Keepaliveinterval

Keepaliveinterval is a parameter used in networking to specify the time between successive keepalive probes sent over a connection to verify that the peer is still reachable. It is part of keepalive mechanisms designed to detect failed peers and, in some environments, to prevent idle connections from being dropped by NATs or firewalls. The value is usually expressed in seconds or milliseconds and can be configured at different levels—per connection, per application, or system-wide—depending on the protocol and platform.

How it works: In a typical keepalive setup, a connection waits an initial idle period before the first probe is sent. If the peer responds, the timer resets. If no response is received after a defined number of probes, the connection is considered dead and may be closed. The keepalive interval is the time between probes; the initial idle time (keepalive time) and the probe count are related parameters that vary by stack. Not all implementations expose all three knobs, and defaults differ.

Usage and configuration: Linux exposes the trio net.ipv4.tcp_keepalive_time, tcp_keepalive_intvl, and tcp_keepalive_probes. Windows uses KeepAliveTime and KeepAliveInterval in the TCP/IP stack registry. The exact values depend on the operating system, network environment, and application requirements, as well as on any middleware or network devices involved.

Considerations: Shorter keepalive intervals provide faster failure detection and better resilience to timeouts but generate more periodic traffic and can interact poorly with some intermediaries. Longer intervals reduce overhead but may delay detection of a dead peer. Applications should balance responsiveness, bandwidth, and device behavior when choosing values.

Note: This concept is distinct from application-layer keepalives such as HTTP keep-alive or WebSocket pings, which operate at higher protocol levels.