Machine learning approach for dynamic adjustment of bfd timers in sd-wan networks
AssigneeCisco Technology, Inc.
InventorsJean-Philippe Vasseur, Vinay Kumar Kolar, Grégory Mermoud
Life of this patent
Estimating the term…
Abstract
In one embodiment, a device obtains performance data regarding failures of a tunnel in a network. The device generates a failure profile for the tunnel by applying machine learning to the performance data regarding the failures of the tunnel. The device determines, based on the failure profile for the tunnel, whether the tunnel exhibits failure flapping behavior. The device adjusts one or more Bidirectional Forwarding Detection (BFD) probing timers used to detect failures of the tunnel, based on the determination as to whether the tunnel exhibits failure flapping behavior.
Claims20
A method, comprising:
- obtaining, by a device, performance data regarding failures of a tunnel in a network;
- generating, by the device, a failure profile for the tunnel by applying machine learning to the performance data regarding the failures of the tunnel;
- determining, by the device and based on the failure profile for the tunnel, whether the tunnel exhibits failure flapping behavior, wherein the failure flapping behavior is indicative of oscillation between up and down states;
- and
- adjusting, by the device, one or more Bidirectional Forwarding Detection (BFD) probing timers used to detect failures of the tunnel, based on the determination as to whether the tunnel exhibits failure flapping behavior by reducing, based on a determination that the tunnel does not exhibit failure flapping behavior, a timeout timer that controls when the tunnel is deemed to have failed after a BFD probe is sent via the tunnel and was unacknowledged.
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:
- obtain performance data regarding failures of a tunnel in a network;
- generate a failure profile for the tunnel by applying machine learning to the performance data regarding the failures of the tunnel;
- determine, based on the failure profile for the tunnel, whether the tunnel exhibits failure flapping behavior, wherein the failure flapping behavior is indicative of oscillation between up and down states;
- and
- adjust one or more Bidirectional Forwarding Detection (BFD) probing timers used to detect failures of the tunnel, based on the determination as to whether the tunnel exhibits failure flapping behavior by reducing, based on a determination that the tunnel does not exhibit failure flapping behavior, a timeout timer that controls when the tunnel is deemed to have failed after a BFD probe is sent via the tunnel and was unacknowledged.
A tangible, non-transitory, computer-readable medium storing program instructions that cause a device to execute a process comprising:
- obtaining, by a device, performance data regarding failures of a tunnel in a network;
- generating, by the device, a failure profile for the tunnel by applying machine learning to the performance data regarding the failures of the tunnel;
- determining, by the device and based on the failure profile for the tunnel, whether the tunnel exhibits failure flapping behavior, wherein the failure flapping behavior is indicative of oscillation between up and down states;
- and
- adjusting, by the device, one or more Bidirectional Forwarding Detection (BFD) probing timers used to detect failures of the tunnel, based on the determination as to whether the tunnel exhibits failure flapping behavior by reducing, based on a determination that the tunnel does not exhibit failure flapping behavior, a timeout timer that controls when the tunnel is deemed to have failed after a BFD probe is sent via the tunnel and was unacknowledged.
Dependencies are inferred from the claim text.