1, further comprising: computing, by the service, the mobility paths in the network in part by clustering trajectories of the client devices between the access points in the network. ### Claim 9 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: receive data indicative of roaming failures along mobility paths in the network, wherein the mobility paths represent ordered series of wireless access points via which wireless clients have accessed the network over time; use, based on the data indicative of the roaming failures, a machine learning-based model to associate mobility path failure metrics with portions of the mobility paths; receive a threshold mobility path failure metric from the particular client device; identify, for a first mobility path, an alternate mobility path that has a lower mobility path failure metric than that of the first mobility path based on an output from the machine learning model; and trigger a mobility path reroute for a particular client device in the network on the first mobility path to the alternate mobility path based in part on a determination that the mobility path metric associated with the alternate mobility path is lower than the received threshold. ### Claim 10 (depends on 9) The apparatus as in claim 9, wherein the apparatus triggers the mobility path reroute by: sending data indicative of the alternate mobility path to the particular client device via an access point in the network, wherein the data indicative of the alternate mobility path causes the particular client device to roam to an access point along the alternate mobility path. ### Claim 11 (depends on 10) The apparatus as in claim 10, wherein the data indicative of the alternate mobility path comprises at least one mobility path metric associated with the alternate mobility path. ### Claim 12 (depends on 9) The apparatus as in claim 9, wherein the process when executed is further configured to: receive, from the particular client device, an indication that the particular client device experienced a roaming failure along the alternate mobility path; and modify the machine learning-based model based on the received indication that the particular client device experienced a roaming failure along the alternate mobility path. ### Claim 13 (depends on 9) The apparatus as in claim 9, wherein a mobility failure corresponds to a traffic disruption experienced by one of the client devices caused by a failure to roam between wireless access points. ### Claim 14 (depends on 9) The apparatus as in claim 9, wherein the mobility path reroute is triggered in part based on one or more service level agreements (SLAs) associated with one or more active applications on the particular client device. ### Claim 15 (depends on 9) The apparatus as in claim 9, wherein the wireless access points are Wi-Fi access points. ### Claim 16 (depends on 9) The apparatus as in claim 9, wherein the process when executed is further configured to: computing, by the service, the mobility paths in the network in part by clustering trajectories of the client devices between the access points in the network. ### Claim 17 A tangible, non-transitory, computer-readable medium storing program instructions that cause a device to execute a process comprising: receiving data indicative of roaming failures along mobility paths in a network, wherein the mobility paths represent ordered series of wireless access points via which wireless clients have accessed the network over time; using, based on the data indicative of the roaming failures, a machine learning-based model to associate mobility path failure metrics with portions of the mobility paths; receiving a threshold mobility path failure metric from the particular client device; identifying, for a first mobility path, an alternate mobility path that has a lower mobility path failure metric than that of the first mobility path based on an output from the machine learning model; and triggering a mobility path reroute for a particular client device in the network on the first mobility path to the alternate mobility path based in part on a determination that the mobility path metric associated with the alternate mobility path is lower than the received threshold. ### Claim 18 (depends on 17) The computer-readable medium as in claim 17, wherein the process further comprises: receiving, from the particular client device, an indication that the particular client device experienced a roaming failure along the alternate mobility path; and modifying the machine learning-based model based on the received indication that the particular client device experienced a roaming failure along the alternate mobility path.