RTOarvot

RTOarvot, or retransmission timeout values, are timers used by transport protocols to control retransmission of unacknowledged data. They are most commonly associated with TCP, where the sender waits for an acknowledgement before retransmitting a segment that may have been lost in the network. RTOarvot are not fixed; they adapt to observed network conditions through round-trip time measurements.

In TCP, each connection tracks a smoothed RTT estimate (SRTT) and the RTT variation (RTTVAR). The retransmission timeout is derived from these values. A commonly cited formulation is RTO = SRTT + 4*RTTVAR, with safeguards that bound RTO to minimum and maximum values to avoid timers that are too aggressive or too conservative. Different TCP implementations may use slightly different formulas, but the adaptive principle remains the same.

The accuracy of RTT sampling and the presence of jitter affect RTO accuracy. Unstable networks, congestion, or queuing delays can inflate RTTVAR, causing longer timeouts; conversely, noisy samples can temporarily produce too-small RTOs and premature retransmissions. Systems apply limits and periodic resynchronization after long idle periods to maintain reliability.

Operationally, RTOarvot influence throughput and latency. If the timer is too short, the sender may retransmit unnecessarily, wasting bandwidth and triggering congestion control adjustments. If the timer is too long, the application experiences higher latency before data is recovered from loss. Modern stacks tune RTO behavior automatically and variably across platforms.

See also: retransmission timeout, RTT estimation, TCP congestion control.