keepalive

A keepalive is a minimal data packet or message transmitted at regular intervals between two connected devices, typically across a network link or serial connection, to confirm that the link remains active and responsive. In energy and utilities operations, keepalives are critical for supervisory control and data acquisition (SCADA) systems, remote terminal units (RTUs), and telemetry networks where a silent connection cannot be trusted to be a working one.

The core problem keepalives solve is ambiguity: when no substantive data flows between two devices, neither knows whether the other is still online or whether the communication path is intact. Network equipment such as routers, firewalls, and proxies often sever idle connections to free resources, leaving both endpoints thinking they remain connected. A keepalive message, sent every 30 to 300 seconds depending on the protocol and network conditions, forces traffic across the link and prevents intermediate equipment from dropping it. In SCADA systems monitoring power distribution or water infrastructure, losing a connection silently is worse than losing it noisily; a keepalive ensures an operator knows immediately when a remote site goes dark.

Keepalives are deliberately lightweight. A TCP keepalive probe is typically just a few bytes of acknowledgment traffic. Most industrial protocols, Modbus, DNP3, IEC 60870-5-104, incorporate keepalive or heartbeat mechanisms into their application layer rather than relying on TCP alone, because TCP keepalives may be stripped by network security appliances or disabled by default. An industrial device might send a link test frame every 60 seconds, a null message that carries no real-world data but proves bidirectional communication is working.

Configuration and failure modes

The timeout interval and retry behavior must be tuned to suit the network. Satellite links to remote substations may require longer timeouts; cellular backups may need shorter ones to detect dead air quickly. If keepalives themselves are lost due to congestion or flapping routes, a device will declare the link down and may trigger failover to a redundant path, switching SCADA traffic to a backup radio or fiber connection. Conversely, if keepalives arrive but the remote device is actually hung or corrupted, the link appears healthy while control commands silently fail, so keepalives are a necessary but not sufficient check on system state.

The term reflects its function: keeping a connection alive by forcing it to do work. In early serial and dial-up systems, any packet transmission would reset idle timers; in modern networks, explicit keepalive traffic is necessary because the links themselves are always nominally open. Utilities operators rarely think about keepalives directly, but when a remote terminal unit suddenly stops responding to a dispatcher, failure to receive keepalives is usually the first indicator that the communication path has broken.