Adaptive MAC grouping and timeout in EVPN environments using machine learning

An apparatus, comprising:

• one or more network interfaces to communicate with a network;
• a processor coupled to the one or more network interfaces and configured to execute a process;
• and a memory configured to store the process executable by the processor, the process when executed operable to:
• monitor Ethernet virtual private network (EVPN) traffic in the network for a plurality of media access control (MAC) addresses;
• generate a machine learning-based traffic model for the MAC addresses based on the monitored EVPN traffic, wherein the machine learning-based traffic model predicts future traffic patterns for a particular one of the MAC addresses;
• determine a timeout for the particular MAC address based on the future traffic patterns predicted by the machine learning-based traffic model for the particular MAC address;
• and cause the particular MAC address to be timed out from one or more forwarding tables in the network based on the determined timeout.

A tangible, non-transitory, computer-readable medium storing program instructions that, when executed by a device in a network, cause the device to perform a process comprising:

• monitoring, by the device, Ethernet virtual private network (EVPN) traffic in the network for a plurality of media access control (MAC) addresses;
• generating, by the device, a machine learning-based traffic model for the MAC addresses based on the monitored EVPN traffic, wherein the machine learning-based traffic model predicts future traffic patterns for a particular one of the MAC addresses;
• determining, by the device, a timeout for the particular MAC address based on the future traffic patterns predicted by the machine learning-based traffic model for the particular MAC address;
• and causing, by the device, the particular MAC address to be timed out from one or more forwarding tables in the network based on the determined timeout.