flapping
The situation where a resource, a network destination, etc., is advertised as being available and then unavailable (or available by different routes) in rapid succession.
flapping: when a network route keeps failing and recovering
Flapping occurs when a network interface, route, or BGP advertisement oscillates rapidly between up and down states, often multiple times per second. Rather than staying failed or staying healthy, the resource toggles its availability in quick succession, causing routing protocols to continuously recalculate and send update announcements across the network.
The cause is typically instability at the physical layer: a fiber optic cable with a weak connection that intermittently loses signal, a switch port with corroded contacts, or a wireless link affected by interference or movement. Each time the link drops below the threshold for carrier detection, the routing daemon marks the route as unavailable and floods neighbors with withdrawal messages. When the signal recovers momentarily, it advertises the route as available again. This cycle repeats, generating unnecessary control traffic and CPU load on routers.
Flapping wastes bandwidth on routing updates and destabilizes the network. Routers may take longer to converge on stable routes. BGP speakers may accumulate flap history and impose temporary dampening, holding a route down even after it stabilizes. Network management systems should log flapping events and alert on repeated failures of the same interface or prefix, since flapping often precedes complete failure.
The term reflects the mechanical analogy: a relay contact that bounces, making and breaking electrical connection in rapid succession. In networking, flapping is diagnosed by examining syslog messages, interface statistics, and routing table change logs. The interval between state changes is usually measured in seconds or milliseconds, distinguishing flapping from normal failover.
Prevention requires attention to physical infrastructure: securing and cleaning connections, replacing aging cables and transceivers, and eliminating sources of electromagnetic interference. Tuning holddown timers and dampening thresholds in routing protocols can reduce the impact of brief flaps, though they add delay during legitimate network changes.