Sign in Subscribe

RTO retransmission timeout timer

Last updated on 

The RTO (Retransmission Timeout) timer is a mechanism used in computer networks to control the retransmission of lost or delayed packets. When a sender transmits a packet, it expects to receive an acknowledgment (ACK) from the receiver to confirm successful delivery. However, due to network congestion, packet loss, or other factors, ACKs may not arrive within a certain time frame.

The RTO timer is used to determine how long a sender should wait for an ACK before assuming that the packet was lost and retransmitting it. The goal of the RTO timer is to strike a balance between retransmitting too quickly, which can result in unnecessary network congestion, and waiting too long, which can cause delays in data delivery.

Here's a detailed explanation of how the RTO timer works:

  1. Initial RTO: When a connection is established, the RTO timer is initialized to a default value. This default value is typically based on the characteristics of the network, such as the estimated round-trip time (RTT) and the variance in RTT. The RTT is the time taken for a packet to travel from the sender to the receiver and back.
  2. Measuring RTT: As packets are sent, the sender measures the time it takes for each packet to reach the receiver and receive an ACK. This measured time is called the SampleRTT. It represents the observed RTT for a specific packet.

Estimated RTT: To estimate the RTT, the sender calculates a weighted average of the SampleRTT values. The weights assigned to each SampleRTT can vary depending on the implementation. The estimated RTT is denoted as EstimatedRTT.

EstimatedRTT = (1 - α) * EstimatedRTT + α * SampleRTT

The parameter α (alpha) is a smoothing factor that determines how much weight is given to the new SampleRTT compared to the previous EstimatedRTT. It typically ranges between 0 and 1, with values closer to 0 giving more weight to the previous EstimatedRTT.

Deviation in RTT: Alongside the EstimatedRTT, the sender also calculates the deviation in RTT, denoted as DevRTT. It provides an estimate of the variance or fluctuations in the RTT.

DevRTT = (1 - β) * DevRTT + β * |SampleRTT - EstimatedRTT|

Similar to α, the parameter β (beta) is a smoothing factor that determines the weight given to the new difference between SampleRTT and EstimatedRTT. It helps capture the variations in the RTT.

Timeout Interval: Based on the EstimatedRTT and DevRTT, the sender calculates the timeout interval, denoted as TimeoutInterval. The TimeoutInterval determines how long the sender will wait for an ACK before assuming the packet was lost and triggering retransmission.

TimeoutInterval = EstimatedRTT + 4 * DevRTT

The value 4 is often used as a scaling factor to provide a conservative estimate and account for possible variations in the network.

  1. RTO Timer: The RTO timer is set to the TimeoutInterval value whenever a packet is sent. The sender starts the timer and waits for an ACK. If the ACK arrives before the timer expires, the sender resets the timer for the next packet.
  2. Timer Expiry: If the timer expires before an ACK is received, the sender assumes the packet was lost and initiates retransmission. The RTO timer is reset to the new TimeoutInterval value, and the packet is sent again.
  3. Adaptive Adjustments: As more packets are transmitted and ACKs are received, the sender continues to update the EstimatedRTT and DevRTT values. This adaptive adjustment allows the RTO timer to adapt to changes in network conditions, such as increased or decreased congestion levels.

By continuously measuring and adjusting the RTO timer, the sender can optimize retransmission decisions, reduce unnecessary retransmissions, and maintain an efficient flow of data in the network. The precise algorithms and parameter values used for RTO timer calculation may vary between network protocols and implementations, but the fundamental principles remain similar.